A Hybrid Ant Colony Optimization Approach for the Cyclic Antibandwidth Problem

This paper presents a hybrid approach (HACO-CAB) consisting of an ant colony optimization approach and a local search operator for the cyclic antibandwidth (CAB) problem. The CAB problem which is one of various graph layout problems deals with placing an n-vertex of graph into the cycle C n in such a way that the minimum distance (calculated in the cyclic distance) of adjacent vertices is maximized. In this proposed approach, the role of ant colony optimization approach is to search potential breadth-first search spanning trees (solutions) or level structures of a high edge density graph so that upon applying labeling on such trees can lead to high quality solutions (in terms of fitness value) of the CAB problem. Two labeling schemes are applied to label the solution. To further improve the fitness value of current solution, a local search operator is applied. On a set of benchmark instances, HACO-CAB, in terms of computational results, shows its effectiveness in comparison to state-of-the-art approaches.