Proceedings of the 2012 International Conference on Industrial Engineering and Operations Management

Istanbul, Turkey, July 3 – 6, 2012

Scheduling Project Crashing Time using Linear Programming

Technique
Omar M. Elmabrouk
Department of Industrial and Manufacturing System Engineering
Benghazi University, Benghazi-Libya

Abstract

In this paper, some concepts of linear programming and critical path method were reviewed to describe recent
modeling structures that have been of great value in analyzing extended planning horizon project time-cost crashes
problems. A simplified representation of a small project and a linear programming model was formulated to
represent this system. Procedures to solve these various problems formulations were cited and the final solution was
obtained by using LINDO program. The model developed represents many restrictions and management
considerations of the project. It could be used by construction managers in a planning stage to explore numerous
possible opportunities to the contractor and predict the effect of a decision on the construction to facilitate a
preferred operating policy given different management objectives. An implementation using this method was shown
to outperform several other techniques and a large class of test problems. Linear programming shows that the
algorithm is very promising in practice on a wide variety of time-cost crashes problems. This method is simple,
applicable to a large network, and generates a shorter computational time at low cost, along with an increase in
robustness.

Keywords
Linear programming, project crashing, time-cost trade-off, CPM.

1. Introduction
Project management is the process of the application of knowledge, skills, tools, and techniques to project activities
to meet project requirements. That is to say, project management is an interrelated group of processes that enables
the project team to achieve a successful project. The functions of project management include:
a) Planning – Planning the project and establishing its life cycle.
b) Organizing – Organizing resources: personnel, equipment, materials, facilities, and finances.
c) Leading – Assigning the right people to the right job, motivating people, setting the course and goals for
the project.
d) Controlling – Evaluating progress of project and, when necessary, applying hangs to get it back on track.
Performing these functions in an organized framework of processes is the job of the Project Manager. The
project engineer is assigned by an organization the responsibility and authority to manage a technical
project or program in a manner that will result in meeting the project objectives, the three basic objectives
for which the contractor‘s project engineer is responsible are as follows:

a) Deliver a project that meets the requirements of the specification.

b) Deliver a project that meets the requirements of the contract delivery schedule.
c) Meet the company’s profit objectives for the contract.
From a project management perspective, a project is considered a success if:
a) The resulting information system is acceptable to the customer.
b) The system is delivered “on time”
c) The system is delivered “within budget”

Project management has evolved as a new field with the development of two analytical techniques for planning,
scheduling, and controlling of projects. These are the critical path method (CPM) and the project evaluation and
review technique (PERT). These techniques are neither suitable for the scheduling of linear projects nor adequate
for addressing typical challenge related to time-cost trade-off [1].
308
Optimal schedule cost can be determined by trial and error for small project, but in a realistic project consisting of
many activities, such trial-and error becomes extremely tedious and impossible. A very limited number of computer
programs are available but far from perfect. Such programs have a limited capacity to accept time-cost data and at a
very high price. Thus in this paper, LP (Linear Programming) as an optimization technique has been developed to
aid in the quick determination of the minimum cost for every possible value of project duration. Clearly, the use of
this optimization techniques incorporated with time-cost trade-off becomes an economic necessity [1].

2. Literature Review
The cost of the network activities had been optimized for various overall durations. The optimum crash of time
against cost had been made. This approach was an acceptable tool of management and proving to be not only
superior method for planning, scheduling and controlling project progress, but also was very real and valuable assets
to contractors in convincing the owner of their potentials and abilities [2].

Project scheduling under uncertainty by using survey and research potentials was carried out. In that survey they
reviewed the fundamental approaches for scheduling under uncertainty, reactive scheduling, stochastic project
scheduling, fuzzy project scheduling, robust (proactive) scheduling and sensitivity analysis. The potentials of those
approaches for scheduling under uncertainty projects with deterministic network evolution structure were discussed.
They offered a review of the major approaches to deal with scheduling risk and uncertainty. The methodologies for
stochastic project scheduling basically view the project scheduling problem as a multi-stage decision process [3] .

Project crashing and costs laws in the knowledge age was studied. That study seeks to add contributions of the
innovation and industrial economics to more used techniques of crashing in the projects management domain. First,
it presented the brute force method, improved for the use of the MS Project. Second, it developed the models for
determining the earliest crash completion time and for determining a least costly crash schedule. Third, it established
costs laws, which allow inferring that the cost of a project does not depend only on the production rate but
depending also on the time were the first unit of production will be available, on the global volume of production
and on the project completion time. Projects management might considerably help organizations that search for a
better market position, providing a higher significance to the competitive advantages to be developed by the
company along the time [4] .

