Abstract: This paper presents an improved genetic algorithm for the Resource Constrained Project Scheduling Problem (RCPSP). The schedules are constructed using a heuristic that builds active schedules based on priorities that takes into account the degree of criticality for the resources, which is derived from the solution of a relaxed problem with a constraint on accumulative resources. The computational results with instances from the PCPLIB library validate the effectiveness of the proposed algorithm. We have obtain the best average deviations of the solutions from the critical path value. The best known solutions were improved for some instances from the PCPLIB.

Cite: Goncharov E.N.
An improved genetic algorithm for the resource-constrained project scheduling problem
XIII International Conference Optimization and Applications 26-30 Sep 2022