International Journal of Management Science and

Engineering Management

ISSN: 1750-9653 (Print) 1750-9661 (Online) Journal homepage: https://www.tandfonline.com/loi/tmse20

Application of decision theory and bee-inspired

method to railway system route optimization

Kah Huo Leong, Chen Wang, Hamzah Abdul-Rahman, Kamran Shavarebi,

Patrice Boursier & Siaw-Chuing Loo

To cite this article: Kah Huo Leong, Chen Wang, Hamzah Abdul-Rahman, Kamran Shavarebi,
Patrice Boursier & Siaw-Chuing Loo (2019): Application of decision theory and bee-inspired method
to railway system route optimization, International Journal of Management Science and Engineering
Management, DOI: 10.1080/17509653.2019.1604190

To link to this article: https://doi.org/10.1080/17509653.2019.1604190

Published online: 09 May 2019.

Submit your article to this journal

Article views: 45

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tmse20
INTERNATIONAL JOURNAL OF MANAGEMENT SCIENCE AND ENGINEERING MANAGEMENT
https://doi.org/10.1080/17509653.2019.1604190

Application of decision theory and bee-inspired method to railway system route

optimization
Kah Huo Leong a, Chen Wang b
, Hamzah Abdul-Rahman a
, Kamran Shavarebia, Patrice Boursierc
and Siaw-Chuing Loo d
a
Faculty of Science Technology Engineering and Mathematics (STEM), International University of Malaya-Wales, Kuala Lumpur, Malaysia;
b
College of Civil Engineering, Huaqiao University, Xiamen, China; cSchool of Computing and IT, Taylor‘s University, Selangor, Malaysia;
d
Centre of Building, Construction and Tropical Architecture, Faculty of Built Environment, University of Malaya, Kuala Lumpur, Malaysia

ABSTRACT ARTICLE HISTORY

Route planning for multiple destinations via a railway system (RS) is challenging, especially in a Received 17 July 2018
complex network with hundreds of stations and interchanges, resulting in a railway traveling Accepted 3 April 2019
salesman problem (RTSP), which is a variant of the traveling salesman problem (TSP). Limited KEYWORDS
attention has been devoted to solving the RTSP, despite the increase in the number of people Route optimization;
using RSs and the added complexity of network expansions. An optimization algorithm and Travelling salesman
mathematical model based on the foraging behavior of bees and decision theory were used in problem; railway system;
this paper to identify the optimum RS route to multiple destinations before returning to the first bee algorithm; Decision
station. Data collected from RSs in Japan and Malaysia were used to create 200 test cases (100 cases theory; railway traveling
with each dataset), and the results were compared to specific solutions and the official optimum salesman problem; route
route planner to prove that the model is a promising approximation method. The solutions were management; swarm
intelligence
evaluated and verified by comparing results from another 200 brute-force cases generated by the
Wiley TSP Solver. The results obtained from the experiments prove the reliability and capability of
the route planning and optimization solutions for RSs with different complexities and in different
environments without the assistance of information technology.

