Predictive multiobjective genetic algorithm for dynamic multiobjective optimization problems

Dynamic multiobjective optimization problems require an algorithm to continuously track a changing Pareto optimal solutions over time. Therefore, a new predictive multiobjective genetic algorithm(PMGA) is proposed, in which the centroid of Pareto optimal is soluted by clustering. And Pareto optimal solutions are described by applying the centroid points and reference solutions. Then the prediction set is generated by using the inertia predict and Gauss mutation. After an environment changed, the prediction set is incorporated in the current population to increase the population diversity by guided fashion. Finally, experimental studies on dynamic multiobjective optimization problems are carried out. The simulation results show that PMGA can quickly adapt the dynamic environments and track Pareto optimal solutions.