Hybrid Approach to Optimal Packing Using Genetic Algorithm and Coulomb Potential Algorithm

It is difficult and computationally time-consuming to find the best possible solutions for blank packing problems, because they include a lot of underlying combinational conditions. This paper presents two approaches for packing two-dimensional irregular-shaped polygonal elements—a real-encoded genetic algorithm and a hybrid algorithm using a real-encoded genetic algorithm and a local optimization algorithm. The local optimization algorithm presented is a novel one utilizing the Coulomb potential technique.

In the hybrid approach, the real-encoded genetic algorithm generates the order of the polygons while the coulomb potential algorithm determines the embodiment layout under the fixed combinations so as to minimize the scrap. The hybrid genetic algorithm is found to give better results for problems of larger size although it takes more computational time.