Simulated Annealing and Adaptive Dynamic Variable Band Mutation for Structural Optimization by Genetic Algorithms

Genetic Algorithm is known as a general tool for optimization in the field of discretevariables such as the case of structural problems. For such a search to be efficient andpractical, the genetic parameters should be selected of proper type and should be fine tuned. For this purpose, a direct index coding is utilized in this paper and the corresponding mutation operator is improved in two ways. First, it is modified incorporating simulated annealing concepts to form a generation based genetic algorithm. A dynamic variable band mutation is also developed as the second attempt in order to properly adopt mutation operator with the treated problem during optimization. A number of structural examples are treated by the proposed method to investigate its effect on efficiency and convergence rate of the genetic search.