An Improved Genetic Algorithm And Its Application

Genetic algorithm is a random search algorithm which uses the idea of natural selectionand biological evolution to search the optimal solution. It has been used in many fields. Manyresearches find that changing the crossover and mutation rate, designing a reasonable fitnessfunction make a difference in improving the performance of GA. The article proposes animproved GA which is used in function and combination optimization combining with therecent research via studying the basic theory, technical realization, the convergence andcharacteristics of GA deeply. The primary research works of the article are listed as follows:(1)Discusses the emergence theory and characteristics of adaptive GA in depth, analyzesseveral improved methods and summarizes the merit and demerit of the algorithm.(2)Improves the crossover and mutation rate, gives the improved formulas to improve theproblems of the adaptive genetic algorithm and some improved algorithms on the basis ofprevious studies. The crossover rate decreases along with the increase of the fitness function,while the mutation rate increases. The crossover and mutation rate of the individual, whosefitness is greater or less than the average fitness, could adjust adaptively in different regionseparately. The search accuracy and convergence speed are improved to some extent.(3) The fitness function is used to evaluate the individual. Satisfying the fitness function’design condition, the article gives a new fitness function via analyzing its mechanism ofaction in depth and using the logistic curve. Meanwhile, the article brings in the judgement ofthe individual number whose fitness is close to0or1, which is used to prevent theseindividuals occupying the population quickly. In the early evolution, the fitness and copies ofthe individuals with higher fitness are controlled. In the later evolution, the difference of theindividual’s fitness is increased. The competition of the population is enhanced. The problemof early maturing is improved. In the end, a test function is used for testing the performanceof the improved GA, which has better performance compared with other algorithms.(4)The improved GA is used for the function optimization and0-1knapsack problem,which shows good performance in search accuracy, convergence speed and stability.