A Novel Genetic Simulated Annealing Algorithm for the Resource-Constrained Project Scheduling Problem

A novel hybrid meta-heuristic algorithm, entitled as RCPSPGSA, is proposed for solving the resource-constrained project scheduling problem (RCPSP) in this paper. The algorithm incorporates the simulated annealing algorithm (SA) into genetic algorithm in order to improve local searching performance and boost up evolution capability. In each evolution iteration GA generates a new temporary population, and after that SA is used for improving every individual in it and at the mean time the next gap population is generated. For the sake of keeping the same convergence direction and speed of GA and SA, the cooling procedure occurs at the end of each evolution iteration. Simulation experiments are performed on the standard project instance sets of PSPLIB, and orthogonal experiment method is introduced to solve the parameter selection problem. Parameter combinations selected by this method are proved to be outperformed. Experimental results show that RCPSPGSA improves solution quality for J30, J60, J90 sets and not bad for J120.