Title: Diversity guided genetic algorithm to solve the resource constrained project scheduling problem
Authors: Israa' Y. Ismail; Mahmoud A. Barghash
Addresses: Industrial Engineering Department, Islamic University of Gaza, P.O. Box 108, Gaza, Gaza Strip, Palestinian Authority, Palestine. ' Industrial Engineering Department, University of Jordan, Amman 11942, Jordan
Abstract: This paper proposes a diversity guided genetic algorithm to solve the resource constrained project scheduling problem (RCPSP). The proposed algorithm is based on the concept of diversifying the population when it tends to converge in order to widen the search space and overcome the problem of premature convergence. Three sources of population diversity are suggested: multiple point crossover, diversity guided mutation (DGM), and no relative marriage (NRM)-based selection. The proposed algorithm also uses the forward-backward improvement (FBI) to transform a solution to one or more enhanced solutions during the search process. The computational experiments show that the proposed DGGA-FBI significantly outperforms the standard GA and gives competitive results with respect to the state of the art algorithms to solve RCPSP.
Keywords: genetic algorithms; GAs; population diversity; project scheduling; resource constraints.
International Journal of Planning and Scheduling, 2012 Vol.1 No.3, pp.147 - 170
Received: 07 Jun 2011
Accepted: 30 Mar 2012
Published online: 29 Oct 2014 *