Abstract:

In this paper we describe a new Ant Colony Optimization (ACO) algorithm for solving Graph Matching Problems (GMPs), the goal of which is to find the best matching between vertices of multi-labeled graphs. We experimentally compare this new ACO algorithm with greedy and tabu methods on the base of different classes of GMPs. We also show that our algorithm finds the best solutions on a majority of UNINA benchmark instances.