Evolutionary Algorithm for the Novel RCPSP Scheduling Problem

Abstract

This paper proposes and states a novel problem called TDOS-RCPSP, a general case of the classical RCPSP problem. TDOS-RCPSP problem has many applications in science, especially in industrial production lines such as car assembly and textiles. In this study, the TDOS-RCPSP problem is proved to be NP-Hard. Besides, like other scheduling problems, TDOS-RCPSP will be classified and represented by Graham notation. Since the TDOS-RCPSP is NP-hard, a new evolutionary algorithm based on the Cuckoo Search strategy is proposed to find near-optimal schedules in a reasonable computation time. To evaluate the proposed algorithm, we perform an evaluation study using two datasets containing the iMOPSE dataset, the datasets of previous algorithms, and the TNG industrial sewing dataset. The experimental results show that the proposed algorithm has better performance than previous algorithms.