Charter Offers
Charter Offers
Charter Offers
in Shipping
Fuzzy Set Theory and Case-Based Reasoning Approaches
Setyo Nugroho
D83
Concept of Compatibility in Shipping
- Fuzzy Set Theory and
Case-Based Reasoning Approaches -
vorgelegt von
-Dr.-Ing.-
genehmigte Dissertation
Promotionsausschuss:
BERLIN 2005
D83
Dedicated to:
my children,
my wife
and
my parents
i
Summary
This thesis aims at contributing a new perspective on the way living pro-
blems in shipping practice are addressed.
Shipping is the workhorse of the world economy. It connects nearly all
parts of the world. State of the macro economy, weather and political events
affect the sector very much. The magnitude of their influence is unclear
and is difficult to justify objectively. Those aspects are taken into account
when determining of freight rates or charter hire. Therefore shipping can be
viewed as a complex system.
The existing tools including market forecasts and shipping indexes are
less useful for conducting daily shipping practice. Information is available
in abundance, but the useful one is still scarce and expensive.
In spite of the complexity of the living problems in shipping, an experi-
enced shipping practitioner can solve problems quickly and -frequently- sa-
tisfactorily. An experienced ship broker, for example, can assess the value
of ships better and faster than any value assessment method. This merit
leads to the idea, that it would be useful to have a better insight on this
phenomenon that experience and intuition play an important role in practi-
ce and that similar problems tend to have similar solutions. Therefore the
capability to recognize similar problems will be helpful to find the desired
solution.
Compatibility, a concept proposed as an attempt to explain and to em-
brace the way problems-solvings (value assessment and planning tasks)
are conducted in daily shipping practice. The concept of compatibility invol-
ves an act of grouping of similar attributes of similar objects, for example
in the case of a ship replacement. The concept concerns also an act of
matching between objects of different roles or functionalities, for example
between a ship and route, or between a ship and a charter hire value. It
implies that the concept of compatibility means an act of problem-solving.
Fuzzy set theory and case-based reasoning are used, given the capa-
bility for addressing the notions of vagueness and learning respectively,
which mark the concept of compatibility. Three models have been deve-
loped : (a) fuzzy arithmetic-based voyage estimation (b) tailor-made ship-
ping index and (c) case-based stowage planning. The implementation of
the concept delivers a new and promising way of addressing the problems
which are not addressed satisfactorily at present.
ii
Kurzfassung
Mit dieser Arbeit wird versucht, eine neue Perspektive für die Lösung
von Problemen in der Handelsschifffahrtspraxis zu schaffen.
Schifffahrt ist in der Weltwirtschaft der wichtigste Transporteur, vor allem
für Güter. Sie verbindet fast alle Teile der Welt miteinander. Die Welt-
wirtschaft, die Politik einzelner Länder, Einzeleinflüsse wie Ölpreis, Klima
oder Wetterbedingungen beeinflussen den Wirtschaftssektor Schifffahrt in
großem Maße. Die Größenordnungen der Einflüsse sind veränderlich und
schwierig objektiv fassbar. Sie sind der Hintergrund auch für den Wert
eines Schiffes sowie für Fracht- und Charterraten. Schifffahrt ist daher ein
hoch komplexes System.
Existierende Hilfsmittel, wie Schifffahrts-Indexe, Bewertungsmodelle und
Marktprognosen, sind keine hilfreichen Werkzeuge im täglichen Schifffahrts-
geschäft. Viele Informationen sind verfügbar, aber die wirklich nützlichen
sind knapp und teuer.
Trotz der Komplexität der realen Problematik können erfahrene Schiff-
fahrtspraktiker ihre Probleme oft schnell und zufriedenstellend lösen. Zum
Beispiel können Schiffsmakler den Wert von Schiffen besser und schneller
abschätzen als veröffentlichte Bewertungsmodelle. Diese Fähigkeit führt
zu der These, dass es nützlich wäre, eine verbesserte Einsicht in das
Phänomen zu gewinnen, dass Intuition und Erfahrung viel zum täglichen
Problemlösen beitragen und ähnliche Probleme dazu neigen, zu ähnlichen
Lösungen zu führen. Das Erkennen ähnlicher Situationen kann daher zu
wichtigen Hinweisen für die Lösung eines Problems führen. Der Begriff
Kompatibilität wird benutzt, um die Art und Weise zu erklären, in der Prob-
lemlösungen in der Praxis (z.B. Wertermittlung und Beschäftigungsplanung)
erfolgen.
Der Begriff Kompatibilität enthält einen Akt von Gruppierung ähnlicher
Attribute oder ähnlicher Objekte, z.B. im Zusammenhang mit dem Ersatz
eines Schiffes durch ein anderes. Kompatibilität bedeutet ferner auch ein
Zusammenpassen von Objekten unterschiedlicher Funktionalitäten oder
Aufgaben, z.B. zwischen Schiff und Hafen oder zwischen Schiff und Char-
terrate. Es zeigt sich, dass die Feststellung von Kompatibilität einen Akt
der Problemlösung darstellt.
iii
Human experts are not systems of rules, they are libraries of experiences
1
[66]
v
vi
Contents
I Observations 1
1 Introduction 3
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.1 Few scenes from shipping practice . . . . . . . . . . . 3
1.1.2 State of the art of shipping research . . . . . . . . . . 6
1.2 Hypothesis and problem formulation . . . . . . . . . . . . . . 8
1.2.1 Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2.2 Problem formulation . . . . . . . . . . . . . . . . . . . 9
1.3 Scope and structure . . . . . . . . . . . . . . . . . . . . . . . 9
1.3.1 Scope of work . . . . . . . . . . . . . . . . . . . . . . 9
1.3.2 Report structure . . . . . . . . . . . . . . . . . . . . . 10
vii
Contents viii
6 Conclusion 189
B Glossary 205
D Lebenslauf 225
Contents x
List of Figures
xi
List of Figures xii
xv
List of Tables xvi
Observations
1
Chapter 1
Introduction
1.1 Motivation
1.1.1 Few scenes from shipping practice
Chartering is the heart of bulk shipping. Ships are chartered-in and chartered-
out everyday. Bulk shipping plays an important role in the world economy,
as bulk carriers transport most of cargoes world wide. It is a very economi-
cal transport mode.
Intensive exchange of information marks the activity of chartering and
operations departments everyday. A ship practitioner, either in chartering or
operations department, communicates endlessly with brokers, cargo own-
ers, agents and partner shipping companies, through phone conversations,
faxes, e-mails and postal letters. Messages containing cargo owners seek-
ing suitable tonnage or ship owners seeking for employment for their ves-
sels is received and transmitted daily. In many cases such messages are
sent by brokers.
Besides, information from newspaper or television concerning the oil
not suitable
ships
suitable
ships
3
1.1. Motivation 4
price, world economy, events, usually political ones, which may affect the
world trade, fuel price is followed closely. Most of messages received do
not contain something useful. One needs to filter out useless messages
from unreliable sources or messages containing irrelevant topics. The most
useful and compact information is still scarce and expensive.
This section portrays three scenes describing the tasks in ship charter-
ing and operations. Firstly, consider the following conversation in a char-
tering department:
From the pile of faxes and e-mails received, a chartering staff will find nu-
merous ships available for charter. The message is usually brief containing
a summary of the particulars of the ship and its position when available, as
shown in Table 1.1.
From the ships offered to him, he will make a selection. Few can be
viewed as suitable ships and the remaining are not suitable ones, as shown
in Figure 1.1. What has he been doing actually? Can one categorize those
ships simply into suitable and not suitable ships? Can one do that objec-
tively? On which considerations does he base his selection? How well can
current methods address the above situation?
In the following steps, as the negotiation has been started, he has ex-
changed some information with the broker. More details become available
for both parties, i.e. charterers’ and owners’ wishes, expectations or prefer-
ences2 . The size of information available to both parties grows. It becomes
1
a synonym to bulk carrier, a ship carrying dry bulk/ unpacked cargoes
2
The plural forms of ’owner’ and ’charterer’ are often used in communications, i.e. own-
ers/ ship owners and charterers, though those usually denote a single organisation (one
ship owner or one charterer)
5 Chapter 1. Introduction
more suitable
ships
Figure 1.2: More suitable, less suitable and not suitable ships
evident that he needs to narrow his selections. Among the suitable ships,
one can divide into two finer groups, more suitable or less suitable ships,
see Figure 1.2. The process goes on until he finds the most suitable one.
The above tasks teach us the idea of grouping as a part of decision
making task. It is a task of grouping of ships into suitable and not suitable
ships, or in a finer fashion. The concept ’suitable’ reflects the degree of
preference of the chartering practitioner. The immediate striking question
lies before us. The concept ’compatible’ will be used to embrace concepts
such as ’suitable’ 3 .
Consider now the second situation. A ship practitioner is seeking a ship.
His ship broker offers him a ship available for employment, with following
brief particulars:
MV FOREST CHAMPION
LGR, PAN, 1993
26472 DWT ON 9.54 M
G 33917 CBM
14/22 + 2 MDO
ALL ABT
OPEN GLASGOW
hire of about $ 6,000 / day would be appropriate for that ship. In other
words, a charter hire of approximately $ 6,000 is compatible to such a ship.
The third situation. It may appear that a ship, say MV F OREST C HAM -
PION , serving a route regularly has to be docked soon. Another vessel
must replace her. The task is to find the most suitable vessel to replace
her. Ships having similar characteristics to her are strong candidates. But
this criterion is far from adequate. The most compatible ship must comply
with an array of criteria such as having a reasonable charter hire and her
current location must be in the vicinity of MV F OREST C HAMPION ’ S current
sailing area.
demand of ships, exchange rates, fuel price till macro economy [46]. This is
an initial sign that shipping is a complex system. Efforts have been under-
taken to address valuation problems in shipping. A valuation task focusing
on the future is called forecasting. Beenstock, Vergottis, Wergeland and
Tang, to name a few, developed forecasting models, which might help us to
predict the future charter rate movements of a market segment [8, 23, 75].
This market segment means a set of ships of a range of sizes, operating in
few standard main routes. Those forecasts are of use for macro-economic
analysis purposes, but as we will find later in Chapter 2, those forecasts
are of a little use in shipping practice.
Resources such as shipping indexes are published regularly. Indexing
is in fact an attempt to describe the whole shipping market using few main
shipping trades. Those may give an adequate overview of the whole ship-
ping market, but a chartering staff cannot rely very much on those indexes.
Most of routes or trades are very specific, whilst those indexes are too gen-
eral to be practicable. Therefore the shipping indexes are not much of use
in conducting daily chartering negotiations.
Intuition plays an important role in shipping practice. Various tasks con-
cerning the value assessment and decision making in operations, charter-
ing and sales and purchase involve intuitive judgments. Datz adds [16]:
The role of intuition and experience has been widely acknowledged [46,
63]. But efforts to understand and to make use of it seem to be non ex-
istent. Shipping research seems to have been marked with its preference
for quantitative methods, statistics and operation research [23, 79, 75]. Ad-
land and Pace’s efforts to address the valuation tasks in shipping illustrate
this view very clearly [60, 5]. Terms such as intuition and experience are
vague and subjective. Those seem to be less accessible to many. Many
researchers do not consider those aspects seriously.
Decision making is a process of determining an action. In real world, the
situations are far less than ideal. There are many things which are unknown
or partially known. The probabilities of certain events are not always avail-
able. Therefore, Simon argues in his Theory of Bounded Rationality, that
it is hardly possible to make decisions rationally [72, 71]. Goncalves’ ar-
gument that shipowners do not make decisions rationally confirms Simon’s
view [27]. In spite of his advocation for using mathematical approaches,
Adland admits that the results can not yet challenge the existing intuitive
valuation method in shipping [5].
1.2. Hypothesis and problem formulation 8
Findings from
shipping practice
Understanding
Proposed
the concept of
implementations
State of the art compatibility
of shipping research
2.1.1 Definition
From the size of shipment, the carriage of cargoes can be divided into two
main categories, partial or full shipments. In partial shipments, a ship car-
ries a number of cargo units, or cargoes belong to more than one shippers.
Such a cargo is called general cargo or break bulk cargo. In the second
category, cargoes are shipped as a homogeneous mass, usually without
packing and in large quantities or bulk. Its shipment is also known as a
“one ship, one cargo” shipment. This cargo is called bulk cargo. This cargo
can either be liquid, such as crude oil, product oil, liquefied gas or dry, such
as coal and iron ore. An additional class of cargoes, called neo-bulk, are
a combination of both definitions mentioned earlier. Neo bulk cargoes are
physically break bulk cargoes usually packaged; but those are shipped in
large quantities, such as coils or bundled wood products1 . Unless stated
otherwise, the term ’bulk’ in this thesis refers to the dry bulk.
The physical manifestation of the cargo determines the type of ship
used to carry it. Tanker and bulk carrier are ships used to carry liquid and
dry bulk cargoes respectively. A multipurpose dry cargo vessel is used to
transport break bulk cargoes. Other types of ships denoting the combi-
nation of types, such as Ore Bulk Oil (OBO) carrier, or emphasizing spe-
cific feature of the ship, such as self-unloader (for a bulk carrier having
own discharging gear on board) are also used widely in practice. MPTW
stands for a multipurpose tween deck ship, OHBS stands for open hatch
bulk ship). Unfortunately the categorization of cargoes and types of ships
used in statistics and in practice involves some inconsistency [47].
1
For a more details, see [47, 73, 80].
13
2.1. Dry bulk shipping 14
Table 2.3: Major export and import areas of dry bulk commodities [83]
2.1. Dry bulk shipping 16
F IELDS
Week ending
Port of delivery
Port of redelivery
Name of vessel
Year of built
Deadweight
Speed
Fuel consumption
Date of laycan
Terms
Charter rate
Charterer
Table 2.4: Database fields of Time Charter Fixtures 1997-1999 [50, 51,
52]
Number of
ports of call
Deadweight
Bulk carrier is a ship used to carry dry bulk cargoes. Figure 2.1 shows
MV L UISE O LDENDORFF , a Panamax class bulk carrier. She was built at
Samsung Heavy Industries, South Korea, delivered in 1994. She is gear-
less, has a deadweight of 72,873 tdw and seven cargo holds with average
volume of about 11,500 cbm. The vessel’s construction is strengthened and
capable to carry heavy cargoes. Her overall length is 224.95 m, breadth
32.24 m.
MV C AROLINE O LDENDORFF , a handy size bulk carrier of 22,159 is a
tdw, see Figure 2.2. She is log-fitted, built in 1993 at Onomichi Shipyard,
Japan. She is equipped with four cranes of each 30 t lifting capacity. Fur-
ther principal dimensions: overall length is 157.50 m, breadth 25.00 m.
A ship has basically three main functionalities namely: (a) holding car-
goes (b) moving cargoes from port to port (c) transferring from ship to
shore and vice versa (optional). Deriving from the above, the following de-
sign aspects of utmost important: (a) size of cargo spaces (b) power of
main engine, speed of ship and its fuel consumptions (c) cargo gear, such
as a crane or self-unloading belt conveyor.
