A novel swarm intelligence approach for the constrained scheduling problem

The Resource Constrained Project Scheduling Problem (RCPSP) is a prevalent problem in the domains of Project Management and Operations Research. This problem demands a huge number of computational resources as the project's complexity increases. Hence, approximate methods using heuristics and metaheuristics have proven to provide feasible solutions to the problem in acceptable time. This research adapts a prominent Swarm Intelligence (SI) metaheuristic, the Firefly Algorithm (FA), to solve the RCPSP with some modifications. By applying a 2-point crossover for the intensity-based firefly movement, a discrete framework has been provided. Further, to reduce the occurrence of the solution being stuck in the local optima, a swap-based mutation is introduced. The performance of the proposed research has been thoroughly evaluated through extensive experiments conducted on widely recognized benchmark instances. Comparison with seminal works demonstrated the competitiveness of the approach.