Genetic Algorithm And It's Application Research In Control Theory

Genetic Algorithm is a kind of iterative stochastic searching algorithm of integrity optimization, which is built on Darwin's natural selection and genetic mechanism. This algorithm provides an effective new way to solve complex optimization problems that seems difficult to many traditional methods, and also brings new vigor to artificial intelligence control and intelligent control. At present, the genetic algorithm is still lack of a unified, complete theoretical system. Its search efficiency, scopes also need to be further expanded and improved. Meanwhile, genetic algorithms and fuzzy intelligent control integration need to be further developed. Conducting in-depth research in these three areas is of great realistic significance, so this paper has the following specific content:Firstly, genetic algorithms theoretical foundation is improved, and the relation between the genetic algorithm's renewable capacity of first-order, second-order model and control parameters are given. Meanwhile, Mathematical description of crossover and mutation genetic operator is given . Focus on the analysis of cross-operator, then, the necessary and sufficient conditions of cross-operator search scope is given. Based on the analysis of Markov chain in the use of genetic algorithms for two typical convergences, a unified standard genetic algorithm convergence is presented in this paper.Secondly, genetic algorithm performance measurement standards based on the computation time efficiency is proposed. Take this for reference, the size of group, the optimal parameters selection of crossover and variation probability are given, as well as several methods to improve the efficiency of genetic algorithms, including channels of improving the selection, crossover and mutation operator. At the same time with the new population diversity function, a modified genetic algorithm of system of parallel structures is implemented.Thirdly, linear and nonlinear shipping roll model is established, and stochastic simulations are done under simulation sea waves. While select a certain type of stabilizer fins as control object, in order to solve the problem of dimension disaster encountered when looking for the optimal parameters, a initial population of accelerating the convergence of optimization parameters is proposed. And a fuzzy intelligent controller is designed based on the genetic algorithm to make ship rolling less.Finally, Comparison and simulation of linear and nonlinear situations are done respectively under different sea conditions and encounter angles. Compared with the traditional designing method PID, the results show that the designs of genetic - fuzzy controller are better in both robustness and control.