Bi-Level Programming Model in Road Network Programming and Its Genetic Algorithm

After analyzing and summing up the characters of the existing road net programming model's objectives and constraints, this paper establishes a bi-level and multi-objectives programming model: in which the upper-level decision maker is the leader, who thinks over the basic request of the road net programming and real travel-hours, making decisions to affect the lower-level decisions with the variables based on the system optimal. This decision includes many objects, such as the least invest, the largest capacity and the highest technical grade, while the lower-level decision maker is the followers, who establish the Stochastic Users' Equilibrium (SUE) based on the path which is the least travel cost or is regarded the least travel cost by the users, making decisions to optimize their own objectives based on the upper-level model. According to the characteristics of the model, the genetic algorithm as the solution strategy is put forward. At last, this paper validates the model and its algorithm with an example.