Improved Columnar Competitive Model for Travelling Salesman Problem

This paper proposes an improved columnar competitive model (ICCM) to solve the travelling salesman problem (TSP). The ICCM with winner-take-all (WTA) principle can derive solutions for the TSP by providing an improved energy function. It shows that the competitive computational network guarantees the convergence to feasible states, and it is proved that the ICCM can escape from the local minimum. The results are compared with some other methods.