1. Introduction in practice, thereby making the ideas tangible and identify-

ing their limitations for multiple stations in a complex net-
This paper presents solutions that do not require the use of a
work. The study contributes to the re-contextualization of
computer for three different RS routing problems: (1) the
operations research for solving RS routing problems (i.e. the
RTSP, (2) the optimum route in a complex network and (3)
application of bee swarm intelligence in a new human
the TSP. The solutions were inspired by the bee-foraging
context and showing the applicability of the algorithm and
technique and decision theory. In the first experiment, the
mathematical model in a new situation) without the use of
Malaysia RS was used to examine the effectiveness and effi-
information technology. It also aims to provide solutions
ciency of the model in solving the RTSP. The Tokyo Metro
that will aid tourists, project planners and business practi-
(TM) RS was used as a reference in the second experiment to
tioners in planning better RSs.
assess the practically of the solutions in the identification of the
optimum route to a station in a complex network that con-
nects more than 300 stations. Finally, the solutions were used 2. Literature review
to solve 200 TSP cases created by the Wiley TSP Solver. The
The railway traveling salesman problem (RTSP) represents a
results were compared with the exact solutions to verify the
practical extension of the classical TSP by considering a rail-
solutions developed in this research study. Three experiments
way network and train schedules (Hu & Raidl, 2008). In it, a
were conducted to demonstrate the practicality, capability,
salesman may use a railway network to visit various cities to
efficiency and effectiveness of the proposed solutions, which
conduct businesses, starting and ending at the same station
do not require the use of a computer, to route planning
and traveling via the optimal route (Hadjicharalambous, Pop,
problems. The case studies in the experiments will be bene-
Pyrga, Tsaggouris, & Zaroliagis, 2007). Unlike the general
ficial to RS users in Malaysia and Tokyo, especially during the
concept of the TSP, the RTSP allows nodes to be visited
Tokyo Olympics in 2020.
more than once because otherwise it would be inconvenient
Moreover, the developed solution serves as a future
to restrict the usage of some bottleneck stations and to force
reference for research on the use of bee swarm intelligence,
the salesman to take an alternative route (Hu & Raidl, 2008).
collective intelligence, knowledge discovery in databases
The goal of solving the RTSP is to determine which train
(KDD) and decision theory in solving the RTSP problem.
connections would minimize the salesman’s travel time using
This research is also a confirmation and expansion of an
a railway network (Kah Huo, Abdul-Rahman, Wang, & Siaw-
existing bee-inspired optimization algorithm and the use of
Chuing, 2017) and result in the minimum travel cost (Matai,
decision theory in operations research and project manage-
Singh, & Mittal, 2010). Building on the results of the authors‘
ment. This research is the first to implement a theoretical
previous research into bee-inspired railway-system traveling-
principle and framework by showing how it can be applied

CONTACT Kah Huo Leong [email protected] Faculty of Science Technology Engineering and Mathematics (STEM), International University of
Malaya-Wales, Kuala Lumpur 50408, Malaysia
© 2019 International Society of Management Science and Engineering Management
2 K. H. LEONG ET AL.

salesman solutions, the solutions introduced in this paper were updating (RABC) produced stronger exploration ability,
developed by considering common constraints that will better effect on convergence speed, optimization precision,
improve the traveling time via railway system with the intent and keeping good robustness validity (Sun, Chen, & Zhang,
to make the solutions more practical and applicable to RS 2018). This algorithm was able to find the optimum solution
users worldwide (Leong, Abdul-Rahman, Wang, Onn, & faster as it can explore further in searching processes.
Loo, 2016; Wang, Leong, & Abdul-Rahman, 2017). Recently, in 2019, a hybrid of Queue Ant Colony
Optimization and ABC optimization (Ant-Bee) was used
to optimize the task in a mobile cloud computing environ-
2.1. Swarm intelligence and decision theory
ment (Sundararaj, 2019), genetic algorithm with BA in
Swarm-based algorithms are population-based algorithms mobile robot path optimization (Carballo, Morales, &
that can be used to provide low-cost, fast, and robust solu- Trujillo-Romero, 2017) and the combination of ABC with
tions to complex problems (Panigrahi, Shi, & Lim, 2011). Modified Choice Function produced an effective model to
Swarm intelligence has been used by many researchers to solve the TSP problem (Choong, Wong, & Peng Lim, 2019).
solve complex optimization problems (Yuce, Packianather, In addition to finding the optimum solution, BA can be a
Mastrocinque, Pham, & Lambiase, 2013), Over the years part of a process. In classification and prediction, BA is used
various bee-inspired solutions have been proposed to in the feature selection process to choose the best features
solve complex optimizations problems especially in engi- or attributes to enhance the performance of classifying data
neering, operations, and management-related fields and prediction. This was used by several researchers to help
(Nikolić & Teodorović, 2013). For example, the Bee them classify the data efficiently. For example, implement-
Algorithm (BA) was inspired by the use of stochastic search ing ABC as a feature selection algorithm (ABC-FS) on Self-
when bees forage food (Nasrinpour, Bavani, & Teshnehlab, Care Activities Dataset based on ICF-CY (SCADI) that has
2016). Furthermore, Marriage Honey Bees Optimization 206 attributes (Kele & Keles, 2018). It gives higher accuracy
(MBO) and Honey-Bees Mating Optimization (HBMO) than using other feature selection method such as Gain
were based on the mating of bees and their breeding beha- Ration, Info Gain and Chi-Square. It has also been used to
viors (Yuce et al., 2013). Bee Colony Optimization (BCO) find the important concrete components to predict the
and Artificial Bee Colony (ABC) optimization are two of the strength of the concrete (Keleş, Keleş, & Kiliç, 2018).
most well-known bee-inspired solutions that have been Decision-making is the process of making a choice from a
proven to be efficient in solving optimization problems number of alternatives to achieve the desired result (Eisenfuhr,
that have a large dataset (Nasrinpour et al., 2016; Yuce et 2011). Thus, decision theory (DT) is generally concerned with
al., 2013). the problem of making complex decisions. This theory dis-
Instead of using single bee-inspired algorithm alone, it cusses how humans combine and compare the information
can be combined with other algorithms or methods. In from certain problems and make the best decision (Berger,
2017, researchers started to apply this hybridization method 2013). According to Zhu, Crainic, and Gendreau (2014), deci-
to enhance the performance of BA and to make it applicable sions may be taken at different levels to consider the scope,
in various problem areas. For example, a new algorithm, the time horizon, and cost of the project. Figure 1 demonstrates
Directed Bee Colony (DBC) optimization algorithm, was how the bee-foraging process is related to DT and how this
created to solve the Multidimensional 0–1 Knapsack concept is used to develop the model.
Problem (MKP) by using a multi-agent environment and
honey bee swarms‘ techniques (Bole & Kumar, 2017). This
3. Methodology
algorithm is found to be more robust and produces a super-
ior and accurate solution than the traditional methods. In the initial phase of the research, a framework based on
Additionally, a hybridization of ABC and random location the bee-foraging concept and decision theory principles was

