A New Multi Objective Evolutionary Agorithm With Preference

Multi_objective optimization problems(MOP) usually correlate with multiple decision makers' preference and the decision maker's preference can not be quantified by using precise values. In this paper, a novel preference method based on multiple decision maker's preference(MDP) is presented. In the proposed method, stochastic vectors are generated in fixed intervals as the preference vectors, which are used as the weights of the objectives. By integrating preference with genetic operators, a new hybrid genetic algorithm with preference is proposed. The simulation results indicate that the different preferences will result in different optimal solutions, which shows the model is reasonable. There are some practical problems which can be modeled as an MOP with preference. Therefore, this studying is of great value in solving sort multiple decision-making problems.Compared to the evolutionary algorithms, convergence on evolutionary algorithms is relatively rare, a new MOP evolutionary algorithm model is set up and its global convergence is proved in this paper.