ENCE 603 Management Science Applications in Project Management Lectures 5-7

ENCE 603 Management Science Applications in Project Management Lectures 5-7
Project Management LP Models in Scheduling, Integer Programming
Spring 2009 Instructor: Dr. Steven A. Gabriel www.eng.umd.edu/~sgabriel

1 Copyright 2008, Dr. Steven A. Gabriel
Outline
Project Scheduling Critical Path Method (CPM) AON and AOA methods Project Crashing Precedence Diagramming Method (PDM) Gantt Charts
Project Networks
Project activities described by a network Can use the activity-on-node (AON) model Nodes are activities, arrows (arcs) indicate the precedence relationships Could also consider the activity-on-arc (AOA) model which has arcs for activities with nodes being the starting and ending points AON used frequently in practical, non-optimization situations, AOA is used in optimization settings First AON, then AOA Main idea for both is to determine the critical path (e.g., tasks whose delay will cause a delay for the whole project)
Project Networks
Sample project network (AON) (read left to right) Dashed lines indicate dummy activities Key: Activity, Duration (days)
Network Analysis
Network Scheduling: Main purpose of CPM is to determine the critical path Critical path determines the minimum completion time for a project Use forward pass and backward pass routines to analyze the project network Network Control: Monitor progress of a project on the basis of the network schedule Take correction action when required Crashing the project Penalty/reward approach
Activity on Node (AON) Representation of Project Networks
Project Networks
A: Activity identification (node) ES: Earliest starting time EC: Earliest completion time LS: Latest starting time LC: Latest completion time t: Activity duration P(A): set of predecessor nodes to node A S(A): set of successor nodes to node A
Project Networks
In tabular form Activity Predecessor A n/a B n/a C n/a D A E C F A G B,D,E Duration 2 6 4 3 5 4 2
Sample Computations ES(A) =Max{EC(j), j in P(A)}=EC(start)=0 EC(A)=ES(A)+tA=0+2=2 ES(B)= EC(start)=0 EC(B)=ES(B)+ tB=0+6=6 ES(F)= EC(A)=2 EC(F)= ES(F)+4=6, etc. F,4 D,3 end
A,2
start
B,6
G,2
C,4
Copyright 2008, Dr. Steven A. Gabriel
E,5
8
Project Networks
Notation: Above node ES(i), EC(i), below node LS(i),LC(i) Zero project slack convention in force
0,2 A,2 4,6 0,0 start 0,0 0,4 C,4 0,4 0,6 B,6 3,9 4,9 E,5 4,9 Sample Computations LC(F) =Min{LS(i), i in S(F))}=11 LS(F)=LC(F)-tF=11-4=7 etc.
2,6 F,4 2,5 D,3 6,9 G,2 9,11 7,11 9,11 end 11,11 11,11
Project Networks
During the forward pass, it is assumed that each activity will begin at its earliest starting time An activity can begin as soon as the last of its predecessors has finished C must wait for both A and B to finish before it can start Completion of the forward pass determines the C earliest completion time of the project
A B
During the backward pass, it is assumed that each activity begins at its latest completion time Each activity ends at the latest starting time of the first activity in the project network
10
Project Networks
Note: 1=first node (activity),n=last node,i,j=arbitrary nodes, P(i)= immediate predecessors of node i, S(j)= immediate successors of node j, Tp=project deadline time
1 3 2 5 4
P(3)= {1,2} S(3)= {4,5}
Rule 1: ES(1)=0 (unless otherwise stated) Rule 2: ES(i)=Max j in P(i) {EC(j)}
i1 i i2
Why do we use max of the predecessor ECs in rule 2?

11
i3
Project Networks
Rule 3: EC(i)=ES(i)+ti Rule 4: EC(Project)=EC(n) Rule 5: LC(Project)=EC(Project) zero project slack convention (unless otherwise stated for example, see Rule 6) Rule 6: LC(Project)=Tp Rule 7: LC(j) =Min i in S(j) LS(i) Rule 8: LS(j)=LC(j)-tj
j1
Why do we use min in the successor LSs in rule 7?
j2 j3
Project Networks
Total Slack: Amount of time an activity may be delayed from its earliest starting time without delaying the latest completion time of the project TS(j)=LC(j)-EC(j) or TS(j)=LS(j)-ES(j) Those activities with the minimum total slack are called the critical activities (e.g., kitchen cabinets) Examples of activities that might have slack Free Slack: Amount of time an activity may be delayed from its earliest starting time without delaying the starting time of any of its immediate successors. FS(j)= Min i in S(j) {ES(i)-EC(j) Lets consider the sample network relative to critical activities and slack times
CPM-Determining the Critical Path AON

Step 1: Complete the forward pass Step 2: Identify the last node in the network as a critical activity Step 3: If activity i in P(j) and activity j is critical, check if EC(i)=ES(j). If yes activity i is critical. When all i in P(j) done, mark j as completed Step 4: Continue backtracking from each unmarked node until the start node is reached
CPM-Forward Pass Example AON 2,6

0,2 A,2 2,5 0,0 start 0,4 C,4
Activity A B C D E F G Predecessor A C A B,D,E Duration 2 6 4 3 5 4 2
F,4 D,3 11,11 9,11 end B,6 4,9 E,5

Notation: Above node ES(i), EC(i)
0,6
G,2
Sample Computations ES(A) =Max{EC(j), j in P(A)}=EC(start)=0 EC(A)=ES(A)+tA=0+2=2 ES(B)= EC(start)=0 EC(B)=ES(B)+ tB=0+6=6 ES(F)= EC(A)=2 EC(F)= ES(F)+4=6, etc.
15
CPM-Backward Pass Example AON

Notation: Above node ES(i), EC(i), below node LS(i),LC(i) Zero project slack convention in force
0,2 A,2 4,6 0,0 start 0,0 0,4 C,4 0,4 0,6 B,6 3,9 4,9 E,5 4,9 2,5 D,3 6,9 G,2 9,11 Sample Computations LC(F) =Min{LS(i), i in S(F))}=11 LS(F)=LC(F)-tF=11-4=7 etc.
2,6 F,4 7,11 9,11 end 11,11 11,11
CPM-Slacks and the Critical Path AON

Total Slack: Amount of time an activity may be delayed from its earliest starting time without delaying the latest completion time of the project TS(j)=LC(j)-EC(j) or TS(j)=LS(j)-ES(j) Those activities with the minimum total slack are called the critical activities. Examples of activities that might have slack Free Slack: Amount of time an activity may be delayed from its earliest starting time without delaying the starting time of any of its immediate successors. FS(j)= Min i in S(j) {ES(i)-EC(j)} Other notions of slack time, see Badiru-Pulat Lets consider the sample network relative to critical activities and slack times
CPM Analysis for Sample Network AON

Activity A B C D E F G Duration (Days) 2 6 4 3 5 4 2 ES 0 0 0 2 4 2 9 EC 2 6 4 5 9 6 11 LS 4 3 0 6 4 7 9 LC 6 9 4 9 9 11 11 TS 6-2=4 9-6=3 4-4=0 9-5=4 9-9=0 11-6=5 11-11=0 FS Min{2,2}-2=0 Min{9}-6=3 Min{4}-4=0 Min{9}-5=4 Min{9}-9=0 Min{11}-6=5 Min{11}-11=0 Critical Activity? No No YES No YES No YES
0,2 A,2 4,6 0,0 start 0,0 0,4 C,4 0,4

2,6 2,5 F,4 7,11 9,11 G,2 9,11

18
Total Project Slack TS(1)+TS(C)+TS(E)+TS(G)+TS(n)=0
0,6 B,6 3,9
D,3 6,9 4,9 E,5 4,9
11,11 end 11,11
Project Networks
When results of a CPM analysis are matched up with a calendar, then we obtain a project schedule Gantt chart is a popular way to present this schedule Using the ES times from the sample AON project network, we have the following Gantt chart (could also use latest completion times as well, extreme case when all slack times are fully used)
Project Networks
G F E D C B A
1 2 3 4 5 6 7 8 9 10 11
Days
Note, Gantt chart shows for example: Starting time of F can be delayed until day 7 (TS=5) w/o delaying overall project Also, A, D, or both may be delayed by a combined total of four days (TS=4) w/o delaying the overall project B may be delayed up to 3 days without affecting the overall project completion time Can ignore precedence arrows (better20 for large networks)
10
Activity on Arc (AOA) Representation of Project Networks
Project Networks: Activity on Arc (AOA) Representation

