蟻群最佳化應用於旅行商問題之參數研究

近年來醫療資訊逐漸進步，但在面臨重大事故傷亡或是急診病患急救中，有著醫療資源卻無法妥善運用，災難現場負責人無法有效指揮調度可運用的資源，以及傷病患在後送醫院時分布不均的問題等問題的產生。一旦事故再次發生，我們是否能夠有效率地將適當的急救路線或是醫療資源配送做一個完善的路徑規劃，則是一個值得探討的問題。此問題可利用蟻群最佳化來克服；蟻群演算法（Ant Colony Optimization, ACO），又稱螞蟻演算法，是一種用來在地圖中尋找優化路徑的機率型演算法。而在文獻探討中我們會深入了解蟻群演算法及旅行商問題，它由Marco Dorigo於1992年的論文提出，其靈感於螞蟻在尋找食物過程中發現路徑的行為，為啟發式演算法的一種，這類演算法常起源於觀察自然界現象而來，如有模仿生物進化過程的演化演算法(evolutionary algorithms)，也有效法鳥類覓食的粒子群最佳化(particle swarm optimization)等等。由於這些演算法具有高彈性(不限於求解某一特定問題)、實作簡單、求解效率高等特質，目前已成為最佳化演算法中的新興熱門領域。有鑑於此，本論文先探討蟻群最佳化應用於旅行商問題之參數分析，透過此研究以了解在何種參數下可以在不同的路況下找到最佳的路徑，以達到有效率地整合所有可資運用的醫療資源，最後產生結果比較得到最佳的參數結果。

Recently, medical information advances gradually, but medical resources cannot be properly used and the person who is responsible for the disaster site cannot effectively dispatch the resources available when facing a major accident or rescuing casualties. In addition, uneven distribution of patients is also a problem at the time of sending patients into hospitals. Once an accident happens again, whether we can efficiently use the rescue route or medical resources to make a perfect path planning is worth our attention. This problem can be overcome by ant colony optimization (ACO), which is a probability algorithm and often used to find an optimal path in the map. Marco Dorigo proposed the ACO in 1992, which was inspired by ants looking for food. The ACO is a heuristic algorithm, which is originated from observing natural phenomena, such as evolutionary algorithms mimicking biological evolution and particle swarm optimization (PSO) mimicking birds or fish foraging. Since these algorithms possess high elasticity (not limited to solving a particular problem), is easy to implement, and can efficiently find the solutions, it has become an emerging field of research. Therefore, this paper first studied which parameters of the ACO for the traveling salesman problem (TSP) are suitable for searching the shortest route of different situations. Based on this research, we can find the best paths for different routes in order to efficiently integrate medical resources available.