Recurrent Neural Approach for Solving Several Types of Optimization Problems
Ivan Nunes da Silva, Christiane Wagner, Lucia V., Rogerio A. · InTech eBooks · 2008
This chapter presents an approach for solving optimization problems using artificial neural networks. More specifically, a modified Hopfield network is developed and its internal parameters are computed using the valid-subspace technique. The developed approach allows to solve several classes of optimization problems through a unique neural network architecture. The optimization problems treated in this chapter are the combinatorial optimization problems, dynamic programming problems and nonlinear optimization problems. An energy function Eop was designed to conduct the network output