Node i arc(i,j)=activity(i,j) Node j
Nodes represent the realizations of some milestones (events) of the project Arcs represent the activities Node i, the immediate predecessor node of arc(i,j) is the start node for the activity Node j, the immediate successor node of arc(i,j) is the end node for the activity Want to determine the critical path of activities, i.e., those with the least slack
22
11
Activity on Arc (AOA) Representation

The early event time for node i, ET(i), is the earliest time at which the event corresponding to node i can occur The late event time for node i, LT(i), is the latest time at which the event corresponding to node i can occur w/o delaying the completion of the project Let tij be the duration of activity (i,j) The total float (slack) TF(i,j) of activity (i,j) is the amount by which the starting time of (i,j) could be delayed beyond its earliest possible starting time w/o delaying the completion of the project (assuming no other activities are delayed) TF(i,j)=LT(j)-ET(i)-tij The free float of (i,j), FF(i,j) is the amount by which the starting time of activity (i,j) can be delayed w/o delaying the start of any later activity beyond its earliest possible starting time FF(i,j) = ET(j)-ET(i)-tij
AOA Network Structure

The network is acyclic (o/w an activity would precede itself)
1 2 3
Each node should have at least one arc directed into the node and one arc directed out of the node (with the exception of the start and end nodes), why? Start node has does not have any arc into it and the end node has no arc out of it All of the nodes and arcs of the network have to be visited (that is realized) in order to complete the project, why?
24
12
AOA Network Structure

If a cycle exists (due perhaps to an error in the network construction), this will lead to cycling in the procedures More specifically, critical path calculations will not terminate Need a procedure to detect cycles in the project network (e.g., Depth-First Search method)
Rules in AOA Networks

1. Node 1 represents the start of the project. An arc should lead from node 1 to represent each activity that has no predecessors. 2. A node (called the finish or end node) representing completion of the project should be included in the network. 3. Number the nodes in the network so that the node representing the completion of an activity always has a larger number than the node for the start of an activity (more than 1 way to do this). 4. An activity should not be represented by more than one arc in the network. 5. Two nodes can be connected by at most one arc.
13
Small Sample Project

Activity Predecessor Duration A B C D E F G A C A B,D,E 2 6 4 3 5 4 2
Small Sample Project AOA
14
Using Linear Programming to Find a Critical Path

Let xj= the time that the event corresponding to node j occurs Let tij=the time to complete activity (i,j) For each activity (i,j), we know that before node j occurs, node i must occur and activity (i,j) must be completed
x j xi + tij , (i, j )
Let 1 be the index of the start node Let F be the index of the finish node (i.e., when the project is completed) LP objective function is to minimize xF-x1, i.e., the total project time
Min x5 x1 s.t. A) x2 >= x1 + 2 B) x4 >= x1 + 6 C) x3 >= x1 + 4 D) x4 >= x2 + 3 E) x4 >= x3 + 5 F) x5 >= x2 + 4 G) x5 >= x4 + 2 Variables unrestricted in sign
15

Min x5 - x1 s.t. A) x2 - x1 >= B) x4 - x1 >= C) x3 - x1 >= D) x4 - x2 >= E) x4 - x3 >= F) x5 - x2 >= G) x5 - x4 >= end free x1 free x2 free x3 free x4 free x5 2 6 4 3 5 4 2
OBJECTIVE FUNCTION VALUE VARIABLE X5 X1 X2 X4 X3

31
11.00000 REDUCED COST 0.000000 0.000000 0.000000 0.000000 0.000000
VALUE 11.000000 0.000000 6.000000 9.000000 4.000000

Min x5 - x1 s.t. A) x2 - x1 >= B) x4 - x1 >= C) x3 - x1 >= D) x4 - x2 >= E) x4 - x3 >= F) x5 - x2 >= G) x5 - x4 >= end free x1 free x2 free x3 free x4 free x5 2 6 4 3 5 4 2
ROW
SLACK OR SURPLUS A) 4.000000 B) 3.000000 C) 0.000000 D) 0.000000 E) 0.000000 F) 1.000000 G) 0.000000

32
DUAL PRICES 0.000000 0.000000 -1.000000 0.000000 -1.000000 0.000000 -1.000000
16

For each variable with zero value and zero reduced cost there is an alternative optimal solution. For each constraint with zero slack and zero dual variable there is an alternative optimal solution. For each constraint with a dual price of 1, increasing the duration of the activity corresponding to that constraint by one day will increase the duration of the project by one day. Those constraints identify the critical activities.
AOA Project Network: After-Work-Hours Chores

start 1 a c 5 f
Activity a b c d e f g h i j k l m Description Dad, Mom, and son arrive home in the same car Dad and Mom change clothes Son watches TV Mom warns up the food* Dad sets the table Son does homework** Mom fixes salad The family eats dinner Dad loads the dishwasher Mom checks sons homework Son practices (insert musical instrument name here) All go to sons basketball game All wash up and go to bed a a b b c d e,f,g h h h i,j,k l
b 2
3 e
d 4 g 6 h k 9 7
i j
8 dummy arc 10 l
dummy arc
Immediate Predecessors
11
m 12
end
*For politically correct project networks, Mom and Dad are interchangeable. ** In a perfect world, activity f precedes activity c!
17
Dummy Arcs in AOA Networks

Since activities: i (Dad loads dishwasher),j (Mom checks sons homework), and k (son practices musical instrument) all have the same predecessor activity h (family eats dinner) and the same immediate successor, activity l (go to basketball game), this would mean 3 parallel arcs between nodes 7 and 10 An activity network allows only one arc between any two nodes so nodes 8 and 9 are drawn and connected to node 10 via dummy arcs
Finding the Critical Path in an AOA Project Network for Introducing a New Product
Activity a b c d e f Description Train workers Purchase raw materials Produce product 1 Produce product 2 Test product 2 Assemble products 1 and 2 into new product 3 Immediate Predecessors a,b a,b d c,e Duration (Days) 6 9 8 7 10 12
a 1 b 2
3 d
f 5 6
e dummy arc 4
18
Finding the Critical Path in an AOA Project Network for Introducing a New Product min x6-x1
s.t. x3-x1>=6 ! arc (1,3) Why variables free (i.e., not x2-x1>=9 ! arc (1,2) necessarily nonnegative)? x5-x3>=8 ! arc (3,5) x4-x3>=7 ! arc (3,4) When ok, when not? x5-x4>=10 ! arc (4,5) x6-x5>=12 ! arc (5,6) Excel version of this x3-x2>=0 ! arc (2,3) end LP? ! could have variables free or not !free x1 x2 x3 x4 x5 x6
LP OPTIMUM FOUND AT STEP 1 OBJECTIVE FUNCTION VALUE 1) 38.00000
Project completed in 38 days

REDUCED COST 0.000000 0.000000 LP will have many alternate optima all with 38 0.000000 days. In general, the value of xi in an optimal 0.000000 0.000000 solution may assume any value between ET(i) and 0.000000
VARIABLE VALUE X6 38.000000 X1 0.000000 X3 9.000000 X2 9.000000 X5 26.000000 X4 16.000000
LT(i).
ROW 2) 3) 4) 5) 6) 7) 8)
SLACK OR SURPLUS DUAL PRICES 3.000000 0.000000 0.000000 -1.000000 Critical path goes from start to finish node in 9.000000 0.000000 which each arc corresponds to a constraint with dual 0.000000 -1.000000 0.000000 -1.000000 price=-1, i.e., 1-2-3-4-5-6 is a CP (more on dual 0.000000 -1.000000 prices later...) 0.000000 -1.000000 38
19
For each constraint with a dual price of 1, increasing the duration of the activity corresponding to that constraint by delta days will increase the duration of the project by delta days This assumes that the current vertex remains optimal Now we consider a time-cost tradeoff approach to scheduling
Project Crashing in Activity on Arc (AOA) Project Networks
20
Project Crashing and Time-Cost Analysis, Sample Data

