Genetic Algorithm Applies To Optimum Structure Design Based On ANSYS

With the development of the science and the progress of the society, especially, due to the fast development of computers, high effective technology of optimum structure design and intelligent computation are in urgently needed. The theories of optimum structure design have developed for forty years, and many optimum methods are quite mature. In these years, a new optimum method, genetic algorithm, is rapidly developing. The method, originating from natural evolution and genetic mechanism and simulating auto-adapted phenomenon of nature, has merits of no relying on the question model, overall optimization, concealment parallelism, high efficiency, solving different non-linear problems, etc. This paper mainly makes these reseach:(1) This paper analyzes and synthesizes genetic algorithm's basic concepts, theory and its relative application contents in structural optimization. The paper introduces basic theory of structural optimization and other optimum methods; it analyzes and studies some key factors in GA's application, such as, mathematics modeling, processing restraint condition, generation of initial population, choice of controlling parameters, etc.(2) Simple genetic algorithm is characterized by"Premature Convergence (PC)" and weakness in local search. Considering the disadvantage of GA, this paper proposes to improve genetic algorithm, diversifying the initial populations, to avoid PC and improve optimization efficiency and quality.(3) This paper uses ANSYS instead of compiled procedure to optimize structures, to combine genetic algorithm optimization procedure with the structure finite element analysis procedure, making the data transmit between the two procedures. Thus, the combination can save a lot of time and improve the reliability of result.(4) This paper combines genetic algorithm with ANSYS to optimize computation of the two kind typical structures. The result indicates this method, being correct and feasible, can fully display the merits of both ANSYS and the genetic algorithm. All in all, higher efficiency, better result and stranger usability are their merit.