摘要
为了综合评价城市配送车辆限行政策,借鉴多代理模型的理论和方法,在系统分析车辆限行下各利益相关者诉求及行为特征的基础上,构建城市配送车辆限行政策评价模型。模型由车辆路径选择模型和强化学习模型共同组成,通过改进遗传算法克服城市配送中配送方案优化问题,同时引入Q学习算法解决各代理间的博弈问题。通过对无限行政策、区域限行政策及路段限行政策进行仿真模拟,结果表明:多代理模型能够有效刻画不同限行政策下各利益相关者的行为变化,定量计算出限行政策对利益相关者的各种影响,从多个视角对限行政策进行评价,模型的适应性得到一定验证;限行政策都会增加城市配送的总成本和服务延误时间,同时减少污染物排放,但不同限行政策带来的正面效益与负面影响均有所不同;区域限行政策能够大幅降低限行区域内的污染物水平,但带来的负面效应较大,尤其在增加城市配送车辆固定成本和服务延误的同时,还会由于配送车辆绕行等行为提高非限行区域的污染物排放水平,不能降低污染物排放的整体水平;路段限行政策带来的各方面负面影响较少,并且能够更有效地防止污染物聚集,从整体上控制污染物排放,但路段限行政策需具有较强的灵活性,对政策制定部门的动态管控能力要求较高。
In order to comprehensively evaluate the urban distribution vehicle restriction policy, an evaluation model for this policy is established using the theories and methods of multi-agent model based on a systemically analysis of the requirements and behavior characteristics of the stakeholders under such policy. The model combines the vehicle routing model and the reinforcement learning model. An improved genetic algorithm is used to optimize the urban distribution scheme, and the Q learning algorithm is introduced to solve the game problems between various agents. The results of the simulation under unrestricted policy, regional restriction policy and section restriction policy show that (1) The multi-agent model can effectively depict the changes of stakeholders' behaviors under different policies and quantitatively calculate the impact of these policies on stakeholders, which enables us to evaluate the policy from different perspectives and also proves a good adaptability of the model. (2) The policy will increase the total cost of urban distribution and the delay of service, while reduce the pollutant emission, but the positive and negative impacts of different policies are different. (3) Although the regional restriction policy can significantly reduce the level of pollutants in the region, it increases the total cost of urban distribution and the delay of service, and also has greater negative effects on reducing the overall level of the pollutant emission as it rises the pollutant emission level in non- restricted regions due to the behavior of distribution vehicles bypassing. (4) The section restriction policy has less negative effects comparing with the other two policies, it can effectively prevent pollutant accumulation and control the pollutant emissions in general. However, the section restriction policy needs a strong flexibility and requires professional dynamic management and control capabilities of the policy making department.
作者
张圣忠
李继东
徐赛娜
ZHANG Sheng-zhong;LI Ji-dong;XU Sai-na(School of Economics and Management, Chang'an University, Xi'an Shaanxi 710064, China;Jiangsu Weixin Engineering Consultants Co. , Ltd. , Nanjing Jiangsu 210014, China)
出处
《公路交通科技》
CAS
CSCD
北大核心
2018年第6期87-94,共8页
Journal of Highway and Transportation Research and Development
基金
教育部新世纪优秀人才支持计划项目(NCET-11-0716)
关键词
交通工程
政策评价模型
多代理模型
城市配送车辆限行政策
遗传算法
transport economics
policy evaluation model
multi-agent model
urban distribution vehiclerestriction policy
genetic algorithm