Figure 1. Integration of decision theory and bee foraging method.

 INTERNATIONAL JOURNAL OF MANAGEMENT SCIENCE AND ENGINEERING MANAGEMENT 3

developed. After verifying the framework, a new method Table 2. BIM table structure (template).
was introduced to solve the research problems by employ- S SB IC EB DF F UB
ing a table template designed. Three experiments were
designed by using different settings to evaluate the efficiency
of the solutions in solving routing problems. where all the data collected is analyzed. The next level
Experiment 1 was designed to test how efficient the (DF) is inspired by the way bees share information on
solution could be used in solving the RSTP. The Malaysia the ‘dance floor’ to communicate their findings to others.
RS dataset was used in the 100 test cases created, and the This level is considered the decision-making phase, where
results were compared with the exact solutions generated by the unemployed bee (UB) will turn into a follower (F) after
using a brute force technique. The cases in Experiment 2 choosing the route shared on the dance floor and before
were designed by using the TM railway dataset to evaluate moving to the final level and repeating the cycle. The
how the solutions identify the optimum route to a single template of the BIM structure (Table 2) will be used in
station in a complex network. The results were compared every cycle of the process, with one table assigned to one
with the optimum solutions recommended by the TM offi- station stop. All the output in the UB column at the end of
cial planner. In Experiment 3, the TSP problem-solving tool the cycle will be merged as the optimum path to the
from Wiley (a learning portal) was used to create 200 TSP desired stations.
cases to assess the efficiency of the solutions in solving
common TSP cases without using a computer.
3.2. Experiments
Figures 2–5 show how the BIM was applied to identify the
3.1. Proposed algorithm
optimum route to complete a tour. Column SB is used to
Table 1 shows the parameters used and the bee-inspired locate nearest G along the same line as the starting station,
algorithm developed to solve the research problems. The Sn. IC column is used to locate nearest interchange if G is
flowchart and processes of the algorithm are presented in not found in the same line as S. Column EB is used to
Appendix A. compare all possible routes to G and record the shortest
The algorithm is used as the main reference in develop- route identified in column F. Travel time to the station
ing a novel bee-inspired method (BIM), where a standard recorded will be recorded in column T. All the records in
table structure could be used to solve the RS routing pro- column S will be used to generate the optimum route to all
blems (Table 2). The verified method is explained in desired stations in the tour as presented below. ‘X‘ in the
Table 3. table shows the desired destination, G could not be found in
Level 1 (SB) works as a scout bee, where possible routes the process. The circle shows the station involved in the
to location G will be explored. Level 2 (EB) is the phase experiment and the arrow shows the route chosen.

Table 1. Proposed bee algorithm to solve routing problems.

