Assessing the Effects of Several Parameters of an Enhanced Multi-Pass Algorithm for the RCPSP

In recent years, there have been many heuristic methods used to solve the resource-constrained project scheduling problem. Amid these, different multi-pass methods have been proposed. Application of these approaches requires the setting up of many parameters such as the priority rule, the number of iterations and the number of eligible activities retained for each activity selection step. But, the effects of these parameters on the final solution are not yet quantified. Besides, the solution of the multi-pass methods is built from scratch without exploiting the best already obtained solutions. In addition, knowing that the resource-constrained project scheduling problem possesses the same optimal makespan with its symmetric counterpart, the search spectrum could be enlarged by solving the symmetric problem. A multi-pass algorithm using the partial destruction and the symmetric problem solution techniques is proposed. An experimental plan based on the Taguchi method was executed in order to quantify the effects of the original heuristic parameters and to prove the benefits of implementing the proposed techniques