Hardware‐oriented ant colony optimization and its application to dynamic route guidance problem

Abstract

AbstractAnt colony optimization (ACO) is based on the behavior of food gathering of ants and it is a powerful search tool particularly when applying it to combinatorial optimization problems. However, ACO has the inherent problem of substantial processing time, because it requires a lot of repetitive calculations. In this article, we propose novel hardware‐oriented ACO (H‐ACO) to reduce processing time. H‐ACO adopts new integer arithmetic instead of conventional floating point arithmetic in the optimization process and it achieves high‐speed processing while keeping the quality of solutions. In addition, H‐ACO reduces the hardware resources, accelerates the clock frequency, and improves the latency, if it is implemented as a dedicated hardware. Furthermore, we also propose a new route guidance algorithm which combines H‐ACO with Dijkstra's algorithm and it can be applied to a dynamic route guidance problem. Experiments using actual map data demonstrate the validity of the proposed route guidance algorithm. © 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.