Solving constraint satisfaction and optimization problems by a neuro-fuzzy approach

The solution of constrained satisfaction and constrained optimization problems using a Hopfield model requires determination of the values of a certain number of coefficients linked to the surrounding conditions of the problem. It is quite difficult to determine these values, mainly because a heuristic search is necessary. This is not only time-consuming but may lead to solutions that are far from optimal, or even nonvalid ones. So far, there have been no works in literature offering a general method for the search for coefficents with will guarantee optimal or close to optimal solutions. This paper proposes a fuzzy approach which allows automatic determination of Hopfield coefficients.