Priority rule for resource constrained project planning problem with predetermined work package durations

Abstract

 Comparison of priority rules for the RCPSP  Examination of hybrid rules for the RCPSP  Proposition of a new priority rule for the RCPSP Figure A. A project schedule example for a single resourcePurpose: Project planning has a significant role in project management to progress regularly and complete a project on time. The studies for Resource Constrained Project Scheduling Problem, where the available resources are limited, indicate that the choice of priority rule affects the success of planning and there is not a single priority rule for reaching the optimum solution for all problems. In this study, a priority rule that can be used to solve Resource Constrained Project Scheduling Problem with predetermined work packages is proposed. Theory and Methods:In order to explore the most common 8 priority rules PSPLib J30, J60, and J120 datasets are used. Since the success of the rules changed according to the dataset, Maximum Area and Longest Progressing Time rules are selected from the priority rules, which give the most successful results, and examined. Two priority rules, which are Area and Resource Usage and Area and Process Time, has been proposed and tested on the PSPLib J30, J60, J120 dataset. The success of the rules is measured with the number of shortest schedule generation and average deviation for both serial and parallel scheduling. Results:In general, it is observed that the proposed Area and Resource Usage rule gives successful results in both serial and parallel scheduling, and it produces better results than the rules; Maximum Area, Highest Resource Usage, and Longest Progressing Time which are used separately. Also, the proposed rule was one of the two most successful one among the examined rules. It gave 57% better results than the First Eligible Activity rule, and produced the best results in the j120 dataset, where both the sample and the number of activities were higher.The proposed rule benefits from the shortest time generation feature of Maximum Area rule and the advantage of Highest Resource Usage rule that produces the closest to the optimum. In this way, it is able to produce more numbers of schedule which are close to optimum. In addition, the proposed Area and Resource Usage rule produces solutions below the average execution time of the other tested rules. Conclusion:The increase in the number of activities and samples leads to a decrease in the number of shortest-time schedule production for all priority rules and to an increase in the mean deviation. However, it was found that Maximum Area rule was successful in producing the shortest schedule and Highest Resource Usage rule was successful resulting in the lowest average deviation value. Therefore, a hybrid priority rule that takes these two rules into consideration is proposed. The proposed rule is successful when both the number of shortest-term schedule generation and average deviation are taken into consideration. It also performed in a time period below average execution time when compared with others.