Parameters:
S as starting point,
G as desired destinations,
T[] as total traveling time,
IC as interchange station,
G[] as temporary storage of G found and analyzed,
n as number of possible route from S or IC,
m as number of possible route to identified G,
Z as number of G in the tour,
r as possible route to G from S or IC

[Unemployed bee]
The following steps will be repeated for Z times (i = 0; i ≤ Z)

[Scout]
Check line and station type of S
If (IC) Check number of possible route, n
If(n > 6){Explore partial available routes for G} because the exploration to the neighborhood routes can be too large and cannot be completely explored
with n acceptable computational effort (note that the number of variables Xi is the exponential in the number of routes), else {Explore all r}

[Employed bee] – share knowledge, backward and forward passes

If(G found from r){check how many G found and how many r identified that can reach any G, m -> CA1}
else {transit to nearest IC and set IC as S}
Check G in the r connecting S
if (G located){Return to S and check any nearer G via IC in between -> M1} else {Store traveling time, T[], set G as new S and store in G[]}
End
Comparison and analysis 1 (CA1), Calculate time difference to G, T via different routes, Ts
Calculate difference number of station passed before reaching G via 2 different routes, Ns
If Ns>Ts Add 1 minute per station passed as stopping time to T as new temporary time for comparison, t
Select the r with the shortest t, Store traveling time, T[], set G as new S and store in G[]
Else Select r with the shortest T from S to IC to G, denote as T, Store traveling time, T[], set G as new S and store in G[]
End

[Onlooker]
if(Z times (i ≥ Z-1){Display G[]}else{Keep exploring till located all G -> B1}
4 K. H. LEONG ET AL.

Table 3. Table to locate optimum route in RS by using the BIM template.

Level 1: SB Level 2: EB Level 3: DF
Starting station (Locate G) IC (Analyze Gs) (Compare) Chosen Route: F Reset: UB
S1 = S1, Check if any G found Other Gs found connecting to Choose the optimum and G chosen Set G to next S
Check i < Z IC before the identified G drop others
SB: Scout bee | EB: Employed bee | DF: Dance floor | F: Follower | UB: Unemployed Bee

Figure 2. Generating solution by using BIM in Experiment 1.

Figure 3. Solution generated by using BIM in Experiment 1.

3.2.1. Sample case in Experiment 1 3.2.2. Sample case in Experiment 2

Starting station, S = Taman Jaya, Desired destinations, Starting station, S = Kamiyacho, Desired destination,
G = KLCC,Bandaraya,Bank Negara G = Sakuradamon
Optimum route = Taman Jaya -> KLCC -> Bandaraya -> Optimum route = Kamiyacho – > Kasumigaseki – >
Bank Negara -> Taman Jaya Hibiya/Yurakucho – > Sakuradamon – > Kamiyacho

Total Travel time ¼ T1 þ T2 þ T3þT4 Total travel time ¼ T1 þ T2

¼ 1036þ332þ795þ1049 ¼ ð180 þ 60 þ 300 þ 120Þþ660
¼ 3212seconds ¼ 1320seconds
 INTERNATIONAL JOURNAL OF MANAGEMENT SCIENCE AND ENGINEERING MANAGEMENT 5

Figure 4. Generating solution by using BIM in Experiment 2.

Figure 5. Solution and total traveling time identified by using BIM in Experiment 1.

3.2.3. Experiment 3 – TSP possible routes, in Process [15] the travel distance is
Two hundred cases with different number of vertexes and calculated. Last, but not least, in Process [17] the shortest
distance is created by using the Wiley TSP tool. The cases travel distance is chosen and set as a new starting point S.
are picked randomly from 5 to 9 vertexes. We have started Since the shortest travel time is traveling from ‘a’ to ‘j’,
from 5 vertexes because lesser than that are considering as which is 110s, ‘j’ is set as a new starting point, S2 with
simple cases and can be easily see by eyes. In other hand, 110s as its travel time. The processes are repeated by adding
the cases will stop at 9 vertexes because brute force algo- 1 to i for every loop and checking the condition where i is
rithm from the tool can only generate up till 9 vertexes that less than Z. Process [14]–Process [17] will begin when the
was also supported by research done by Awuni in 2014. condition returns true, and the number of possible routes,
Figure 6 shows a case tested in experiment conducted to n, decreases by 1 after a loop. At the 9th loop, i = 7 + 1 = 8
verify the solutions developed and Table 4 shows how the and it is equal as Z = 8; thus, the condition will return false
algorithm is used to solve the problem. The processes are and the system is terminated. The optimal solution
explained further in Appendix A routes = a→j→b→c→d→e→f→i→h→a and the total
This case consists of eight desired destinations, G, with S distance = 110 + 125 + 70 + 67 + 117 + 122
as the starting and ending point. The second process + 92 + 113 + 114 = 930s.
(Process [2]) is initialize i = 0, where i is the number of
loops. Initialization of the looping process is needed to
ensure that all destinations are reached before going back 4. Findings
to the starting point. Next, for Process [3] it is necessary to
check whether the number of loops (i) is smaller than Z. Planning ahead for a journey to multiple destinations is
The total number of Gs in the tour is 8, thus Z = 8, and it is essential, especially for business practitioners and tourists.
a constant. Since i = 0, which is less than Z = 8, this Although it may sound simple, without a proper planner or
condition is true. In Process [14], by referring to the dia- tool, determining the optimal route is not easy and may not
gram, a bee can move along the entire route; thus, the even serve the purpose. In addition, an unplanned journey
number of possible routes is n = 8. After picking the may lead to longer total travel time and increased travel
6 K. H. LEONG ET AL.

