Application of Hopfield neural network routing algorithm in nongeostationary satellite communication networks

The Hopfield neural net routing algorithm can be used in seamless Walker delta satellite communication networks to solve the routing problem. The method of simulated annealing as a measure of global optimization is proved to be a perfect way to solve the problem of oscillation ring and so eliminate the local stable state. However, the Hopfield neural net is a large scale net, which needs a large quantity of neurons and consumes much time for calculating. We propose a neuron-reduced Hopfield neural network routing, which not only needs fewer neurons but also quickens the speed of the routing calculation. We choose the optimal parameters for Hopfield-net in two nongeostationary satellite networks with different altitude, and compare the performance of this new Hopfield routing algorithm with various origination-destination-pairs at different times.