多类型公共自行车调配优化决策问题研究

Public bicycle system, as an emerging public transportation, has been developed rapidly in recent years. However, there are some common problems emerging in the operation, such as "hard rental, hard return". To solve the problem is the key to the success of implementing public bicycle system. In this research project, we are dedicated to study multi-type bicycle repositioning problems in static and dynamic environment. By analyzing the repositioning objects, the flow, the interactions such as substitution property and occupancy property between different bicycle types, we abstract different theoretical problems. They are the multi-type bicycle repositioning problems with static demand, time dependent demand and dynamic demand. The corresponding mathematical models are developed. Furthermore, we develop the mixed tabu search algorithm to solve the problem with static demand and time dependent demand, and the temporal and spatical decomposition method base on updating mechanism with very large scale neighborhood search strategy to solve the problem with dynamic demand. We plan to analyze the developed algorithms theoretically, and conduct extensive computational experiments with different parameters to ensure the effectiveness of the algorithms. By doing this, we can provide the novel optimization theory and methods by obtaining the objectives with maximizing the bicycles balancing at each station as well as reducing the repositioning cost and improving the current actual bicycle repositioning strategies. From the above, the research can not only enrich the problems and methods of combination optimization in operations research, but also has the important practical application value for the development and improvement of urban public transportation.

公共自行车交通系统，作为新兴的公共交通系统，近年来发展十分迅速。但运营过程中“租车难、还车难”问题却制约了该系统健康发展，如何有效解决该问题是其顺利实施的关键。本项目拟对静态和动态环境下多类型公共自行车调配问题展开研究。在对调运对象及其流动性、不同类型公共自行车之间相互影响(可替代性和可占用性)分析基础上，凝练出不同背景(静态需求、时变需求、实时动态需求)下的多类型公共自行车调配优化理论问题，并分别建立数学模型。针对静态需求和时变需求问题，设计混合禁忌搜索算法；针对实时动态需求问题，设计时空分解方法，并采用大规模邻域搜索策略的更新机制。通过理论分析和数值计算研究算法策略和参数对求解的影响，保证算法求解质量优、效率高。在最大限度保障公共自行车站点网络需求平衡的同时，降低调运成本，提高调配管理水平。研究不仅可以丰富运筹学组合优化理论问题和方法，而且对于发展和完善城市公共交通具有重大的实际意义。

公共自行车交通系统,作为新兴的公共交通系统,近年来发展十分迅速。但运营过程中“租车难、还车难”问题却制约了该系统健康发展,如何有效解决该问题是其顺利实施的关键。本项目对静态和动态环境下多类型公共自行车调配问题展开研究。在对调运对象及其流动性、不同类型公共自行车之间相互影响(可替代性和可占用性)分析基础上,凝练出不同背景(静态需求、时变需求)下的多类型公共自行车调配优化理论问题,并分别建立数学模型。针对静态需求和时变需求的多类型公共自行车调配问题提出混合遗传算法和混合禁忌搜索算法求解。针对基于实时动态需求的多类型公共自行车调配问题提出了时空分解方法，利用基于VLSN邻域优化策略进行实时更新，VLSN搜索技术能够探索出有效的搜索策略。通过理论分析和数值计算研究算法策略和参数对求解的影响,保证算法求解质量优、效率高。在最大限度保障公共自行车站点网络需求平衡的同时,降低调运成本,提高调配管理水平。本项目中设计的算法还成果应用于其他优化问题。研究不仅可以丰富运筹学组合优化理论问题和方法,而且对于发展和完善城市公共交通具有重大的实际意义。