城市道路单行系统布局优化的双层规划模型和混合算法

Bilevel programming model and hybrid solution algorithm for the configuration of one-way streets

研究了基于出行者路径选择行为的单行道布局优化问题.借助于双层规划思想,以最小化研究区域内的总旅行时间为交通管理者的决策目标,建立了单行道布局优化的混合整数非线性规划模型,用0-1变量表征路段单行与否,用Logit型随机用户均衡网络模型刻画在交通管理者确定的某一单行道布局方案下的网络均衡流量模式.设计了GA-MSA组合式算法,其中遗传算法求解上层问题,MSA算法求解在上层给定的单行布局方案下的路段均衡流量模式.为使初始化和遗传操作得到的染色体可行,设计了相应的染色体修复程序.算例分析验证了用定量化方法优化单行道布局的必要性,参数敏感度分析解析了参数取值对优化结果的影响趋势和程度.

One-way traffic is a cost-effective and efficient strategy in urban transportation management. This paper conducts the optimization of one-way street layout while taking into account drivers＇ route choice behaviors. A binary mixed integer nonlinear bilevel programming model, with the objective to minimize the total system travel time encountered by all users, is first formulated to depict this Stackelberg game. A logit-type stochastic user equilibrium model is adopted to describe drivers＇ route choice behaviors reacting to a given layout scheme. A hybrid GA-MSA solution algorithm is then presented to solve this proposed model. The Genetic Algorithm （GA） with a special chromosome repairing approach is to solve the proposed upper level sub-problem while Method of Successive Averages （MSA） is to obtain an optimal link flow pattern under a given one-way street layout scheme from GA. Finally, a numerical experiment is included to demonstrate the necessity of optimizing one-way street layout and the robustness of the proposed methodology, and the sensitivities of critical parameters in the proposed algorithm are also analyzed to depict the effects of parameter values on the optimal scheme and to obtain their empirical ranges.