A neural network for solving the travelling salesman problem on the basis of city adjacency in the tour
暂无分享,去创建一个
A neural network for solving the traveling salesman problem (TSP) is proposed. The network is a modified version of the network suggested by J.J. Hopfield and O.W. Tank (1985) In the network of Hopfield and Tank, a neuron Ux,i denotes city x occupying position i in the tour. This results in a network that, in general, is incapable of performing a shift in position for a number of adjacent cities, since this would temporarily increase the energy of the system. In the proposed network, a neuron Ux, y indicates whether or not cities x and y are adjacent in the tour. An extra layer is added for the detection and elimination of closed subtours. This approach seems to have four major advantages: (a) computer simulations show a faster and more natural convergence of the network towards a configuration that represents a solution; (b) larger instances of the TSP can be solved readily; (c) the energy function is simplified; and (d) a general hardware implementation is possible because the distance terms occur only in the external inputs
[1] J J Hopfield,et al. Neurons with graded response have collective computational properties like those of two-state neurons. , 1984, Proceedings of the National Academy of Sciences of the United States of America.
[2] H. Szu,et al. Fast TSP algorithm based on binary neuron output and analog neuron input using the zero-diagonal interconnect matrix and necessary and sufficient constraints of the permutation matrix , 1988, IEEE 1988 International Conference on Neural Networks.
[3] Sweet. Determination of parameters in a Hopfield/Tank computational network , 1988 .