Optimal Design Of A Part Of High-speed Underwater Vehicle Based On Genetic Algorithm

Currently, high-speed of underwater weapons has been needed by the military higher by higher. However paradoxically, the great resistance has been brought by the high-speed underwater vehicle as well. The main method of resistance-reducing of supercavitation motion body is making the movement formed around the cavitation. The cavitation device is an important component of cavity formation with high speed underwater. The quality of the design space is not only with the drag reduction effect of the underwater vehicle, but also has the great influence on the shape of the cavity. Genetic algorithm, based on natural selection and population genetics, is an iterative, evolution, broad applicability search method with global convergence.The development of the optimization design, the solutions of the problems and the development of Genetic algorithm has been introduced. Then the basic theory of optimal design and several conventional methods and genetic algorithm optimization theory has been described. And referred to simply introduce a genetic algorithm to remove invalid ideological constraints.In this paper, the high-speed underwater voyage has been analyzed through the dynamics to establish the dynamic equation. The optimization equation has been established through the boundary conditions of the cavity shape, which the semi-cone angle cavitator is the variable displacement, and the goal is the speed. Unlike the traditional optimization method,, the continuous differentiable is not required by the fitness function of genetic algorithm.Boundary conditions can be consider by the method of penalty function. Genetic algorithms is a very suitable method to solve the fitness function, in the process of solving with MATLAB, the population size of the genetic algorithm and the hereditary algebra should be set smaller to find the optimal results repeatedly of which the scope of the initial population is the result to increase the number of species and genetic algebra. In less time then we can find the best results. At the same time the local optimum into the genetic algorithm can be avoided.