A scale-free based memetic algorithm for resource-constrained project scheduling problems

Abstract

The resource-constrained project scheduling problem (RCPSP) is a popular problem that has attracted attentions of many researchers with various backgrounds. In this paper, a new memetic algorithm (MA) based on scale-free networks is proposed for solving RCPSPs, namely SFMA-RCPSPs. In SFMA, the chromosomes are located on a scale-free network. Thus, each chromosome can only communicate with the ones that have connections with it. In the experiments, benchmark problems, namely Patterson, J30 and J60, are used to validate the performance of SFMA. The results show that the SFMA performs well in finding out the best known solutions especially for Patterson and J30 data sets, besides, the average deviations from the best known solutions are small. Therefore, SFMA improves the search speed and effect. © 2013 Springer-Verlag.