Evolutionary Computation and Mathematical Programming
Fernando Jiménez, José Luís Verdegay · 2001
A problem-independent evolutionary computation technique for general constrained optimization problems is proposed in this paper. Mathematical programming problems such as linear, nonlinear, integer, boolean and mixed programming problems can be solved by using this technique. Some important characteristics of the Evolutionary Algorithm are a natural representation of solutions, tournament selection, complete generational replacement and elitism strategy. Simulation results show a good performance by this method. An extension of this technique to find multiple nondominated solutions to constrained multiobjective optimization problems is also approached.