Research For Cellular Genetic Algorithm

The nature is the source of inspiration for human. For hundreds years, the application of biology science in the practical problems is proved successfully ,and filtered in several subjects. Genetic Algorithms is a kind of global optimization random search algorithm based on the principles of natural selection and genetic evolution, and it has become one of the most important part of Evolutionary Algorithms (EA). Currently, GA is widely applied because of its simple structure, strong robustness and excellent capability of solving nonlinear function optimization compared to other traditional search algorithms.However,EA typically ignore the complex interactions between individuals(local populations) in spatially structured environments and speciation events.It not only shows that the natural disturbances have a strong effect on dynamices of the population,but also shows that the individuals of the population do not change,such as death,resurrection or migration.This is not the fact with the acctual biological evolution.So, it exists some flaws: slow convergent rate , bad global convergent performances etc.In order to improve the performances,this paper describes a novel evolutionary algorithm inspired by the nature of spatial interactions in ecological systems based on the cellular automata,called CGAD.The introduction of natural disturbances is similar with the regeneration after extinction in the natural environment,which will not take the population to degrade ,but to make them get rid of being slow state in evolution.This paper proves these expressions through testing some functions optimization, the algorithm improves its convergent rate and global convergent performances, and overcomes the premature. What's more, through numerical experiments data and figures, proved that the new algorithms has the better convergence rate and global convergence ratio than SGA.Cellular Automata is a discrete and dynamic model that has provided effective virtual laboratory in the field of large-scale simulation computation for studying the behavior of systems.The core of the cellular automata is the evolution rule,and different rule will lead to different results.In order to improving the life reproduction and the probability of survival,a new improving evolution rule is proposed by changing the cell state in the local neighbourhood based on the rule of'the game of life'introduced by Conway. Experiments conducted show that the cellular genectic algorithm with the improving evolution rule can be used for solving the optimisation problem of complicated function, which gives promising results.