Improving RCPSP solutions quality with Stacking Justification – Application with particle swarm optimization

• We present a new justification technique, Stacking Justification, for solving RCPSP.
• Using multiple justification techniques was proven to be effective.
• Using combined priority rules will properly spread swarm particles.

Justification is a simple technique which was presented for improving the quality of schedules generated with algorithms for solving RCPSPs. Few researches implemented justification with meta-heuristics and proved that algorithms optimization results using justification outperforms the performance of the same algorithms without justification. In this paper, Stacking Justification is presented for further improvement of generated schedules quality. For testing the performance of this new justification technique, the multiple justification particle swarm optimization (MJPSO) algorithm was developed. Test results show that implementing Stacking Justification in a multiple justification approach was having a significant effect on solutions quality, and the performance results were very effective when compared to best performing algorithms in literature. Additionally, combined priority rules (CPR), a suggested approach for proper spread of swarm particles over several good quality areas within search space during swarm initialization, showed a considerable improvement on results especially for small number of generated schedules which is more suitable for practical applications.