Solving Resource-Constrained Project Scheduling Problem with Particle Swarm Optimization

This paper presents a Particle SwarmOptimization (PSO) algorithm for solving Resource-Constrained Project Scheduling Problems (RCPSP). The PSO model is a new population based optimization strategy introduced by Kennedy and Eberhart in 1995. The PSO is a cooperative and competitive algorithm who belongs to the class of the evolutionary algorithms. We here specialize the algorithm of PSO to the particular case of the RCPSP. We have redefined the PSO operators (the addition of two velocities, the difference between two positions, the external multiplication between a scalar and a velocity). We propose in this paper, an extension of the PSO system that integrates a new displacement of the particles and we highlight a relation between the coefficients for each dimension between the classical PSO algorithm and the extension. The experiments on instances from the PSPLIB show that the proposed PSO algorithm is able to solve to optimality most problems of the series and close to the best known solutions on the hardest instances. This seems to indicate that this general algorithm is a good candidate for solving scheduling problems.