本研究以Chen and Chang (1999)所提出負效用軟時窗動態用路人最佳化出發時間/路徑選擇模型為基礎，加入路段容量限制，構建路段容量限制軟時窗動態用路人最佳化出發時間/路徑選擇模型。以拉氏法進行求解，並探討拉氏乘數的意義及在演算法中的更新方式。 結合周鄭義(1999)所提出動態號誌時制最佳化之研究-雙層規劃模型之應用，在下層模型以軟時窗動態用路人最佳化出發時間/路徑選擇模型取代動態用路人均衡模型，成為號誌控制及軟時窗動態用路人出發時間/路徑選擇雙層規劃模型，以卓訓榮(1991)提出之廣義反矩陣方法獲得敏感度分析資訊，進行模式求解。最後，在這個雙層模型的下層模型中加入容量限制，也就是下層模型為路段容量限制軟時窗動態用路人最佳化出發時間/路徑選擇模型，進行構建及求解。 The thesis is based on dynamic user-optimal departure time/route choice problem with soft time-windows by Chen and Chang(1999), embedding in link inflow constraints, and constructing dynamic capacitated user-optimal departure time/route choice problem with soft time-windows. The thesis uses the Lagrangian method to solve the problem and analyzing the meaning of the Lagrangian multipliers and its updating in the algorithm. The thesis combines dynamic network signal control system and substitutes dynamic user equilibrium route choice modal for dynamic user-optimal departure time/route choice problem with soft time-windows in the lower level. In accordance with variational inequality sensitivity analysis theory attain sensitivity analysis information to solve the problem. At last in the bi-level model embeds in link inflow constraints in the lower level namely the lower level is dynamic capacitated user-optimal departure time/route choice problem with soft time-windows. Constructs the model and solves it.
