交通管理單位為維護道路鋪面，需長期支出鉅額預算以確保其品質，而陸運車種中，對鋪面破壞最嚴重的車種為卡車，且卡車業者又常以超載的方式，試圖最大化其營運利潤，因此，管理者需針對超重卡車制定出一套管理準則，用以規範卡車業者，並減緩卡車對路面環境與鋪面破壞情形。目前國外多使用動態地磅做為管理超重卡車的策略，而現今由於科技進步，卡車業者易於尋找最小成本路徑，且實務上除了動態地磅之外，還可增添更多道路管理策略，故本研究考量將路段禁行概念，納入規劃管理超重卡車的範疇，並對新增的超重卡車管理策略做探討，使將來決策者能有效率地規劃不同策略之位址。 本研究為模擬實務中交通管理單位與卡車業者交互之行為，利用數學規劃方法與網路流動技巧，發展雙層數學規劃模型，目標為最小化管理單位維護道路破壞成本，與道路服務水準降低所致之管理成本。模式求解方面，本研究為模擬管理者選址之時決策方式，上層使用數學規劃模式並透過套裝軟體CPLEX求解，下層則採用最短路徑演算法求解，將兩者結合後，發展多回合啟發解法。輸入資料部分，本研究參考內華達州之模擬路網進行範例測試，測試結果顯示，部分道路禁行或關閉部分車道後，管理者所需支出之道路維護成本，相較以往僅考量設置地磅策略時多半較低，而求解效果良好，並對輸入資料的參數進行調整，進行敏感度與方案分析，期能供未來管理者於不同情境下，在輸入參數部分，做出有效之調整與後續決策。;As far as road administration is concerned, the overweight trucks could result in serious damage to roads. Road administration need to reduce the negative costs caused by truckers; as a consequence, the government should enact relevant laws. To solve the problem, weight-in motion has been widely discussed in recent years. As a result of the well developed technology of today, truckers would drive on the minimum cost roads easily instead of k-th shortest path. This thesis presented multiple strategies of road traffic management based on the concept of restricted traffic lane and road closed.
This thesis employs mathematical programming methods and network flow techniques to develop a Bi-Level mathematical programming method. The upper level aims to minimize the damage caused by overweight truckers’ evasive behavior and the additional cost incurred from the decrease of level of service. The lower level model simulates truckers will choose the minimum cost roads according to the deployment of several strategies of road traffic management. Due to the difficulty in solving the mathematical programming method, this thesis performed heuristic to solve the bi-level model efficiently. This thesis not only evaluated model though imitated Nevada network but conducted different stages of sensitivity analysis. The result showed that the combined three strategies outperform using weigh-in-motion. It is hoped that the model will help road administration to make decision efficiently in the future.
