A neural network to solve discrete dynamic programming problems
R A Francelin, Fernando A. C. Gomide · 2002
An artificial neural network to solve discrete dynamic programming problems is proposed. This approach presents some advantages with regard to alternative approaches because of the inherent parallelism of the neural networks. Some important applications are addressed to illustrate the usefulness of the approach proposed. An algorithm for dynamic programming neural network (DPNN) implementation and an equivalence between the DPNN and the dynamic programming procedure are presented.>