Speeds of Handymax bulk carriers vary between 10 and 19 knots. The
speed of Handymax bulk carriers vary between 10 to 19 knots. Few op-
erators prefer to have speedier ships in return for more potential carrying
capacity (in ton miles annually).
According to its size, bulk carrier sector can be divided into three classes:
Handy size, Panamax and Cape size. Handy size bulk carriers have a size
of 10,000-50,000 dwt. Panamax bulk carriers have the maximum size for
passing the Panama Canal, which is restricted to a width of 32.25 m. Their
typical deadweight is approximately 60-80,000 dwt. Cape size is the largest
2.2. Chartering and Sale & Purchase 20
25000
Time Charter Hire (US$ /day)
20000
15000
10000
5000
0
01/97 04/97 07/97 10/97 01/98 04/98 07/98 10/98 01/99 04/99 07/99 10/99
Date
ship’s class. All ships beyond the Panamax class are categorized into this
class, their deadweight is usually over 100,000 dwt.
Figure 2.4 shows the development of time charter fixtures 1997-1999. This
section highlights the time charter fixtures for all bulk shipping segments,
Handy size, Panamax and Cape size. From the chartering and opera-
tions viewpoints, the way the chartering negotiations, the operations are
managed, there is basically no difference between all the above-mentioned
segments. A separate detailed analysis for each segment is less relevant
and goes beyond the scope of the research. Figures 2.5 and 2.6 show the
sizes (deadweight) and speeds of vessels deployed in charter market.
Ships are employed until approximately they reach the age of 27 years,
see Figure 2.7. Figure 2.8 shows the relation between the age of the ves-
sels and their charter hire. Whilst the relations between the fuel consump-
tions of the main engine and time charter rate and ship’s deadweight are
shown in Figures 2.9 and 2.10.
21 Chapter 2. Nature of Dry Bulk Shipping
200000
Deadweight (ton)
150000
100000
50000
0
01/97 01/98 01/99 01/00
Date
18
17
16
Speed (knots)
15
14
13
12
11
10
01/97 01/98 01/99 01/00
Date
35
30
25
Age (years)
20
15
10
0
01/97 01/98 01/99 01/00
Date
25000
Time Charter Hire (US$/day)
20000
15000
10000
5000
0
0 5 10 15 20 25 30 35 40
Age (years)
Figure 2.8: Age versus Time Charter rate [50, 51, 52]
23 Chapter 2. Nature of Dry Bulk Shipping
20000
Time Charter Hire (US$/day)
15000
10000
5000
0
10 20 30 40 50 60 70 80
Fuel consumption (ton/day)
Figure 2.9: Main engine fuel consumption versus T/C rate [50, 51, 52]
70
60
Fuel consumption (ton/day)
50
40
30
20
10
0
0 50000 100000 150000 200000 250000
Deadweight (ton)
35
30
25
Age (years)
20
15
10
0
0 50000 100000 150000 200000 250000
Deadweight (ton)
25000
Time Charter Hire (US$/day)
20000
15000
10000
5000
0
0 50000 100000 150000 200000 250000
Deadweight (ton)
18
16
Speed (knots)
14
12
10
8
0 50000 100000 150000 200000 250000
Deadweight (ton)
260
250
Time Charter Hire (US$ /day)
240
230
220
210
200
190
01/97 04/97 07/97 10/97 01/98 04/98 07/98 10/98 01/99 04/99 07/99 10/99
Date
2.2.1 Chartering
Chartering is a basic form in shipping. A ship is hired from a shipowner
to a charterer either directly or through one or more brokers. A charter
party is an agreement which states the details of the terms and conditions
concerning the operations of the ship.
Two types of Charter Parties dominate the chartering market4 : (a) Time
Charter consisting of Trip Time and Period Time Charters and (b) Voyage
Charter. In a Time Charter Party, a vessel is hired for a specified period of
time for payment of a daily, monthly or annual fee. In its first variant, Trip
Time Charter Party, the vessel is chartered on the basis of a time charter
period of a specified voyage and for the carriage of a specified cargo. The
shipowner earns a lump-sum per day for the period determined by the voy-
age. Whilst in its second variant, Period Time Charter Party, the ship earns
a daily hire, paid monthly or semi monthly in advance. The shipowner re-
tains possession and mans and operates the ship under instructions of the
charterer who pays voyage costs, including bunkers.
In the Dry Voyage Charter Party, the ship earns freight per ton of cargo
transported on terms, set out in the charter party, which specifies the pre-
cise nature and volume of cargo, the ports of loading and discharge and
the lay-time and demurrage. All costs are paid by owners.
In order to conduct chartering, a party which helps smoothen the ne-
gotiation between shipowners and charterers is called a broker. There are
three types of broker, (a) owners’ broker (b) cargo broker (c) competitive
broker [81]. An owners’ broker acts in the best interest of the shipowners
and is often a part of a shipping company. A cargo broker acts in the direct
interest of the owner of the cargo ad is often a part of its organization. A
competitive broker has no links with either the owner of the ship nor the
owner of the cargo. He acts as a pure intermediary with the only objec-
tive of concluding a deal, a fixture and earns his 1.25% commission of the
freight revenues.
Shipping costs
There are four cost categories distinguished in the running of ships. These
are [73, 81]:
1. Capital costs, which cover interest and capital repayments and are
determined by the way in which the ship has been financed.
2. Operating costs, which constitute the expenses involved in the day-
to-day running of the ship such as including manning costs, stores
and maintenance.
4
Other types of charter parties, such as Bareboat Charter Party, are not elaborated in
this research, as this type of Charter Party does not occur frequently.
27 Chapter 2. Nature of Dry Bulk Shipping
3. Voyage costs, which comprise the variable costs associated with the
actual sailing of the ship, such as bunker, port charges and canal
dues.
4. Cargo handling costs, which are the costs for stowing, loading and
discharging of the cargo.
The first two costs, namely capital and operating costs, do not depend
on whether a sea voyage is made or not (i.e. voyage independent costs).
In contrast to the former, voyage and cargo handling costs are voyage de-
pendent costs, i.e. the costs arise only when a voyage is made.
Handy size bulk carriers contain numerous and non-standardized niches.
Market reports give an overview on the tendencies in the charter market.
Relevant events, such as oil prices, or some political events somewhere
are well considered. When it comes to a real negotiations, the market re-
ports are far from adequate. An aspect which plays an important role in the
assessments of shipping players (consultants, brokers and charterers) is
that the assessment is heavily affected by qualitative factors.
Shipment values
Back-haulage cargoes
When fixing a contract, a vessel can be delivered in any place in the world.
It is not unusual that a vessel has to perform a carriage to remote places,
meaning far away from cargo centers, i.e. ports or areas from which car-
goes are easily to be obtained. Therefore anticipating the period after the
redelivery of the vessel is important. At which level, a shipowner will be
willing to charter out his ship, depends upon the market situation.
2.2. Chartering and Sale & Purchase 28
This shows that a ship’s geographical position determines its charter hire
value, besides the aspect of time. More precisely, the distance to the near-
est cargo center could be decisive in determining the charter rate of a ves-
sel.
are without ballast bonus. In Madagascar area, they believe, they would
face difficulty to find back-haulage cargoes. Therefore they would charge
a higher rate for redelivery in such an area. Though MV Z I Y UA S HAN is
larger than MV S EA B AISEN, the time charter rate of MV Z I Y UN S HAN
is lower than the later due to its age. Qualitative aspects such as their
perception on flags and reputation of the shipowners are not considered.
The Atlantic desk was represented by Sebastian Dohrendorff who has
been working at EO for six months at the time when the survey was con-
ducted. Two situations are considered: (a) Delivery: Hamburg, redelivery:
Le Havre via: Norfolk with a voyage length of about 45 days (b) Delivery:
Hamburg, redelivery Norfolk with a voyage length of about 25 days. A gen-
eral remark he made is that redelivery of the previous voyage in Japan
would make an unrealistically high charter hire or ballast bonus if the ves-
sel is to be put in service in Atlantic. Assumption is made that those ships
are to be redelivered in the vicinity of Hamburg in the previous voyage. His
assessment is shown in Table 2.9.
Dohrendorff suggests that the ones of the Concept Carriers (CC)11 must
know better the estimated values of all unanswered spaces. All have filled
out the questionnaire quickly, they do not need to use any calculator. It
shows that the capability of estimating charter values involves reoccur-
rence. The above valuation assessment task does not involve intensive
calculation processes. The one who has never been in touch with a simi-
lar case, would have a difficulty to assess the charter value. Recognizing
similar situations apparently have contributed very much to conduct an as-
sessment task intuitively and quickly.
Figures 2.15, 2.16 and 2.17 show the charter transactions of the Con-
cept Carriers (CC) in 2000. The vessels are time chartered-in vessels by
the Concept Carriers and chartered out, in this case, as period time charter
11
a subsidiary of Egon Oldendorff, see 2.5.
31 Chapter 2. Nature of Dry Bulk Shipping
0.35
Unit time charter hire ($/dwt_day)
0.3
0.25
0.2
0.15
0.1
0.05
20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
The market values of the ship are estimated to decrease $ 500,000,- till
$ 1,100,000 annually, see Figure 2.21. Some still believe that the Greeks
are well-known for their outstanding knowledge on the movements of ships’
prices. The shipping community keeps watching closely when the Greeks
33 Chapter 2. Nature of Dry Bulk Shipping
0.35
Unit time charter hire ($/dwt_day)
0.3
0.25
0.2
0.15
0.1
0.05
0 5 10 15 20 25 30 35 40
Period (months)
0.35
0.3
Unit time charter hire ($/dwt_day)
0.25
0.2
0.15
0.1
0.05
-0.05
1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998
Year of built
EU vessels
0.5
0.4
Unit income ($/dwt_day)
0.3
0.2
0.1
0.3
0.25
Unit operating cost ($/dwt_day)
0.2
0.15
0.1
0.05
0
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
1000
900
800
Unit market value($/dwt)
700
600
500
400
300
200
100
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
2500
Market value depreciation (x$1000/year)
2000
1500
1000
500
0
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
It implies that more details are needed to obtain more accurate assess-
ment16 .
Trade-off in freight rate might take place. For example in grain seasons
in South America, EO is willing to discount freight rate to South America,
for the sake of positioning the vessel in that region17 .
There are a lot of market niches in the handy bulk shipping. There
are many ships with specific properties. It covers properties such as the
suitability for log carriage or for container carriage, gear type and capacity.
It is difficult to estimate or assess charter rates reasonably. 18
Pistorius believes that Panamax on the other hand shows a cyclical pat-
tern. It has never reached a $12,000 level19 . An important main question
concerning running those vessels is how to route this type of vessel. Two
12
Peer Gröpper, interview, 22 May 2003
13
For example: www.bnp-paris.com and http://www.shipvalue.net/default.asp, down-
loaded 29 September 2003
14
Fax from Allship to EO, 2 June 2000
15
This implies, that if a site inspection were made, the values stated in the sheet might
differ.
16
From methodological point of view, see also Chapter 4, sub-section 4.4.1 and Table 4.5
17
Patrick Hutchins, interview
18
Mark Pistorius, interview on 4 July 2000
19
Compare with Wijnolst’s findings in [79]
37 Chapter 2. Nature of Dry Bulk Shipping
8
Estimated market value ($ million)
2
15000 20000 25000 30000 35000 40000 45000 50000 55000 60000 65000 70000
Deadweight (ton)
8
Estimated market value ($ million)
2
2 4 6 8 10 12 14 16 18
Age (years)
alternatives are available either to stay at a port waiting for the right time
for fixing cargo at a reasonable rate or to position to more promising areas
20 .
2.3 Operations
2.3.1 Running
The chartering department represents the commercial face of the company.
The initiative and the main role in negotiation for ship’s employment are per-
formed by this department, see Figure 2.24. The duo between chartering
and operations department seems indispensable in running the shipping
20
Mark Pistorius, interview, 4 July 2000
21
Very Large Crude Carrier
39 Chapter 2. Nature of Dry Bulk Shipping
Performance at port
In handy size sector, charterers play a very important role for achieving a
reasonable ship performance at ports of call. An owner’s protecting agent
is usually appointed to deal with charterer’s agent in order to achieve the
above goal. Saving is not that much at port and agency expenses, but
more on time-related costs, namely fixed costs (capital costs, fixed operat-
ing costs) 30 .
Time is an important factor which determines the profitability of a ship
operation. It is not easy to estimate port time. Port congestion, cargo han-
27
Mark Pistorius, interview, 4 July 2000; Gunnar Eisenecker & Jan-Henning Asmussen,
interview
28
Mark Pistorius, interview, 4 July 2000.
29
Mark Pistorius, interview, 4 July 2000.
30
Mark Pistorius, interview, 4 July 2000
2.3. Operations 42
dling rate, weather, labor, to name a few, play an important role. Published
information can be helpful but it is far from adequate. Shipowners rely very
much on information provided by their local agent, since they have local
knowledge of the conditions 31 .
Tariffs at few ports are much more expensive. At these ports the ship
operators watch and guide the operations of their ship very carefully, since
they will pay in cash for any delay. Meanwhile at other cheaper ports,
operation staffs are ’allowed’ to be less careful. Most costs are actually
caused by the time-related fixed costs. Saving on unnecessary extra due
at expensive ports can be substantial:
Every country and every port has its own typicality. A shipment to a partic-
ular port may experience things which make ship operators, shippers and
consignees care more about their operations in future.
There was an issue in the year 2000, concerning the Asian gypsy moth.
Gypsy moth caterpillars decimate forests. The caterpillars feed on the
leaves of trees and shrubs. Masters/ owners of vessels which have called
at Russian far-east ports, China (north of latitude 30 degrees) Korea and
Japan (Hokkaido) between September and October should arrange for a
thorough inspection of their vessel to ensure that there are no egg masses
of larvae 34 . In relation to an on-going chartering negotiation a confirmation
of applying the above the so-called Gypsy Moth Clause can be necessary,
as illustrated by the following:
OWNS CONFIRM VSL HAS NOT CALLED CIS PACIFIC PORT(S)
BETWEEN 40 AND 60 DEGREES NORTH
BETWEEN JULY AND SEPTEMBER IN LAST TWO YEARS
GYPSY MOTH CLSE APPLY
35
Operation evaluation
A brief evaluation of 22 voyages, the estimated values when fixing the char-
ter parties and the actual values after the operations are entirely completed
are compared. The focus is paid to the predictability in terms of time (port
and sailing time) or money (income and result). Those are the main con-
cerns for all parties (chartering, operations and technical departments) in-
volved in a charter party agreed.
Port days are in general not easy to predict, in contrast to it, sailing
days are in general more predictable. Figure 2.25. Panamax vessels tend
to have longer voyages than the smaller Handy size bulk carriers. The
smaller vessels have more ports of choice to call, see Figure 2.26. Handy
size bulk carriers call more various types and sizes of ports. Their perfor-
mances concerning cargo handling speed, claims, waiting time and tariff
vary considerably. The types of cargoes are less uniform, involvement of
’new’ ports agents appears to be more frequent than in the case of Pana-
max vessels. It mounts in the form of low predictability of port time. Handy
size vessels’ port expenses are therefore less predictable than those of the
Panamax vessels, see Figure 2.27.
