Using elitist particle swarm optimization to facilitate resources leveling optimization analysis

The present study developed a new algorithm to find the implementation plan that had the smallest variance of resources while meeting a fixed duration. And the proposed algorithm modified a population-based search procedure, particle swarm optimization, by adopting an elite archiving scheme to store nondominated solutions and by aptly using members of the archive to direct further search. The termination of the proposed algorithm was set to that no new members have entered into the elite archive. In addition, this paper also established an evaluation function between the resource variance and the actual start time of activities by introducing a virtual variable. Through a practical network, the proposed algorithm is shown effective and efficient in conducting resource optimization analysis. Future applications of the proposed algorithm are suggested in the conclusions.