Research On Multi-Agent Genetic Algorithm Based On Uniform Design And Application In Game

The multi-objective optimization problem means that there are multiple optimization objectives,and each optimization objective tends to be in conflict.If only one of the objectives is considered,the optimization results of other objectives are reduced.This type of problem has always been a hotspot in the neighborhood of optimization research.problem.The traditional mathematics planning method has many deficiencies in solving multi-objective optimization problems.For example: relying on the mathematical characteristics of the problem,the lack of space for the problem solution search,and the high computational complexity of the algorithm.As a kind of intelligent algorithm,genetic algorithm is widely applied to multi-objective optimization problems due to its high robustness and weak dependency on the property of the problem.In this paper,the uniform experimental design and multi-agent technology are combined.With the help of co-evolution,a multi-agent genetic algorithm based on uniform design is proposed to solve multi-objective optimization problems.And the proposed algorithm is applied to the mobile phone tower defense game.According to the characteristics of the tower defense game,a numerical model of game elements is established to prove the application value of the algorithm.This thesis mainly works as follows:1.Introduce the research background,research content and organizational structure of this paper.2.Briefly introduce relevant theories,classical algorithms and development history of genetic algorithms and multi-objective optimization problems.3.Based on uniform experimental design,multi-agent technology,and co-evolutionary thinking,a new multi-agent genetic algorithm based on uniform experimental design is proposed.The new algorithm evenly distributes the initial population by uniform population initialization.Then through the field of competition,uniform crossover,mutation,self-learning and other operators to complete the evolution of the population.Finally,by co-evolutionary operators,two independently evolved populations exchange and cooperate with each other,thereby increasing the diversity of the population.4.Apply the new algorithm proposed in this paper to mobile phone tower defense game.According to the game design theory and player characteristics,design the numerical model of the game element and abstract the game problem model.The processing of the algorithm according to the characteristics of the game means that it can be better applied to the game.