On mixed-discrete nonlinear optimization problems: A comparative study

Modified genetic algorithms are developed and presented in this paper. Principles of genetics and natural selection are adapted into the search procedure for mixed-discrete nonlinear optimization problems. Such classes of global optimization algorithms are based on a randomized selection of design space that yields an improvement in the objective function. An implementation of the approach to a series of test problems in engineering design optimization with diversity of variable representations and demonstrated nonconvexities are discussed, and the results are compared with other algorithms. Results show that genetic algorithms are able consistently to provide efficient, fine quality solutions that are robust to genetic parameters and provide a significant capability for mixed-discrete constrained nonlinear optimization problems.