Figure 6. Red solution by using proposed algorithm in red.

Table 4. Application of algorithm in solving TSP-related problem.

Starting Point, S1 = a
Desired destinations, G = b, c, d, e, f, h, i, j
T[] as total travel distance where[] represents as an array
G[] as temporary storage of G found and analyze.
n as number of route from S or IC
p as number of possible route to identified G
Z as number of G in the tour
r as m as number of possible route to G from S or IC

Process [1]: Initialization

Case 1 consists of 8 desired destinations, G with s as starting and ending point.

Process [2] Initialize i = 0

Initialization of the looping process is needed to ensure all destinations are reached before going back to the starting point.

Process [3] Check i < Z is true

Number of G in the tour is 8, thus Z = 8, Since i = 0, is less than Z = 8, this condition is true.

Process [14]:
Referring to the diagram bee can move to the entire route, thus the number of possible route, n, is 8. n > 6, hence pick randomly possible routes

Process [15]: Calculate travel distance,

Travel time from a to b = 204, Travel time from a to e = 224
Travel time from a to f = 220, Travel time from a to h = 114
Travel time from a to i = 137, Travel time from a to j = 110

Process [17]:
Choose the shortest travel distance and set as new starting point S, j is set as new starting point, S2 with 110

Repeat
i = i + 1, Number of G in the tour is 8, thus Z = 8, i = 0 + 1 = 1 is less than Z = 8
Condition will return true. Referring to the diagram, bee can move to 7 possible routes. Hence n = 7, n > 6, thus pick randomly possible routes.

Calculate travel distance

Travel time from j to b = 125, Travel time from j to c = 138
Travel time from j to h = 180, Travel time from j to f = 175

Choose the shortest total distance and set as new starting point b is set as new starting point, S3 with 125
Repeat
i = i + 1, Number of G in the tour is 8, thus Z = 8, i = 1 + 1 = 2 and it is less than Z = 8
Condition will return true. Referring to the diagram, bee can move to 6 possible routes. Hence n = 6, n > 6, thus pick randomly the possible routes.

Calculate travel distance

Travel time from b to c = 70, Travel time from b to d = 104
(Continued)
 INTERNATIONAL JOURNAL OF MANAGEMENT SCIENCE AND ENGINEERING MANAGEMENT 7

Table 4. (Continued).
Travel time from b to e = 191, Travel time from b to f = 285

Choose the shortest distance and set as new starting point c is set as new starting point, S4 with 70
Repeat
i = i + 1, Number of G in the tour is 8, thus Z = 8, i = 2 + 1 = 3 and it is less than Z = 8
Condition will return true. Referring to the diagram, bee can move to 5 possible routes. Hence n = 5, n < 6, thus calculate all the possible routes.

Calculate travel distance

Travel time from c to d = 67, Travel time from c to e = 183
Travel time from c to f = 297

Choose the shortest distance and set as new starting point d is set as new starting point, S5 with 67
Repeat
i = i + 1, Number of G in the tour is 8, thus Z = 8, i = 3 + 1 = 4 and it is less than Z = 8
Condition will return true. Referring to the diagram, bee can move to 4 possible routes. Hence n = 4, n < 4, thus calculate all the possible routes.

