Negative Learning in Ant Colony Optimization: Application to the Multi Dimensional Knapsack Problem

Abstract

In this paper we continue our recent work on the development of a negative learning component for ant colony optimization, which is a metaheuristic algorithm that is mostly based on positive learning, that is, on learning from positive examples. In particular, we apply our approach to the well-known multi dimensional knapsack problem as a test case. The obtained results show that our negative learning approach significantly outperforms the standard ant colony optimization approach.