A new artificial neural network approach to solving knapsack problems

The knapsack problem is a classical combinatorial optimization problem which plays an important role in network assignment. Artificial neural networks have been effective methods in other large-scale optimal problems, and in this paper artificial neural networks for the knapsack problem are explored. A new neural network to solve knapsack problem is presented. The proposed neural network is proved to be completely stable. The simulation examples show its efficiency and the validity of the algorithm in finding the optimum solution to knapsack problems.