Architecture for high-speed Ant Colony Optimization

The traveling salesman problem (TSP) is one of the most important problems in combinational optimization. Many works have done for this problem using ant colony optimization (ACO). The ACO is one of the most powerful optimization methods that combines distributed computation, auto-catalysis (positive feedback) and constructive greedy heuristic in finding optimal solutions for combinational optimization problems. Most of these previous works deal with software processing. However, ACO has the inherent problem of requiring substantial processing time. Therefore, the dedicated ACO hardware becomes important when applying ACO to combinational problems. In this paper, we propose a new hardware architecture for ACO. No previous studies have, to our knowledge, applied ACO hardware to TSP, as this study does using the proposed architecture. Experimental results to evaluate the proposed algorithm show improvement comparison with software processing.