The difficulty to estimate the length of voyage is evident. Years of built of
vessels do not affect very much on the sailing time predictability very much,
see Figure 2.28. But the Handy size vessels’ sailing time predictability is
better than that of Panamax vessels, Figure 2.29. There is no guarantee
that younger vessels produce a higher result (gross profit per voyage), Fig-
ure 2.31. Voyage lengths of Panamax vessels are frequently longer than
estimated. The higher contribution of ballast voyages plays an important
role here, since the duration of ballast voyage is difficult to predict, see
Figure 2.34.
Figures 2.32 and 2.33 show that the actual result is less predictable
than the time charter equivalent. These confirm the low predictability of the
voyage duration and the need for having a capability to better estimate the
performance of ship at port, see also sub-section 2.3.3.
37
Joern Jentzsch, e-mail dated 4 July 2000 and Loss Prevention Bulletin, 146 - 06/00. A
member means a shipping company which hold a membership of the association issuing
the above-mentioned bulletin.
45 Chapter 2. Nature of Dry Bulk Shipping
Sailing days
Port days
2.5
Actual/estimate of sailing and port days
1.5
0.5
0
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
80
70
60
Actual length of voyage
50
40
30
20
10
0
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
1.5
Actual/estimate of port expenses
0.5
0
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
2
Actual/estimated length of voyage
1.5
0.5
0
1975 1980 1985 1990 1995 2000
Years of built
2
Actual/estimated length of voyage
1.5
0.5
0
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
200000
Actual result ($/voy)
-200000
-400000
-600000
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight
200000
Actual result ($/voy)
-200000
-400000
-600000
1975 1980 1985 1990 1995 2000
Years of built
5
Actual/estimation
-5
-10
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
5
Actual/estimation
-5
-10
1975 1980 1985 1990 1995 2000
Years of built
2.5
Actual/estimate of ballast days
1.5
0.5
0
10000 20000 30000 40000 50000 60000 70000 80000
Deadweight (ton)
To ensure that all parties, at least starting internally, appointed agents and
later the broker or partner charterer/shipowner, are performing the agree-
ment accordingly one may need to influence and reeducate others. Since
this could be a very important factors. Few terms may differently under-
stood and interpreted, with financial consequences. As an operation direc-
tor, Mark Pistorius, may need to interfere in very fundamental operational
and legal aspects in shipping.
The number of vessels handled, the number of parties and persons in-
volved increase the volume of information handled dramatically. In any
38
Ian Barclay, Posting to ’Cargo Care’, 20 June 2000.
39
E-Mail form Mark Pistorius to Department of Chartering, on 29 June 2000
40
E-Mail form Mark Pistorius to Department of Chartering, on 16 June 2000
51 Chapter 2. Nature of Dry Bulk Shipping
stage of chartering negotiations and ship operations, the accuracy and the
completeness of information is an important aspect. Few past information
for a particular port can be used safely, but for some other ports are not
the case. The need for information on the actual performance of a ves-
sel, cargo handling speed in relation to the cargoes handled, the conges-
tion situation and actual port-related costs is huge. An intensive exchange
of information between the chartering and operations departments are in-
evitable. A documentation on the past experience employing a vessel for a
particular trade is extremely useful, as illustrated by the following:
Security
This aspects can be a grey area. How secure is secure? How far is a
company willing to let their ship trade in a particular area, which is said to
be insecure. Which information sources should we trust?
Usually a very close cross-check actions and close co-operations with
the other departments and the company agents are arranged. They do not
solely depend on the official report either from local governments or from
international organizations. Intensive communications are made, and fre-
quently by phone42 . From various sources, one may build a better assess-
ment and his conviction on the real situation in the trading area in concern.
It is likely that all these information, including hearsay and gossips are as
important as the scarce and not timely available official information43 . Usu-
ally the company has a general policy regarding this44
41
E-Mail form Mark Pistorius to Department of Chartering, on 18 May 2000
42
“We have to be cautious if we enter new areas. We rely fully upon the information of
our agents. We put additionally few days for safety reasons. The more insecure there area
is, the higher this safety margin would be”, Patrick Hutchins, interview, 9 May 2000.
43
More reports and data, within reason, cannot hurt, but much more important are real-
life stories meaningful to the manager. Much of an executive’s daily time, according to
Henry Mintzberg’s careful observations of managerial behavior, is spent seeking just that.
Mintzberg observes that businessmen prefer concrete information, even gossip, specu-
lation, and hearsay, to the abstracted summary information contained in routine reports
flooding their offices. Page 164 [18]
44
Security is important but generally the troubled regions are the ones that we do not
regularly trade anyway so this does not come up often. Viveka Mansukhani, E-Mail and
interview, 31 May 2000
2.4. Multifaceted aspects of shipping 52
premium
Price supplier
Price
standard
supplier
commodity
supplier
Quantity
Quantity
closer than in manufacturing sector, see Figure 2.36. This properties ap-
ply to shipping as well. Maritime business is largely based on personal
relationships, Panayides notes [62].
The quality is difficult to asses objectively, moreover in this service sec-
tor. What practically exists is the perceived quality. It is usual that a com-
pany consciously maintains their good name. It is sometimes costly, for
example, preferring not to carry low-paying cargoes over a temporary bal-
last voyage or idle time. In shipping index, such a segmentation is not
clearly visible.
Such a perception on other companies’ reputation is a company-wide
knowledge. Except in the case of black-listed companies, see Chapter 2,
it is hardly possible to formalize it, for example, in numerical scores or a
black-and-white policy. Within this ’allowed area’ there is still a huge gray
area which is subject to subjective interpretation.
The voyage charter hire, time charter hire and selling price of a ship de-
pend upon not only the physical aspects of the ship itself, but other factors
may play a more important role, see also Chapter 3. Those decisive factors
are the reputation of the ship manager, the potential of obtaining cargoes
in the voyages to follow, the assessment of consequences of international
political events. These are not objectively verifiable.
Few success stories are marked with the success of shipping compa-
nies of obtaining contracts of affreightment in difficult time or fixing a longer
time charter party just short before the rates are tumbling down. The capa-
bility or, some may entitle it a luck, for achieving an astonishing return, is
2.4. Multifaceted aspects of shipping 54
evident.
and that suit the ships I am working specifically or that are the
size/area that we usually work. Generally the most important
e-mails come to my personal box as then I see it immediately,
can respond immediately etc and the brokers/ clients I work with
closely, always send their messages to my personal box.
Reliability of a broker is important. Charterers that we have
worked with before, who have a track record and we have a
charter party with them and we know them well, always get pri-
ority. 47
Such qualitative aspects are definitely not to be overlooked in real-world
shipping practice 48 .
2.4.3 Strategy
In general the Sale and Purchase Department of Egon Oldendorff49 acts
very actively in the sale and purchase market. It follows the rule:
If a vessel is attractive and its price is low then buy it. Then
charter it out.50
They bought two Ro-ro vessels recently, and both are immediately chartered-
out. The expertise of the organization does not make it possible to operate
both vessels by themselves. The aspect of skill at the organization level is
considered seriously. End of 1999, four vessels of Hoegh Lines (Norway)
were bought by Egon Oldendorff. Those are multipurpose ships serving
worldwide, Asia-West Europe-North America. Hoegh Lines and Egon Old-
endorff agree to transfer the operations later in 2000, in order to enable
learning from Hoegh Lines’ current operations. Since 2000 Egon Olden-
dorff operates the vessels after setting up offices in several countries. Its
service is called Indotrans Service. In 2003 the vessels were sold to the
China Navigation Company Limited. Another rule is:
If the company badly needs a vessel but the price is high, the
company prefers waiting. 51
47
Viveka Mansukhani, E-Mail and interview, 31 May 2000
48
“I would be working with brokers, whom I know very well or they work exclusively with
certain charterer”, Sebastian Dohrendorff, interview, May 2000
49
A shipping company located in Lübeck, see later in sub-section 2.5.
50
Peer Gröpper, interview, 22 May 2000. Sale and purchase of ships is basically carried
out solely by Peer Gröpper. Since it involves a huge amount of money and longterm strategy
of the company, he consults closely with Henning Oldendorff, Chairman. And the final
decision lies at the later as well.
51
Peer Gröpper, interview, 22 May 2000. Sale and purchase of ships is basically carried
out solely by Peer Gröpper. Since it involves a huge amount of money and longterm strategy
of the company, he consults closely with Henning Oldendorff, Chairman. And the final
decision lies at the later as well.
2.4. Multifaceted aspects of shipping 56
.. and with bunker price still rising the interest in time charter
rather than voyage charter seems apparent 56 .
On center of cargoes:
In the Med there seems a small activity at the moment for stan-
dard types, and again Charterers are trying to push rates lower.
Some cement and mineral stems were open for standard Handy
types to both Continent and US Gulf destinations.
“If the fuel price is high, then the voyage charter rate is also
high. Time Charter becomes preferable”.
“If the rate is low and the technical department approves it,
then low paying cargoes, such cement clinkers or scraps, would
likely be accepted”.
“A ship laid up for a long time tends to have a lower value”.
“Our company’s fixtures tend to be beyond the market level”.
There must e a long list of rules, if one wants to express them all. It be-
comes hardly possible to make it practicable. One cannot perform his job
well using the above rules only. There is still some vague area which is
difficult to express without ambiguity or which needs to be understood in
context, for example:
Furthermore the knowledge is changing over time. The views on ideal ships
may change over time, for example:
57
Mark Pistorius, e-mail of 14 August 2000
59
Education
Reports say that the Atlantic rates today are firm. It is not en-
tirely correct. To be more specific, there are no cargoes coming
from East Coast of both North and South America. It is in fact
difficult to rely fully upon market reports, since those are not
always relevant for daily negotiations64 .
Freight rates continue at a slow rise but the modern Eco types
do attract good numbers. There is a good activity activity with a
steady flow of new orders. We feel that the expected improve-
ment on the freight level has still not reached its anticipated
level. 65
Large and small Handy sizes are there. No clear and sharp distinction
between those groups is necessary to have the texts easily understood, by
the readers.
J.E. Hyde Freight Market Comment (6 June 2000):
What was clear was that the lack of fixing being done is not
because there is a lack of business around but due to the dis-
crepancy between Owners’ and Charterers’ respective ideas in
freight values.
The transportation sector, which shipping belongs to, connects ports, ar-
eas and countries all over the world. All these possess its own situations
and typicalities, in terms of macro economic, politics and cultural typicali-
ties. The shipping faces these situations and react upon them immediately.
Shipping companies have the necessity to have a close contact with their
partner shipping companies, brokers, shippers, consignees and govern-
ment officials. Their interpretations are important, since these may play a
very important role in determining the selling price of their product, charter
hire, see also ??. The following is noted by Maersk Broker:
O LDENDORFF C ARRIERS
Self-unloading bulk trade
Panamax bulk trade
Handy size bulk trade
CONCEPT steel trades (Handymax parcelling Black
Sea/Asia, Asia/Med)
Breakbulk service South East Asia/Europe and North
America
AUSTRANS bulk parcel service
Multi purpose vessels
Open Hatch Box Shaped Vessels
Bulk Logistics (Transhipment, Lightering, Topoff, Self-
unloading barge)
Sale and Purchase, Newbuilding brokerage
Logistics Projects
they operate some 200 vessels, 80 are owned, and 120 are chartered-in
vessels. Its fleet performs 8,000 port calls in 120 countries, and carries
some 80 million tons of bulk and unified cargoes across the ocean77 .
The above company description does not suggest that a company of
a completely different organisation style would not work well. The above
is not the recipe to ideally conduct shipping business either. The above
description emphasized the importance of knowledge, the formal and infor-
mal one, in an organisation. The formal knowledge is well-documented and
clear, can be obtained from books and, usually from, formal education. The
informal knowledge comes from own or someone else’s experience. The
usage of informal knowledge marks the role of intuitive actions and contex-
tual understanding becomes evident, especially when the environment is
complex, see also Chapter 4.
3. Shipping practice does not use any computer programs for assessing
the charter value or selling price of ships assessment as a foundation
for their decision. They rely very much upon experience and intuition.
This chapter elaborates the views and perspectives of the past research on
shipping problems, and their ways addressing problems. A special atten-
tion is paid to the tasks which support the decision-making processes. In
order to be able to make a decision satisfactorily, a decision makers relies
on two things, namely resources and tools. Resources include information
on the state of the shipping, world economy, forecasting, S+P and charter
reports. Tools embrace methods or programs used to assist decision mak-
ing or to generate information such as forecasting models or ship’s value
assessment methods.
69
3.1. Chartering negotiations 70
Swazi Sugar Association let Galbraith London find a ship to carry sugar
from Maputo to London. Galbraith informed Mansukhani, that they needed
a vessel for the above-mentioned purpose.
31 May 2000. She offers US$ 31.50 per ton, and adds extra clauses
concerning demurrage, US$ 300 and US$ 5,000 per day at loading and
discharging ports respectively. She responds:
3.1. Chartering negotiations 72
MV W ORLDSTAR
On 2 May 2000 Ernst Russ, a broker, offered MV W ORD S TAR to Egon Old-
endorff. Mansukhani was interested to charter-in this vessel. The ship was
3.1. Chartering negotiations 74
2 May 2000 16:45. Ernst Russ adds some additional information re-
quired by Mansukhani concerning the load capacities on tanktops, weather
decks, hatch covers and tween deck. This information can be of use, in
case Mansukhani intends to use the vessel for transporting heavy cargoes.
8 May 2000 13:30. Ernst Russ informs that the shipowners would ex-
pect a charter hire of US$ 4,800 per day. They gives a hint to Mansukhani,
if she is really serious about the vessel, she would very likely get the vessel
at US$ 4,700 daily 1 . Mansukhani comments on the fax sheet, that she
expects a charter hire of US$ 4,650 per day.
1
This method is also used to show the clients, how the broker works hard for them.
75 Chapter 3. Existing Resources and Tools
8 May 2000 14:28. Ernst Russ informs Mansukhani that the owners
insists a charter hire of US$ 4,800 daily. More details pertaining a.o. al-
lowed trading area are provided. A standard Charter Party form is pro-
posed by the owners, namely the New York Produce Exchange (NYPE)
Charter Party form. This phase is still not binding yet; a clause states that
all agreement is subject to approval of the board of directors.
