Preventive Feedback PCNN Model and Its Application in the Combinatorial Optimization Problems

An improved pulse coupled neural network (PCNN) model is proposed to solve combination optimization problem with help of PCNN auto-wave characteristic. Based on Tri-state cascading pulse coupled neural network (TCPCNN), a preventive feedback method by using the triangle inequality theorem is introduced. In the process of searching solutions, all solutions are judged by the triangle inequality theorem and solutions of poor quality are removed. Therefore, the solution space complexity of combinatorial optimization problems decreases and the efficiency and accuracy are improved. This algorithm is applied to the shor test path (SP) and the traveling salesman problem (TSP) simulations. The results show that the proposed algorithm can effectively reduce space complexity and further improve the searching speed.