The development of genetic algorithms for the finite capacity scheduling of complex products, with multiple levels of product structure

In this paper, the development of a genetic algorithms based scheduling tool that takes into account multiple resource constraints and multiple levels of product structure is described. The genetic algorithms include a repair process that rectifies infeasible chromosomes that may be produced during evolution process. The algorithm includes problem encoding, chromosome representation and initialisation, genetic operation, repair process, fitness measurement and chromosome selection. The data structure and algorithm are detailed step by step. The tool generates schedules that minimises the penalties caused by early and late delivery of components, assemblies and final products. The method is applied using data obtained from a collaborating company that manufactures complex capital goods. It is demonstrated that the schedules produced perform significantly better than those produced by the company using a conventional planning method.