Evolutionary Algorithm Based On Knowledge Representation,Extraction And Influence

The knowledge evolution algorithm extracts unknown,potentially useful knowledge in the process of evolution,it see the problem solving as a process of interaction between evolutionary algorithms,optimization problems,evolutionary knowledge.According to the generated interactive information,the algorithm can search more excellent feasible solutions,and it can enhance the ability of algorithm to deal with complex optimization problems.However,in practical applications,there are some problems in extracting and using knowledge,and the evolutionary algorithms used are not the same,but it can improve the problem solving ability of knowledge evolution algorithm and reduce the difficulty of solving the problem to solve the optimization problem effectively.The cloud model is used to represent the knowledge in the evolution process in this paper,it analyzes the process of knowledge extraction in detail and studies the similarity between optimization problems,and it combines different evolutionary algorithms in order to analyze the difficulty of optimization problem.This paper builds the simplest cloud model in the process of optimization as knowledge and estimate the problem according to the knowledge.It breaks the traditional coding methods based on individuals and links knowledge information with individual information,and this method simplify the process of knowledge extraction.The contents and results of this paper are described as follows:(1)Basic concepts of cloud model are introduced,cloud model is used to represent knowledge,the specific method of extracting knowledge is given,and the change of knowledge in the course of evolution is analyzed according to the experiment.(2)This paper combines evolutionary algorithm with cloud model,then it constructs similarity theory based on cloud model,and defines the concept of optimization at the same time.The search purpose of evolutionary algorithm is extended to find the simplest cloud model for corresponding problem.This paper proves that the original optimization problem is similar to the simplest cloud model and the influence of similarity on the difficulty is analyzed,in the meantime.Experiments show that the proposed method can reduce the difficulty of problem effectively.(3)A multi-population evolutionary algorithm based on cloud knowledge is researched.This algorithm extracts the knowledge of the simplest cloud model in the process of evolution,it estimates the optimization problem based on this knowledge,and the evaluation operator and the regional division operator are designed,on this basis,the solution space is divided into regions which can reduce the search area of population effectively.Experimental analysis proves that the algorithm can find global optimal solution quickly,and it can reduce the difficulty of optimization problem.(4)This paper proposes an adaptive differential evolution algorithm in the cloud coding environment.This algorithm uses the knowledge representation of cloud model.It changes the traditional coding to redundant coding,and this method allows knowledge information to evolve at the same time as individual genetic information.It is closer to reality than traditional coding and makes the process of using knowledge easier.Population individuals become cloud model individuals,each cloud model individual produces cloud droplets randomly,and excellent cloud droplets are preserved according to knowledge devouring strategy.