MV WORLD STAR/ ACCT EO
DELIVERY AFPS ST PETERSBURG ATDNSHINC
LAYCAN 12/15 MAY 2000 0000-2400 HRS
ONE TCT VIA GD AND SF PORT(S)/BERTH(S)/ANCHORAGE(S)
FM BALTIC TO USEC ALWAS AFLOAT ALWAYS WITH IWL, IN/OUT GEO
ROTATION, MULTIPLE LOADING
TRADE EXCLS ISRAEL/TURKISH/CYPRUS OCC/LEBANON/
EX YUGO/NORWAY/SWEDEN/DENMARK/FINLAND/CUBA
HIRE USD 4800 DAILY INCLOT PAYABLE EVERY 15 DAS IN ADV
REDELIVERY DLOSP 1 SP BOSTON/GALVESTON RGE PICO ATDNSH-
INC
DURATION ABT 40 DAYS WOG
SUB CARGO EXCLUSIONS
BOD ABT 320 MT IFO AND ABT 130/150 MT MDO SAME ON REDELY
SUB CHARS BOD APPROVAL TO BE LIFTED 1800 HRS COB TIME 9 MAY
2000
SUB FURTHER TERMS/DETAILS AS PER CHRTS EXECUTED NYPE
She does not agree with point 9. She wants a US$ 3,500 lumpsum as an
incentive for crew for hold cleaning.
30 May 2000 15:40. Clarckson wants a daily charter hire of US$ 7,3006 .
Viveka scratches her thoughts, she believes that a fixture of $7,500 is fea-
sible. Time is a critical factor. This seems to be the decisive factor:
The J.E. Hyde Shipping Index uses the size of ship, namely deadweight,
as the only attribute representing a ship’s characteristics. Eleven main bulk
trades represent the main world’s bulk trades. It becomes evident that such
an index is inadequate to address the need of the industry. The negotiations
MV L UCY O LDENDORFF /Shinwa and MV L UCY O LDENDORFF/Swazi are
not covered by those indexed adequately.
How does she then know certain appropriate charter hire values for
the ships, for those trades? Shipping Index gives too rough information
about the industry; and a chartering practitioner cannot do about it. On
should listen to views of brokers, cargo owners, media and all possible
sources. From recent charter fixtures one can see the tendencies of the
market. Such charter fixtures, published by brokers, are a collection of
specific fixtures, see 3.2.2.
Panocean and Shinwa are willing to pay a much higher charter higher for
MV L UCY O LDENDORFF, i.e. $7,500/day and $ 7,850/day respectively,
whilst the average time charter hire for its class, a 22,000 dwt bulk car-
rier is about $ 5,900 /day.
J.E. Hyde reported9 that the fuel price tend to increase. Increasing fuel
price drives the shipowners to release their ship in a time charter nego-
tiation, in order to avoid risks involving the voyage costs volatility due to
the uncertain fuel price. At the moment of negotiation the Pacific market
is firmer than the Atlantic market. Therefore she is reluctant to let the ship
trade the Atlantic market after the end of the charter party, even though the
time charter equivalent is highest among the others, see the negotiation
code 1.
The last reasoning, says Mansukhani, is that the negotiation has reached
an advanced stage. More details related to the negotiated charter party
have become known to her. Those details include the delivery date, pos-
sible background information concerning the reputation of the charterers,
potentials and risks related to the development of fuel price and ideas for
the next charter party 10 .
The above negotiation shows that the objective of the chartering ne-
gotiation is to obtain an acceptably satisfying result, instead of an optimal
result. Probabilities of events of all available alternatives are not quantifi-
able.
8
Viveka Mansukhani, E-Mail and interview on 31 May 2000
9
J.E. Hyde report dated 30 May 2000
10
The company expects from the staff to be able to make decision quickly too [21]
SHIP LO LO LO LO LO LO LO LO AO
Swazi Pan- Doo Doo
CHRS - Toepfer - - Shinwa
Sugar ocean Yang Yang
C/P V/C T/C T/C V/C V/C V/C V/C T/C V/C
3.1. Chartering negotiations
3.2.1 Tasks
Maritime forecasting and market research studies are the most common
applications of maritime economics [73]. Maritime forecasting is concerned
with predicting events on the shipping market as a whole, the prospects as
a broad category of ship types and the overall level of supply and demand,
see for examples Beenstock and Vergottis, Wergeland or Tang [8, 81, 75].
Maritime forecasting has a little relevance with the daily tasks of chartering
and operations at operational management level.
Shipping market research is concerned with studying the economic ac-
tions of individuals or companies within the market or a market segment. It
means a study of a specific ship, ship type, trade flow or business unit, gen-
erally relating to a specific business decisions [73]. Those matters concern
the chartering and operations departments very much. The staffs of those
departments are concerned with the prevailing rates of specific routes of
interest for their ships.
Tasks addressed in the market research are a.o. [73]:
TIME CHARTER
“MV WASHINGTON T RADER” 2000 74000 dwt dely S Korea 9/15
June trip via Queensland redel Far East $ 1250 daily - NYK
“MV G LOBAL F” 1998 73729 dwt dely Hong Kong 15/25 June
trip via Indonesia redel Japan $ 12500 daily - MOL-Navix
GRAIN
“MV L ARA” 6000/5 hss US Gulf/Egypt + options end June $16.50
fio 10000/7000 - Cargill
COAL
“MV C APE M ARIA” 150000/10 Richards Bay/ Hadera 1/10 July
appprox $9.30 fio scale/20000sx - NCSC
MISC
“MV AVRA” 55000/10 phosrock Casablanca/Lasero Cardenas
mid June $11.85 fio 12000sc/8000sc
Such fixtures contain specific chartering situations. Those are helpful, and
often used as a reference during chartering negotiation, to estimate the
prevailing rates for a specific trade.
12
Wonsild & Son A/S Shipbrokers, dated 06 August 2000
3.2. Usable resources 84
to June and has now reached a post-war high of yen 4,260 bil-
lion. Most of debt is though attributed to the bankruptcy of the
retail chain Sogo totalling yen 2,290 bn.
To match the recent recovery in demand for steel, major steel
makers have been increasing output at their main production
plants, and the crude steel output is expected to reach the psy-
chologically important figures of 100 million tons this fiscal years
for the first time in three years.15
Fairplay Forecast
... forecast for the next ten years based on annual growth rate of
2.6 percent, and has assumed a sustained recovery in Asia - in-
cluding stronger grain imports and a much improved economic
performance in Japan. MSI has also excluded any major eco-
nomic recessions during the coming decade and a significant
upwards shift in energy prices.16
In the supply side, assumptions are made based upon the MSIS’s belief
that the old tonnage of over 20 years old will grow. Their assumption on
the supply side is that the key factor in assessing future development is the
size of 25+ year-old fleet that is potentially available for scrapping. Their
further assumptions:
1. Increasing scrapping
2. Vessels of over 30 years old will still be allowed to sail
3. New-building tonnage’s delivery is higher than the scrapping.
On the demand side they assume that steam coal will play a more im-
portant role in the coming decade, as India and South East Asia have been
building new coal-fired power plants. West European stem coal demand is
also expected to rise, continuing a revival that started after a major inter-
ruption in the first half of the 1990s, caused mainly by the UK’s ’dash for
gas’.
15
Japan Market Update by Maersk Broker, Copenhagen, fax dated 25 September 2000
16
Fairplay Market Forecast - Dry Bulk 2000
3.2. Usable resources 86
Last This
Route / Deadweight (t) Charter Outlook
week week
Duration min max type US$/d US$/d
Trans 45000 47000 V/C 8250 8250 steady
Atlantic r/v 42000 43000 V/C 7500 7500 steady
37000 38000 V/C 6750 6750 steady
28000 32000 V/C 6250 6250 steady
USG/ 45000 47000 V/C 10250 10250 steady
Far East 42000 43000 V/C 9750 9750 steady
37000 38000 V/C 9000 9000 steady
28000 32000 V/C 8250 8250 steady
Pacific r/v 45000 47000 V/C 8400 9300 softening
42000 43000 V/C 8800 8700 softening
37000 38000 V/C 7400 7400 softening
Continent/ 45000 47000 V/C 10000 10000 steady
Far East 42000 43000 V/C 9500 9500 steady
37000 38000 V/C 8500 8500 steady
28000 32000 V/C 7500 7500 steady
Mediterra 45000 47000 V/C 10250 10250 steady
nean/ 42000 43000 V/C 9500 9500 steady
Far East 37000 38000 V/C 8500 8500 steady
28000 32000 V/C 7500 7500 steady
Far East/ 45000 47000 V/C 9000 9000 softening
Europe 42000 43000 V/C 8400 8400 softening
37000 38000 V/C 7500 7500 softening
12 months 45000 47000 T/C 10000 10000 steady
T/C 42000 43000 T/C 9400 9400 steady
37000 38000 T/C 8200 8200 steady
28000 32000 T/C 7500 7500 steady
Last This
Route/duration Cargo Charter Outlook
week week
(ton) type US$/d US$/d
hss Gulf/Continent 55000 V/C 15.75 16.00 firm
hss Gulf/Japan 52000 V/C 23.75 24.00 firm
Trans Atlantic r/v - T/C 12000 12250 firm
Pacific r/v - T/C 11500 11500 easing
Trip Atlantic/ Far
- T/C 12000 12250 firm
East
Trip Far East / At-
- T/C 11400 11250 easing
lantic
12 months T/C - T/C 11000 11250 firm
per year, reaching 240M tonne at the end of the decade. As imports by
developing countries rise, so will the incentive to develop cargo-handling
facilities, which will be the greatest benefit to Panamaxes, MSI believes.
But sub-50,000 DWT handy tonnage will remain the main workhorse of the
grain trade, accounting for almost 60 per cent of required DWT demand in
2010.
Lorentzen & Stemoco publishes the Weekly Market Outlook for both dry
and wet bulk cargo. The dominant markets are Pacific and Atlantic markets,
which are highlighted and commented, for example :
Better rates could appear for handy size tonnage plying the
Black Sea to Far East steel trade. Since there is a shortage
of suitable vessels in the area, this could eventually have an
effect on the trade.17
17
Lorentzen & Stemoco Weekly Market Outlook, on Handy size sector of the dry bulk
shipping, 11 August 2000.
Charter Delivery/ Redelivery/
Route Deadweight (t) Cargo/ Remarks
type Loading ports Disch. ports
min max
1 35,000 45,000 trip time Continent Far East -
2 25,000 - voyage US Gulf Algeria 5% HSS
3.2. Usable resources
Transpacific
3 40,000 45,000 trip time Singapore Japan
round voyage
Transatlantic
4 25,000 30,000 trip time Skaw Passero
round voyage
5 30,000 35,000 trip time South Africa Continent -
6 20,000 - voyage US Gulf Venezuela 10% HSS
7 30,000 35,000 voyage USNH South Korea scrap
Boston-
8 35,000 40,000 trip time Singapore -
Galveston
9 20,000 25,000 voyage Black Sea China steels
Antwerp-
10 25,000 35,000 voyage Brazil Hamburg any grain (sf 55’)
range
11 20,000 23,000 voyage Queensland Japan bulk sugar
Handy size tonnages are divided into few classes, according their size
and trading area. Lorentzen & Stemoco uses a ’standard’ bulk carrier of age
less than 15 years old and having cranes on board. The sizes of ships of
handy size class are clustered into five groups, see Table 3.5. For Panamax
class bulk carriers, all ships are groups into one, see 3.7. Those definitions
on the ship are apparently adequately accepted and understood by the
shipping practitioners. The outlook is expressed in qualitative terms, such
as stable, firm, easing or softening.
3.3 Tools
Tools comprise methods or programs assisting shipping practitioners to
conduct their job. To this date the tools used basically do not differ very
much from those used in the past. Apart from the fact that computers have
become an important part of running ship business, the way the things
are conducted are still manual. Many assessment and planning tasks are
still performed manually. No computer program supports the employee for
assessing the values of ship or to generate forecasts. The shipping practi-
tioners rely much on the skill to conduct their job.
The next section elaborates a simple and very useful tool to help assess
the merit of a voyage, called the voyage calculation. Then it will be elab-
orated on methods used shipping research to address shipping problems,
particularly concerning valuation and decision making tasks.
9000
Time Charter Hire (US$ /day)
8000
7000
6000
5000
4000
01/99 02/99 03/99 04/99 05/99 06/99 07/99 08/99 09/99 10/99 11/99 12/99
Date
30
Freight Rate (US$/ton)
25
20
15
10
5
01/99 02/99 03/99 04/99 05/99 06/99 07/99 08/99 09/99 10/99 11/99 12/99
Date
3.3.2 Statistics
There are two categories of statistics, descriptive and inferential statistics
[26]. Descriptive statistics is used to organize and summarize the data in
samples and populations. This includes procedures of ordering and group-
ing data into distributions and procedures for representing the data graph-
ically. Inferential statistics is used to make educated guesses (inferences)
about populations based on random samples from the populations. The
ings are still open, approximations are useful information, as the company’s management
wants to be kept informed on the merits of the operations of their ships.
3.3. Tools 92
Wijnolst and Bartelds [79] endeavored to answer the question “Are there
bulk carriers consistently over-performing or under-performing the overall
market ?” using the period time charters and grain charter index of Pana-
max bulk carriers 1989 - 1994.
After some indexing procedures, middle values are removed, leaving
the data split into two poles, ships with low and high charter rates. Cor-
relation techniques are then applied for both groups, to find out whether
design characteristics including age of vessels play inherently a role in its
95 Chapter 3. Existing Resources and Tools
commercial success.
A factor which plays an important role is the quality of the ship manage-
ment. This is in fact the perceived quality or reputation of the ship manage-
ment. And this is a subjective factor. An assessment on the value of ship
are expectedly to produce a better result, if the subjective perception on
the way the ship is managed is the same. This can be achieved when the
valuation is made from the view of one party, i.e. one shipping company
with its own philosophy and its own way to conduct the business.
He reveals that there are no specific design characteristics resulting in
an outstanding economic performance. He concludes:
concludes from the indexes within the 1985-1988 period are not cyclical,
but they move randomly. That means that high prices are not necessarily
followed by lower prices.
Goncalves develops a method for determining the optimal policies for
ship chartering [27]. He believes that forecasting models are useful, but
inadequate to solve the problem. Nevertheless forecasting is at he first
step towards a more rational decision making process 20 . His views on the
method development for obtaining optimal chartering policies “.. to con-
sider a stochastic process of freight rates with some sort of mean rever-
sion. In this way, the effects of shipping cycles and short term seasonality
would be included. For this case, I expect a more complex mathematical
development in particular with respect to optimal policies and parameter
estimation. Possibly, the discrete time version with simpler policies will be
a starting point”.
Forecasting has a poor reputation in maritime circles. The argument
that forecasts are never right is often put forward with a great conviction
by shipowners who have been far too successful in business to have their
opinions taken rightly [73].
Evans evaluates the performance of shipping markets [22]. In short
term he believes that shipowners are profit maximizers and that market
rates are equal to marginal costs. It is an evidence that the existence of
allocative efficiency of resources. He relies his analysis on the assumption
that the knowledge of the market is perfectly known. The exchange of infor-
mation between ship brokers in local, regional or international exchanges
ensures that the current state of supply and demand is known with some
degree of certainty.
