The Research On Dynamic Multi-objective Algorithm Based On Preference Information

There are many multi-objective optimization problems(MOPs)in real life,that people have many expectations for them.For example,the classic path planning problem: the problem proposer hopes to obtain a path plan,that can ensure that the destination can be reached,also the shortest route,the shortest time consuming,and the least fuel consumption can be achieved as much as possible at the same time.At the same time,the environment and expectations of many problems in multi-objective problems may change over time.For example,in the path planning problem,the road condition information will change at different times,and the degree of importance the decision makers attach to each goal will also change under different circumstances.This paper refers to this kind of problems as dynamic multi-objective optimization problems based on preference information,or dynamic preference problems for short.And further put forward that the key to solving this kind of problem lies in how to reasonably optimize the dynamic algorithm and the preference algorithm according to the characteristics of the problem,and propose a strategy method that effectively combines them.In order to solve this kind of problems,this article mainly proposes two strategies,namely DAR strategy and PPM strategy.The DAR strategy dynamically changes the guiding strength of preference information for population evolution by adding adaptive adjustment of parameters.While ensuring the speed of convergence,this strategy effectively reduces the loss rate of boundary solutions in the evolution process,ensures the integrity of the final preference region,and improves the performance of the preference algorithm in a dynamic environment.The PPM strategy,by proposing a prediction algorithm for the preference area,reduces the lag of the preference area relative to the change of the preference point to a certain extent.Therefore,when the preference information changes rapidly,the algorithm can quickly and accurately obtain the solution set closer to the ideal preference region.In the experimental part of this paper,3 kinds of changing preference test problems are designed,and combined with 17 kinds of classical dynamic environment test problems to become 51 kinds of dynamic preference test problems.In these 51 test questions,the two strategies of DAR and PPM were compared and tested.The comparison of indicators and the experimental results show that the two strategies proposed in this paper have obvious optimization effects on 46 test problems.Experiments prove that the algorithm proposed in this paper is efficient in most cases.