A new method for determining which project activities to crash was presented. The calculation of resource elasticity
required no new information, and was computed using the project manager’s conventional estimated of normal and
crash duration times as well as normal and crash costs. Instead of assuming a linear relationship between crash
spending and time saved, a linear relationship between the additional resources used per day and the time saved
were assumed. That assumption allowed computing a resource elasticity measurement that tolled which activities
would respond most significantly to increases in the resources allocated to them. That elasticity measure was a
practical, easily implemented metric that allows the manager to identify which activities’ duration times were more
or less sensitive to daily spending rates. That metric could be used for determining which activities should be
crashed, which should be slowed (if the money for crashing had to come from another activity), and for determining
which activities need to be most closely monitored [5].

The simulation approach for optimization project cost and schedule was one of a variety of tools that could use to
bring projects back under control and reinforce the use of project management in organizations. The use of
simulation to crash project management networks in order to reduce time and cost overruns was a worth endeavor.
The project manager, in collaboration with the IT division, could routinely submit each developed network to
crashing (using the simulation program), before major work commences [6]. Hence, the optimization of time and
cost process technique could be incorporated as a standard procedure for every project was concluded, the time
spent on the actual crashing was minimal and the project management schedule could be reduced to a minimum
optimum level to save time and money [6].

3. Linear Programming Technique

Linear programming is a tool for decision making under certain situation. So, the basic assumption of this approach
is that we have to know some relevant data with certainty. The basic data requirements are as follows:
309
a) We have to know the project network with activity time, which can be achieved from PERT and CPM.
b) To what extent an activity can be crashed.

c) The crash cost associated with per unit of time for all activities. Before formulating the model, let us define some
relevant terms. We know, a project is the combination of some activities, which are interrelated in a logical sequence
in the sense that the starting of some activities is dependent upon the completion of some other activities. These
activities are jobs which require time and resources to be completed. The relationship between the activities is
specified by using event. As an event represents a point in time that implies the completion of some activities and
the beginning of new ones, the beginning and end point of an activity are thus expressed by two events. To reduce
the time to complete the activity, more resources are applied in the form of additional personnel and overtime. As
more resources are applied, the duration is shortened, but the cost rises. The maximum effort is applied so that the
activity can be completed in the shortest possible time. The equation for the cost slope is

CC − C n
Ui = (1)
Tn − Tc

Where:
Ui,Cc and Cn are the cost slop, the crash cost and normal costs, respectively. Tc and Tn are the crash and normal
times, respectively. The cost slope shows by how much the cost of the job would change if activities were speed up
or slowed down. Before formulating the model, some relevant terms will be defined. It is very well known, a
project is the combination of some activities, which are interrelated in a logical sequence in the sense that the
starting of some activities is dependent upon the completion of some other activities. These activities are jobs, which
require time and resources to be completed. The relationship between the activities is specified by using event. As
an event represents a point in time that implies the completion of some activities and the beginning of new ones, the
beginning and end point of an activity are thus, expressed by two events.

Now let’s define the variable of the problem.

Yi = The time when an event i will occur, measured since the beginning of the project,
where i = (1, 2, 3,…,n).
Xi = Amount of times (measured in terms of days, weeks, months or some other units) that each activity i will be
crashed, where i = (1, 2, 3…n).

The objective is to minimize the cost of crashing total project via minimize the durations of crashing activities that
multiplied by their associated costs slope, then adding the resultant cost to the normal cost of project completion.
The LP objective function will be:
n
min z = ∑U X
i =1
i i (2)

This objective function is subject to some constraints. These constraints can be classified in to three categories.

Crash time constraints: We can reduce the time to complete an activity by simply increasing the resources or by
improving the productivity, which also requires the commitment of additional resources. But, it is not possible to
reduce the required time to complete an activity after a certain threshold limit. Strive for such intention will result in
superfluous resources employment which will be an inefficient approach. That is why, the allowable time to crash an
activity has a limit. Constraints unfolding the network: These set of constraints describe the structure of the
network. As we mention earlier that the activities of a project are interrelated, the starting of some activities is
dependent upon the completion of some other activities; we must have to establish research sequence of the
activities through constraints.

