Research On Membrane Calculation Models Of Two Biological Evolutionary Algorithms

In recent years,with the development of artificial intelligence?AI?technology,the scale of AI-related problems has become larger and larger.However,the speed of development of computing capabilities of electronic computer models is gradually slowing down,making it more and more difficult to solving large-scale problems in AI technologies.To break this limitation,more and more scientists are beginning to look for new computing models with more powerful computing capabilities.Inspired by information processing mechanisms in nature,natural computing with natural distributed parallel computing features is one of them.Membrane computing,as a new branch of natural computing,is a distributed parallel computing framework proposed by Romanian scientist,Prof.P?un.Nowadays,more and more researchers are paying attention to it.And concepts related to membrane computing have been widely used in combinatorial optimization,economics,computer graphics and linguistics and other fields.Membrane computing is also called P systems,its related researches mainly includes two aspects of computational model research and P system design.In the computational model research,researchers have proposed three basic computational models of cell-like P system,tissue-like P system and neural-like P system.In the design of P systems,researchers have also designed a series of P systems for problems in many fields,including numerical P systems that deal with problems in the economics field,arithmetical P systems that implement basic arithmetic operations,and P systems for solving NP-hard problems and so on.In previous studies,many researchers have used evolutionary algorithms such as genetic algorithms and ant colony algorithms in their designed P systems,but the algorithms were only used as sub-algorithms or abstract calculation process.Acturally,these evolutionary algorithms have not been implemented by completely using evolutionary rules in P systems.Based on the characteristics of maximum parallelism and non-determinism of the membrane computing model,this paper combines two representative biological evolution algorithms,the genetic algorithm and the ant colony algorithm,with membrane computing models,and proposes the design of genetic P systems and ant colony P systems.The working in this paper is shown as follows.?1?The genetic P system?_GA is designed based on the cell membrane system model,including its membrane structure and evolutionary rules.This paper also analyzes the computational performance,parallel characteristics,correctness and practicability of?_GA.?2?The ant colony P system?_ACS is designed based on the cell membrane system model,including its membrane structure and evolutionary rules.This paper also analyzes the computational performance,parallel characteristics,correctness and practicability of?_ACS.The proposal of?_GA and?_ACS not only extends the computational researches of membrane computing from basic operations to complex computations based on algorithm frameworks,but also provide new ideas for parallel implementation of genetic algorithms and ant colony algorithms.