Optimal Neural Networks for the Shortest Path Computation

One of the key algorithm research areas for unpiloted aerial vehicles team cooperative control is solution of network optimization models. Computational speed is an important success criterion as much of the cooperative work is done in real time. In this paper a neural network (NN) approach to solving the shortest path problem in the context of cooperative control problems will be introduced. We prove the optimality of our approach mathematically. Computational time will be used as a comparison and it will be shown that significant time savings result from the NN approach.