高速公路上由事件所造成的非重現性延滯約占總延滯的13%至30%，因此減少事件應變所花費的時間、提高整體應變效率，能改善高速公路的延滯狀況。本研究提出以多場站車輛途程問題為基礎之動態事件應變指派模型及其運作流程，目標為最小化整體事件之應變時間，同時考量潛在事件發生的隨機性，並建構禁忌搜索演算法作為求解方式。案例分析顯示本研究提出之模型可將求解時間縮短至40 秒內，並可改善事件的應變效率，尤其在高事件發生頻率之情境，其事件應變時間改善效益可達14%至42%。此外，相較於不考量事件隨機性的狀況，模型在考量隨機性下對總應變時間可增加約3.79%之改善，但若過度考量隨機性反而會導致總應變時間增長。

The non-recurrent delay caused by the incident accounts for around 13% to30% of the total freeway delay. Hence, shortening the time spent on incidentresponse and improving response efficiency are critical for reducing the delay ofthe freeway. This study proposes a dynamic dispatch model of emergencyresponse teams for freeway incidents based on the multi-depot vehicle routingproblem and the online operation procedure of the proposed model. The majorobjective is to minimize the total incident response time while taking intoaccount of the stochasticity of potential incident occurrence and constructs atabu search algorithm as the solution method for the mathematical model. Theresults of the case study indicate that the proposed model can shorten thesolution time to less than 40 seconds and improve the efficiency of the incidentresponse. Especially in the scenario where the frequency of incident occurrenceis relatively high, the improvement can reach 14% to 42%. In addition,compared with the situation that does not consider the stochasticity of incident,the proposed model provides additional 3.79% improvement in terms of totalresponse time when considering the stochasticity. However, over-emphasis onthe stochasticity may lead to an increase in the total response time because ofrelatively conservative strategies.