Calculate travel distance

Travel time from d to e = 117, Travel time from d to f = 235
Travel time from d to h = 161, Travel time from d to i = 288

Choose the shortest distance and set as new starting point e is set as new starting point, S6 with 117
Repeat
i = i + 1, Number of G in the tour is 8, thus Z = 8, i = 4 + 1 = 5 and it is less than Z = 8
Condition will return true. Referring to the diagram, bee can move to 3 possible routes. Hence n = 3, n < 6, thus calculate all possible routes.

Calculate travel distance

Travel time from e to f = 122, Travel time from e to h = 215
Travel time from e to i = 126

Choose the shortest distance and set as new starting point f is set as new starting point, S7 with 122
Repeat
i = i + 1, Number of G in the tour is 8, thus Z = 8, i = 5 + 1 = 6 and it is less than Z = 8
Condition will return true. Referring to the diagram, bee can move to 2 possible routes. Hence n = 2, N < 6, thus calculate all possible routes.

Calculate travel distance

Travel time from f to h = 148, Travel time from f to i = 92

Choose the shortest distance and set as new starting point i is set as new starting point, S8 with 92
Repeat
i = i + 1, Number of G in the tour is 8, thus Z = 8, i = 6 + 1 = 7 and it is less than Z = 8
Condition will return true. Referring to the diagram, bee can move to 1 possible route. Hence n = 1, n < 6, thus calculate all possible routes.

Calculate travel distance

Travel time from i to h = 113
Choose the shortest distance and set as new starting point h is set as new starting point, S9 with 113

Repeat
i=i+1
Number of G in the tour is 8, thus Z = 8, i = 7 + 1 = 8 and it is equal as Z = 8

Condition will return false. Thus it will go to process return trip, [3.B]
Solution Obtained:
Solution Routes = a→j→b→c→d→e→f→i→h→a
Total distance = 110 + 125 + 70 + 67 + 117 + 122 + 92 + 113 + 114 = 930s

costs. The results obtained from the test cases showed how matched results obtained shows that the solutions are effi-
the proposed model was capable of generating optimal cient in solving complex railway routing problems. The
routes that were similar to exact solutions in 95 of the 100 solutions also solved 182 TSP cases generated by the TSP
cases tested. Therefore, the model may generate highly reli- tool. With accuracy of 91%, we concluded that the solutions
able results in cases with varying complexities. The findings presented in this paper are efficient in solving simple TSP
have significant implications for understanding how swarm without using any computer.
intelligence can be used to solve complex problems without Although the research has successfully demonstrated the
a super computer. The results may also serve as the founda- efficiency of the solutions developed, these results are sub-
tion for future RTSP studies. The results obtained in ject to a few limitations. External factors that may affect the
Experiment 2 shows 90% of the solutions generated reliability of the results were eliminated by defining con-
matched the optimum routes proposed by TM planner. stant values in the solutions. If external factors such as
Only 10 cases, representing 10% of the results, did not travel time between stations and congestion at stations are
match, but they still achieved the shorter travel time com- taken into consideration, the suggested optimal route gen-
pared to the planner’s results. The higher percentage of erated by the algorithm might differ from the results
8 K. H. LEONG ET AL.

