| Title |
A Study of Optimized Three Dimensional Steel Frame DesignBased on the Improved Genetic Algorithm |
| Keywords |
genetic algorithm ; optimum design ; convergency performance ; steel framework ; discrete variables |
| Abstract |
A Genetic Algorithm(GA) is a strategy that models the mechanisms of genetic evolution based on the principles of survival of the fittest and adaptation. The optimization using a GA is readily adaptable to discrete problems without any approximation and offers an open format for constraint specification. GAs have been successfully applied to many structure design problems.
But GAs have demerit of late convergence. In this study, various methods are used to complement this demerit of convergency performance. These methods are Simple Genetic Algorithm(SGA), Parallel Genetic Algorithm(PGA), Binary String Genome(BSG), Interger Array Type Genome(IAG), Roulette Wheel Selection(RWS), Tournament Selection(TS), Power Law Scaling(PLS) and Sigma Truncation Scaling(STS).
A design procedure incorporating GAs is developed for discrete optimization of three-dimensional steel framed structures. This procedure conforms to the Allowable Stress Design(ASD) method. The objective function considered is total weight(or cost) of the structure. The objective function is minimized subjected to strength requirements. Special features of proposed method include discrete design variables, the standard steel sections provided by AISC manual, and the upper methods. |