本研究旨在改進蟻群演算法的性能，以解決專案排程中的最優解搜索問題。傳統蟻群演算法受限於過早收斂和局部最優解的問題。為了克服這些限制，本研究提出一種混合蟻群演算法，結合了經典螞蟻、脫線螞蟻和固執螞蟻的行為，同時引入階段式排程的概念，以減少局部最佳解對最終結果的影響。透過使用小型專案驗證混合蟻群演算法的性能，從中觀察到每次獨立執行時，混合蟻群演算法都能成功達到最優解，而傳統蟻群演算法的成功率則較低。在增加活動數量的情況下，混合蟻群演算法仍能保持較高的成功率，而傳統方法則表現更差。這些研究結果顯示，混合蟻群演算法在尋找最優解方面具有相當的優勢，展現出良好的穩定性和高效性，對於解決複雜專案排程問題具有潛力。這些成果對於工業界的業務決策和學術界的研究發展具有重要價值，為未來相關研究提供了實用的解決方案。;The aim of this study is to improve the performance of ant colony algorithm to solve the optimal solution search problem in project scheduling. The classical ant colony algorithm presents the issues of premature convergence and local optimization. To overcome these limitations, this study proposes a hybrid ant colony algorithm that combines the behaviors of classic, deviated, and persistent ants, and also introduces the concept of phased scheduling to minimize the impact of local optimal solutions on the final results. The performance of the hybrid ant colony algorithm is verified by using a small-scale project, in which it is observed that the hybrid ant colony algorithm can successfully achieve the optimal solution in each execution, while the success rate of the classical ant colony algorithm is lower. The hybrid ant colony algorithm maintains a higher success rate when the number of activities is increased, while the classical method performs worse. According to these results the hybrid ant colony algorithm has a considerable advantage in finding the optimal solution, exhibits good stability and efficiency, and has the potential to solve complex project scheduling problems. These results are valuable for business decision-making in industry and research development in academia, and provide practical solutions for future related research.
