本研究採用最短路徑網路攔截問題為基礎,探討國土防衛縱深作戰的兵力部署策略,以雙層規劃建構「防禦者-侵略者」模型。同時提出兩種巢式序列架構的求解方法,上層採用進化式演算法(簡群演算法或基因演算法)求解伏擊點設置的最佳化問題,下層則以Dijkstra算法求解最短路徑問題,並透過實例問題進行實證分析與演算法之比較。
This study applied the shortest path network interdiction problem to optimize the force deployment for homeland defense in depth, and constructed a "defender-attacker" model by using bilevel programming. For solving the proposed model, two nested sequential approaches were proposed. The upper level uses an evolutionary algorithm (simplified swarm optimization or genetic algorithm) to optimize the selection of ambush points, and the lower level uses the Dijkstra algorithm to solve the shortest path problem. Numerical simulations on 18 real cases for model analysis and verification of methods were given.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。