Thesis (BS Computer Science) -- University of the Philippines Mindanao, 2007

Integer Programming (IP) is a technique that finds numeric solutions to problems wherein variables take integer values. IP problems are NP-complete and so far, no efficient solutions algorithms have been found to solve these types of problems. Heuristic methods have been applied to solve these types of problems. This study tested the efficiency of a hybrid algorithm, Particle Swarm Optimization and Simulated Annealing in solving Integer Programming benchmark test problems. The pure PSO algorithms was combined with a non-population based local escape hill-climbing heuristic Simulated Annealing (SA). To test the hybrid algorithm, different iteration values and SA parameters were used. Results of the test showed that for large numbers of iterations such as 25000, large values of Simulated Annealing temperature are needed to have a good solution quality and a better running time. But for smaller iterations, smaller temperatures are more suitable to use but a large geometric ratio would be better for both large and small iteration values. The results of this study showed that PSO-SA is a promising method in solving IP problems