Nonnegative constraints: All decision variables must ≥ 0. So, the constraints are:

310
Crash time constraints: Xi≤ Allowable crashing time for activity i measured in terms of days, weeks, months or
some other units.
Constraints unfolding the network: There will be one or more constraints for each event depending on the
predecessor activities of that event. As the event 1 will start at the beginning of the project, we begin by setting the
occurrence time for event 1 equals to zero. Thus, Y1 = 0. The other events will be expressed as follows:

Start time of this activity (Yi) = (start time + normal duration -crash duration) for this immediate predecessor.

Project completion constraints: Ym≤ project deadline after being stretched, where m indicates the last event of that
project. This constraint will recognize that the last event (completion of last activities) must take place before the
project deadline date.

3. A Prototype Example
In order to illustrate how to implement linear programming technique in scheduling project crashing time Al jabal
construction company, road infrastructure project has been selected. The project deals with the construction of a rod
of 463.875 meter long and it is consists of two phases, the first phase is known as foul and storm drainage works
phase and the second phase is known as rods and side walks phase. Table 4.1a and Table 4.1b show the list of the
various activities needed for each phase. The first column represents the activity identification number, the
description of each activity is shown in second column, and the third column provides the activity duration. Table
4.2 shows a project with hypothetical normal time - cost data and crash time –cost data.

Table 4.1a Project activities with duration for the first phase
Activity description Duration
ID
(Day)
Road (463.875 m)
1 Excavation foul 10
2 Foul pipe installation, backfilling & testing 10
3 Foul manhole construction 10
4 Foul laterals & hours connections 10
5 Excavation storm 10
6 Storm pipe installation, backfilling & testing 10
7 Storm manhole construction 10
8 Storm pipe junctions backfilling 10

Table 4.1b Project activities with duration for the second phase
Activity description Duration
ID
(Day)
Road (463.875 m)
9 Excavation to formation 30
10 Water house connections 30
11 Cable crossing ducts 30
12 Gully & gully connection to storm 30
13 Road endings 30
14 Compaction to formation 30
15 Sub base laying & compaction 30
16 Granular base coarse laying & compaction 30
17 Asphalt binder to roads 10
18 House concrete steps and ramps 12
19 Sidewalk backfill & base coarse 10
20 Sidewalk wearing 10
21 Asphalt wearing to roads 10
22 Finish roads & sidewalks road 0

311
 Table 4.2 Activities list with hypothetical normal and crash data
Crash Normal cost Crash
Activity Depends on Normal time (days)
time(days) $ cost $
code
1 - 10 9 4,400.00 4,840.00
2 1 10 9 3,035.00 33,385,00
3 1 10 9 4,200.00 4,620.00
4 2,3 10 8 87,600.00 105,120.00
5 4 10 9 4,400.00 4,840.00
6 5 10 9 30,350.00 33,385.00
7 5 10 9 4,200.00 4,620.00
8 5 10 9 2,400.00 2,640.00
9 6,7,8 30 28 7,650,00 8,160.00
10 9 30 28 10,950.00 11,680.00
11 9 30 28 12,570.00 13,408.00
12 9 30 28 10,400.00 11.093.00
13 10 30 28 29,643.00 31,620.00
14 13 30 28 20,161.00 21,505.00
15 14 30 28 28,292.00 30,178.00
16 15 30 28 33,133.00 35,342.00
17 16 10 10 33,416.00 33,416.00
18 16 12 10 5,460.00 6,370.00
19 17 10 10 7,400.00 7,400.00
20 11,12,19 10 10 12,600.00 12,600.00
21 18,20 10 10 38,567.00 38,567.00
22 21 0 0 0 0

Now, the question arises what will be or what will be the minimum completion date of the project after being
crashed. To determine this extent, we have to develop a CPM based on crash time of the activities that will provide
us the minimum time beyond which the project cannot be crashed. So, the adjusted deadline of the project must be
greater than or equal to the project.

The total duration for the completion of the project is 97 days based on the critical path. Assume, the project
manager decides to complete the project within 91 days. Now, the project manager should know how to crash the
activities of the project (with minimum additional cost) so that the total cost will be minimized.

Figure 4.9 shows AON network with maximum reduction time for each activity. The cost-time slopes of the
activities are shown in Table 4.3

Figure 4.9 Maximum reduction time of the network activities.

