Thesis (BS Computer Science) -- University of the Philippines Mindanao, 2006
Constraint handling is considered one of the most complicated parts of engineering design optimization. Real-world limitations frequently introduce multiple, non-linear and non-trivial constraints on a design. Due to this complexity and unpredictability, a general deterministic solution is hard to find. In recent years, several evolutionary algorithms, search techniques, heuristic and meta-heuristic methods have been proposed for constrained engineering optimization problems. Hu et.al. (2003) used Particle Swarm optimization (PSO) in solving such problems. However, despite the good results, they also found out some limitations to the study. To avoid those restrictions and create a more efficient algorithm that would still generate favorable results, this paper presents an embedded hybrid of PSO and Simulated Annealing (SA) for solving engineering optimization problems. PSO is a heuristic type of algorithm that generate solutions which are near optimal while SA is a generic probabilistic meta-algorithm for the global optimization problems, namely locating a good approximation to the global optimum of a given function in a large search space. Four benchmark engineering problems with constraints were tested namely, (1) pressure vessel design problem, (2) welded beam design problem, (3) minimization of the weight of the tension/compression spring, and (4) Himmelblau?s nonlinear optimization n problems. The best solution of the above-mentioned method is better compared to all other algorithms previously reported in the literature