An Efficient Particle Swarm Optimizer with Application to Man-Day Project Scheduling Problems

Ruey-Maw Chen,Frode Eika Sandnes

doi:10.1155/2014/519414

Abstract

The multimode resource-constrained project scheduling problem (MRCPSP) has been confirmed to be an NP-hard problem. Particle swarm optimization (PSO) has been efficiently applied to the search for near optimal solutions to various NP-hard problems. MRCPSP involves solving two subproblems: mode assignment and activity priority determination. Hence, two PSOs are applied to each subproblem. A constriction PSO is proposed for the activity priority determination while a discrete PSO is employed for mode assignment. A least total resource usage (LTRU) heuristic and minimum slack (MSLK) heuristic ensure better initial solutions. To ensure a diverse initial collection of solutions and thereby enhancing the PSO efficiency, a best heuristic rate (HR) is suggested. Moreover, a new communication topology with random links is also introduced to prevent slow and premature convergence. To verify the performance of the approach, the MRCPSP benchmarks in PSPLIB were evaluated and the results compared to other state-of-the-art algorithms. The results demonstrate that the proposed algorithm outperforms other algorithms for the MRCPSP problems. Finally, a real-world man-day project scheduling problem (MDPSP)—a MRCPSP problem—was evaluated and the results demonstrate that MDPSP can be solved successfully.

Highlights

The results demonstrate that the proposed algorithm outperforms other algorithms for the multimode resource-constrained project scheduling problem (MRCPSP) problems
The well-known resource-constrained project scheduling problem (RCPSP) is a combinatorial optimization problem where activities are scheduled such that the makespan is minimized, while satisfying given precedence constraints between the activities and resources
The solution quality is measured by evaluating the ratio of optimal solutions (OPT) found which is calculated using (12); “besti” represents the best solution found for instance i

Summary

Introduction

The well-known resource-constrained project scheduling problem (RCPSP) is a combinatorial optimization problem where activities are scheduled such that the makespan is minimized, while satisfying given precedence constraints between the activities and resources. The scheduling target of MRCPSP is to find an adequate mode assignment for each activity and determine a satisfactory activity priority while satisfying the constraints, minimizing the makespan. Scheduling problems such as job-shop, flow-shop, and vehicle routing have been studied intensively, and are confirmed to be NP-complete in their general forms. Different man-day combinations are considered as a different operation modes This manday project scheduling problem (MDPSP) can be regarded as a MRCPSP problem.

Scheduling Problems

Experimental Results and Comparisons

Conclusions