Cooperative discrete particle swarms for multi-mode resource-constrained projects

In this paper, the multi-mode resource-constrained project scheduling problem (MRCPSP) is considered for makespan minimization, which leads to even utilization of machine capacity. According to the characteristics of the considered problem, a discrete particle swarm optimization (DPSO) is adapted, based on which a cooperative optimization method with multiple discrete particle swarms (CPSO) is proposed for MRCPSP. Two swarms are distinctively adopted to the mode assignment and activity sequencing sub-problem. Once a mode list and an activity list are obtained, the two swarms cooperatively search for better solutions. In other words, one swarm updates the mode list in terms of the given activity list and the activity list is improved by the updated mode list by the other swarm. As the same time, a local search procedure is investigated to balance the exploration and exploitation. Computation results of Project Scheduling Problem Library (PSPLIB) sets show CPSO is efficient to solve complicated combination optimization problems.