需求稀少情况下随机动态装卸混合问题研究

Study on stochastic dynamic pick-up and delivery problem with light traffic

针对随机动态装卸混合问题中存在的排队现象,运用排队论推导出需求稀少情况下随机动态装卸混合问题期望系统时间的下界;提出了一种实时优化策略——多车场随机队列中位策略;推导出需求稀少情况下,多车场随机队列中位策略和实际应用中广泛采用的随机队列中位策略的期望系统时间,并分析了期望系统时间的渐近性.模拟计算结果表明,需求稀少情况下,多车场随机队列中位策略明显优于随机队列中位策略;当服务强度趋于零时,多车场随机队列中位策略近似为最优策略.

According to the queuing phenomena exiting in the stochastic dynamic pick-up and delivery problem,the lower bound of expected system time of the stochastic dynamic pick-up and delivery problem with light traffic is deduced by applying queuing theory.A real-time optimal policy called multi-depot stochastic queue median policy is put forward.The expected system time of this policy is deduced in the case of light traffic,and the asymptotic properties of the expected system time are analyzed.The same deduction ...