Parallel Genetic Algorithm Research And Application

Genetic Algorithm(GA) is a new subject. Its application began to flourish since the eighties of the 20's century. It exists natural parallel characteristic, Parallel Genetic Algorithm (PGA) is an important branch of GA and more and more specialists pay attention to it. In this paper, we discussed the theory of PGA more throughly. At the same time we presented PGA for solving three problems in the application achievement under the environment of the PVM. The first is finding roots of complex functional equation which is sponsored by the fund of AnLlui province ministry of education. The number of item is 99JL00006. We presented it by master-slave control networks PGA based on simulated annealing method, and the result is satisfactory. In this paper, we researched and discussed the mathematical theory and the main technology about the implementation of this algorithm, analyzed the efficiency of PGA . The algorithm is better than the iterative method and the downhill method etc. The second is the train control problem for saving energy, it's sponsored by the fund of the national natural science. The number of item is 69874001. We presented several train control instances: the level track, the gradient track, the limited speed etc. And the result is satisfactory too. In this paper, we discussed a mathematical model of the train control problem for saving energy~ the technology of application in the implementation of PGA. At last we gave some examples relatively and analyzed the efficiency of PGA. The last is the identical machine scheduling problem, we often fall across it in the engineering application. In this paper, we presented a kind of PGA which is based on master-slave control networks for solving identical parallel machine scheduling problem with constraint, and computational results show that this kind of PGA is efficient and fit for large scale identical parallel machine scheduling problem. We don't find the report about the PGA for solving the forward three problems. At the same time, the forward three applications show that PGA has 4 wide application foreground. PGA combining with other optimization method can improve the efficiency of algorithm.