Adland [5] investigated the performance of theoretical models for vessel
evaluation and investment decisions in shipping. He introduced non para-
metric freight rate models in maritime economics, a methodology that is
often used in empirical financial research but has yet to be applied to ship-
ping. The research can possible uncover the efficiency of the second-hand
market or can yield superior return from asset play.
He shows a relationship between time charter rate and the values of a
vessel. Freight rate and time charter rate have a very close relationship.
These represent long-term expectation of earning. Second hand values lie
between newbuilding and scrap values. The price of new vessels generally
represent the steel, outfitting material, labor costs and yard’s profit. Like-
wise, the scrap value is the price the demolition yard is willing to pay for the
recoverable steel and material content of the ship.
Adland believes that the capacity adjustment is the main driving force
20
Rational means a. relating to reason; not physical; mental. b. agreeable to reason; not
absurd, preposterous, extravagant, foolish, fanciful, or the like; wise; judicious; as, rational
conduct; a rational man [1].
97 Chapter 3. Existing Resources and Tools
1. Maximin rule: which is based on finding the smallest payoff for each
action and the choosing the action for which this is largest.
Maximin (Ai ) = max(min(xi1 , xi2 , xi3 ))
2. Maximax rule finds the largest possible payoff for each action ad
chooses for which this the largest.
Maximax (Ai ) = max(max(xi1 , xi2 , xi3 ))
3. Minimax rule is based upon the idea to make a very cautious decision
based which is least harmful, i.e. least loss. This rule finds the max-
imum loss for each action and then chooses the action where this is
the smallest
Minimax (Ai ) = max(min(yi1 , yi2 , yi3 ))
Pay- Situations
Loss Situations
off
Action Θ1 Θ2 Θ3
Action Θ1 Θ2 Θ3
A1 y11 y12 y 13
A1 x11 x12 x13
A2 y21 y22 y23
A2 x21 x22 x23
A3 y31 y32 y 33
A3 x31 x32 x33
A4 y41 y42 y43
A4 x41 x42 x43
(b) Loss table
(a) Pay-off table
forks. The probabilities and expected payoffs at each chance fork are com-
puted and at each decision fork the action with the largest expected payoff
is chosen.
Remarks
The decision theory requires that all information must be input and pre-
sented numerically. And the degrees of importance among the variables
must clearly / numerically be defined. All decision-making problems con-
tain similar features [23], see also Chapter 4, sub-section 4.2.1:
but they do not force themselves into quantifying the probabilities of events.
They are comfortable with they way it is carried today. And it is the natural
way of decision making. Formalizing today’s natural way of decision ma-
king to a mechanistic way laid out by the decision making theory would slow
down the process of decision making and it would simplify the complexity
of decision making very much.
Expressing a strategic alliance consideration, which may allow a low
time charter equivalent, in the way it is expressed qualitatively, is better
and more understandable than in the form of scores. Applying the method
would mean degrading the expertise degree of the expert [19].
To know the market and its tendencies in the overall field is the
professional talent of the broker, acquired after long experience.
There may be particular trades where the specialist broker may
be the greater authority. The assessment of the ’trend of the
market” is no exact science, it is rather a knack of the profes-
sional man, the broker.
Norman (1979) argues that shipowners are too strongly influenced by cur-
rent spot rates in their assessments of future market conditions, and that
101 Chapter 3. Existing Resources and Tools
Pay-off Situations
Action Θ1 Θ2 Θ3
Swazi x11 N/A N/A
Panocean x21 x22 x23
Doo Yang x31 N/A N/A
Toepfer x41 N/A N/A
P(Θj ) N/A N/A N/A
this is the reason why owners are seemingly irrational in their contracting
decisions21 [27].The valuation process in practice follows the following pro-
cedure: first is look at similar vessels sold recently or currently offered for
sale [5]. Then compare its age. A rule of thumb: ship is depreciated over
15-20 year. the depreciation is 5-6% annually . Third, speed. 2% value
difference per knot. Clearly valuation process is quite simplistic and sub-
jective, and is not directly related to the expectations of future earnings .
Adland observes that the shipowners do not always decide rationally
[5]. This adds to the irregularity of the market. "Since shipowners are
constantly trying to second-guess the cycle, crowd psychology is believed
to play an important role in this game and adds to the irregularity". In spite
of his advocation for using mathematical models, he admits The results can
not yet challenge the existing valuation method in shipping.
input output=f(input)
output (b)
between input and output ’more visible’, Figure 3.9 (b), e.g. using statistical
techniques. Or, sometimes the system is treated as a causal process, see
Figure 3.9 (c), e.g. using econometric modelling.
A lot of research has been undertaken and has contributed very much
for conducting macroeconomic analysis. But shipping index, forecasts and
forecasting models, for example, are of less use for those who conduct
daily shipping practice.
Datz suggests that manager should decide based on facts and not on
intuition. On the other hand, he believes that in order to make forecasts
useful it is imperative to include the user’s view on the developing situa-
tions, as he sees it [16]. Apparently this seems a general view of the ship-
ping research agenda today, see also for example Beenstock and Vegottis,
and Evans and Marlow [8, 23].
The fact that rational decision making procedure does not apply well in
many decision making tasks in daily shipping practice seems to be over-
looked, see Section 3.1. Intuitive way of accomplishing tasks has not been
considered as a research subject seriously. The paradox between the
views on the way shipping is run and the fact that rational decision mak-
ing does not work satisfactorily in daily shipping practice, on one hand, and
the selection of topics and the ’mechanistic’ views on the shipping system
and the choice of methods conducted in shipping research to this date, on
the other hand, confirms the necessity to have a new way of looking at the
103 Chapter 3. Existing Resources and Tools
shipping tasks. By doing so, the development of tools usable for shipping
practice can be implemented.
4. The decision making in practice does not follow the rational decision
theory. The objective is not to obtain the optimal solution, but to obtain
an acceptable solution.
105
Chapter 4
4.1 Introduction
Chapter 2 showed us that ship practitioners rely very much upon their in-
tuition. Their experience, information and views from brokers, shipping
agents and newspapers are very important sources for assessing generally
prevailing charter hire values and sale & purchase prices of ships more ac-
curately. Published indexes and statistics can be useful too. They are
viewed as a rough guide. But they are not always of direct relevance to
actual daily negotiations. Many trades and ships, with their specific design
characteristics, are not covered adequately by those index. The indexes
are too general to be practicable.
The shipping research community has been working hard to support
the shipping industry by developing e.g. forecasting tools. Shipping market
models have been developed in order in order to understand the process,
the functioning of the market, and in turn it can be used for a.o. predicting
the shipping market tendencies. To conduct daily tasks in shipping, such
models are of little use.
Shipping practice is still marked by decision-making processes which
are fast and intuitive. Ship’s value assessment tasks or forecasting tasks
rely very much upon the capability of the managers to assess, and they do
not rely on fancy tools.
The gap between both worlds, that of research and shipping industry
respectively, in terms of the way they view the problems, is probably not
that small. Findings shown earlier in Chapter 3 show that rational decision
makings are not necessarily practicable. Intuitive actions which mark many
decision-making actions in the industry, seem to be less appreciated by the
107
4.1. Introduction 108
4.2 Foundations
4.2.1 Bounded Rationality
A classical view on decision making
Bounded Rationality
Chapter 2 and 3 have taught us that many decisions in shipping practice are
not made rationally. Objective probabilities of future events of, such as bad
weather, high fuel prices or magnitude of cargo damage, are not explicitly
known. Shipping practitioners may use some subjective probabilities, but
they do not quantify them numerically. As shipping connects nearly all parts
of the wold, this sector is volatile to exogenous factors, such as weather
and political events of a country. The influence of exogenous factors on the
shipping system is tremendous. This enhances the size of complexity of
the system. The uncertainty plays an important role.
In his theory of bounded rationality, Herbert Simon3 argues that the ca-
pacity of a human is limited and it is too small to address the complexity
of the real world rationally. The complexity of the world is marked with
the incompleteness and inadequacy of human knowledge, the inconsisten-
cies of individual preference and belief, the conflicts of value among people
and groups of people. Nevertheless a human is able to solve the complex
problems by simplifying the problem formulation drastically. By applying
approximate and heuristic techniques, the complexity of the problem can
be downsized [72].
Optimization techniques, under the umbrella of the operations research,
are a prominent content in management science. Linear and dynamic pro-
gramming are among a few which are frequently used to solve managerial
problems, see [30]. Simon argues that the usage of optimization tools will
only be of use if the uncertainty is absent or full probability distributions for
uncertain events are available. The classical theory of decision making will
3
Nobel Prize Laureate in Economics 1978
4.2. Foundations 112
ship
Shiping company
never handle any real problem satisfactorily, unless the uncertainty does
not play a central role [70].
the command. This department becomes the main contact with the charter-
ers. Depending on situations, the relationships may vary from one case to
another, see Figure 4.2. The way all parties act reflects the quality of their
relationships. Between close long-time partners, the parties may allow late
payments or being friendly to claims. The other extreme exists as well, a
new partner, or company which is supposed to have a bad reputation. The
space between them is huge. These relationships will be known after one
knows the preferences of the management. Since it is hardly possible to
express it sharply, a consultation between managers is necessary.
Many variables affect the shipping industry, from fuel price until political
events. Their effects to the industry is difficult to determine. Many informa-
tion exists in qualitative form and is subject to interpretation. The complex-
ity of a system arises when information is unreliable, or irrelevant,or when
there are too many interacting variables which possess some non-linearity
properties [9, 68, 84]. A system can be viewed as a complex system when
it is composed of sub-systems and its nature of relationships is imperfectly
known [74]. Shipping can therefore be viewed as a complex system.
In order to handle such a complex system, the size of the problem must
be cut down, Simon argues . By doing so problems can be solved more
easily. Using approximation or heuristic heuristic techniques, among a few,
may open a way to better handle complex problems [72].
ship
Shipping company
Cargo
Broker Shipper
owner
ship
Shipping company
Cargo
Broker Shipper
owner
Amount Payment
terms
Trading
area
Amount Payment
terms
Trading
area
Amount
Payment
terms
Trading
area
(3) Detailed
lengths to viewers, see Figure 4.6. The viewers can estimate well the aver-
age length of the bars. But they have a difficulty to estimate the total length
of the bars. Higgins uses the concept ’accessibility’, which means the ease
with which particular mental contents come to mind . Average is more ac-
cessible than the sum. They can estimate the average length of the bar not
by calculating the sum of bars’ lengths divided by the number of bars. The
sum operation is not an intuitive act, it is a deliberate act, and it is effortful
and much slower compared to the intuitive act [35].
Another finding of Kahneman concerns the way a human perceives
events. Figure 4.7 shows a graphic containing the degrees of pain of a
patient. Viewers believe that patient B suffers more than patient A. But the
patient A feels he is more suffering. The worst moments and the values in
the end episode are weighted heavier.
Kahneman suggests that intuition is automatic, effortless, associative
and difficult to control or modify. Reasoning, on the other hand, is slow,
effortful and deliberately controlled. Intuitive thinking operates on basic
representations with relatively little elaboration or extra computation. The
basic representation of sets includes average values of features, but does
not normally include their sums [35].
Experience is the accumulation of knowledge or skill that results from
direct participation in events or activities [3]. Remembering past successes
are useful. The ways those problems were solved, can be reused or copied
to solve current problems. Past failures can be useful too. Those failures
mean warnings. Hints and warnings are very powerful knowledge, heuristic
knowledge. Remembering all experiences are not practicable and are not
possible, due to limited human capacity to store those. Selectively forget-
ting unuseful experiences are necessary.
Every chartering negotiation is unique. The parties, ship’s and cargo
conditions, weather, macro-economic and political situations are different.
The details of Charter Parties are therefore always different from previous
ones. But one realizes very much that a big part of the Charter Party is
standard. It is efficient if one does not have to draw a Charter Party from
scratch every time he concludes a chartering negotiation. A standard Char-
ter Party form is helpful to address this problem.
It is helpful to expedite accomplishing the chartering tasks, and to ease
preventing from forgetting or overlooking essential clauses. There are a
number of standard Charter Party forms available, such as those of the
New York Produce Exchange (NYPE) or of the Baltic and International
Maritime Council (BIMCO). This is an act of reusing good clauses which
are applicable for current case. The practice has gone a bit further. Not
only the use of standard Charter Party is widely popular, one uses also
past Addendum or Rider’s Clauses.
The above method is efficient, and it works. Inappropriate or even in-
correct data can easily be modified, when necessary. The Charter Party for
123 Chapter 4. Towards the Concept of Compatibility
MV LUCY OLDENDORFF
ACCT: SHINWA KAIUN KAISHA. LTD., TOKYO
DEL DLOSP SINGAPORE ATDNSHINC
LAYCAN 0000HRS 11 JUN/ 2400HRS, 2000 LOCAL TIME
FOR ONE TCT VIA SA(S)/SB(S)/SP(S) AUSTRALIA TO SE ASIA AA AWIWL
WITH LAWFUL CGO INTENTION GRAIN DURATION ABT 30-40 DAYS WOG.
REDEL DLSOP 1 SP JAPAN/SPORE INCL. S.KOREA/MALAYSIA/FULL IN-
DONESIA/PIPICAO ATDNSHINC
HIRE USD 780 PDPR INCLOT PAYABLE 15 DAYS IN ADVANCE
BOD TO BE ABOUT 400-650 MTS IFO AND 50/80MTS MDO. OWNERS
GUARANTEE VSL HAS SUFFICIENT BUNKER TO PERFORM CHRTRS IN-
TENDED VOYA EX ESPERANCE, WEST AUSTRALIA TO JAPAN. BUNKER
ON REDELIVERY TO BE ABT SAME QTTY AS ON DELIVERY PRICES AS
AGREED IFO 170/MDO 230 PMT BENDS.
Second group: Improvement of past Charter Party, the following lines are
to be added:
The above form is also presented in one of the following shorter forms:
If A and B, then C
A & B →C
The antecedent of the rule is the premise or left-hand side (LHS) and the
consequent is the action on the right-hand side (RHS).
A chained mechanism can be illustrated as follows:
Rule 1 If A, then B
Rule 2 If B, then C
Now we have input data equal to A. The consequent or the result is C. This
is called forward chaining or data-directed inference. The fact that the data
is known (A), it drives to deduce a conclusion (C). In full form the above is
written as follows
125 Chapter 4. Towards the Concept of Compatibility
Knowledge Expert
Expert engineer system
Data, Knowledge
texts base
Rule 1 If A, then B
Rule 2 If B, then C
Data A
Conclusion C
RULE 200
IF 1. The site of the culture is blood, and
2. The stain of the organism is gramneg, and
3. The morphology of the organism is anaerobic, and
4. The portal of entry of the organism is GI
RULE 01
IF Fuel price is high
THEN Time Charter Party is preferred over Voyage Charter
Party (certainty degree = 0.9)
RULE 02
IF 1. Ship is smaller, and
2. Ship calls more ports
THEN Profitability is less accurate
RULE 03
IF Charter rates are firming
THEN Less paying cargoes like scraps have difficulties to
find carriers
An expert system can be implemented well if the experts can can ex-
press the domain knowledge unambiguously and consistently [40]. It be-
comes evident soon, that applying expert system in ship chartering is im-
practicable, for the following reasons: a. it is difficult, when possible, to ex-
press the chartering knowledge sharply, unambiguously, and consistently.
b. the rules will hardly be complete c. modifying and adding rules are man-
ual and laborious. Nevertheless, an expert system can still be of use for a
limited scope of a very specific area.
