Research And Application Of Adaptive Multi-objective Particle Swarm Optimization Algorithm Based On Elitist Strategy

Most of the decision-making and planning problems which we face in real life are often affected by many factors.Although these factors may conflict with each other,but need to be considered comprehensively,which is similar to solving a multi-objective optimization problem(MOP).At present,the intelligent optimization algorithm has made great progress in solving MOPs,but there are still many shortcomings.For example,when a multi-objective optimization algorithm faces some complex problems,premature convergence and uneven solution distribution may occur.For different MOPs,the solution effects are often uneven.Based on a lot of in-depth research on MOPs,this paper explores the performance and application of multi-objective particle swarm optimization algorithm(MOPSO).The main contents include:(1)Firstly,aiming at the problem of insufficient convergence and uneven distribution of solutions in the multi-objective optimization algorithm based on Pareto dominated method,a new MOPSO algorithm is proposed,which combines the idea of particle velocity constraint and the diversity information of non-dominated solutions.Firstly,the velocity constraint factor is introduced into the particle velocity update process and a new velocity guidance direction is integrated to adjust the particle velocity to meet the needs of the global search ability in the early stage and the local development ability in the later stage;then the external archive is updated by using the non-dominated solution diversity evaluation mechanism based on the elite library to ensure the diversity of the nondominated solution distribution;finally,the external archive is used According to the distribution of non-dominated solutions in the archives,the best elites in the external archives are selected to guide the particles in the current state,so as to improve the efficiency of algorithm optimization.(2)PSO algorithm prematurity occurs when face some complex issues,which is due to not fully utilize the population to search for information to guide the search direction.In PSO algorithm,the process of particle optimization is affected by both global optimization and individual optimization,and tends to converge to the global optimal position quickly.In this way,the lack of diversity may lead to the failure to jump out of the local optimal.In order to avoid this situation,we proposes a MOPSO algorithm based on elitist strategy with global diversity and cross generation competition guidance.Based on the idea of competition strategy,a cross generation elite competition strategy is designed to guide particles to approach the Pareto front(PF).The efficient elite guidance mechanism based on global diversity is introduced.By evaluating the diversity of the current algorithm,the efficient elite guidance particles are selected to form a new speed guidance direction,and the competitive elites guide the particles together.(3)In this paper,the application of MOPSO algorithm is explored based on the constraints in the process of Mars probe braking acquisition.First of all,the collision factors of fuel consumption and orbit accuracy in the process of detector braking and capturing are taken as the objectives to form a MOP.Then,two improved MOPSOs that we proposed are used to optimize the problem.The simulation results show that the optimized Mars probe can not only achieve the target orbit accuracy requirements,but also reduce the fuel consumption to a relatively minimum.