Project Duration T=11 T=10 T=10 T=10 T=9 T=9 T=9 T=9 T=8 T=8 T=8 T=7 Crashing Strategy S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 Description of Crashing Activities at normal duration Crash F by 1 unit Crash C by 1 unit Crash E by 1 unit Crash F and C by 1 unit Crash F and E by 1 unit Crash C and E by 1 unit Crash E by 2 units Crash F, C, and E by 1 unit Crash F by 1 unit, E by 2 units Crash C by 1 unit, E by 2 units Crash F and C by 1 unit, and E by 2 units
Total Cost $2,775 $2,800 $3,025 $2,875 $3,050 $2,900 $3,125
If c crashable activities, there are 2c possible crash strategies, why? Suppose we can crash 6 of the 7 activities 26=64 possible crash strategies There are 12 of the 64 strategies shown here
$2,975 $3,150 $3,000 $3,225 $3,500
Project Crashing and TimeCost Analysis
21
Project Crashing and Time-Cost Analysis A Specific Example

Define the variables: A= # of days by which activity a is reduced (unit cost =$10) B= # of days by which activity b is reduced (unit cost =$20) C= # of days by which activity c is reduced (unit cost =$3) D= # of days by which activity d is reduced (unit cost =$30) E= # of days by which activity e is reduced (unit cost =$40) F= # of days by which activity f is reduced (unit cost =$50) We have the following LP
Project Crashing and Time-Cost Analysis -An Example

min 10A+20B+3C+30D+40E+50F s.t. A<=5 B<=5 C<=5 D<=5 E<=5 F<=5 x3-x1+A>=6 ! arc (1,3) x2-x1+B>=9 ! arc (1,2) x5-x3+C>=8 ! arc (3,5) x4-x3+D>=7 ! arc (3,4) x5-x4+E>=10 ! arc (4,5) x6-x5+F>=12 ! arc (5,6) x3-x2>=0 ! arc (2,3) x6-x1<=25 ! at most 25 days end ! could have variables free or not !free x1 x2 x3 x4 x5 x6
Excel version of this LP?
44
22
Project Crashing and Time-Cost Analysis -An Example

9 0 1 b 2 dummy arc 9 4 0 1 b 2 dummy arc 4 4 6 a 3 d e c 4 16 a 3 d e c 26 5 f 38 6
Normal AOA network Total project time=38 days CP=1-2-3-4-5-6 Numbers by nodes are solution
13 5
25 6
Crashed AOA network Total cost = $390 Total project time=25 days CPs=1-2-3-4-5-6, 1-3-4-5-6 Numbers by nodes are solution Crash variables: A=2,B=5,C=0,D=5,E=3,F=0
Precedence Diagramming Method in Activity on Arc (AOA) Project Networks

23
Precedence Diagramming Method (PDM)

Normal CPM assumptions are that a task B cannot start until its predecessor task A is completely finished PDM allows activities that are mutually dependent to be performed partially in parallel instead of serially The usual finish-to-start dependencies are relaxed so that the performance of the activities can be overlapped The result is that the project schedule can be compressed (like project crashing in that sense)

The time between the finishing or starting time of the 1st activity and the finishing or starting time of the 2nd activity is called the lead-lag requirement between the two activities Four basic lead-lag relationships to consider: Start-to-Start Lead (SSAB) This specifies that activity B cannot start until activity A has been in progress for at least SS time units SS Example?
A
1.
B
24

2. Finish-to-Finish Lead (FFAB) This specifies that activity B cannot finish until at least FF time units after the completion of activity A Example?
B FF

3. Finish-to-Start Lead (FSAB) This specifies that activity B cannot start until at least FS time units after the completion of activity A (CPM takes FSAB=0) Example?
B FS
25

4. Start-to-Finish Lead (SFAB) This specifies that there must be at least SF time units between the start of activity A and the completion of activity B Example?
A SF
Can also express the leads or lags in percentages (instead of time units) Can also use at most relationships as well as the at least ones shown above
51

An example: 3 activities done in series project duration of 30 days using conventional CPM method
0,10 A,10 0,10 10,20 B,10 10,20 20,30 C,10 20,30
C B
A
5 Copyright 2008, Dr. Steven A. Gabriel 10 15 20 25 30 35 40 45 50 52
Days
26

The same 3 activities done in series but with lead-lag constraints project duration of 14 days, a 16 day speedup over the conventional CPM method
0,10 A,10 0,10 SS(AB)=2 FF(AB)=2 2,12 B,10 2,12 4,14 C,10 4,14 SS(BC)=2 FF(BC)=2
C B
A
2 Copyright 2008, Dr. Steven A. Gabriel 4 10 53 12 14
Days

Must be careful about possible anomalies in PDM Example:
0,10 A,10 0,10 10,20 B,10 FF(AB) 10,20 =10 20,30 C,10 SS(BC) 20,30 =10
Now crash B and reduce the duration of task B from 10 days to 5 days You would think that the total projection duration would decrease from 30 to something lower However, the SS(BC) constraint forces the starting time of C to be shifted forward by 5 days project duration actually increases even though Bs duration has decreased!
54
27

FF(AB)=10 A 0 10 B 20 C 30 SS(BC)=10 FF(AB)=10 A 0 10 15 B 20 25 C 35
Before Bs time is reduced
SS(BC)=10
After Bs time is reduced As a safeguard, may want to perform one activity change at a time and record the result
Precedence Diagramming Method Example

You are a planner at the National Aeronautics and Space Administration (NASA) planning the next major rocket development, production, and launching to the planet Neptune. Due to the particular positioning of the planet Neptune relative to Earth and the other planets in between, the rocket must be within 100,000 kilometers of the planet Saturn somewhere between 120 and 125 months from today in order to make it to Neptune in a reasonable amount of time. If this time window is not satisfied, the cost of reaching Neptune skyrockets dramatically (no pun intended). For example, if the time is greater than 125 months, it is estimated that $100 million more are needed to reach Neptune due to additional engineering considerations. Consider the following set of activities related to this project shown in the following table.
28
PDM-AOA Project Network for Example

