Research On Techniques For Continous Function Optimization Based On Genetic Algorithm

Population-based genetic algorithm is a kind of random searching method using C.R.Darwin's evolutionary theory and G.J.Mendel's genetic theory.Its application predominance lives in higher-dimensional, multimodal , non-linear and multiobjective complicated problems,which are difficult for traditional searching methods.With the development of computer technology,genetic algorithm which is one of key technologies of computational intelligence in 21^st century have been applied to computer science ,engineering technology, management science,social science and so on.Continous function optimization is very important for us and is also an important research domain of genetic algorithm's research and its application.In the real world,continous function optimization,due to the complexity of problem itself ,simple genetic algorithm cannot solve such problems.After analysising existing genetic algorithm with related technologies,two kind of genetic algorithms are presented,which can solve single objective continous function optimization problems and multiobjective continous function optimization problems respectivedly.The first genetic algorithm which uses multi-parent to generate offspring in order to maintain the diversity of population and uses elitism preservation can improve the convergence of genetic algorithm.The second genetic algorithm,local search and neighborhood search mechanism are introduced,can improve the searching efficiency and can maintain the diversity of population through fitness sharing using larger population,which can find the near-optimal and near-complete Pareto front of problems.Numerical experiments have shown that two algorithms can both gain better performance.