A Cooperative Machine Learning-Based Algorithm: The Case of Max-Min Knapsack Problem with Multiple Scenarios
摘要
The paper introduces an algorithm based on machine learning and a descent method to address a variant of the knapsack, namely the max-min knapsack. The resulting cooperative algorithm combines the following phases: learning, exploitation, and exploration strategies, working together to provide high-quality solutions. The algorithm’s effectiveness is demonstrated through a computational analysis on instances extracted from the literature, and the provided results are compared to existing methods, highlighting the superiority of the novel method.