The Optimization Of Differential Evolution Algorithm And Its Application Research

Differential Evolution (DE) algorithm is a simple and efficient global optimizationuncertainty searching method, proposed by Rainer Storn and Kenneth Price in1995.Because the basic principles of DE algorithm is simple, has less parameters and itssearching is random, parallel and global. It has been widely used in machine intelligence,pattern recognition, and other fields, and achieved good results. The DE algorithm isoutstanding compared to other algorithms, but the basic DE algorithm also has thedisadvantage of the other intelligent algorithms. The DE algorithm suffers from thecontradiction between convergence speed and accuracy, and the problem of prematureconvergence; additionally, it also suffers from the stagnation problem, where the searchprocess may occasionally stop proceeding toward the global optimum even though thepopulation has not converged to a local optimum or any other point; finally, DE algorithmis sensitive to the choice of the parameters and the same parameter is difficult to adjust todifferent problems.Therefore, study and analyze the performance of the algorithm on the basis of theexisting research, and improve the basic algorithm and applied it to the actual.First, review the background and purpose of the research. Through reading andanalyzing a lot of literature, summarize the main algorithms and their characteristicswithin the field of optimization research; present the background, characteristics anddisadvantage of the differential evolution algorithm.Second, introduce of the basic principle of differential evolution algorithm andanalyze the improved direction according to the characteristics of differential evolutionalgorithm, also summarize the main research achievements that have been achieved inrecent years.Third, through the analysis of existing literature and research, propose two improvedalgorithms. To achieve the improved performance of the algorithm improvements, thesetwo methods are improved by improving the evolutionary modes and parameters. Toimprove the search capabilities of the algorithm, two evolutionary modes are combined; to enhance the adaptability of the algorithm, and to some extent reduce the sensitivity of thealgorithm parameters, the adaptive parameters is used in the evolutionary process.Through typical benchmark function tests, the results show that the algorithm has highersearch accuracy and convergence speed, and has a strong ability of global optimization.Finally, according to the characteristics of differential evolution algorithm, thealgorithm was applied to the sun automatic tracking control. In floating power supplysystem, it is necessary to track the sun in order to make full use of solar radiation and toreduce the weight of the aerostat. The simulation analysis shows that: the parameters thatare optimized by differential evolution algorithm can make the tracking system achieverapid and accurate tracking.