Optimal Design Of Pre-Stressed Steel Structures By Improved Genetic Algorithms

Pre-stressed steel structure is a novel structural system and has been widely used in all kinds of buildings such as bridge, factory workshop, sport building and so on. The theory of its optimal design has yet to be solved in the follow aspects: Optimal design variables are treated as continuous, the case of discrete variable has not been considered; The load is only divided by person and its optimization has not been taken into account. In view of engineering practice, A highly efficient global optimization algorithms is introduced into optimal design of pre-stressed steel structure with discrete variables in this paper.To overcome the premature of simple genetic algorithm (SGA), the calculating efficiency of SGA is improved by the following two ways; Firstly, in order to ensure convergence of algorithm and reduce influence factor of convergence, elitist selection and adapted crossover and mutation operator are adopted. Secondly, because this study deals with finding minimize volume(weigh) of structure, to quicken convergence of this problem, in line with mechanics criterion the paper carry out mutation to some genes in chromosome according to variations of internal force. This will let algorithms to search answer with our hoping directions. The searching ability of improved SGA presented in this paper is verified by optimal design example using other algorithms.When applying genetic algorithm for optimal design of pre-stressed steel structure, ways of penalty coefficient is adopted for resolving stress and displacement restrains in this paper. Solution will convergent to optimal point of feasibility through penalty of solution outside feasible region again and again. In order to quicken convergence in this paper, The optimal process of pre-stressed steel structure is divide into several parts. At the end of this paper, load optimization to optimal designof pre-stressed steel structure is carried out. The value of every parts of load can be decided on the base of optimization.The feasibility and searching ability of improved SGA presented in this paper is verified by optimal design example of single pre-stressed steel structure and multiple pre-stressed steel structure.Thanks for the support of the National Nature Science Foundation of China to this paper.