Sets
A crisp set is marked by its sharp boundary expressing ’an object either
belongs to this set or not’. Suppose Ā is a complement of a set A, defined
as a set of all elements in the universal set which are not in A, see Figure
4.9. X is defined as the universal set. Its important properties are:
127 Chapter 4. Towards the Concept of Compatibility
Union A ∪ Ā = X
Intersection A ∩ Ā =
ship A1
company B1
ship A2
ship A3 company B2
ship A4
Relations
Crisp relations describe the presence or absence between elements. Map-
ping is a method to represent the relations visually. Figure 4.12 shows ’be-
longs to’ relations, between elements of the sets ’ships’ and ’companies’.
Another representation is using coordinates, like: R={<ship A1, company
B1>, <ship A2, company B2>, <ship A3, company B1>]. Those relations
can be then presented in a Cartesian diagram too [37]. Furthermore an-
other way of presentation is also available, using a matrix or a crisp relation
4.3. A survey on methods 130
ship A1
0.8
0.7
0.2 company B1
ship A2 1.0
1.0
0.3
0.2
ship A4
Fit to Ships
mbv≥ 1.0 mbv≥ 0.9 mbv≥ 0.6 mbv≥ 0.3
B1 A3 A3 A1, A3 A1, A3, A4
B2 A2 A2 A1, A2 A1, A2, A3
None A1, A4 A1, A4 A4 -
Fuzzy numbers
What is the meaning of approximately 14 knots? This implicitly contains
two things:
4.3. A survey on methods 132
This is the underlying reason for applying the fuzzy set theory11 for ad-
dressing the notions of imprecision. An approximative number, like approx-
imately x, is called a fuzzy number x eis a set of numbers
e. A fuzzy number x
around an exact number x. When a vessel sails at an actual speed of
13.97 knot, which deviated very slightly from the can be treated fully the
same as a meaning a set of exact numbers around 14, which may have
various degrees of truth. This degree of membership, also called member-
ship value, has a value between 0 and 1, meaning totally wrong and fully
true respectively.
An exact number we use everyday, called also a crisp number, is a
special case of a fuzzy number. An exact number can defined as a fuzzy
number having only one member with a degree of membership 1. Degrees
of truth are defined using a function, called a membership function. This
membership function, may take several forms, and can be expressed in
straight mathematical functions12 , see Figure 4.14.
For the sake of computation efficiency, a trapezoidal function will be
chosen, see Figure 4.15. This trapezoidal membership function is defined
as follows:
0, x≤a
x−a
b−a , a ≤ x ≤ b
membership value(x) = 1, b≤x≤c
d−x
, c≤x≤d
d−c
0, d≤x
DSW Algorithm
Arithmetical operations with fuzzy numbers are operations with sets. First
of all a fuzzy number is to represented in such a way which can easily
11
The fuzzy set theory was introduced by Lotfi Zadeh in 1965.
12
This function may remind us to a probability function. The probability theory deals with
expectation of a future event, based on something known now[37]. From its point of view,
the question addressed would be “What is the probability that 13.50 can be viewed as
14.00?”, which makes less sense.
The uncertainty resulting from the imprecision of meaning of a concept is addressed by
the fuzzy set theory in different way. The question will be “How true is it to view 13.50 as
14.00?” Further extensions to nun numerical values can easily be implemented, such as
“How true or acceptable is it to view a 35,000 dwt as a big ship?”
133 Chapter 4. Towards the Concept of Compatibility
14
f 0+ = [12.25, 15.75]
14
f 0.5 = [13.00, 15.00]
14
f 1.0 = [13.75, 14.25]
Fuzzy numbers
1.4
About 14
About 5
About 19
1.2
1
Membership value
0.8
0.6
0.4
0.2
0
5 10 15 20
Number
14
f+e f where
5 = 19,
19
f 0+ = [12.25, 15.75] + [4.00, 6.00] = [16.25, 21.75]
19
f 0.5+ = [13.00, 15.00] + [4.45, 5.55] = [17.45, 20.55]
mbv
exact
value
Defuzzification
known
similar Problem
problem space
Retrieve
current
problem
known
Solution
solution
space
Adapt current
solution
reminding of
similar problem
Current problem Old problem
Old problem’s
Solution
solution
lems he faced personally, but those can also be problems faced by his
colleagues. The problems are indexed according to their similarities to the
new problem. The way (one of) the most similar problem was solved can
be borrowed to solve the new problem. By doing problem solving can be
accomplished efficiently.
The root of CBR is found in the works of Roger Schank on dynamic
memory and the functioning of reminding of earlier situations (episodes or
cases) in 1982 [4]. CBR has benefited very much from earlier works on phi-
losophy, in particular that of Wittgenstein on the meaning of language. He
stated that the meaning of any object is hardly possible to define precisely.
The meaning of an object depends also on the context. And, comparing
objects is easier than defining the properties of an object. An effective way
of describing an object is by presenting it in association of or in a set of
other objects [48, 78].
In the reality of shipping, it is difficult to define ’a good ship’. Framing it
within a context of a certain shipping company and a certain trading route,
the difficulty of defining ’a good ship’ decreases. And this task becomes
much easier, when we associate it with an existing ship.
The following definition is applied within the CBR. A case is an object
which represents specific knowledge. It contains knowledge at operational
level [39]. A case contains three major elements, and it can be expressed
as a tuple as follows:
Retrieve
New Indexed
problem cases
Case- Reuse
base
Retain
Confirmed Proposed
solution solution
Revise
advance, for example the most similar one is, or n most similar ones are
automatically chosen. There are two types of similarity value calculations,
local or global similarity calculations. Local similarity deals with the values
of a single attribute or a feature. The global similarity calculations represent
the holistic view of the case. The relative importance of the attributes is
reflected by their weights. For more details, see below.
R EUSE. The chosen case acts as a guide for solving current problem.
The problem is solved by borrowing the idea how the chosen case was
solved. To do this, one can simply copy or adapt the solution concept of the
chosen case, whenever necessary.
R EVISE. This phase is complementary to the previous one. If the pro-
posed solution is not satisfactory, then the solution is modified.
R ETAIN. Retain phase is the learning phase of the system. After com-
pleting a session, e.g. problem-solving or a planning session, the result
is added with a remark. The result is then saved into the casebase. This
session is now instantly available for the next task session.
Similarity measurement
Q : query
C : case
i : individual attribute
f : similarity function for attribute i
w : weighting of attribute i
The function f is the attribute similarity function in the form of,
for example, a trapezoidal function or a fuzzy associative map,
as illustrated in Figure 4.22 and Table 4.6 respectively.
0.0 0.0
3 6 9 12 75% 94% 106% 125%
months
0.0 0.0
3 6 9 12 50% 87.5% 112.5% 150%
miles
new mortgages and to evaluate the current value if mortgage packages that
may be purchased. The usual way of appraising the value is by inspecting
the site and assessing it manually. This process is costly, about $500 per
object, it lasts about three to four days.
The Table 4.5 shows that such a manual method produces the least
error compared to other methods. Another finding shown there is that the
increasing number of attributes used leads to an increasing accuracy of
the assessment [10]. The site inspection producing the best assessment
accuracy may be interpreted as an act which incurs more attributes; their
size and details are unidentifiable.
Bonissone et al. at General Electric developed a Fuzzy CBR - based
method, called P ROFIT (Property Financial Information Technology) for tack-
ling this task. After entering the attributes of the property in question,
P ROFIT retrieves similar cases form the casebase. Six attributes are con-
sidered: address, date of sale, living area, number of bathrooms and bed-
rooms. Similarity calculations are performed between the property in ques-
4.4. Examples from our neighbours 142
Case
1 2 3 4 5 6+
1 1.00 0.50 0.05 0.00 0.00 0.00
2 0.20 1.00 0.50 0.05 0.00 0.00
Query 3 0.05 0.30 1.00 0.60 0.05 0.00
4 0.00 0.05 0.50 1.00 0.60 0.20
5 0.00 0.00 0.05 0.60 1.00 0.80
6+ 0.00 0.00 0.00 0.20 0.80 1.00
tion (query) and the cases stored. The similarity value of each attribute
is calculated using the a trapezoidal fuzzy membership function, see 4.22.
The attribute similarity value of two other attributes, i.e. number of bed-
rooms and bathrooms, is calculated using a fuzzy relation, see 4.6. The
combined similarity value is obtained using weighted aggregation. Similar
cases are then ranked.
The best four to eight cases are selected. The relative of importance of
each case is determined from its total similarity value. The values of those
cases are than used to calculate the price of the house in concern, using
weighted sum method.
1.00 very
near
0.75
near
0.50
slightly
0.25
near
-c 0 c
solving a new problem by analogy. This resembles the idea of the CBR
methodology [28].
Hansen developed the WIND-1 system, a CBR-based forecasting model
[29]. This system contains two parts, first a large database of weather ob-
servations containing over 300,000 hourly observations. The second part
is the fuzzy k-nearest neighbors algorithm. This algorithm applies the CBR
methodology. The fuzzy set theory is applied to conduct the similarity mea-
surement. Expert forecasters determine the fuzzy relationships between
attributes. Three types of basic measurements: (a) difference of values
(x1 − x2 ) (b) ratio of values (x1 /x2 ) and (c) nominal attributes. Figures 4.23
and 4.24 illustrate the usage of the difference and ratio for determining the
similarity value. The similarity value between nominal attributes is deter-
mined using a table, called also the fuzzy associative memory, see Table
4.7.
The relative importance of the case is determined from its similarity.
Then the cloud ceiling heights and horizontal visibilities are estimated using
weighted sum method.
before using them, those components are treated, or called cured, in a large
oven, called an autoclave. If curing is not successful, the components must
be discarded.
The components of various sizes are placed in the autoclave simultane-
ously, see Figure 4.25. In practice the operators relied upon the drawings
of previous layouts to inform how to layout the autoclave. Since the com-
ponents are rarely identical the risk of failure always exists. The curing
characteristics of the autoclave are unknown, and each component has its
own curing characteristics too.
Rule-based expert system was developed. It failed, the operators were
unable to formulate the rules for curing the components adequately. Ther-
modynamic modeling failed too. Learning from the success of the natu-
ral way the operators solved the layout problems, the following strategy is
taken: reuse previously successful loadings.
The process is described schematically by Figure 4.26. Parts or compo-
nents are categorized according to their priority. For any given list of parts,
Clavier seeks to retrieve past layouts that successfully cured the highest
4.5. Concept of compatibility 146
whilst the latter concerns with the question if a vessel fits to a certain charter
hire.
Being ’compatible’ means being capable of existing in harmony or be-
ing capable of bein in combination with others14 . The above definition em-
braces two important concepts, namely grouping and matching. Grouping
concerns with entities having the same role to serve a certain purpose, for
example. Matching concerns entities having different roles to serve a cer-
tain purpose. Entities are object in concern, for example ships. Purpose
of having a ship is associated with operating a ship, e.g. coal carriage. A
decisive factor is not only the physical properties of the ship, e.g. design
characteristics and age, but the capability to fulfill her role to accomplish
a task or job or to serve a purpose. The above leads us to two types of
compatibility: (a) entity compatibility, and (b) relation compatibility.
Suppose ship A carries coal, fixed for a long term contract. Due to
collision she is damaged, ship A must be replaced by a compatible one
immediately. The contract states that the ship owner guarantees the coal
carriage, and any problems related to the ship, e.g. damage, the owner
must find another ship to maintain the service. Her sister ship B is located
far away from the traded area of ship A. It would involve high mobilization
costs, long, non-operational time, loss of income and possible breach of
contract, if ship B is to be employed. A plausible solution is to take another
ship, say ship C, having less similar characteristics, located in the vicinity
of the traded route. She can be put into service within shorter time. Ship
C is more compatible to ship A to serve the purposes (due to her shorter
mobilization time in maintaining the existing service).
The above showed the application of the concept to an entity of a com-
plex structure such as a ship. The concept applies as well for an object
of a simple structure; that is an object which can sufficiently be described
using one variable only, such as time, deadweight, or length. For example:
length of port stay of 5.05 days and 5 days are regarded as similar or even
equivalent in practice.
Entity compatibility can be viewed as an entity similarity measurement.
This can be either numerical, textual, graphical or of other form. The fuzzy
set theory may accommodate this matter well, by viewing the entity com-
patibility as proximity relations.
Examples of problems which are viewed to belong to the entity compat-
ibility are:
In case MV G RETKE O LDENDORFF has been fixed earlier for another Char-
ter Party, he will be offered another ship which more or less comparable
15
E-Mail Viveka Mansukhani, dated 10 July 2004
149 Chapter 4. Towards the Concept of Compatibility
fully compatible
less compatible
less compatible
not compatible
The above implies this: if the problems are similar (ship’s sizes are similar),
their solutions (their charter hires) are likely to be similar too.
4.5.4 Properties
The concept of compatibility involves the capability of seeing the relation-
ships between elements. The following aspects mark the concept:
(a) Compatibility may change over time.
The suitability of, e.g. a ship, to serve a certain purpose may change
over time. Her compatibility depends on trade and on the ship owners too,
see Figure 4.30.
(b) Compatibility involves vagueness. It is a simple fact that is no longer
sufficient to describe things in a black-or-white fashion. The definition of,
e.g. a good ship and a poor ship, is a matter of degree. The boundary of ’a
good ship’ is vague, unsharp, see Figure 4.31. To be noted here, that when
it comes to the final decision, decisions are usually sharply defined: either
yes or not. This involves a defuzzification process.
151 Chapter 4. Towards the Concept of Compatibility
rules. This method of solving problem can be very efficient for a lim-
ited scope of problems. Real problems grow in size in the course of
time. It is the source of difficulty. The program is not getting smarter
because of having solved more problems. The program must be fed
manually with new rules. Modification and maintenance of rules is
laborious and difficult.
5. Fuzzy set theory is a superset of the crisp set theory. The fuzzy set
theory aims at addressing the notion of imprecision marking the ship-
ping system. This fuzzy set theory, in combination with other methods
when necessary, may contribute to address the entity compatibility
problem.
Proposed Implementations
157
5.1. Fuzzy arithmetic-based voyage estimation 158
analysis.
To address the interpretation complexity, it would be of use to loose
the intention of using high precision. One may cut down the complexity
using approximation, see Simon et al in sub-section 4.2.2. If the precision
is higher, the complexity increases exponentially [Ross1996].
The following voyage estimation example is taken from Evans and Mar-
low [23], with some necessary additional information and modifications for
demonstration purpose. MV U NIVERSE C ARDIFF, a bulk carrier, is to carry
25,500 ±5%ton scrap from New Orleans to Yokohama. The freight rate
is US$ 25.00/ton fio2 , loading and discharging speeds are 3,750 ton/day
average, 7 day respectively, SHEX3 .
