Research And Application Of The Intelligent Algorithm

Because of its simple operation, good performance and easy realization,intelligent algorithm has been widely applied to various fields in real life,and it has been successful solved the complex optimization problems, compared with the traditional optimization algorithm, intelligent algorithms are concerned widely.Intelligent algorithm based on the population is a new method obtained by the understanding and research of social species behavior, this method is based on the exchange of information between groups, can quickly find the optimal solution of the problem. First introduce the origin of several intelligent algorithms, and then present the biological mechanism and the steps of the algorithm in detail,finally the mainly improved direction of the algorithms are summarized.According to the research of algorithms,this paper introduces a new intelligent algorithm –backtracking search optimization algorithm,this algorithm is a new evolutionary algorithm based on population that is introduced by Civicioglu P in 2013.First introduce the new algorithm operation in detail,and then for the algorithm’s shortcoming of convergence slowly, a novel mutation scale factor based on Maxwell-Boltzmann distribution and a crossover strategy with greedy property were introduced to improve it. Use Maxwell-Boltzmann distribution to generate mutation scale factor, which can enhance search efficiency and convergence speed. Adopting mutation population learning from outstanding individuals in less exchange-dimensional crossover strategy, add greedy property to crossover as well as fully ensure population diversity, that avoid easily trapping into local minima most algorithm have when added greedy property.Under the full study of the backtracking search optimization algorithm, for Artificial Bee Colony Algorithm’s shortcomings on convergence slowly and trapping into local minima easily,a hybrid search strategy based on multi-dimensional search and one-dimensional search in the employment bee search is presented, which can improve convergence rate of the algorithm by overcoming the shortcoming of convergence slowly under single one-dimensional search strategy. Moreover, put forward a new selection strategy for follow bee,to enhance the diversity of population and strengthen global searching ability.The PID controller parameters optimization problem as the research object,the backtracking search algorithm was applied to the parameter optimization problem. First produce the search area according to the previous parameter tuning method,the vector consisting of three parameters of the controller as the intelligent algorithm individual, then set up the appropriate fitness function, from the step response curve we can see that the algorithm can obtain better optimization effect in the practical application than the genetic algorithm.