The Study And Application Of Adaptive Genetic Algorithm Based On The Gradient Operator Genetic

Genetic algorithm is a new intelligent optimization algorithm, due to theirversatility, without specific question types restricted search space, parallelprocessing, speed is fast , efficient and practical, not subject to search space iscontinuous, differentiable, whether the advantages and restrictions, for practicalproduction optimization problem provides a generic solution algorithm. At present,the genetic algorithm in image processing, automatic control, pattern recognition,function optimization, data mining, planning and design, robotics and the socialsciences and other fields has been widely used in and have achieved good results.First,this paper introduces the theory of genetic algorithm, and discuss theprinciple and the method, and analyses the shortcomings of the basic geneticalgorithm. For example: slow convergence, premature convergence, and is easy tofall into local optimal solution and other defects. Then, this paper introduces in detailthe niche genetic algorithm, including the niche genetic algorithm theory, nichegenetic algorithm in practical application and analysis, and puts forward thedeficiency of niche genetic algorithm. Therefore, this paper introduces the isolationtechnology and the gradient operator, formed a kind of isolation technology andgradient operator based on niche genetic algorithm. And it can be applied to functionoptimization problems, theory and practice show that, the new algorithm for multi-modal function optimization problem is better than basic niche geneticalgorithm.The main advantage of this algorithm is as follows:(1) adopt isolation technology. In accordance with the natural geographicalisolation technique, the initial population is separated into several subgroups, eachsubgroup evolved independently, without mutual influences. Each level of evolutionand its magnitude depends on the average fitness of. After isolation, on the subgroups of the evolutionary process of individual flexible control.(2) this paper adopts improved fitness function, the purpose is to accelerate theconvergence speed.(3) the paper into a new optimization operator - gradient optimization operator.Gradient operator has a fast optimization feature. Algorithm for each one, then agradient operator, so that individual to the local optimal solution is near, thisincreased understanding of the accuracy.