Genetic algorithms hybridized with the self controlling dominance to solve a multi-objective resource constraint project scheduling problem

The Resource Constraint Project Scheduling Problem (RCPSP) is one of the most challenged scheduling topics. Compared to the other scheduling problems, the RCPSP pays special attention to the consumable resources with limited capacities, which is the major issue that industry has to cope with. In our study, we tackle a Multi-Objective RCPSP with minimization of the makespan, the total job tardiness and maximization of the workload balance level. Non-dominated Sorting Genetic Algorithm II (NSGAII) and NSGAIII are applied at first to find approximated Pareto fronts. In particular circumstances, decision makers would prefer preselected propositions than the whole Pareto front. For this reason, we have integrated in our study, the Self Controlling Dominance Area of Solutions (SCDAS) in our algorithms find more fine-grained Pareto fronts, and solutions with good qualities on all objectives. Small, medium and large size instances, featured by different parameters of jobs and resources are tested. A comparative study is carried out where the hypervolume and the metric-C are used to evaluate the performances of different methods. The improvements brought by the SCDAS are proved regarding both metrics.