Combining artificial neural networks and evolution to solve multiobjective knapsack problems

Abstract

The multiobjective knapsack problem (MOKP) is a combinatorial problem that arises in various applications, including resource allocation, computer science and finance. Evolutionary multiobjective optimization algorithms (EMOAs) can be effective in solving MOKPs. Though, they often face difficulties due to the loss of solution diversity and poor scalability. To address those issues, our study [2] proposes to generate candidate solutions by artificial neural networks. This is intended to provide intelligence to the search. As gradient-based learning cannot be used when target values are unknown, neuroevolution is adapted to adjust the neural network parameters. The proposal is implemented within a state-of-the-art EMOA and benchmarked against traditional search operators base on a binary crossover. The obtained experimental results indicate a superior performance of the proposed approach. Furthermore, it is advantageous in terms of scalability and can be readily incorporated into different EMOAs.