Activity A B C D E Activity Description Develop trajectory plan for rocket Generate specifications for rocket design Request funding from Congress Begin initial search for contractors Prepare modified budget (using suggestions from Congress) Select rocket contractor Build and test rocket Create simulation model for rocket trajectory Prepare rocket launch & launch rocket Proceed towards Jupiter and then perform gravitational slingshot maneuver around Jupiter Achieve Saturn orbit (within 100,000 kilometers) Immediate Predecessors A Duration (months) 5 12
A,5
B,12
3
B B C 12 6 12
D,6
5
E,12
F,12
C,12
6 4
G,48
H,24
8 7 9
F G H
D,E F F
12 48 24
dummy arc precedence arc
I,12
10
J,36
11
K,1
I J
G,H I
12 36
12
PDM-Example
Compute the critical paths and project duration by formulating and solving an appropriate LP model to capture the precedence relationships between the activities. Let xi be the time for node i. The associated LP model is thus: min x12-x1 s.t. A) x2-x1 >= 5 ! A 1 A,5 B) x3-x2 >= 12 ! B 2 B,12 C) x4-x3 >= 12 ! C D,6 F,12 3 5 D) x5-x3 >= 6 ! D C,12 6 H,24 E) x5-x4 >= 12 ! E 4 E,12 8 G,48 F) x6-x5 >= 12 ! F 7 G) x7-x6 >= 48 ! G 9 H) x8-x6 >= 24 ! H I,12 10 DU1)x9-x8 >= 0 ! dummy arc dummy arc J,36 11 DU2)x9-x7 >= 0 ! dummy arc precedence arc I) x10-x9 >= 12 ! I K,1 12 J) x11-x10 >= 36 ! J K) x12-x11 >= 1 ! K 58
29
PDM-Example
1
OBJECTIVE FUNCTION VALUE 1) VARIABLE X12 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 ROW A) B) C) D) E) F) G) H) DU1) DU2) I) J) K) 150.0000 VALUE 150.000000 0.000000 5.000000 17.000000 29.000000 41.000000 53.000000 101.000000 77.000000 101.000000 113.000000 149.000000 SLACK OR SURPLUS 0.000000 0.000000 0.000000 18.000000 0.000000 0.000000 0.000000 0.000000 24.000000 0.000000 0.000000 0.000000 0.000000 REDUCED COST 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 DUAL PRICES -1.000000 -1.000000 -1.000000 0.000000 -1.000000 -1.000000 -1.000000 0.000000 0.000000 -1.000000 -1.000000 -1.000000 -1.000000
A,5
B,12
3
C,12
D,6
5
E,12
F,12
6 4
G,48
H,24
8 7 9
dummy arc precedence arc critical path arc
I,12
10
J,36
11
K,1
12
We see that the project time is 150 months which is too high (greater than 125 months).
59
PDM-Example
Question: Due to various reasons, it is believed that activities A, B, C, and D can be sped up as follows. Activity B can start as soon as 1 month after A starts. Activity C can start as soon as 1 month after activity B starts. Activity D can start as soon as one month after activity C starts. The total cost for this acceleration is $5,000,000. Modify the project network from part a (i.e., the uncrashed one) to allow for these possibilities. What is the total project time allowing for these changes? Note: Could also try project crashing to speed things up, not considered here.
30
PDM Combined with LP
Create one start and one end node for each activity that has a PDM rule. Insert arrows to enforce the new relationships Solve as previous cases
PDM-Example
Answer: We can modify the AOA network to include two nodes for A, namely A1 and A2 when activity A starts and when it finishes, respectively. The same modification can be applied to activities B, C, and D. We need to make sure that the earliest that B1 can start is 1 month after A1, the earliest that C1 can start is 1 month after B1, and the earliest that D1 can start is 1 month after C1. The resulting new project network is as follows. Note: there is some arbitrariness in connecting A2, B2, and D2, other A,5 slight variations are possible.
A1 B1
A2 B2
1
B,12
1 1
E,12
F,9
C,12
6
G,38
H,24
C1
D,6
C2 D2
dummy arc precedence arc ti
8 7 9
I,12
D1
10
J,36
11
K,1
12
ti= time that node i is visited
31
PDM-Example
A,5
A1
A2
B,12
1 1 1
B1 C1
D,6
New model is thus: min x12-xa1 s.t. A) xa2-xa1>= 5 ! B) xb2-xb1>= 12 ! C) xc2-xc1>= 12 ! D) xd2-xd1>= 6 !
A B C D
B2 C2 D2
dummy arc
E,12
F,9
C,12
6
G,38
H,24
SS1)xb1-xa1>= SS2)xc1-xb1>= SS3)xd1-xc1>= A2) B2) C2) D2) xb2-xa2>= xc2-xb2>= x4 -xc2>= x5 -xd2>=
1 1 1 0 0 0 0 12 12 48 24 0 0 12 36 1
! SS(A,B)=1 ! SS(B,C)=1 ! SS(C,D)=1 ! ! ! ! ! ! ! ! ! ! ! ! ! A2 B2 C2 D2 before before before before B2 C2 4 5
8 7 9
I,12
D1
10
J,36
precedence arc ti
11
K,1
12
E) x5 -x4 >= F) x6 -x5 >= G) x7 -x6 >= H) x8 -x6 >= DU1) x9 -x8>= DU2) x9 -x7>= I) x10-x9 >= J) x11-x10>= K) x12-x11>=
E F G H dummy arc dummy arc I J K
PDM-Example
OBJECTIVE FUNCTION VALUE 1) VARIABLE X12 XA1 XA2 XB2 XB1 XC2 XC1 XD2 XD1 X4 X5 X6 X7 X8 X9 X10 X11 135.0000 VALUE 135.000000 0.000000 14.000000 14.000000 1.000000 14.000000 2.000000 26.000000 3.000000 14.000000 26.000000 38.000000 86.000000 86.000000 86.000000 98.000000 134.000000 REDUCED COST 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 ROW A) B) C) D) SS1) SS2) SS3) A2) B2) C2) D2) E) F) G) H) DU1) DU2) I) J) K) SLACK OR SURPLUS 9.000000 1.000000 0.000000 17.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 24.000000 0.000000 0.000000 0.000000 0.000000 0.000000 DUAL PRICES 0.000000 0.000000 -1.000000 0.000000 -1.000000 -1.000000 0.000000 0.000000 0.000000 -1.000000 0.000000 -1.000000 -1.000000 -1.000000 0.000000 0.000000 -1.000000 -1.000000 -1.000000 -1.000000
Total project time is 135 months, still too big, will need to consider crashing the project.
32
PDM Example
A,5
A1
A2
B,12
OBJECTIVE FUNCTION VALUE 1) 135.0000 VALUE 135.000000 0.000000 14.000000 14.000000 1.000000 14.000000 2.000000 26.000000 3.000000 14.000000 26.000000 38.000000 86.000000 86.000000 86.000000 98.000000 134.000000 REDUCED COST 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000
1 1 1
B1 C1
D,6
B2 C2 D2
dummy arc
E,12
F,9
C,12
6
G,38
H,24
8 7 9
I,12
D1
10
J,36
precedence arc ti
11
VARIABLE X12 XA1 XA2 XB2 XB1 XC2 XC1 XD2 XD1 X4 X5 X6 X7 X8 X9 K,1 X10 X11
12
Integer Programming
All linear programming problems so far assumed that fractional answers were acceptable In practice not always ok, why? Certain classes of LPs we studied will have integer solutions, which ones and why? Want to explore modeling aspects when we specify certain variables must be integer-valued Why this is a MUCH harder problem to solve in general Interesting applications of binary variables for encoding logic in mathematical programs
33
The Toy Problem Revisited

Recall the toy production problem from before
Complete LP Max 3x1+2x2 s.t. 2x1+x2 <=100 x1+ x2 <=80 x1 <=40 x1 >=0 x2 >=0 (Objective function) (Finishing constraint) (Carpentry constraint) (Limited demand constraint on soldiers) (Nonnegativity constraint on soldiers) (Nonnegativity constraint on cars)
Optimal solution: x1=20, x2=60 , with an optimal objective function value of z=$180
34

According to LP theory, a solution (if it exists) must be at one of the vertices (also called extreme points)
LP feasible region
(0,0)
(0,80)
(20,60)
(40,20)
(40,0)
In this case, all vertices are integer-valued (i.e., whole numbers) This is fortunate since we want to produce a whole number of toys and soldiers What if this were not the case? That is, what if the the solution were not integer-valued?
69

Modified Complete LP Max 3x1+2x2 (Objective function) s.t. 2x1+x2 <=99.7 (Finishing constraint) x1+ x2 <=83.5 (Carpentry constraint) x1 <=40 (Limited demand constraint on soldiers) x1 >=0 (Nonnegativity constraint on soldiers) x2 >=0 (Nonnegativity constraint on cars)
New optimal solution: x1=16.199997, x2=67.300003 , with an optimal objective function value of z=$183.2 How has the feasible region changed? But this fractional answer really doesnt really make sense, we dont want to produce a fractional number of toy soldiers or cars (no one would buy them).
35
We add integer constraints