generated by the exact method. This difference will affect logistics, tourism, aerospace, transportation, engineering
the overall result and cause the verification of the algorithm and manufacturing.
to fail. Non-availability of the latest travel time is another
issue that may greatly affect the accuracy of the output. A
further limitation of this research is the absence of a plat- Disclosure statement
form to test how well the swarm intelligence-based algo-
No potential conflict of interest was reported by the authors.
rithm can perform on a server and generate output to
multiple users at the same time. Since the research was
limited to a maximum of nine stops in a tour, it is not
ORCID
possible to verify and check the reliability of the algorithm
when the case becomes more complex. However, this issue Kah Huo Leong http://orcid.org/0000-0001-7266-631X
could be addressed if the solution was converted to an Chen Wang http://orcid.org/0000-0001-7892-3575
application that enables the computer to perform the ana- Hamzah Abdul-Rahman http://orcid.org/0000-0003-2177-2109
Siaw-Chuing Loo http://orcid.org/0000-0003-2528-1817
lysis. The research was not specifically designed to evaluate
how the novel solutions developed perform compared to
other methods that use a computer. As no general dataset References
for RS could be found online, benchmarking and evaluation
Berger, J. O. (2013). Statistical decision theory and Bayesian analysis.
on how effective the solutions are in solving the research
New York: Springer Science & Business Media.
problems comparing to other available methods could not Bole, A. V., & Kumar, R. (2017). Multidimensional 0–1 Knapsack using
be performed. directed bee colony algorithm. In 2017 IEEE International Conference
on Intelligent Techniques in Control, Optimization and Signal
Processing (INCOS) (pp. 1–10). doi:10.1109/ITCOSP.2017.8303080
Carballo, E. A. S., Morales, L., & Trujillo-Romero, F. (2017). Path
5. Conclusion planning for a mobile robot using genetic algorithm and artificial
bee colony. In 2017 International Conference on Mechatronics,
This study focused on the development of solutions using a Electronics and Automotive Engineering (ICMEAE) (pp. 8–12).
novel method based on the bee-foraging technique and IEEE. doi:10.1109/ICMEAE.2017.16
decision theory principles. These solutions will enable Choong, S. S., Wong, L.-P., & Peng Lim, C. (2019). An artificial bee
users to identify their ideal routes using the designed table colony algorithm with a modified choice function for the traveling
salesman problem. Swarm and evolutionary computation (Vol. 44).
template. All the solutions were verified using experiments
doi:10.1016/j.swevo.2018.08.004
with different datasets. The experiments were designed to Eisenfuhr, F. (2011). Decision making. New York, NY: Pringer.
assess the RTSP, TSP and optimum route identification Hadjicharalambous, G., Pop, P., Pyrga, E., Tsaggouris, G., & Zaroliagis,
solutions. The results demonstrate that the solutions gener- C. (2007). The railway traveling salesman problem. Optimization,
ated the optimal route in most cases, which proved the 00104, 264–275.
Hu, B., & Raidl, G. R. (2008). Solving the railway traveling salesman
efficiency of the solutions for the research problems. These
problem via a transformation into the classical traveling salesman
findings have significant implications for understanding problem. Proceedings - 8th International Conference on Hybrid
how swarm intelligence, such as in a bee colony, can be Intelligent Systems, HIS 2008 (pp. 73–77). doi:10.1109/HIS.2008.30
used to solve complex optimization and operations pro- Kah Huo, L., Abdul-Rahman, H., Wang, C., & Siaw-Chuing, L. (2017).
blems without complex solutions or tools. With the applica- Bee inspired route management approach and use of internet of
things. The Journal of Modern Project Management, 5, 2.
tion of minor modifications, the developed solutions can be
Kele, M. K., & Keles, U. (2018). Artificial bee colony algorithm for
applied to any type of RS. The experimental studies showed feature selection on SCADI dataset. UBMK 2018-3rd International
that the presented solutions were highly competitive in Conference on Computer Science and Engineering, 2019 January (pp.
comparison to the exact solutions investigated in the 463–466). doi:10.1109/UBMK.2018.8566287
research. Keleş, M. K., Keleş, A. E., & Kiliç, Ü. (2018). Prediction of concrete
strength with data mining methods using artificial bee colony as
Despite these promising results, the current bee
feature selector. In 2018 International Conference on Artificial
inspired route optimization research study is limited. In Intelligence and Data Processing (IDAP) (pp. 1–4). doi:10.1109/
the future, it would be interesting to assess how bee- IDAP.2018.8620905
inspired route optimization algorithm can be converted Leong, K. H., Abdul-Rahman, H., Wang, C., Onn, C. C., & Loo, S.-C.
to a learning machine to generate better, smarter decisions (2016). Bee inspired novel optimization algorithm and mathemati-
cal model for effective and efficient route planning in railway
by using database and web technology. Further research
system. PLoS ONE, 11(12), e0166064.
may also be conducted to identify other optimization Matai, R., Singh, S. P., & Mittal, M. L. (2010). Traveling salesman problem:
techniques that may enhance the algorithm by simplifying An overview of applications, formulations, and solution approaches.
the complex equations involved in the decision-making London, UK: InTechOpen. Retrieved from http://www.intechopen.
process. This will significantly reduce the time required com/books/traveling-salesman-problem-theory-and-applications/travel
ing-salesman-problem-an-overview-of-applications-formulations-and-
to generate ideal solutions and improve the overall perfor-
solution-approaches
mance of the system when it is used within a server setup. Nasrinpour, H. R., Bavani, A. M., & Teshnehlab, M. (2016). Grouped
Future research may also investigate the integration and bees algorithm: A grouped version of the bees algorithm. Computers,
concurrent functionality of multiple optimization methods 6(1), 5.
to solve different combinatorial optimization problems, Nikolić, M., & Teodorović, D. (2013). Transit network design by bee
colony optimization. Expert Systems with Applications, 40, 5945–5955.
such as vehicle routing, knapsacks, vertex, assignment
Panigrahi, B. K., Shi, Y., & Lim, M. H. (Eds.). (2011). Handbook of
and clique problems. Highly customizable optimization swarm intelligence: Concepts, principles and applications (Vol. 8).
solutions will be useful in related industries such as Berlin, Heidelberg: Springer.
 INTERNATIONAL JOURNAL OF MANAGEMENT SCIENCE AND ENGINEERING MANAGEMENT 9

