模糊型動態用路人最佳化變動需求／出發時間／路徑選擇模型暨演算法

Fuzzy Dynamic User-Optimal Variable Demand/ Departure TimelRoute Choice Model and AIgorithm

本研究由人類認知的角度，探討不完全資訊下的模糊型動態用路人最佳化變動需求/出發時間/路徑選擇問題模化為變分不等式模型。由於人類主觀認知的旅行時間具有模糊、不真確的特質，故本研究假設其可以可能率分配描述之。進而採用可能率規劃的技巧，即先對模糊路段旅行時間取α-cut再取代表值的方式，將模糊型動態用路人最佳化變動需求/出發時間/路徑選擇模型轉換為明確型動態用路人最佳化變動需求/出發時間/路徑選擇模型。而對等於此行前決策模型的模糊/動態用路人最佳化均衡條件則色括路徑選擇行為與旅次需求函數兩部份。模型求解方面，則透過修正時空路網的手法，將變動需求問題轉換為固定需求問題，再應用巢化對角演算法求解之。最後，利用簡單的測試路網來分析模型的求解結果。

The fuzzy/dynamic user-optimal variable demand/departure time/route choice problem is formulated using a variational inequality approach. This problem assumes that the link travel times are imprecise/vague, and the fuzziness of subjective link travel times can be characterized by a possibility distribution. by adopting the representative believed link travel time at each α-cut level, the fuzzy/dynamic user-optimal variable demand/departure time/route choice problem can therefore be transformed into a crisp dynamic user-optimal variable demand/departure time/route choice problem. This pretrip model complies with the fuzzy/dynamic user-optimal equilibrium conditions, both on the route choice behavior and on the trip demand functions. By adopting appropriate network representation, the fuzzy/dynamic user-optimal νariable demand/departure time/route choice problem can be transformed into the fuzzy/dynamic user-optimal departure time/route choice problem, which can be solved by the nested diagonalization method. Finally, numerical examples are provided for demonstration.