Research On Improved Genetic Algorithm Based On Fixed Point Theory

Genetic algorithm is a random search method and it is the most widely method of optimization. But the genetic algorithm existing some deficiency. For example, the "immature convergence"and the convergence precision are not higher.Aiming at these problems, the fixed point theories introduced into genetic algorithm using fixed point algorithm of"Simplicial triangulation - Integer label". Propsing three kinds of improved genetic algorithms to solve optimization problem.Firstly, an improved genetic algorithm is based on triangulation, which will introduce the triangulations theory into the genetic algorithm. Firstly, it introduces the optimization problem into fixed point problem. With the relative coordinates of concept design the genetic coding, crossover operator and mutation operator, in order to improve genetic algorithm. Through this method overcomes the faults that with triangulations solving fixed point when using artificial label. We introduced gradually accurate triangulation thoughts into genetic algorithm, making the search range gradually decreased, improved the efficiency of the algorithm of search. Secondly, the improved genetic algorithm will introduce J ₁ triangulation and integer label into the genetic algorithm, the design of individual coding; using coding classify classifies individuals. The individual which carries on crossover operation according to the individual category is limited. If two individuals do not satisfy then the crossover limited, and one individual carries on mutation operation, this makes genetic operation more actual effect. Thirdly,an improved genetic algorithm based on hJ ₁ triangulation which the hJ ₁ triangulation is introduced into genetic algorithm, the algorithm increased increasing dimension operator, which find all the whole standard simplex as objective convergence criteria.In this paper, the fixed point algorithms were combined with genetic algorithm, guarantee the population diversity. It is solved the genetic algorithm convergence problem, and then through the test function to simulation experiments, the results show that the proposed improved genetic algorithm compared with other genetic algorithm has higher global and effectiveness.