Complete Integer Program (IP) Max 3x1+2x2 (Objective function) s.t. 2x1+x2 <=99.7 (Finishing constraint) x1+ x2 <=83.5 (Carpentry constraint) x1 <=40 (Limited demand constraint on soldiers) x1 >=0 (Nonnegativity constraint on soldiers) x2 >=0 (Nonnegativity constraint on cars)
x1, x2 integer-valued New optimal solution: x1=16, x2=67 , with an optimal objective function value of z=$182 Note: Just rounding to the nearest integer worked in this case but in general, it wont even produce a feasible solution. Not a good way to solve a MIP 71 What does the feasible region look like in this case?
The Geometry of the Toy Problem with Integer Constraints (Excel output)
90 80 70 60 50 x2 40 30 20 10 0 0 5 10 15 20 x1 25 30 35 40 45
16,67 x1+x2<=83.5
2x1+x2<=99.7
x1<=40
36
The Geometry of the Toy Problem with Integer Constraints (MATLAB output)
16,67 x1<=40
x1+x2<=83.5
2x1+x2<=99.7

LINDO Formulation
Max 3x1+2x2 s.t. 2x1+x2 <=99.7 x1+ x2 <=83.5 x1 <=40 x1 >=0 x2 >=0 end gin x1 gin x2 (Objective function) (Finishing constraint) (Carpentry constraint) (Limited demand constraint on soldiers) (Nonnegativity constraint on soldiers) (Nonnegativity constraint on cars) (gin stands for general integer variable)
Special case of binary variables (=0 or 1) to be used later, the command to make the variable x a binary variable is inte x or inte x
37

LP OPTIMUM FOUND AT STEP 3 OBJECTIVE VALUE = 183.199997 NEW INTEGER SOLUTION OF 182.000000 BOUND ON OPTIMUM: 182.0000 ENUMERATION COMPLETE. BRANCHES= AT BRANCH 0 PIVOTS= 5 0 PIVOT 5
LAST INTEGER SOLUTION IS THE BEST FOUND RE-INSTALLING BEST SOLUTION... OBJECTIVE FUNCTION VALUE
Why is the objective function worse?
1) 182.0000 VARIABLE VALUE REDUCED COST X1 16.000000 -3.000000 X2 67.000000 -2.000000 ROW SLACK OR SURPLUS DUAL PRICES 2) 0.699997 0.000000 3) 0.500000 0.000000 4) 24.000000 0.000000 5) 16.000000 0.000000 6) 67.000000 0.000000 NO. ITERATIONS= 5 BRANCHES= 0 DETERM.= 1.000E 0
75
The Knapsack Problem
38

Suppose there are n items to be considered for inclusion in a knapsack Each item has a certain per unit value to the traveler who is packing the knapsack Each item has a per unit weight that contributes to the overall weight of the knapsack There is a limitation on the total weight that can be carried Objective: Maximize the total value of what is packed into the knapsack subject to the total weight limitation We can use IP to solve this problem Why a trivial solution is not apparent Generalizations of this problem beyond a knapsack

Definitions: For j=1,2,...,n let vj>0 be the value per unit for item j wj>0 be the weight per unit of item j W be the total weight limitation xj is the number of units of item j included in the knapsack
max s .t .
v x
j j =1 j j
w x
j =1
W , x j 0 ,integer, j = 1,K ,n
78
39
The Knapsack Problem: Project Selection Example

Definitions: For j=1,2,...,n let vj>0 be the value if project j is selected cj>0 be the cost of selecting project j B be the total budget available limitation xj =1 if project is selected, 0 otherwise
max s.t.
v x
j j j j
c x
j
x j {0,1} , j = 1, 2, K , n
The Knapsack Problem: Project Selection Example

Project Value 1 25.99 2 17.56 3 21.33 4 14.34 5 24.37 6 24.37 7 21.33 8 11.65 9 25.27 10 21.33 11 18.46 12 11.65 13 25.27 14 17.56 15 21.33 Cost 13.69 12.31 15 12.73 13.69 12.31 15 12.73 13.69 12.31 15 12.73 13.69 12.31 15
15 projects, total budget of 100 Why not just fund all 15? Total cost is 202.2, therefore, need the right subset Dont just pick the least costly ones, want high value ones too Cherry-picking solution is not always the best Use Excel to solve this integer program (IP)
40
Scheduling Under Limited Resources Using Integer Programming

Suppose there are n items to be considered for inclusion in a knapsack Each item has a certain per unit value to the traveler who is packing the knapsack Each item has a per unit weight that contributes to the overall weight of the knapsack There is a limitation on the total weight that can be carried Objective: Maximize the total value of what is packed into the knapsack subject to the total weight limitation We can use IP to solve this problem

Definitions: For j=1,2,...,n let cj>0 be the value per unit for item j wj>0 be the weight per unit of item j W be the total weight limitation xj is the number of units of item j included in the knapsack
max s.t.
c x
j j =1 j j
w x
j =1
W , x j 0, integer, j = 1,K ,n
82
41

The Knapsack Problem for scheduling
Each activity to be scheduled at a specific instant is modeled as an item to be included in the knapsack The composition of the activities in a scheduling window (certain amount of time) is viewed as the knapsack Note: for activity scheduling, only one unit of each activity (item) can be included in the schedule at any given scheduling time; in general, cant schedule the activity twice at the same time! The knapsack problem for activity scheduling is done at each and every scheduling time t The objective is to schedule as many activities of high priority as possible while satisfying precedence relationships w/o exceeding the resources

Definitions: zt =overall performance of the schedule generated at time t pj= the priority value for activity j t= current time of scheduling xjt= binary variable, =1 if activity j is scheduled at time t, =0 o/w St=set of activities eligible for scheduling at time t k= number of different resource types rij=units of resource type i required by activity j Rit=units of resource type i available at time t
max zt = s.t.
p x
j jS t jt
jt
r x
ij jS t
Rit , i = 1,K ,k , x jt {0,1}, j S t for all t

84
42

Note that the next scheduling time, t, for the knapsack problem is given as the minimum of {the finishing times of the scheduled and unfinished activities} Now lets consider a specific example (see page 303) Note that the priority values do not change from time one time period to the next (fixed prioritization scheme) Can also consider knapsack problem with changing priority values (variable prioritization scheme) Can use specialized methods to solve this problem
43
Example from Badiru-Pulat

Activity A B C D E F G Id 1 2 3 4 5 6 7 Duration Priority R1 2.17 55.4 6 100 3.83 72.6 2.83 54 5.17 88 4 66.6 2 75.3 R2 3 5 4 2 4 2 6 0 4 1 0 3 7 2 Predecesor
A C A B,D,E
Resource Availability 1 10 2 15
t=0
S0={A,B,C}
max 55.4xa+100xb+72.6xc s.t. R1)3xa+5xb+4xc<=10 R2)0xa+4xb+1xc<=15 end int xa int xb int xc OBJECTIVE FUNCTION VALUE 172.6000 XA = 0.000000 XB = 1.000000 XC = 1.000000
44
t=3.83
S3.83={A,E}
max 55.4xa+88xe s.t. R1)3xa+4xe<= 5 R2)0xa+3xe<=11 end int xa int xe
B is ongoing and uses 5 units of R1 and 4 of R2
OBJECTIVE FUNCTION VALUE 88.00000 XA = 0.000000 XE = 1.000000
t=6.0
S6={A}
max 55.4xa s.t. R1)3xa<= 6 end int xa
E is ongoing and uses 4 units of R1 and 3 of R2
OBJECTIVE FUNCTION VALUE 55.00000 XA = 1.000000
45
t=8.17
S8.17={D,F}
max 54xd+66.6xf s.t. R1)2xd+2xf<= 6 R1)0xd+7xf<=12 end int xd int xf
E is ongoing and uses 4 units of R1 and 3 of R2
OBJECTIVE FUNCTION VALUE 120.00000 Xd = 1.000000 Xf = 1.000000
t=9
S9={} Cant do anything! D is ongoing and uses 2 units of R1, F is ongoing and uses 2 units of R1, and 7 units of R2
46
t=11
S11={G}
max 75.3xg s.t. R1)6xg<= 8 R1)2xg<= 8 end int xg
F is ongoing and uses 2 units of R1 and 7 of R2
OBJECTIVE FUNCTION VALUE 73.00000 Xg = 1.000000
Some Other Uses of Integer (Binary) Programming

Fixed charge problems Either-or constraints If-then constraints
47
Fixed Charge Problems: An Example

Manufacturing project involving 3 products (1, 2, 3) Each product requires that an appropriate type of machinery be available Rental rates for machines: Product 1 machine: $200/week Product 2 machine: $150/week Product 3 machine: $100/week Also raw materials and labor required for each product Labor (hours) Product 1 Product 2 Product 3 3 2 6
Raw Materials (lbs) 4 3 4
Fixed Charge Problems

Each week 150 hours of labor and 160 lbs of raw materials are available Also need to consider the variable unit cost and selling price for each product Want an IP whose solution will maximize the weekly net profits Variables?
Sales price Product 1 Product 2 Product 3 $12 $8 $15
Variable Cost $6 $4 $8
48
Variables Let xi= number of units of product i manufactured in a given week (i=1,2,3) Let yi= 1 if any product i is manufactured, =0 o/w (i=1,2,3) (note: if xi>0 yi=1 and if xi=0 yi=0) Constraints? Objective function?

