A Hopfield network for solving a complex scheduling problem in behavioral synthesis

Hopfield type neural networks have shown success in solving hard optimization problems. As problems become more complex, it is increasingly difficult to find a suitable encoding of constraints into an energy function, which will effectively guide the network to a desired solution. This paper describes how a difficult scheduling problem was encoded into a Hopfield type neural network, and reports the results. The neural network devised is one of the most complex ones to date.