地面運補任務之路徑規劃實屬典型的車輛路徑問題(Vehicle Route Problem, VRP)， 由於VRP 最佳化為組合最佳化問題，其求解計算所需時間，會隨需求點數的增加呈指 數成長。本研究針對地面運補任務利用基因演算法(Genetic Algorithm, GA)與反應曲面模 式(Response Surface Methodology, RSM)，規劃最佳化運補路徑。首先採用傳統啟發式演 算法之最鄰近點(Nearest Addition, NA)法嵌入掃描法(Sweep Algorithm, SA)中，建立優良 的GA 初始染色體族群，並利用實驗設計法中的RSM，以有系統的實驗規劃與實驗數 據解析，適配反應曲面模式，以獲得求解「地面運補任務」之GA所需最佳交配率及突 變率組合，以有效逼近運補的全域最佳解，以提升該補給庫運補作業的成效。

The optimization of vehicle routing problem (VRP) is a combinatorial problem. The resolving time exponentially grows with the enlargement of demand points in a VRP. This study aims at local active distribution of army determining the optimal transportation path using response surface methodology and genetic algorithm (GA). The developed optimization procedure simultaneously consider the axial cluster and radius cluster for demand points locations relatively with logistic center by embedding the nearest addition method into sweep algorithm to establish a superior genetic chromosomes population. The response surface model of resolved CVRP path on crossover and mutation is established to determine the optimal crossover and mutation probability settings and thereby enhance the GA searching ability. Finally, the developed optimization procedure was employed to resolve a real case regarding the mission of local active distribution from army to determine the optimal transportation path.