共享单车系统中车辆的损坏会严重影响顾客使用体验。为此,运营方需要投入运维资源(如维修工、运载车等)对坏车进行回收、修复和重新投放。由于运维资源有限,如何优化资源的配置是当前共享单车管理实践中待解决的重要问题。本文结合实际...共享单车系统中车辆的损坏会严重影响顾客使用体验。为此,运营方需要投入运维资源(如维修工、运载车等)对坏车进行回收、修复和重新投放。由于运维资源有限,如何优化资源的配置是当前共享单车管理实践中待解决的重要问题。本文结合实际运维特点,构建了一个封闭排队网络模型,并在运维资源有限的条件下,以最小化顾客损失比例为目标提出决策问题。本文基于前述排队网络模型,提出基于连续时间马尔可夫过程的状态稳态概率和基于离散事件系统仿真,两种求解系统性能指标的方法。针对决策问题解空间有限且离散的特点,本文结合前述仿真方法,采用基于排序择优(ranking and selection)的仿真优化算法来求解。实验算例结果显示离散事件系统仿真可有效估计出系统性能指标;提升维修工和运载车的工作速率或增加数量可改善系统性能表现,但改善效果边际递减。此外,本文采用的排序择优算法可有效求解决策问题,为共享单车的运营管理决策提供参考。展开更多
This work is devoted to stochastic systems arising from empirical measures of random sequences(termed primary sequences) that are modulated by another Markov chain. The Markov chain is used to model random discrete ev...This work is devoted to stochastic systems arising from empirical measures of random sequences(termed primary sequences) that are modulated by another Markov chain. The Markov chain is used to model random discrete events that are not represented in the primary sequences. One novel feature is that in lieu of the usual scaling in empirical measure sequences, the authors consider scaling in both space and time, which leads to new limit results. Under broad conditions, it is shown that a scaled sequence of the empirical measure converges weakly to a number of Brownian bridges modulated by a continuous-time Markov chain. Ramifications and special cases are also considered.展开更多
文摘共享单车系统中车辆的损坏会严重影响顾客使用体验。为此,运营方需要投入运维资源(如维修工、运载车等)对坏车进行回收、修复和重新投放。由于运维资源有限,如何优化资源的配置是当前共享单车管理实践中待解决的重要问题。本文结合实际运维特点,构建了一个封闭排队网络模型,并在运维资源有限的条件下,以最小化顾客损失比例为目标提出决策问题。本文基于前述排队网络模型,提出基于连续时间马尔可夫过程的状态稳态概率和基于离散事件系统仿真,两种求解系统性能指标的方法。针对决策问题解空间有限且离散的特点,本文结合前述仿真方法,采用基于排序择优(ranking and selection)的仿真优化算法来求解。实验算例结果显示离散事件系统仿真可有效估计出系统性能指标;提升维修工和运载车的工作速率或增加数量可改善系统性能表现,但改善效果边际递减。此外,本文采用的排序择优算法可有效求解决策问题,为共享单车的运营管理决策提供参考。
基金supported by the Air Force Office of Scientific Research under Grant No.FA9550-15-1-0131
文摘This work is devoted to stochastic systems arising from empirical measures of random sequences(termed primary sequences) that are modulated by another Markov chain. The Markov chain is used to model random discrete events that are not represented in the primary sequences. One novel feature is that in lieu of the usual scaling in empirical measure sequences, the authors consider scaling in both space and time, which leads to new limit results. Under broad conditions, it is shown that a scaled sequence of the empirical measure converges weakly to a number of Brownian bridges modulated by a continuous-time Markov chain. Ramifications and special cases are also considered.
