Abstract
Most traffic delays during regional evacuations occur due to a large number of vehicles are grouped for a short time. Thus, the reasonable assignment of a vehicle route is one strategy for reducing these delays. To identify an optimal vehicle route assignment plan, this paper presents a multi-source and multi-destination traffic network model. This model is an extension of the minimum cost flow problem and it improves the definition of the travel time based on the traffic volume. It can be used to formulate vehicle routing plans that alleviate traffic congestion and reduce the evacuation time. A genetic algorithm is modified to find the solution for this model. The solution trades off the total vehicle travel distance against traffic jams. An application is presented for Wenzhou city, Zhejiang province, China.