An efficient teaching-learning-based optimisation algorithm for the resource-constrained project scheduling problem

This work proposes a teaching-learning-based optimisation algorithm as an alternative metaheuristic to solve the resource-constrained project scheduling problem (RCPSP). A precedence feasible activity list is employed for encoding the solutions whereas serial schedule generation scheme (SGS) is used as the decoding procedure to derive the solutions. In order to have good initial population, we employ a regret-based sampling method with latest finish time (LFT) priority rule. In addition to teacher and learner phase in basic TLBO, the proposed work also applies two additional phases namely self-study and examination for improving its exploration and exploitation capabilities. The algorithm is tested on well-known instance sets from literature. The performance of the algorithm is found to be competitive with the existing solution approaches available to solve this problem.