Lagrange programming neural network for solving constraint satisfaction problem and its extended problem

The constraint satisfaction problem (CSP) is to find a variable-value assignment which satisfies all given constraints. Because of its well-defined representation ability, it can compactly represent many problems in AI. To solve the CSP, we proposed a neural network called LPPH-CSP. In this paper, we extend LPPH-CSP to deal with the CSP which has an objective function and linear inequality constraints.