Sun, L., Chen, T., & Zhang, Q. (2018). An artificial bee colony algo- systems. Proceedings of the Institution of Civil Engineers -
rithm with random location updating. Scientific Programming, Transport, 1–13. doi:10.1680/jtran.16.00113
2018, 1–9. Yuce, B., Packianather, M. S., Mastrocinque, E., Pham, D. T., &
Sundararaj, V. (2019). Optimal task assignment in mobile cloud com- Lambiase, A. (2013). Honey bees inspired optimization method:
puting by queue based ant-bee algorithm. Wireless Personal The bees algorithm. Insects, 4(4), 646–662.
Communications, 104(1), 173–197. Zhu, E., Crainic, T. G., & Gendreau, M. (2014). Scheduled service
Wang, C., Leong, K. H., & Abdul-Rahman, H. (2017). Bee colony network design for freight rail transportation. Operations Research,
optimisation of the travelling salesman problem in light rail 62(December), 383–400.
10 K. H. LEONG ET AL.

Appendix A. Flowcart and processes of algorithm

Process Description
1 Initialization S = starting point and ending point, desired destinations = G, station names and travel time will be stored in the list of T, n denotes as
the number of bees spread, total number of desired destinations = Z, number of possible route = m/r.
2 initialize i as zero.
3 Check conditional statement, repeat until i less than number of desired destination (i < Z).
3.A Checking line of S
3.B When all of the processes from [3] are finished, display the list of station that is stored in list T and total time travel calculated.
4 Check whether S is interchange.
5 Check number of possible routes, denote as m, Check any possible route.
6 If any possible route, then check whether m is greater than or equal to 6 lines
6.A If m ≥ 6 lines, then choose partial possible routes randomly, except for G identified
6.B If m < 6, then choose all possible routes
7 Spread n bees according to explore better routes
8 Initialize j = 0
9 Repeat until j equal to number of possible routes (j = n)
10 Check whether any G along the route
10.A Check if any G found in the previous path
11 If any G along the route, then check whether more than 1 possible route connected to any G/multiple G
11.A If more than 1 possible route connected to any G/multiple G, then calculate travel time from IC to G
11.A.1 Check number of station between IC to G/multiple G
11.A.2 Calculate travel time difference for routes, denote as Ts
11.A.3 Calculate the difference in number of station for routes, denote as Ns
11.B Choose the previous route which G is found
12 Check whether Ns > Ts
12.A.1 If Ns > Ts, then add 1 minute as train stopping time for each route
12.A.2 Calculate temporary travel time for routes from IC to G, denote as t
12.A.3 Compare the temporary travel time
12.A.4 Choose shortest temporary travel time
12.A.5 Calculate travel time (S to IC to G) being chosen, denote as T2
12.B.1 If Ns ≤ Ts, then ignored the train stopping time
12.B.2 Calculate travel time from S to IC to G, denote as T2
13 Compare T1 and T2
14 Choose shortest travel time
15 Check nearest Interchange station IC
15.A If need to move to another line, then add 5 minute as walking time, denote as T_W
15.B If NOT needed to move to another line, then neglected 5 minutes as walking time
15.C Calculate travel time from S to IC
(Continued)
 INTERNATIONAL JOURNAL OF MANAGEMENT SCIENCE AND ENGINEERING MANAGEMENT 11

(Continued).
Process Description
15.D Set IC as temporary S
16 If S is not an interchange, check whether G is found on the same line
16.A If G found on the same line as S, then calculate travel time, denote as T1
17 Check whether IC found between G and S
17.A If IC found between G and S, then check possible routes for G/multiple G in different lines
17.B.1 Record travel time, denote as Tn
17.B.2 Set new starting point as new S