這篇論文專注於多機器人追逐單一目標的追趕問題。在已知的地 圖上， 目標為最大化目標物的偵測機率。此研究提出以聚類演算法與 貪婪演算法來解決這個問題， 在目標函數為次模函數的情況下， 貪 婪演算法帶有理論保證值。模擬證實了提出的演算法比起另一個演算 法更好。;This research focuses on pursuit-evasion(PE) games with multi-pursuer and an evader. Given the map, the objective function is to maximize the probability of the evader detection. The proposed algorithms compute search path of pursuers via cluster algorithm and greedy algorithms. Since the objective function is submodular, the algorithms have theoretical guarantees. The simulations demonstrate that the proposed method outperforms benchmark approaches.
