AN APPLICATION OF GENETIC ALGORITHM TOWARD SOLVING THE RELIABILITY PROBLEM OF MULTIOBJECTIVE SERIES-PARALLEL SYSTEMS

Since developing an appropriate solution for reliability optimization problem with mathematical programming methods has been considered as difficult techniques, the heuristic approaches increasingly has been applied. Multiobjectve Genetic Algorithm (MGA) has been among heuristic methods that was developed to find solutions for series-parallel systems to obtain maximum reliability, and minimum cost and weight at the system level. These are very common problems in engineering design such as mechanical and electrical systems. It has been shown that the Multiobjectve Genetic Algorithm offers proper results to these problems while it respects to the several objective functions such as reliability, cost and weight. This paper presents the combination of probabilistic search, and one of the decision making methods called Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The Multiobjectve Genetic Algorithm, allows us to achieve a proper design solution while it saves a considerable time compared with some other approaches. At the same time as the reliability, cost and weight were chosen as objective functions, the results obtained by this method showed an overall improvement in comparison to the existing GA method considering cost and weight as constraints.