312
 Table 4.3 Cost-time slope of the activities
Activity time that Additional Cost Crash cost/time
Activity
can be crashed day [2] for crashing $ [3] [4] = 3/2 $/day
Code[1]
1 1 440 440
2 1 3,035.00 3,035.00
3 1 420,00 420,00
4 2 17,400.00 8,700.00
5
1 440,00 440,00
6 1 440,00 440,00
7 1 420,00 420,00
8 1 240,00 240,00
9 2 510,00 255,00
10 2 730 365,00
11 2 438,00 419,00
12 2 692,00 346,00
13 2 1,976.00 988,00
14 2 1,344.00 672,00
15 2 1,884.00 942,00
16 2 2,208.00 1,104.00
17 0 0.00 0.00
18 2 910,00 455,00
19 0 0.00 0.00
20 0 0.00 0.00
21 0 0.00 0.00
22 0 0.00 0.00

4. Linear Programming Model for Crashing the Hypothetical Project

Using the normal times from Table 4.5, the duration of each activity should be given by the following formula:
Duration of activity i = (ESj – Esi) – Xi (4.1)
To illustrate these relationships, consider activity 13 in the project network (Fig. 4.8)
Activity 13 represents immediate predecessor of activity 14:
Activity 13, which has Esi = 20/2/2012.
Activity 14, which has ESj= 25/2/2012.So, duration of activity 13= (25 – 20) – Xi = 5 – Xi. That is to say, activity
14 start after 5 days from starting activity 13.

By including this relationship for all the activities as constraints, the complete linear programming model is given
below.

Minimize Z=440 X1 + 3035 X2 + 420 X3 + 8760 X4 + 440 X5 + 3035 X6 +420 X7 + 240 X8 + 255 X9 + 365 X10
+ 419 X11 + 346.67 X12 + 988.13 X13 + 672.06 X14 + 943.09 X15 + 1104.46 X16 + 0 X17 + 455 X18 + 0 X19 +
0 X20 + 0 X21

Subject to :
a. Maximum reduction constraints :

X1 <= 1 , X2 <= 1 , X3 <= 1 , X4 <= 2 , X5 <= 1 , X6 <= 1 , X7 <= 1 , X8 <= 1 , X9 <= 2, X10 <= 2 , X11 <= 2 ,
X12 <= 2 , X13 <= 2 , X14 <= 2 , X15 <= 2 , X16 <= 2 , X17 <= 0, X18 <= 2 , X19 <= 0 , X20 <= 0 , X21 <= 0

b. Nonnegativity constraints:

X1 >= 0, X2 >= 0, X3 >= 0, ……………….. , X21 >= 0.

313
Y1=0 , Y2 >= 0 , Y3 >= 0 , ……………………, Y21 >= 0.

c. Start time constraints:

Y2 + X1 >= 2
Y3 + X1 >= 2
Y4 – Y2 + X2 >= 10
Y4 – Y3 + X3 >= 10
Y5 – Y4 + X4 >= 10
Y6 – Y5 + X5 >= 2
Y7 – Y5 + X5 >= 2
Y8 – Y5 + X5 >= 2
Y9 – Y6 + X6 >= 10
Y9 – Y7 + X7 >= 10
Y9 – Y8 + X8 >= 10
Y10 – Y9 + X9 >= 4
Y11 – Y9 + X9 >= 8
Y12 – Y9 + X9 >= 9
Y13 – Y10 + X10 >= 4
Y14 – Y13 + X13 >= 5
Y15 – Y14 + X14 >= 2
Y16 – Y15 + X15 >= 2
Y17 – Y16 + X16 >= 22
Y18 – Y16 + X16 >= 22
Y19 – Y17 + X17 >= 4
Y20 – Y11 + X11 >= 30
Y20 – Y12 + X12 >= 30
Y20 – Y19 + X10 >= 5
Y21 – Y20 + X20 >= 5
Y21 – Y18 + X18 >= 12
YFinish – Y21 + X21 >= 10

d. Project duration constraint:

YFinish <= 91, which is means that the manager need to crash the project time 6 days.

5. Analysis and Results

The linear programming model will not only take into account the activities on the critical path, but will also
consider the noncritical activities, which in their turn become critical as the project time decreases. Using LINDO
software, the solution of the model is presented in table 5.1, which indicates that the adding cost (crash cost) by
crashing critical activities to reduce the project duration within 91 Days is $. 2120 and the cost to complete the
project in normal duration (97 Days) is $. 418,145. So, the final project cost is computed by adding crash cost of the
crashing critical activities $2,120 to the normal cost $418,145. So, the final project cost is $420,265.

