An improved genetic algorithm for the optimal design of large trusses

An improved genetic algorithm is presented for the optimal design of large trusses. A method based on violations of normalised constraints with gradually increasing penalty parameters is used to convert the constrained optimisation problem into an unconstrained one. Advanced genetic operators such as two-point cross-over, niche, cloning and fitness scaling are used to obtain an improved genetic algorithm which is very efficient. The use of non-binary coding, which is useful especially in dealing with problems involving a large number of design variables, is also demonstrated. Numerical examples and comparison with known results from literature are presented to illustrate the efficiency and accuracy of the proposed method.