Project Scheduling: A Memetic Algorithm with Diversity-Adaptive Components that Optimizes the Effectiveness of Human Resources

Abstract

In this paper, a project scheduling problem is addressed. This problem supposes valuable assumptions about the effectiveness of human resources, and also considers a priority optimization objective for project managers. This objective is optimizing the effectiveness levels of the sets of human resources defined for the project activities. A memetic algorithm is proposed for solving the addressed problem. This memetic algorithm incorporates diversity-adaptive components into the framework of an evolutionary algorithm. The incorporation of these components is meant for improving the performance of the evolutionary-based search, in both exploitation and exploration. The performance of the memetic algorithm on instance sets with different complexity levels is compared with those of the heuristic search and optimization algorithms reported until now in the literature for the addressed problem. The results obtained from the performance comparison indicate that the memetic algorithm significantly outperforms the algorithms previously reported.