交通运输工程与信息学报

在停车设施最大允许停车数限制条件下,建立了停车需求分布的优化模型。通过定义广义行程时间概念,使得停车设施的选择概率不仅依赖于常用设施的停车吸引力系数,也与设施处的停车收费价格、停车相关步行时间和生成点与设施间的实际行程时间相关。利用部分增广拉格朗日乘子算法将供给限制约束转化为目标函数项,从而使得转化后的具有单纯形特征约束的模型可以利用Frank-Wolfe算法有效求解。数值算例验证了新模型与算法的有效性,研究结果可以为停车需求管理中的停车供需匹配分析提供新的分析思路和方法。

Given the constraints associated with maximum allowable parking berths at parking facilities, an optimization model of parking demand distribution is proposed in this study. Based on the generalized travel time, the probability of selecting a parking facility not only depends on the attractiveness of the facility, but also on the price charged for parking, the related walking time, and the actual travel time between the generation point and parking facility. After transferring the restriction to supply into the objective function items using the partial-augmented Lagrange multiplier method, the resulted model with simplex form constraints can be solved effectively using the Frank-Wolfe algorithm. The numerical example illustrated in this paper verified the effectiveness and efficiency of the new model and method. The results of this study uncovered some insights into matching parking demand and supply in parking demand management.