Let xi= number of units of product i manufactured in a given week (i=1,2,3) Let yi= 1 if any product i is manufactured, =0 o/w (i=1,2,3) (note: if xi>0 yi=1 and if xi=0 yi=0) Let M1, M2, M3 be 3 large positive numbers
max 6x1 + 4x2 + 7x3 200y1 150y2 100y3 Net Revenues rental costs s.t. 3x1 + 2x2 + 6x3 150 Labor constraint 4x1 + 3x2 + 4x3 160 Raw materials constraint x1 M1 y1 x2 M 2 y2 x3 M3 y3 x1,x2 ,x3 0, integer y1,y2 ,y3 {0,1}
Constraints that ensure if xi>0 yi=1 note that if xi=0 yi=0 or =1 but we get yi=0 at an optimal solution (for cost reasons) max x values give M values: 40, 53, 25
98
49

max 6x1+4x2+7x3-200y1-150y2-100y3 s.t. 3x1+2x2+6x3<=150 4x1+3x2+4x3<=160 x1-40y1 <=0 x2-53y2<=0 x3-25y3<=0 end gin x1 gin x2 gin x3 int y1 int y2 int y3
Solution: Optimal net profit of $75 x3=25 y3=1 all other variables =0
Either-Or Formulation
Lets modify the manufacturing problem from before If any of product 1 produced, then it must be at least 25 units, i.e., if x1>0 x1>=25 or equivalently either x1<=0 or x1>=25 If any of product 2 produced, then it must be at least 26 units i.e., if x2>0 x2>=26 or equivalently either x2<=0 or x2>=26 If any of product 3 produced, then it must be at least 27 units i.e., if x3>0 x3>=27 or equivalently either x3<=0 or x3>=27 More general setting, we have two constraints of the form: f(x1,x2,..., xn)<=0 and g(x1,x2,..., xn)<=0 We want to ensure that at least one of these constraints is satisfied For N a large enough positive number and z a binary variable, this is ensured with the following two constraints
f ( x1 , x 2 , K , x n ) Nz g ( x1 , x 2 , K , x n ) N (1 z )
50
max 6x1 + 4x2 + 7x3 200y1 150y2 100y3 Net Revenues rental costs s.t. 3x1 + 2x2 + 6x3 150 Labor constraint 4x1 + 3x2 + 4x3 160 Raw materials constraint x1 M1 y1 x2 M 2 y2 x3 M3 y3 x1 N1z1 25 x1 N1 (1 z1 ) x2 N2 z2 26 x2 N2 (1 z2 ) x3 N3 z3 27 x3 N3 (1 z3 ) x1,x2 ,x3 0, integer y1,y2 ,y3 , z1,z2 ,z3 {0,1}
Copyright 2008, Dr. Steven A. Gabriel 101
Constraints that ensure if xi>0 yi=1 note that if xi=0 yi=0 or =1 but we get yi=0 at an optimal solution (for cost reasons) max x values give M values: 40, 53, 25
Either-or constraints, N values need to be chosen suitably large, for example take all Ns to be equal to 100
LINDO Formulation and Solution to Either-Or Problem

max 6x1+4x2+7x3-200y1150y2-100y3 s.t. 3x1+2x2+6x3<=150 4x1+3x2+4x3<=160 x1-40y1 <=0 x2-53y2<=0 x3-25y3<=0 x1-100z1<=0 -x1+100z1<=75 x2-100z2<=0 -x2+100z2<=74 x3-100z3<=0 -x3+100z3<=73 end gin x1 gin x2 gin x3 int y1 int y2 int y3 int z1 int z2 int z3
Solution: Optimal profit of $62 x2=53 y2=1 z2=1 all other variables =0
102
51
If-Then Formulation
Lets modify the original manufacturing problem one more time (taking away the modifications from before) We want the restriction: if the sum of products 2 and 3 exceed 24 units (true for both cases considered so far), then at least 30 of product 1 must be manufactured (union rules?) More generally, we will be considering
if f ( x1 , x 2 , K , x n ) > 0 then g ( x1 , x 2 , K , x n ) 0 if f ( x1 , x 2 , K , x n ) 0 then g ( x1 , x 2 , K , x n ) 0 or < 0
If-Then Formulation
We can use the following constraints where N is a suitably large positive value and z is a binary variable
g ( x1 , x 2 , K , x n ) Nz f ( x1 , x 2 , K , x n ) N (1 z )
In our example we can take
x1 30 = g ( x1 , x 2 , K , x n ) x 2 + x 3 24 = f ( x1 , x 2 , K , x n )
52
LINDO Formulation and Solution to If-Then Problem

max 6x1+4x2+7x3-200y1-150y2100y3 s.t. 3x1+2x2+6x3<=150 4x1+3x2+4x3<=160 x1-40y1 <=0 x2-53y2<=0 x3-25y3<=0 -x1-100z<=-30 x2+x3+100z<=124 end gin x1 gin x2 gin x3 int y1 int y2 int y3 int z
Solution: Optimal profit of $68 x3=24 y3=1 z=1 all other variables =0
A Production -Inventory Problem (Periodic Review Model):
Multiperiod Production-Inventory Problem

Time is broken up into periods: present period ---- period 1 next period ---- period 2 . last period ---- period T At start of each period, firm must determine how much should be produced (production capacity at each period is limited) Each periods demand must be met on time from inventory or current production
106
53

During any period in which production occurs, a fixed cost as well as a variable per unit cost is incurred There is limited storage capacity
Limit on end-of-period inventory
A per unit holding cost is incurred on each periods ending inventory The firms goal is to minimize the total cost of meeting on time, the demands for periods 1,2,,T

production at period t
Typical Picture
Inventory at period t Inventory from period t -1 demand at period t
54

Month Demand Demand schedule schedule 1 2 3 4 Number of units (tons) 1 3 2 4 At start of each month, how much to produce during the current month (questions for company)
Production Production (if (if production production occurs): occurs): Max Max of of 5 5 units/period units/period to to be be produced produced Set Set up up cost cost $3 $3 Variable Variable cost cost $1/unit $1/unit produced. produced. Inventory Inventory :: Holding Holding Costs Costs (at (at the the end end of of the the month): month): $0.50/unit $0.50/unit Capacity Capacity limitations: limitations: max max of of 4 4 units, units, initial initial inventory=0 inventory=0 Want: Want: Production Production schedule schedule that that will will meet meet all all demands demands on on time time and and will will minimize minimize the the sum sum of of production production and and holding holding costs costs for for the the 4 4 months. months. (Assume (Assume inventory inventory at at start start of of month month 1 1 is is 0 0 units units ) )
109 Variables, Variables, constraints, constraints, objective objective function? function? Copyright 2008, Dr. Steven A. Gabriel

Production schedule (try without LP/IP first)
x1 x1 =? x2 x2 =? x3 x3 =? x4 x4 =?
Inventory i1=0
1 i2=?
2 i3=?
3 i4=?
4 ?
Demand =1
Demand =3
Demand =2
Demand =4
Can Can relate relate to to a a shortest shortest path path problem problem in in a a network network as as follows. follows.
55

