Research And Realization Of The Genetic Algorithm For Dynamic Optimization Problems

Dynamic optimization problems exist in all kind of work in modern world. Commonly, the optimal solution of dynamic optimization problems will change at any time because of the change of objective functions, environmental parameters and constraint conditions. It makes factories encounter a lot of problems and losses in production scheduling and other work areas. In order to reduse the losses and track the moving optimal solution quickly and efficiently which is affected by the above reasons,this thesis proposes a new algorithm whichis based on a mature algorithm-genetic algorithm.The main tasks are the following four aspects in this thesis:1. Give a specific description of different evolutionary algorithms in dynamic environment. Elaborate on the elements of dynamic optimization problems. Describe the reseach progress and strategy of evolutionary algorithm for dynamic optimization problems.2. Introduce the idea of genetic algorithm and make further study of its basic principle, theoretical basis and technology implementation. Genetic algorithm is a robust evolutionary algorithm.3. Research on the Primal-Dual genetic algorithm (PDGA) and propose a new adaptively algorithm, due to the shortcoming of traditional genetic algorithm for dynamic optimization problems which called diversity loss. Test result demonstrates that the effectiveness of the proposed algorithm is better than the former one.4. Propose a novel double-probability primal-dual genetic algorithm, which integrates the greedy approximation method and double-probability primal-dualism scheme for solving dynamic knapsack problem. The result of the tests confirms the effectiveness of the proposed algorithm for dynamic optimization problems.