複数許容解の記憶と交換子に基づくアントコロニー最適化

摘要

In this study, we propose an Ant Colony Optimization (AGO) algorithm based on a memory of plural acceptable solutions and exchange operators. In the proposed method, the intensification of solution search is realized by memorizing acceptable solutions searched by the ACO and introducing exchange operators to them. Also, the diversification of solutions is realized by using plural acceptable solutions. The proposed method is applied to the Traveling Salesman Problem, and its effectiveness is verified by the numerical simulations.%本研究では，複数の許容解の記憶と交換子に基づくアントコロニー最適化（ACO）を提案する．提案手法はACOで探索した許容解を保存し，これらに対して交換子の概念を導入することにより，探索の集中化を図る．また，複数の許容解を用いることで解の多様化も図る．提案手法を巡回セールスマン問題に適用し，数値実験によって提案手法の有効性を確認する．