Production schedule IP
Min 1x1+1x2+1x3+1x4 ! Production variable costs + 3y1+3y2+3y3+3y4 ! Production fixed costs +0.5i1+0.5i2+0.5i3+0.5i4 ! Inventory costs s.t. d1=1 ! Demand for period 1 d2=3 ! Demand for period 2 d3=2 ! Demand for period 3 d4=4 ! Demand for period 4 i1=0 ! Initial inventory i2<=4 i3<=4 i4<=4 ! Inventory capacity x1<=5 x2<=5 x3<=5 x4<=5 ! Production capacity i1+x1-d1 -i2=0 i2+x2-d2 -i3=0 i3+x3-d3 -i4=0 i4+x4-d4 =0 ! Period 1 material balance ! Period 1 material balance ! Period 1 material balance ! Period 1 material balance x1-100000y1 <=0 ! Consistency between production and set-up cost varibles x2-100000y2 <=0 ! Consistency between production and set-up cost varibles x3-100000y3 <=0 ! Consistency between production and set-up cost varibles x4-100000y4 <=0 ! Consistency between production and set-up cost varibles end ! Nonnegativity implied by LINDO inte y1 inte y2 inte y3 inte y4

20 20 is is the the minimum minimum cost cost for for the the 4 4 months months optimal optimal schedule. schedule.
x1 x1 1 Inventory i1=0 1 i2=0 2 i3=2 3 i4=0 4 0 =1 x2 x2 =5 x3 x3 =0 x4 x4 =4
Demand =1
Demand =3
Demand =2
Demand =4
Can Can relate relate to to a a shortest shortest path path problem problem in in a a network network as as follows. follows.
56

Representation of Inventory Example
1,0 2,0 3,0 4,0
(month, (month, state) state)
1,1
2,1
3,1
4,1
1,2
2,2
3,2
4,2
5,0
Month 5
1,3
2,3
3,3
4,3
1,4
Month 1
2,4
Month 2
3,4
Month 3
4,4
Month 4
Assignment Problems (You are now a Project Manager)

Want to optimally assign workers to tasks Suppose we have n tasks to be performed by n workers The cost of worker i performing task j is given as cij Remarks: 1. If the number of tasks to be done is greater than the number of workers, we add dummy workers to balance the problem 2. If the number of workers is greater than the number of tasks, we add dummy tasks to balance the problem 3. If no problem of overlapping a workers time, the time can be split between projects and each segment can be modeled as a separate resource (can consider partial allocation of resource units to multiple tasks)
57
The Assignment Problem

CPM can be used to control the project duration Such methods do not however, assign resources to project tasks Now we consider the assignment problem , a formulation to optimally assign workers to tasks Suppose we have n tasks to be performed by n workers The cost of worker i performing task j is given as cij Remarks: 1. If the number of tasks to be done is greater than the number of works, we add dummy workers to balance the problem 2. If the number of workers is greater than the number of tasks, we add dummy tasks to balance the problem 3. If no problem of overlapping a workers time, the time can be split between projects and each segment can be modeled as a separate resource (can consider partial allocation of resource units to multiple 115 tasks)

min s.t.
c x
i =1 j =1 ij
ij ij
minimize total costs
x
j =1 n
= 1, i = 1,2,K , n ith worker must get assigned to

exactly 1 task
x
i =1
ij
= 1, j = 1,2,K , n jth task must get exactly 1

worker restriction satisfied indirectly
116
xij 0, i, j = 1,2, K , n nonnegative variables, binary
58

Consider the following cost matrix for an assignment problem with n=5 Select the cheapest workers by task first, will this work?
Task 1 Task 2 Task 3 Task 4 Worker 1 Worker 2 Worker 3 Worker 4 Worker 5 $200 $400 $300 $100 $700 $400 $700 $900 $300 $100 $500 $800 $800 $500 $200
Task5 $400 $500 $500 $400 $200
$100 $1,100 $1,000 $100 $100

min 200x11 + 400x12 + 500x13 + 100x14 + 400x15+ 400x21 + 700x22 + 800x23 + 1100x24 + 500x25+ 300x31 + 900x32 + 800x33 + 1000x34 + 500x35+ 100x41 + 300x42 + 500x43 + 100x44 + 400x45+700x51 + 100x52 + 200x53 + 100x54 + 200x55 s.t. x11+x12+x13+x14+x15=1 ! worker 1 x21+x22+x23+x24+x25=1 ! worker 2 x31+x32+x33+x34+x35=1 ! worker 3 x41+x42+x43+x44+x45=1 ! worker 4 x51+x52+x53+x54+x55=1 ! worker 5 x11+x21+x31+x41+x51=1 ! task 1 x12+x22+x32+x42+x52=1 ! task 2 x13+x23+x33+x43+x53=1 ! task 3 x14+x24+x34+x44+x54=1 ! task 4 x15+x25+x35+x45+x55=1 ! task 5 ! and all variables nonnegative
59

Optimal solution at a cost of $1,400 is as follows
Task 1 Task 2 Task 3 Task 4 Worker 1 Worker 2 Worker 3 Worker 4 Worker 5

Task5
Solving Integer Programs

Certain classes of LPs we studied will have integer solutions so dont need to enforce integrality restrictions Otherwise, how can we solve integer-constrained problems? Many approaches, will give just two mentioned here
Enumeration Branch-and-Bound (pure IP example)
60
Solving Integer Programs Using Enumeration

For small enough problems, can just enumerate all feasible solutions Then pick the one(s) with the best objective function value When this method will work, when it wont
Solving Integer Programs Using Branch-and-Bound for Pure IPs
Basic Idea: Solve a sequence of linear programming relaxations (in the form of a tree structure) to solve original IP Elementary observation: if you solve the LP relaxation of a pure IP and get a solution which has just integer answers, then LP optimal is also IP optimal Example (from Winston) max z=3x1+2x2 s.t. 3x1+x2<=6 x1,x2 >=0, x1,x2 integer Feasible region? LP and IP solutions?
61
Sample Problem: Production of tables and chairs 1 table needs 1 hour of labor & 9 square board feet of wood, $8 in profit 1 chair needs 1 hour of labor & 5 square board feet of wood, $5 in profit Currently: 6 hours of labor, 45 square board feet available IP to maximize profit?

let x1= number of tables produced x2= number of chairs produced
max 8x1+5x2 x1 s.t. x1+x2<=6 9x1+5x2<=45 x1,x2>=0, x1, x2 integer
x2 0 0 0 0 0 0 0 1 1 1 1 1 1 2 2 2 2 2 3 3 3 3 4 4 5 0 1 2 3 4 5 6 0 1 2 3 4 5 0 1 2 3 4 0 1 2 3 0 1 0
z 0 5 10 15 20 25 30 8 13 18 23 28 33 16 21 26 31 36 24 29 34 39 32 37 40
7 6 5 4 Series1 3 2 1 0 0 1 2 3 4 5 6 7
62
Solving Integer Programs Using Branch-andBound for Pure IPs Lets see how to get this solution with the Branchand-Bound Technique 7 LP subproblems to solve
Solving Integer Programs Using Branch-andBound for Pure IPs

original SP0 IP x1=5,x2=0, z=40
x1>=4 x1=4,x2=1.8 z=41 x1>=4, SP4 x2>=2 infeasible fathom node
SP1
no integer constraints, LP relaxation x1=3.75,x2=2.25, z=41.25 IP upper bound is 41.25

x1<=3, x1=3, x2=3,z=39 SP6 has better z-value, fathom this node
SP2 SP5
x1>=4, x2<=1,
SP3
x1=4.44,x2=1,z=40.556
SP6 x1>=4, x2<=1, x1>=5, x1=5,x2=0,z=40 IP lower bound, z=40

