10.5.1 Genetic Algorithm as Automatic Structural Design Tool

The simple GA while powerful, is perhaps too general to be efficient and robust for structural design problems. First, function (or, fitness) evaluations are computationally expensive since they typically involve finite element analysis. Second, the (feasible) design space is at times disjointed with multiple local minima. Third, the design space can be a function of boolean, discrete and continuous design variables. The use of GA to find the optimal solution(s) of engineering design problems is still an open research area. Experience with GA has indicated that more often than not, tuning the GA strategy and parameters can lead to more efficient solution process for a class of problems. Researchers have proposed modifications, such as parameters-pace size adjustment and adaptive mutation for continuous problems, which focus on refining the searching space adaptively, niching genetic algorithms that emphasizing on repeating the fitter individuals and special modification for construction time-cost optimization problems. Research has also made it possible to combine genetic algorithms and gradient-based techniques for handling constraints for aerodynamic shape optimization problems. In this section, let's discuss how GA has been used as an automatic structural design tool.

10.5.1.1 Formulation of the Design Problem

The design of three-dimensional frames can be stated as follows.

where x is the design variable vector, f (x) is the objective function, ni is the number of inequality constraints, ne is the number of inequality constraints, nb is the number of...