Multi-objective Co-evolutionary Algorithm And Its Application In Portfolio

There are many complex optimization problems in the field of engineering and scientific research,which exist more than two conflicting optimization goals.However,these optimization goals cannot be solved simultaneously to obtain the optimal results.Therefore,these optimization problems become an important puzzle in the optimization field.Since the evolutionary algorithm is based on population and it can obtain multiple Pareto optimal solutions after single run,which is very suitable for solving multi-objective optimization problems,it is of great significance to utilize the evolutionary algorithm to study multi-objective optimization problems.The main task of this paper is to study the multi-objective evolutionary algorithms based on the theory of multi-objective optimization and evolutionary algorithms by reading a large number of relevant literatures at home and abroad.In addition,we also design an efficient and effective strategy for obtaining the optimal solution of multi-objective optimization problems and portfolio problems.Finally,we also carry out corresponding experiments and theoretical analysis.The main contributions of this paper are as follows:First of all,aiming at multi-objective optimization problem,a Novel Multi-objective Co-evolutionary Algorithm Based on Decomposition Approach(MCEA)is proposed in this paper.A dynamic resource allocation strategy is designed in MCEA,which is used in the co-evolutionary algorithm framework.Since the complexity of computing each goal is different,the corresponding required computing resources are also different.Therefore,a dynamic resource allocation strategy is proposed for the rational allocation of computing resources.Besides,the differential evolution(DE)has a high global search capability.Therefore,in this paper,a powerful adaptive DE selection operator is designed to operate between subpopulations and archives to improve the population's convergence and diversity.Experimental simulation results show that compared with the relevant algorithms,MCEA performs best,and the proposed strategy can fully improve the performance of MCEA.Secondly,for the problem of portfolio optimization,multi-objective evolutionary algorithm is used to solve the optimal ratio of funds allocation.A multi-objective co-evolutionary algorithm for constrained portfolio optimization(CMCPO)is designed in this paper.First,in order to search for boundary solutions of portfolio problem,a multiple populations for multiple objectives mechanism is introduced to search for each target of the portfolio problem in the CMCPO.Besides,a parameter-adaptive DE operator is designed in the CMCPO so that the algorithm has better convergence and diversity.Finally,the experimental simulation results show that CMCPO has better performance compared with the comparison algorithms,and the parameter-adaptive DE operator can improve the performance of CMCPO.Finally,we draw a conclusion for this paper and provide an outlook of the future research.