Optimization of a mechanical design problem with the modified bacterial foraging algorithm

This paper presents the solution of a real-world constrained bi-objective mechanical design problem in presence of a dynamic constraint by using the Modified Bacterial Foraging Optimization Algorithm. This algorithm, originally designed to solve single-objective optimization problems, is adapted to include in its processes Pareto dominance as selection criterion, an external archive for elitism, and crowding distance as a diversity mechanism. The results obtained are compared with those provided by two evolutionary algorithms. The bacterial-based approach is able to provide a highly competitive performance and the type of solutions found are more suitable, based on the opinion of the expert mechanical designer. Presentado en el XII Workshop Agentes y Sistemas Inteligentes (WASI) Red de Universidades con Carreras en Informática (RedUNCI)