Modified Differential Evolution Algorithm for Solving Multi-Skill Resource-Constrained Project Scheduling Problem

The multi-skills resource-constrained project scheduling problem (MS-RCPSP) is assigning tasks to employees while considering predecessor tasks and multi-skills constraints to create a feasible schedule with both the shortest completion duration. Recently, Differential Evolution with multidimensional real-valued functions (DEM) was introduced to address the MS-RCPSP. The experimental results demonstrated that DEM outperformed GreedyDO, HAntCO, and GA. However, DEM lacked a technique to handle local optimum problem, resulting in unsatisfactory search results. Hence, this research proposes a modification to DEM by enhancing the mutation technique, executing it twice per generation, to increase population diversity and obtain better outcomes. The proposed algorithm was tested on twenty-three test cases from the MS-RCPSP iMOPSE datasets and yielded more satisfactory results compared to the competing algorithms.