Neural algorithms of two-dimensional packing

A detailed review of mechanism and method of packing´s neural network algorithms is made. Through a set of heuristic rules, a two dimension rectangular packing problem is mapped to Hopfield neural network. Research on application of Boltzman machine and Cauchy machine to solve packing problems is carried out, and firstly the chaos neural network is applied to two-dimensional packing problem. Then through simulation, characters of different neural network models and influence factors are studied. The experimental results indicate that compared with traditional packing algorithms, the results obtained by packing neural network are more stable and quicker due to the infinite parallelism