Preference Multi-objective Optimization Based On Reference Vector And Its Application

The ultimate goal of the multi-objective optimization algorithm is to obtain a set of optimal solutions that are evenly distributed over the entire Pareto plane.In dealing with the specific problems in practice,the decision makers often only need to concentrate on the solution in a certain area,and the solution outside the area is irrelevant to the decision maker's decision.Searching these useless solutions wastes a lot of computing resources.In order to effectively deal with such problems,the researchers proposed a preference multi-objective optimization algorithm.The preference multi-objective optimization algorithm is based on the multi-objective optimization algorithm to join the decision-maker's preference information through a certain entry point,which can reduce the computational complexity of the algorithm and more quickly obtain the preferred solution that makes the decision maker satisfied.Therefore,this paper studies the preference multi-objective optimization algorithm,proposes a preference multi-objective optimization algorithm based on reference vector,and validates the algorithm by using a large number of test functions.Then,the mathematical model is established for the distillation process and the paper is used.The proposed preference method optimizes the two goals of energy consumption and production of distillation.The main works of the paper is summarized as follows:1.Firstly,the research on preference multi-objective optimization algorithm is carried out,which illustrates the key points when designing the preference multi-objective evolutionary algorithm,including how to ensure the controllability and distribution of the preference region.2.A Reference Vector Of Non-Dominated Sorted Genetic Algorithm-?(RV-NSGA-?)algorithm is proposed.In order to solve the defects of the reference point method,this paper introduces the reference point method.With reference vector,a reference vector based preference introduction method is proposed,and its performance is verified by experiments on a series of test functions.3.Model optimization of the rectification process.Using the preference algorithm proposed in this paper to optimize the energy consumption and yield of the rectification process,it can be seen from the optimization results that the optimization with the multi-objective optimization algorithm with preference will be better.