Subtracting the crash-time amounts from the normal completion, the result indicates that activity 1 should be
crashed one day. That is to say, activity 2 is starting after activity 1 by one day. Activity 5 crashed one day, 9
crashed two days and 10 crashed two days. The linear programming solution also provides a valuable sensitivity
analysis. As shown in the computer output in Table 5.1, the "reduced cost" column represents the save cost if other
activities "other than the activities that have small slope" were chosen at the same project deadline.

The total cost for crashing will be $ 2120 from this approach. First of all, the manual approach of crashing time is a
time consuming erroneous process. It requires trial and error method to get the optimal result. Linear Programming
model solution also gives us some flexibility by providing sensitivity report of the mathematical model. The linear
programming analysis carried out to determine the optimal policy of investing in extra resources in order to meet the
deadline is obtained. It is important for project manager to recognize the flexibility of the system that can be used to
explore numerous possible opportunities to the contractor. Moreover, this approach allows the user to easily
314
manipulate different project networks of various difficulties representing real world applications, and to study the
effectiveness of the model in the case of large projects. The implementation of the developed model showed more
efficient and reliable results and generated a considerable computational savings along with an increase in
robustness.

Table 5.1. Solution of the model

Objective Final value Reduced
value $ cost $
Z 2120.00 -
X1 1.00 0.00
X2 0.00 2,362
X3 0.00 420
X4 0.00 8,087.93
X5 1.00 0.00
X6 0.00 2,362
X7 0.00 420
X8 0.00 240
X9 2.00 0.00
X10 0.00 0.00
X11 0.00 419
X12 0.00 346.67
X13 0.00 316.07
X14 0.00 0.00
X15 0.00 217.03
X16 0.00 432.39
X17 0.00 0.00
X18 0.00 455
X19 0.00 0.00
X20 0.00 0.00
X21 0.00 0.00
Y1 0.00 0.00
Y2 1.00 2.00
Y3 1.00 0.00
Y4 11.00 0.00
Y5 21.00 0.00
Y6 22.00 0.00
Y7 22.00 0.00
Y8 22.00 0.00
Y9 32.00 0.00
Y10 34.00 0.00
Y11 40.00 0.00
Y12 41.00 0.00
Y13 36.00 0.00
Y14 41.00 0.00
Y15 43.00 0.00
Y16 45.00 0.00
Y17 67.00 0.00
Y18 69.00 0.00
Y19 71.00 2.00
Y20 76.00 0.00
Y21 81.00 0.00
YFinsh 91.00 0.00

315
6. Conclusion
The data needed for crashing project activities by means of linear programming technique are the time and cost for
each activity when it is done in the normal way and then when it is fully crashed (expedited). The project manager
can investigate the effect on total cost of changing the estimated duration of the project to various alternative values.
Using linear programming model, the project manager will be able to determine how much (if any) to crash each
activity in order to minimize the total cost of meeting any specified deadline for the project. An algorithmic model
based on linear programming incorporated with a minimal time-cost crash in a construction project was introduced.
The format of the model lends itself to a wide range of variables and considerations. The introduced modeling
strategy which showed the resources of this interactive approach including a bulk of data to completely analyze the
project is easily possible. It allowed a great number of parameters to simulate project conditions and contractor's
preference and provided potentially useful tool for decision making on project scheduling.

References
1. Senouci A.B. and Eldin N.N., 1996, “A time-cost trade-off algorithm for nonserial linear projects”,
Canadian J. of Engineering, vol. 23,pp. 134-149.
2. Archibald, R. D. and Villoria, R. L., 1966, Network-based management systems (PERT/CPM), Wiley, New
York.
3. Herroelen, W., and Leus, R., 2004, “Project Scheduling Under Uncertainty: Survey and Research
Potentials”, European Journal of Operational Research, pp. 289-306.
4. Lima, M.B., Silva, L.B., and Vieira, R.J., 2006, “Project Crashing and Costs Laws in the Knowledge Age”,
The 3rd International Conference on Production Research.
5. Tibben-Lembke, R.S., and Mitchell, T., 2007, “Activity Resource Elasticity: A New Approach to Project
Crashing”, University of Nevada.
6. Suri, P.K., and Bhushan, B., 2008, “Simulator for Optimization of Software Project Cost and Schedule”,
Journal of Computer Science, vol. 12, pp. 1030-1035.

316