Hybrid meta-heuristic algorithms for the resource constrained multi-project scheduling problem

Thesis Type: Doctorate

Institution Of The Thesis: Orta Doğu Teknik Üniversitesi, Faculty of Engineering, Department of Civil Engineering, Turkey

Approval Date: 2014


Supervisor: RİFAT SÖNMEZ


The general resource constrained multi-project scheduling problem (RCMPSP) consists of simultaneous scheduling of two or more projects with common resource constraints, while minimizing duration of the projects. Critical Path Method and other scheduling methods do not consider resource conflicts and practically used commercial project management software packages and heuristic methods provide very limited solutions for the solution of the RCMPSP. Considering the practical importance of multi-project scheduling and the fact that resource constraints impact the schedules and costs significantly, achieving an adequate solution to the problem is crucial for the construction sector. In this research, we present a new hybrid algorithm which is based on genetic algorithm, simulated annealing, backward forward improvement heuristics. The performance of the algorithms is compared with the performances of the known heuristic procedures and commonly used software packages using test instances particularly developed for multi-project environment. Effectiveness of the developed algorithm is further improved with the application of parallel computing strategies with a Graphical Processing Unit (GPU). Results revealed that effective resource management is a vital process but it is ignored by practitioners, heuristic methods and current software packages. Proposed algorithm showed significant improvements on the state of the art algorithms. It is also shown that parallel computing strategies with a GPU has high potential for meta-heuristic applications specifically for construction management research area in which there is a significant gap in the GPU research.