混合交通网络设计的双层模型及遗传算法求解

Bi-level programming model for mixed transportation network design and genetic solution algorithm

根据混合交通网络设计问题的特点,利用双层规划模型和遗传算法对该问题进行求解。对交通网络中的路段进行分类,通过限定决策变量的取值范围,将混合交通网络离散化。建立混合交通网络设计的双层模型。其中,上层模型以方案总投资额最小为目标函数,以路段负荷度和可行域为约束条件;下层模型为交通流分配的用户均衡模型。根据所建模型的离散特性,研究其遗传算法解法,并给出算法的具体实现步骤。以一个抽象的交通网络为例,给定网络中的路段属性、OD交通量等参数,利用MATLAB软件对模型编程求解,能够获得满意的交通网络设计方案,表明双层模型和遗传算法是一种研究混合交通网络设计问题的有效方法。最后,对该模型存在的不足及改进方向进行了探讨。

A bi-level programming model and a genetic algorithm are used to design problems. To make the mixed network discrete, the road is classified solve mixed transportation network and the range of decision-making variables restricted. A bi-level programming model for mixed transportation networkdesign is presented based on previous research results. In the upper level, minimization of total investment is used as the objective, and feasible region and degree of traffic congestion of the road in the network as the constraint conditions. The user equilibrium assignment is chosen as the lower level programming model. Genetic algorithm is used to solve the bi-level model based on the discrete attribute of the model, and the specific measures of the algorithm are also presented. Taking a simplified transportation network as example and setting up parameters such as OD traffic flow, road attributes, etc., a relatively satisfactory network design scheme is obtained by programming the bi-level model with MATLAB. The results indicate that the bi-level programming model and the genetic algorithm are effective for mixed transportation network design. The shortcomings and the improvement directions of the model are discussed.