Research On The Selection Strategy Of Many-objective Optimization Algorithm

Many-objective optimization problems(Ma OPs)have been widely used in social engineering.However,it has become a difficult issue in the optimization field because of the frontier sensitivity for Ma OPs.With the increase of the number of objectives,there are some problems such as the conflict between diversity and convergence,lack of selection pressure,and insufficient maintenance of diversity.In view of the above problems,this paper aims to explore and design mating selection strategies and environment selection strategies with stronger search capabilities and efficiencies,two kinds of manyobjective optimization algorithms(Ma OEA),namely Ma OEA-MS and Ma OEAES,are proposed and applied in the cluster head optimization problem of wireless sensor networks with low energy adaptive cluster stratification.This paper is mainly introduced from two aspects: mating selection strategy and environment selection strategy.Aiming at the conflict between convergence and diversity caused by the increase of the number of objectives in the many-objective optimization algorithm.This paper proposes a many-objective optimization algorithm based on mating selection strategy(Ma OEA-MS).The algorithm uses the dynamic balance function as the mating selection strategy,and analyzes the impact of the four different parameter designs on the balance function.At the same time,the improved dynamic penalty-based boundary intersection function and the nondominated sorting-reference point strategy are combined as the environment selection strategy,which increases the selection pressure while maintaining the population diversity.Simulation results show that Ma OEA-MS has better performance.In view of the problem of insufficient diversity maintenance and the lack of Pareto selection pressure caused by the increase of the number of objectives in the many-objective optimization algorithm.Based on the dynamic balance function matching selection strategy,this paper further studies the environment selection strategy and many-objective optimization algorithm based on the environment selection strategy(Ma OEA-ES)is proposed.The algorithm designed the improved Tchebycheff function as the maximum sorting strategy,the perpendicular distance from the individual to the ideal point toward the reference vector is used to enhance the convergence ability of the population,and the maximum perpendicular distance of each reference vector is calculated to maintain the population distributed.At the same time,the selection pressure is increased by using the ideal point distance strategy.The improved Ma OEA-ES was tested on 4 to 15 objectives of DTLZ and WFG test suites respectively.Comparing with the state-of-art many-objective optimization algorithm.The experimental results show the superiority of Ma OEA-ES.In order to further verify the effectiveness of the algorithm,this paper applies the proposed algorithm to the cluster head optimization problem of low energy adaptive clustering hierarchy(LEACH)for wireless sensor networks,and builds a many-objective wireless sensor network energy balance model.The four objectives with base station distance,cluster distance,overall network energy consumption and network energy consumption load balancing are optimized simultaneously to solve the problem of excessive energy consumption in wireless sensor networks due to clustering non-uniformity and randomness of cluster head selection.Simulation results show that,compared with commonly used manyobjective algorithms,Ma OEA-MS and Ma OEA-ES algorithms have better advantages in terms of the number of surviving nodes and remaining network energy.In other words,it can verify the effectiveness and feasibility of the proposed algorithm.