On the Capacity of Hopfield Neural Networks as EDAs for Solving Combinatorial Optimisation Problems

摘要: Multi-modal optimisation problems are characterised by the presence of either local sub-optimal points or a number equally optimal points. These optima can be considered as point attractors for hill climbing search algorithms. It is desirable to able model them avoid mistaking optimum global one allow discovery multiple solutions. Hopfield neural networks capable modelling patterns which learned from known patterns. This paper shows how network based on non-optimal samples an objective function. The resulting shown and generate solutions up certain capacity. capacity, method extending it studied.