Incest prevention and multicore combinated evolutionary algorithms for the job shop scheduling problem

Abstract

Evolutionary algorithms (EAs) have been successfully applied to scheduling problems. Current improvements towards convergence issues in EAs include incest prevention and multiplicity features. A multiplicity feature allows multiple recombination on multiple parents [7, 8, 9, 10]. The method was successfully applied to multimodal optimization problems. As a consequence of this approach it was detected that all individuals of the final population are much more centred on the optimum. This is an important issue when the application requires provision of multiple alternative near-optimal solutions confronting system dynamics as in production planning