Professional Documents
Culture Documents
to Schedule Activities of
Constrained Multiple Resource Projects
Lucy C. Morse, John O. McIntosh
and Gary E. Whitehouse
University of Central Florida,
Orlando, Florida
roject managers face ongoing challenges in
managing projects. A project must be
accomplished in minimum time, with minimum cost, and limited resources. Because of economic factors, the project manager must make
the best use of these limited resources within a
specific time frame. Project managers use a variety of techniques to accomplish this scheduling.
The most common approaches to project
scheduling are two traditional techniques that
developed in the late 1950s. One is the Critical
Path Method (CPM), the other, Project Evaluation and Review Technique (PERT). Although
CPM is strictly deterministic and with PERT
there is an uncertainty in the activity time estimates, the methods within these techniques are
quite similar. Both are computer-oriented, define
arrow network diagrams, and define the concept
of a critical path. For these reasons, PERT and
CPM are referred to jointly in this paper as
PERT/CPM.
Network analysis is a decision-making tool for
all levels of management. It may be used at various
stages of project management, from initial planning or analyzing of alternatives to scheduling and
controlling activities that comprise the project.
March 1996
35
Formulation of Algorithm
The project networks were scheduled using an
expanded version of Activity-on-Node Network
Analysis published in Project Management: IIE
Microsoftware [18]. Each of the 108 network
problems was first solved independently using
the ten heuristic methods previously described.
With a matrix of problem durations for each
of the ten scheduling heuristics, an attempt now is
made to find what combination of those heuristics gives the best results. Best is defined as
the minimum project duration. The combination
is a solution subset of two or more heuristics with
groups lowest mean, not necessarily the group of
heuristics with the lowest individual mean.
A heuristic computer algorithm was developed
to find which combination of heuristics minimizes
the durations of constrained resource project networks. An important difference between this
algorithm and other methods used to find the top
performing heuristics for allocating constrained
resources is that the objective here is to find
which combination or subset of heuristics minimizes project duration and meets this objective by
finding the average of the minimum durations.
Results
Using the computer algorithm to find the combination of heuristics giving the best answer was
applied to the durations of a group of 108 network problems. The means of the percentage
increase above optimum for each individual
heuristic and each combination of heuristics is
given in Table 1. As can be seen, the combination
of heuristics performed best. ACTIM with the
lowest mean percentage increase is a 5.2 percent
increase above optimum compared to the combination of two heuristics, which is 3.8 percent or
the combination of four heuristics, which is 2.9
percent above optimum.
March 1996
1*
ACTIM
ACTIM
ACTIM
ACTIM
ACTIM
ACTIM
ACTIM
LFT
LFT
LFT
LFT
LFT
LFT
ROT
ROT
ROT
ROT
ROT
ACTRES
ACTRES
ACTRES
ACTRES
MEF
MEF
MEF
FCFS
FCFS
SJF
Percent Increase
Above Optimum
5.2
3.79
3.23
2.91
2.77
2.71
2.68
37
Conclusion
This study presents a combination of heuristics
that find the average of the minimum durations
for the constrained resource problem. This combination not only supports previous research on
heuristic methods for setting priorities for constrained resource problems, but also produces
results that are significantly better than those
obtained by single heuristics. Due to the variable
nature of network structure and resource availability, different heuristics are more applicable to
certain network problems. For this reason using a
combination of heuristics gives a better chance of
approaching the optimum duration.
Combination of Heuristics. The main objective of this study was to examine constrained
multiple-resource single-project networks using
combinations of two or more heuristic methods
to allocate these resources and minimize the
total project delay. Ten simple rule-of-thumb
heuristics were selected and applied to a series of
networks. The heuristics are Shortest Job First
(SJF), First Come First Served (FCFS), Latest
Finish Time (LFT), Minimum Slack First (MSF),
38
shortest duration. Also, since the heuristics considering resources gave better results for some
problems, more resource-related heuristics should
be investigated.
Although this study was developed to find the
combination of heuristics giving the average of
minimum project durations, the program could
be changed to minimize the maximum error, if
that is the objective.
Combinations of heuristics to schedule project
activities give more optimal solutions than other
methods previously used. This combination of
heuristics not only supports the previous research
on successful simple heuristic methods that set
the priorities for constrained resource problems,
but also produces results that are significantly
better than those obtained by single heuristics.
References
1. Bedworth, David D. 1973. Industrial Systems:
Planning, Analysis, and Control. New York: The
Ronald Press Co.
2. Brown, James Taylor. 1995. Priority Rule Search
Technique for Resource Constrained Project Scheduling. Unpublished Ph.D. Dissertation, University of
Central Florida.
3. Cooper, Dale. 1976. Heuristics for Scheduling
Resource-Constrained Projects: An Experimental Investigation. Management Science, 22 (Jul.), 11861194.
4. Davis, E.W. 1974. Networks: Resource Allocation. Industrial Engineering (Apr.).
5. Davis, E.W. 1973. Project Scheduling Under
Resource Constraints: Historical Review and Categorization of Procedures. AIIE Transactions (Dec.), 297313.
6. Davis, E.W., and Heidorn, G.E. 1971. Optimal
Project Scheduling Under Multiple Resource Constraints. Management Science, 17 (Aug.), 803816.
7. Davis, E.W., and Patterson, J.H. 1975. A Comparison of Heuristic and Optimum Solutions in
Resource-Constrained Project Scheduling. Management
Science, 21 (Apr.).
8. Elsayed, E.A. 1985. Analysis and Control of Production Systems. Englewood Cliffs, N.J.: Prentice Hall
9. Holloway, Charles A., Nelson, Rosser T., and
Suraphongschai, Vicht. 1979. Comparison of a Multi-
39
40
March 1996
This material has been reproduced with the permission of the copyright owner.
Unauthorized reproduction of this material is strictly prohibited. For permission to
reproduce this material, please contact PMI.