The above statement has legal and operational consequences. From
operational viewpoints, the above can expressed as approximations, as
follows:
1. The vessel will carry approximately 25,000 ton, ranging between 24,225
and 26,775 ton of scrap.
2. The loading and discharging speeds are about 3,750 ton/day. In prac-
tice these speeds may deviate from this value due to weather, avail-
ability of cargo and performance of cargo gears. This value of about
3,750 is interpreted as 3,750 ±5%, meaning all values ranging from
3,563 - 3,938 ton/day.
3. Speed of the vessel will be about 14.5 knots. Weather play an im-
portant role for the deviation of speed. A speed of about 14.5 knots
2
free in/out
3
Sunday Holiday Excluded
159 Chapter 5. Proposed Implementations
This sub-section concerns with the last point 7, as it covers all previous
sensitivity calculations. Figures 5.1, 5.2 and 5.3 show the input of amount
of cargo to be transported, vessel’s speed and cargo handling speed for all
27 scenarios.
Amount of cargo
27000
26500
26000
Amount of cargo (ton)
25500
25000
24500
24000
0 5 10 15 20 25 30
Scenario
Vessel’s speed
15.4
15.2
15
14.8
Vessel’s speed (knots)
14.6
14.4
14.2
14
13.8
13.6
0 5 10 15 20 25 30
Scenario
3900
3850
Cargo handling speed (ton/day)
3800
3750
3700
3650
3600
3550
0 5 10 15 20 25 30
Scenario
Voyage duration
68
66
64
Voyage duration (days)
62
60
58
56
0 5 10 15 20 25 30
Scenario
80000
70000
60000
Gross profit per voyage (ton)
50000
40000
30000
20000
10000
-10000
0 5 10 15 20 25 30
Scenario
5400
5200
Time Charter Equivalent (US$/day)
5000
4800
4600
4400
4200
4000
3800
0 5 10 15 20 25 30
Scenario
Unit
T/C Hire
Capital &
Unit Fuel Cons Operating costs Gross Profit
at Sea per Voyage
Distance
Sea Time
Total costs
Speed
Voyage
Length
Waiting Time Voyage costs
Cargo
Volume Port Time
Cargo Handling Cargo Handling
Speed Time
Unit Fuel Cons
Cargo
at Port Time Charter
Weight
Equivalent
Port Charges
Canal Dues
Amount of cargo
0.8
0.6
mbv
0.4
0.2
0
22000 22500 23000 23500 24000 24500 25000 25500 26000 26500 27000
Amount of cargo (ton)
Figures 5.8, 5.9 and 5.10 show the amount of cargo, vessel’s speed
and cargo handling speed represented as fuzzy numbers using a triangular
fuzzy membership function. The approximative values are represented by
these fuzzy numbers.
Figures 5.11, 5.12 and 5.13 show the likely voyage duration, total costs
and income of the charter party. Finally, Figures 5.14 and 5.15 show the
results, i.e. gross profit and time charter equivalent. In words the gross
profit will likely be about US$ 40,000 and its time charter equivalent ap-
proximately US$ 4,650/day.
5.1.4 Discussion
Vessel’s speed
0.8
0.6
mbv
0.4
0.2
0
13 13.5 14 14.5 15 15.5 16
Vessel’s speed (knots)
0.8
0.6
mbv
0.4
0.2
0
3300 3400 3500 3600 3700 3800 3900 4000 4100
Cargo handling speed (ton/day)
Voyage duration
0.8
0.6
mbv
0.4
0.2
0
56 58 60 62 64 66 68 70
Voyage duration (days)
Total costs
0.8
0.6
mbv
0.4
0.2
0
500000 520000 540000 560000 580000 600000
Total costs per voyage (US$)
Total income
0.8
0.6
mbv
0.4
0.2
0
520000 540000 560000 580000 600000 620000 640000
Total income per voyage (US$)
0.8
0.6
mbv
0.4
0.2
0
-20000 0 20000 40000 60000 80000 100000
Gross Profit per Voyage (US$)
0.8
0.6
mbv
0.4
0.2
0
4000 4500 5000 5500 6000
TCE (US$/day)
Chapter 2 and 3 have shown that shipping indices are not adequate to as-
sist shipping practitioners to conduct their daily job in chartering. Using the
Time Charter fixtures used in Chapter 2, a quick glimpse on the coverage of
the J.E. Hyde Index can be made. Using a trapezoidal fuzzy membership
function, ships having similar deadweight to a route of J.E. Hyde Index is
determined. Furthermore the Index describes that the index is appropriate
for ’modern ships’. The definition of modern ships is usually based on age.
More precise definition on ship’s characteristics are not available. Another
important property of an index is the route. The charter fixtures record the
delivery and redelivery ports. Since no port database is available, deter-
mining the ’route compatibility’ using the charter fixtures used in Chapter 2
is impossible.
The coverage of the J.E. Hyde Index is measured on two variables,
size and age. For Route 1, the ship’s size is defined as (a=25,000 b=27,000
c=43,000 d=50,000 dwt) using a trapezoidal membership function, see sub-
section 4.3.2.The age of ship is defined a=0 b=0 c=5 d=12 years.
From 7719 fixtures having complete information on ship’s deadweight
169 Chapter 5. Proposed Implementations
and age, 7377 having membership value (mbv) of 0.0, meaning that the
fixtures are not covered by the Index at all. 342 fixtures (4.6%) having an
mbv ≥ 0.0, 315 fixtures (4.2%) with an mbv ≥0.25, 226 fixtures (3.0%)
having an mbv≥ 0.50 and 103 fixtures (1.3%) ≥>=0.75. Only 47 fixtures
(0.6%) have an mbv =1.00, meaning that only 47 fixtures are well covered
by the Index. Some of them are illustrated in Figures 5.16 and 5.17.
The fixtures contain important geographical aspects which affect the
charter hire very much, namely the delivery and redelivery ports of the
ship. Measuring the geographical similarity is ideally based on the sail-
ing distances, and not based on the straight line distances. Due to the size
of the data and the unavailability of complete sailing distances, geograph-
ical similarity calculations cannot be performed. Therefore the coverage
figures mentioned above are still a too optimistic figure since other criteria,
such as the geographical criterion (trade route), are not included. There-
fore adding other ten indexes of J.E. Hyde would not affect the conclusion
on the coverage of the index, as confirmed by findings in shipping practice,
see also 3.1.2.
25000
Time Charter Hire (US$ /day)
20000
15000
10000
5000
01/97 04/97 07/97 10/97 01/98 04/98 07/98 10/98 01/99 04/99 07/99 10/99
Date
(a)
Compatibility JE Hyde Index - Route 1
30000
lambda >=0
25000
Time Charter Hire (US$ /day)
20000
15000
10000
5000
01/97 04/97 07/97 10/97 01/98 04/98 07/98 10/98 01/99 04/99 07/99 10/99
Date
(b)
Figure 5.16: Coverage measurement of J.E. Hyde Index - Route 1
171 Chapter 5. Proposed Implementations
25000
Time Charter Hire (US$ /day)
20000
15000
10000
5000
01/97 04/97 07/97 10/97 01/98 04/98 07/98 10/98 01/99 04/99 07/99 10/99
Date
(c)
Compatibility JE Hyde Index - Route 1
30000
lambda =1.00
25000
Time Charter Hire (US$ /day)
20000
15000
10000
5000
01/97 04/97 07/97 10/97 01/98 04/98 07/98 10/98 01/99 04/99 07/99 10/99
Date
(d)
Figure 5.17: Coverage measurement of J.E. Hyde Index - Route 1 (con-
tinued)
5.2. Tailor-made shipping index 172
1
Normalized index
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50 60 70
Week
3. Then the weights are determined based on: (a) the similarities be-
tween indexes, see previous points, and (b) the recency of the trans-
actions. The recency of the transactions measures the relevance of
a transaction, i.e. older transactions are to be forgotten.
4. Then using weighted sum method, the value of the new index is cal-
culated using weighted sum method.
Figure 5.18 shows the normalized indexes. The performance of the new
index is compared with its reference, i.e. Route 4 of J.E. Hyde Shipping
Index, is shown in Figure 5.19.
5.2.3 Discussion
The above tailor-made index shown above produces an error of approxi-
mately 10%. When the difference between the peaks and troughs is small,
the values match well. If it is not the case, the estimated value misses the
reality. The method probably still does not match the capability of an ex-
perienced shipping practitioner. Nevertheless the usage of the method is
far cheaper than the usage of an experienced estimator. Therefore for the
usage in normal situations, which are believed to be dominating, is still of
advantage.
173 Chapter 5. Proposed Implementations
7000
6500
Charter hire (US$/day)
6000
5500
5000
4500
4000
0 10 20 30 40 50 60 70
Week
This tool can also be of use to roughly reconstruct the past, meaning
to assess for example the charter rate of the past. This can be useful in
case of insurance claims. Secondly, this tool can be useful, for educational
purposes, for trainees, less-experienced junior managers or experienced
managers just moved in from other business sectors.
ping are generally quite similar to those in other shipping sectors, such as
container shipping.
This section elaborates an implementation of the concept of compat-
ibility in addressing stowage planning problems in container shipping, as
mentioned briefly in sub-section 4.2.4. Stowage planning problems are as
old as the history of the ship itself. It is the task to allocate containers to be
loaded to slots on board the ship. The product of a stowage planning task
is called stowage plan, that is a two-dimensional diagram showing the po-
sitions of containers stowed into a container ship. Increasing efficiency of
terminal and ship operations and increasing size of container ships become
a pressure to stowage planners. Those make the task prone to human er-
ror.
Before the arrival of a container ship, the planning department receives
a container loading list (CLL) and the expected arrival conditions of the
ship. A fundamental problem is the inaccuracy and imprecision of the CLL.
It is not unusual that the CLL is still changing until very short, two hours
or even shorter, before the departure of the ship. That means during the
discharging and loading, new containers can still be added or some other
containers are cancelled.
In spite of the advancement of the information technology, and comput-
ers have become indispensable to conduct daily planning tasks, the core
process of planning, namely creating the plan itself is still performed man-
ually. The planner still has to create the plan himself, then when the plan
is finished he checks, if the criteria are met or not. Figure 5.20 shows the
way the stowage planning task is performed today.
Today stowage planning software provides powerful modules which en-
able the planner to obtain information on the state of the ship and its car-
goes concerning stability, strength, dangerous cargoes, crane split, draught,
trim and visibility check. Those modules expedite the calculation tasks. If
the plan shows that the ship does not have a sufficient static stability, for
example, the user is not assisted with any help as to what to do, which
containers have to be moved and to which slots. A number of efforts to
automate the stowage planning have been undertaken over the past three
decades, see Figure 5.21.
5.3.2 Foundation
In view of the slow research progress made in the last decades, it is of
necessity to take few steps back, to rethink what we understand under
stowage planning. Stowage planning is a search problem and a constraint
satisfaction problem [53]. It is a chained process too, a modification in a
stowage plan of a loading port will change all plans of the following ports.
And stowage planning covers more than the technical aspects of planning,
it involves some aspects of authority with the decision making process in
175 Chapter 5. Proposed Implementations
the company. It is not unusual that important clients want that their cargoes
are stowed in specific slots. At this point, technical stowage planning con-
siderations alone are not adequate to address the problem. Commercial
considerations play a more important role. Therefore the sales or market-
ing department’s decisions will contribute in determining the end look of the
stowage plan.
Every ship, every route is specific. Each has its own typicality. Even sis-
ter ships trading the same route under different ship operators, may have
different stowage plans. Furthermore situations may change, some are
predictable, some other are not. Therefore we believe it hardly possible to
formalize the knowledge contained in the stowage planning objectively in
terms of mathematical formulas or if-then rules. We should find an alterna-
tive direction to address the problem effectively.
Creating a stowage plan means also accepting a fact that the container
loading list (CLL) is less correct and less precise. The loading list may
change during loading and until short before the departure of the ship, the
degree of precision and correctness improve in the course of time. The
knowledge for creating a stowage plan is not stored centrally. It is spread
among many, i.e. books (concerning the technical aspects of planning
such as stability, strength or dangerous cargoes), human (stowage plan-
ning, sales and operations departments) and past stowage plans. The type
of knowledge can be formal, such as stability, or informal, such as subjec-
tive preferences or rules of thumb. It is hardly possible to define desirable
properties of a stowage plan precisely and consistently. In daily practices
one knows if a stowage plan is favorable or less favorable, after seeing it
and comparing with his expectations. But a planner cannot define precisely
what we understand under a good stowage plan7 .
As we have learned form the history of technology advancement, the
nature has been our great inspiration for solving various problems. In
stowage planning, the success story is not far away out of reach. A human
planner solves planning problems everyday. He succeeds addressing plan-
ning problems satisfactorily. Planning is in fact an easy task for a human
planner, but it is still difficult to automate [56]. Viewing the stowage plan-
ning as a search process, a planner does not apply a brute search force,
exploring all possible combinations of slots. He does not need reinvent the
wheel. From experience he has collected numerous problem-solving ses-
sions. Therefore he is in position now to use this knowledge effectively. He
can select only one or a few planning concepts which may likely lead to
an acceptable solution. In fact he reuses old solution concepts, with few
modifications or improvements when necessary.
The strategy must be pragmatic, in order to make it useful and appli-
7
Survey at a Hamburg-based shipping company and two container terminals in Hamburg
and Bremerhaven, 2001.
5.3. Case-based stowage planning 176
Procedure
Input: Input:
Loading List Arrival cond.
Stowage planning
modules
Stability
User creates
Check it
Strength a stowage plan
Visibility
No
OK?
Hazardous
cargoes
Yes
User uses it
for current
planning task
Input: Input:
Loading List Arrival cond.
Retrieval Casebase
The system
proposes
Stowage planning
few plan concepts
modules
User chooses
Trim & draught a concept
Stability
User modifies
Check it the plan guided
Strength by the chosen case
Visibility
No
OK?
Hazardous
cargoes The system stores
Yes the plan into
Casebase
with its corresp.
remarks
User uses it
for current
planning task
similarity measurement
graphical numerical
similarity similarity
membership
value
1.0
query
0.0
a b c case d
Criterion 2: Being within the same partbay in hold -> similar Criterion 5: Being rotated 90 degr. -> NOT similar
Criterion 3: Being mirrored against deckline -> NOT similar Criterion 6: Being rotated 180 degr. -> NOT similar
Chapter 5. Proposed Implementations
The new problem is compared with all problems stored in the casebase.
There are two types of data to be compared: numerical (CLL) and graphical
data (stowage plan). The numerical similarity value is computed using a
trapezoidal fuzzy membership function, see Figure 5.24.
To calculate the graphical similarity value the graphical properties of the
plan must be taken into account, longitudinally and transversally. The cross
section of the ship, the bay, is sub-divided into a few partbays. Containers
placed with the same partbays are considered equivalent or its similarity
value is 1.0. Containers stowed symmetrically along the center line are
considered equivalent. Figure 5.25 shows the graphical similarity criteria.
