| As a search algorithm, after coding, fitness function, reproduction, crossover, mutation, and other proper design and operation ,genetic algorithm can make the global search and local search. It doesn't depend on the specific area, it is self-adapt, global optimization, and implicit paralle.It has a strong ability to solve the problem.Therefore, the genetic algorithm was continually applied to different areas, meanwhile it faces more and more challenges. When solve the problem of multiparameter optimization, the usual binary code string is quite long. A long string of binary choices will bring inconvenience to choice and cross-operation, and result in a waste of the computer's memory and inefficient use, and the lower the search efficiency. If a string of short code did not meet the required precision. In solving the multi-dimensional matrix, will face such a problem. Clearly, to solve the problem of multi-parameter optimization asked multi-dimensional matrix and other issues,the coding strategy should be improved.To address the above issues, this paper presents a binary coding method -matrix coding which can improve the binary coding method. Based on the number of parameters, the number of rows and columns are determined, in accordance with the actual requirements of the accuracy the length L matrix encoded string is determined. 1×m-dimensional binary code will be changed into a string of n×h matrix vector. In this coding method,the coding string of length L is reduced effectively .Selection, crossover operating capacity will be improved .The share of the computer's memory will be eased. The search efficiency of genetic algorithm will be further enhanced.The feasibility of matrix coding genetic algorithm is analyzed from the theory and application. This paper mainly has done the following works:1) To further study the genetic algorithms, genetic algorithm emergence,development,basic principles and characteristics,the common problems of genetic algorithm, genetic algorithm theory research and application are introduced.2) In analyzing the genetic algorithm, the matrix coding genetic algorithm is proposed. The coding strategies, genetic manipulation, fitness function and other related theories are disscused in details based on genetic algoritnm coded matrix.3) The genetic algorithm based on the matrix coding are applied to reality. The feasibility of matrix coding genetic algorithm is proved using PID parameters.Multiple Input n-dimensional continuous-time linear system pole placement and the problem of least squares model identification is the essences of multi- parameter optimization, multi-dimensional matrix operations. Therefore, it applies to the application area of matrix coding genetic algorithm. Matrix coding based on the genetic algorithm to achieve, the experiment shows that matrix coding genetic algorithm can solve such problems.In short, this article describes a genetic algorithms based on matrix coding to solve multi-parameter, muti-dimensional matrix. The basic principle of this method is to use parameters shown by matrix series.To reduce the length of the encoded string at the same time simplify the genetic operation and effectively increase the diversity of population and improve the search efficiency. It is reasonable and practicable to show this method by analying the theory and application, a feasible and effective method of sloving the muti-parameter, muti-dimensioned matrix solution. The theories of genetic algorithms and its application are further expanded. |