SP7
x1>=4, x2<=1, x1<=4 x1=4,x2=1, z=37, fathom node

126
63
Original problem
max 8x1+5x2 s.t. x1+x2<=6 9x1+5x2<=45 end gin x1 gin x2
Subproblem 1 remove integer constraints

max 8x1+5x2 s.t. x1+x2<=6 9x1+5x2<=45 end
7 6 5 4 3 2 1 0 0 1 2 3 4 5 6 7 optimal solution: x1=5, x2=0 optimal obj. fcn value: z=40
x1=3.75,x2=2.25, z=41.25 IP upper bound is 41.25
Conclusion: optimal z-value for IP<=optimal z-value for LP relaxation Upper bound for IP is 41.25 Next step, arbitrarily choose a fractional variable (say x1) and try 2 LPs with the rounded values Subproblem 2: x1>=4, subproblem 3, x1<=3 (branch on x1) Why cant a feasible solution to the IP have 3<x1<4? The point x1=3.75 will be avoided this way (cant return to this solution), why?
127

x1>=4
SP1

x1<=3
SP2
SP3
64
Original problem max 8x1+5x2 s.t. x1+x2<=6 9x1+5x2<=45 x1,x2>=0, x1, x2 integer
Subproblem 2 add new constraint x1>=4 max 8x1+5x2 s.t. x1+x2<=6 9x1+5x2<=45 x1,x2>=0, x1>=4
x1=4,x2=1.8,z=41
Conclusion: integer solution not obtained Next step: arbitrarily choose a fractional variable (branch on x2) and try 2 LPs with rounded values Subproblem #4: x2>=2, subproblem #5: x2<=1

x1>=4 x1=4,x2=1.8 z=41 x1>=4, x2>=2
SP1

x1<=3
SP2 SP5
x1>=4, x2<=1
SP3
SP4
65
Subproblem 4 add new constraint x2>=2 max 8x1+5x2 s.t. x1+x2<=6 9x1+5x2<=45 x1>=4, x2>=2 x1,x2>=0
Conclusion: Subproblem 4 is infeasible, fathom this node Next step: try subproblem 5

SP1

x1<=3
SP2 SP5
x1>=4, x2<=1
SP3
66
Subproblem 5 add new constraint x2<=1 max 8x1+5x2 s.t. x1+x2<=6 9x1+5x2<=45 x1,x2>=0 ,x1>=4, x2<=1 x1=4.44, x2=1,z=40.556
Conclusion: feasible solution, still fractional though so need to branch again Next step: branch on x1, Subproblem 6: add x1>=5, subproblem 7: add x1<=4 Could also try subproblem 3 but we are using a LIFO rule (LIFO=last in first out)

SP1

x1<=3
SP2 SP5
x1>=4, x2<=1,
SP3
x1=4.44,x2=1,z=40.556
x1>=4, x2<=1, x1>=5
SP6
SP7
x1>=4, x2<=1, x1<=4

134
67
Subproblem 6 add new constraint x1>=5 max 8x1+5x2 s.t. x1+x2<=6 9x1+5x2<=45 x1,x2>=0, x1>=4, x2<=1, x1>=5
x1=5,x2=0,z=40 IP lower bound, z=40
Conclusion: candidate solution IP lower bound is now 40 Next step: try remaining node relating to subproblem 7

SP1

x1<=3
SP2 SP5
x1>=4, x2<=1,
SP3
x1=4.44,x2=1,z=40.556

SP7
x1>=4, x2<=1, x1<=4

136
68
Subproblem 7 add new constraint x1<=4 max 8x1+5x2 s.t. x1+x2<=6 9x1+5x2<=45 x1,x2>=0, x1>=4, x2<=1, x1<=4
x1=4,x2=1,z=37 IP lower bound, z=37
Conclusion: further branching on subproblem7 cannot yield a feasible integer solution>37, why? Next step: fathom this node and try subproblem 3

SP1

x1<=3
SP2 SP5
x1>=4, x2<=1,
SP3
x1=4.44,x2=1,z=40.556

SP7

138
69
Subproblem 3 add new constraint x1<=3 max 8x1+5x2 s.t. x1+x2<=6 9x1+5x2<=45 x1,x2>=0, x1<=3
x1=3,x2=3, z=39
Conclusion: Not better than the current lower bound of 40 from subproblem 6 Fathom this node No nodes left to try- done!

SP1

x1<=3, x1=3, x2=3,z=39 SP6 has better z-value, fathom this node
SP2 SP5
x1>=4, x2<=1,
SP3
x1=4.44,x2=1,z=40.556

SP7

140
70

ENCE 603 Management Science Applications in Project Management Lectures 5-7

Uploaded by

Document Informationclick to expand document information002

Document Informationclick to expand document information

Copyright:

Available Formats

ENCE 603 Management Science Applications in Project Management Lectures 5-7

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ENCE 603 Management Science Applications in Project Management Lectures 5-7

Uploaded by

Copyright:

Available Formats

ENCE 603 Management Science Applications in Project Management Lectures 5-7

Project Management LP Models in Scheduling, Integer Programming

Spring 2009 Instructor: Dr. Steven A. Gabriel www.eng.umd.edu/~sgabriel

2 Copyright 2008, Dr. Steven A. Gabriel

4 Copyright 2008, Dr. Steven A. Gabriel

Activity on Node (AON) Representation of Project Networks

6 Copyright 2008, Dr. Steven A. Gabriel

Copyright 2008, Dr. Steven A. Gabriel

P(3)= {1,2} S(3)= {4,5}

Rule 1: ES(1)=0 (unless otherwise stated) Rule 2: ES(i)=Max j in P(i) {EC(j)}

Why do we use max of the predecessor ECs in rule 2?

Copyright 2008, Dr. Steven A. Gabriel

Why do we use min in the successor LSs in rule 7?

12 Copyright 2008, Dr. Steven A. Gabriel

CPM-Determining the Critical Path AON

CPM-Forward Pass Example AON 2,6

F,4 D,3 11,11 9,11 end B,6 4,9 E,5

Copyright 2008, Dr. Steven A. Gabriel

CPM-Backward Pass Example AON

2,6 F,4 7,11 9,11 end 11,11 11,11

CPM-Slacks and the Critical Path AON

CPM Analysis for Sample Network AON

0,2 A,2 4,6 0,0 start 0,0 0,4 C,4 0,4

2,6 2,5 F,4 7,11 9,11 G,2 9,11

Total Project Slack TS(1)+TS(C)+TS(E)+TS(G)+TS(n)=0

0,6 B,6 3,9

D,3 6,9 4,9 E,5 4,9

11,11 end 11,11

19 Copyright 2008, Dr. Steven A. Gabriel

Copyright 2008, Dr. Steven A. Gabriel

Activity on Arc (AOA) Representation of Project Networks

21 Copyright 2008, Dr. Steven A. Gabriel

Project Networks: Activity on Arc (AOA) Representation

Copyright 2008, Dr. Steven A. Gabriel

Activity on Arc (AOA) Representation

AOA Network Structure

Copyright 2008, Dr. Steven A. Gabriel

AOA Network Structure

25 Copyright 2008, Dr. Steven A. Gabriel

Rules in AOA Networks

26 Copyright 2008, Dr. Steven A. Gabriel

Small Sample Project

Small Sample Project AOA

28 Copyright 2008, Dr. Steven A. Gabriel

Using Linear Programming to Find a Critical Path

Using Linear Programming to Find a Critical Path

Using Linear Programming to Find a Critical Path

OBJECTIVE FUNCTION VALUE VARIABLE X5 X1 X2 X4 X3

11.00000 REDUCED COST 0.000000 0.000000 0.000000 0.000000 0.000000

VALUE 11.000000 0.000000 6.000000 9.000000 4.000000

Copyright 2008, Dr. Steven A. Gabriel

Using Linear Programming to Find a Critical Path

SLACK OR SURPLUS A) 4.000000 B) 3.000000 C) 0.000000 D) 0.000000 E) 0.000000 F) 1.000000 G) 0.000000

DUAL PRICES 0.000000 0.000000 -1.000000 0.000000 -1.000000 0.000000 -1.000000

Copyright 2008, Dr. Steven A. Gabriel

Using Linear Programming to Find a Critical Path

33 Copyright 2008, Dr. Steven A. Gabriel