Longitudinally, the vicinity criterion plays a role, i.e. containers stowed close
to each other in longitudinal direction are considered equivalent or similar,
see Figure 5.26.
183 Chapter 5. Proposed Implementations
query
bays located
1 in the same hold
case
similar
query
bays located
2 in neighbouring holds
case
not similar
query
3 bays located
far from each other
case
not similar
Output
After computing the combined similarity values, the cases are then in-
dexed and ranked according to the similarity values. Few most similar
cases are proposed to the user. Now the user has the opportunity to choose
the solution concept which he find the most desirable or favorable one. The
chosen case will act as a guide to modify the arrival plan to the departure
plan. Its way of solving the problem can be applied, i.e. borrowed or even
copied, to solve the new planning problem.
Figure 5.29 shows the usage of the chosen stowage planning concept
to help solve the new planning problem. The upper row is the chosen case
consisting of three elements, from left to right, (a) the arrival plan prior
to loading (b) CLL and (c) the departure stowage plan. The lower rows
are of the new problem, (d) the arrival plan and (e) CLL. To create the
departure plan, the way of allocating containers, modifying an arrival plan
into a departure plan (c) can be reused, imitated or copied. A planning
problem may have more solutions; a stowage planning task may also have
more than one solutions (f) and (g).
The finished plan is then checked using available modules, e.g. stability,
strength etc, if it meets all those criteria. Both favorable and less favorable
plans can be useful for future planning sessions. All information concerning
the stability, strength etc can be included in to the remarks. In general
remark contains information on the quality of the plan either favorable, less
favorable or neutral, textually or numerically. The newly created plan is then
stored into the casebase, and it is immediately available for use in the next
planning session. This last part is the learning mechanism of the system.
The CLL is usually changing over time. It is usual too that for a particular
voyage more planning sessions are performed. The produced departure
plans may be of various qualities. Those all may contribute in enhancing
the richness of the casebase’s contents, see Figure 5.30.
5.3.4 Properties
The system is scalable. It can be applied to any sizes of vessels and for
any routes with a minimum adaptation effort. The system has the capability
of learning. Every newly created stowage plan is stored into the casebase.
This increases automatically the capability of the system to propose so-
lution concepts to the user. This approach assists the human planner in a
natural way. It enhances the remembering capability of a human, by storing
and retrieving cases accordingly whenever required. The system respects
subjective preferences which mark the daily stowage planning tasks by pro-
viding the planner to choose a solution concept which best suits him.
The system is applicable for addressing stowage planning problems
of the whole ship or a part of it, for example for refrigerated containers.
A central aspect of the method is its capability to recognize and retrieve
similar problems stored in the casebase.
187 Chapter 5. Proposed Implementations
Similarity 1.0
value 0.9
0.8
0.7
0.6
0.5
New planning problem
0.4
Positive: guide
0.3
Neutral : additional info and limitations
0.2 Negative: warning, to be avoided
0 1 2 3 4 5 6 7 8 9 10
Voyage number
Conclusion
189
190
Outlook
Forgotten research agenda. The result of this research may concern three
areas: shipping research, industry and education. The findings show us
that shipping practice involves aspects, such as imprecision, intuition, ex-
perience and task-orientedness, which seem to be forgotten by today’s re-
search agenda in shipping. Today’s research agenda seems to be marked
with its preference for addressing macro-oriented problems and its prefer-
ence for having process-oriented views on addressing problems.
More research addressing living problems in daily shipping practice is
necessary. It is also of use for having a wider methodological horizon.
Learning from nature proves to have contributed very much to solve many
of our problems. This includes learning from failures and successes, and
from ourselves, as a human. Having an eye on Artificial Intelligence in
research agenda may contribute towards better developing useful tools for
shipping.
Towards implementation. The concept of compatibility can be imple-
mented in addressing various tasks in shipping and related areas, such
as ship design or fleet scheduling. In areas where knowledge can not be
obtained and formulated satisfactorily, recycling (reusing) past solution con-
cepts and modifying them where necessary are usually an effective way to
address problems.
AI in shipping education. Better understanding soft aspects such as
imprecision and intuition is not something taken for granted in the educa-
tion of shipping, logistic, traffic engineering and marine technology. Intro-
ducing the AI views to them, beside mathematics, statistics and operation
research, would contribute very much for understanding the complexity of
real problems.
Bibliography
191
Bibliography 192
[17] Yuval Davidor and Moshav Avihail. Method for Determining a Stowage
Plan. page 11, 15 1998. US Patent number 6809 489.
[19] Hubert L. Dreyfus and Stuart L. Dreyfus. Mind over Machine - The
Power of Human Intuition and Expertise in the Era of Computer. Basil
Blackwell, 1986.
[37] George J. Klir, Ute H. St.Clair, and Bo Yuan. Fuzzy Set Theory -
Foundations and Applications. Prentice-Hall International, 1997.
[50] New Jersey Maritime Research Inc. Chartering Annual 1997. Maritime
Research Inc., New Jersey, 1998.
[51] New Jersey Maritime Research Inc. Chartering Annual 1998. Maritime
Research Inc., New Jersey, 1999.
195 Bibliography
[52] New Jersey Maritime Research Inc. Chartering Annual 1999. Maritime
Research Inc., New Jersey, 2000.
[53] William Mark. Machine Learning Methods for Planning, chapter Adapt-
ing Plan Architectures, pages 435 – 465. Morgan Kauffmann, 1993.
[56] Steve Minton and Monte Zweben. Machine Learning Methods for
Planning, chapter Learning, Planning and Scheduling: An Overview,
pages 1–27. Morgan Kaufmann, 1993.
[63] Kaj Pineus. Ship’s Value. Lloyd’s of London Press Ltd, 1986.
[67] Denis Riordan and Bjarne K. Hansen. A Fuzy Case-Based System for
Weather Prediction. Engineering Intelligent Systems, 10(3):139–146,
September 2002.
[68] Timothy Ross. Fuzzy Logic and Its Engineering Applications. McGraw-
Hill Inc., 1995.
[71] Herbert A. Simon. The sciences of the artificial. MIT Press, Cam-
bridge, Massachusetts, second edition, 1981. iindexSimon, H. A.
[77] Heike Vogeley, Peter Horstkorte, Haidi Qiu, and Frank Laue. Opti-
mierung der gewichtsverteilung zur Minimierung der globalen Längs-
festigkeitsbelastung. 26 2003.
[80] Niko Wijnolst and Frans Waals. Shipping Industry Structure. Delft
University Press, 1999.
[81] Niko Wijnolst and Tor Wergeland. Shipping. Delft University Press,
1996.
[82] I.D. Wilson, P.A. Roach, and J.A. Ware. Container stowage
pre-planning: using search to generate solutions, a case study.
Knowledge-Based Systems, (14):137–145, 2001.
199
Appendix A
MV Lucy Oldendorff
201
202
speed/consumption:
abt 14 kn on abt 19,5 mts heavy fuel oil (180 cst RME25) plus abt
1,5 mts MDO (DMB), (DMB),
vessel burns MDO with main engine whilst on roads, rivers, anchor-
ages, in ports, canals etc.
port consumption:
abt 2,6 mts MDO with cranes working,
abt 1,5 mts MDO without cranes working
gear:
4 x Mitsubishi electro hydraulic deck cranes of
30 mts each (max jib radius 3 x 24 m + 1 x 22 m)
hoisting speed 30 mts x 18,5 m/min,
12 mts x 37,0 m/min,
bunker capacity abt 725 mts (96 %) IFO/abt 253 mts (96 %) MDO,
water ballast abt 7.458 mts, freshwater evaporator (abt 15 mts/day),
stanchion heights:
hold 1: abt 7,00 m
holds 2-4: abt 8,00 m
Glossary
205
206
• Logic: (1) the science or art of exact reasoning, or of pure and for-
mal thought, or of the laws according to which the processes of pure
thinking should be conducted; the science of the formation and ap-
plication of general notions; the science of generalization, judgment,
classification, reasoning, and systematic arrangement; correct rea-
soning [1]. (2) the branch of philosophy that analyzes inference. (3)
reasoned and reasonable judgment; "it made a certain kind of logic".
(3) the principles that guide reasoning within a given field or situation.
(4) a system of reasoning [3]
• Rational: (1) relating to reason; not physical; mental. (2) Having rea-
son, or the faculty of reasoning; endowed with reason or understand-
ing; reasoning. (3) Agreeable to reason; not absurd, preposterous,
extravagant, foolish, fanciful, or the like; wise; judicious; as, rational
conduct; a rational man [1]. (4) consistent with or based on or using
209 Appendix B. Glossary
1 AA Always Afloat
AAAA Always Accessible Always Afloat
AAOSA Always Afloat or Safe Aground. Condition for a ves-
sel whilst in port
AARA Amsterdam-Antwerp-Rotterdam Area
ABAFT Toward the rear (stern) of the ship. Behind.
ABOARD On or within the ship
ABOVE DECK On the deck (not over it - see ALOFT)
ABT About
ADCOM Address Commission
ADDENDUM Additional chartering terms at the end of a charter
party
AFSPS Arrival First Sea Pilot Station (Norway)
AFFREIGHTMENT The hiring of a ship in whole or part
AFT At or towards the stern or rear of a ship
AGROUND Touching or fast to the bottom
1
http://www.uq.net.au/~zzksteph/, downloaded 11 January 2005 and http://www.m-i-
link.com/dictionary/acronym_IJK.asp, downloaded on 16 January 2005
211
212
BI Both Inclusive
BIMCO The Baltic and International Maritime Council
BL Bale
BL (Bill of Lading) A document signed by the carrier
which acts as a Contract of Affreightment, a receipt
and evidence of title to the cargo.
BM Beam
BN Booking Note
BOB Bunker on Board
BOFFER Best Offer
BOW The forward part of a ship
BROB Bunkers Remaining on Board
BROKERAGE Percentage of freight payable to broker (by owners
in c/p’s) or applicable to sale or purchase
BSS Basis
BSS 1/1 Basis 1 Port to 1 Port
BT Berth Terms
BULKHEAD A vertical partition separating compartments
BUNDLING This is the assembly of pieces of cargo, secured into
one manageable unit. As a rule of thumb it is to
present cargo at a size easily handled by a large (20
tonne) fork lift.
BUNKERS Name given for vessels Fuel and Diesel Oil supplies
(Originates from coal bunkers)
BUOY An anchored float used for marking a position on the
water or a hazard or a shoal and for mooring
BWAD Brackish Water Arrival Draft
CAF Currency Adjustment Factor
CBM Cubic Metres
CBFT (or CFT) Cubic Feet
CFR (or C&F) Cost and Freight
CHART A map used by navigators
CHOPT Charterers Option
CHTRS Charterers
CIF Cost, Insurance & Freight. Seller pays all these
costs to a nominated port or place of discharge.
CKD Completely knocked down
COA Contract of Affreightment - Owners agree to accept
a cost per revenue tonne for cargo carried on a spe-
cific number of voyages.
CIP Carriage and Insurance paid to...
COACP Contract of Affreightment Charter Party
COB Closing of Business
COBLDN Closing of Business London
214
Lebenslauf
A USBILDUNG
1984-1985 Studium Schiffs- und Meerestechnik an der Technis-
chen Universität Sepuluh Nopember in Surabaya.
1985-1986 Sprachkurs Niederländisch in Jakarta und Utrecht.
Stipendium von der N UFFIC /Niederlande.
1986-1993 Studium Seeverkehr an der Technischen Universität
Delft, Niederlande, Abschluß: ir. / MSc. Stipendium
von der N UFFIC/Niederlande und von der B APPE -
NAS /Indonesien
1999 Sprachkurs Deutsch in Surabaya und in Dresden.
1999-2005 Promotion am Fachgebiet Seeverkehr an der Tech-
nischen Universität Berlin. Stipendium vom DAAD
(1999-2001).
B ERUFSERFAHRUNG
1993 Mitarbeiter, Wijsmuller Engineering, IJmuiden,
Niederlande. Projekt: “Kohletransport in Ostkali-
mantan mit Bargen”.
225
226
D IVERSES
EDV Matlab, Pascal, Postgresql, Qt, C++, Python.
Sprachen Javanisch (Muttersprache), Indonesisch, Englisch,
Deutsch, Niederländisch.
Interessen (a) Seeverkehr: Planung, Management und Inno-
vation (b) Anwendungen der künstlichen Intelligenz,
vor allem aus den Bereichen Fuzzy Logik und Fall-
basiertem Schließen.
Freizeit Spielen mit Kindern, Debian Linux, Jogging und Tis-
chtennis.
PATENT
227 Appendix D. Lebenslauf
W ISSENSCHAFTLICHE A RBEITEN
1 Setyo Nugroho. Addressing Management of Uncertainties in Ship
Chartering - A Fuzzy Logic Approach. X Spanish Conference on Fuzzy
Logic and Technologies, 20-22.September 2000, Sevilla.
2 Setyo Nugroho, Jens Heyer and Peter Horstkorte. Conceptual Design
of Box-Type Barges. Technische Universität Berlin, 2002. Unveröffen-
lichter C OMSTAU -Projektbericht.
3 Setyo Nugroho. Voyage Scenario. Technische Universität Berlin, 2002.
Unveröffenlichter C OMSTAU -Projektbericht.
4 Csaba Piller and Setyo Nugroho. Stauung von Gefahrgut. Technische
Universität Berlin, 2002. Unveröffenlichter C OMSTAU -Projektbericht.
5 Horst Linde and Setyo Nugroho. Assessing A Stowage Plan.
Technische Universität Berlin, 2003. Unveröffenlichter C OMSTAU -
Projektbericht.
6 Setyo Nugroho. Case-Based Stowage Planning: A Proposal.
Technische Universität Berlin, 2003. Unveröffenlichter C OMSTAU -
Projektbericht.
7 Setyo Nugroho. Preliminary Experiments with A Case-Based Stowage
Planning System. Technical report, Technische Universität of Berlin,
2003. Unveröffenlichter C OMSTAU -Projektbericht.
8 Setyo Nugroho. Case Representation of a Round Voyage in a Case-
Based Stowage Planning System. Presentation at the Technische
Universität Hamburg-Harburg. June 11, 2003. Unveröffenlichter
C OMSTAU -Projektbericht.
9 Setyo Nugroho. Case-Based Stowage Planning for Container Ships:
An Outline. Technical report, Technische Universität Berlin, 23 February
2004. Unveröffenlichter C OMSTAU -Projektbericht.
10 Setyo Nugroho. Case-Based Stowage Planning for Container Ships.
International Logistics Conference, 2-3.Dezember 2004, Izmir, Turkey.
11 Setyo Nugroho. Case-Based Stowage Planning System. PORTS 2005
Conference, 20-22.April 2005, Barcelona, Spain. Maritime Engineering
and Ports III, Volume 80 of WIT Transactions on the Built Environment.
ISSN 1743-3509. WIT Press.