基于随机需求的接运公交时刻表的鲁棒性优化方法研究

接运公交需求的随机性会对接运公交时刻表的编制与投入运营产生扰动影响. 为降低此不利影响,基于鲁棒性思想研究了随机需求下接运公交时刻表的优化方法. 随机接运公交需求的每一组实现为一个情景,情景目标值综合考虑乘客成本与运营成本的最小化. 以接运公交需求为随机变量,建立了接运公交时刻表的鲁棒性优化模型,模型综合考虑情景的期望值与偏差期望值最小化. 模型具有组合优化与随机规划特性,选用遗传算法进行求解. 最后给出算例验证了模型与算法的有效性. 算例结果表明随着偏差权重系数的提高,鲁棒性模型更倾向于降低随机需求带来的扰动.

Methodology of Robust Optimization for Feeder Bus Timetable Based on Stochastic Demands

In order to decrease disturbance influence of stochastic demand for feeder bus timetable design and put into operation, a methodology based on robust optimization for feeder bus network timetable under stochastic demand was proposed. A scenario was identified as a specific evolution of stochastic feeder bus demand. The objective value of scenario minimized passenger cost and operation cost. Considered feeder bus demand as stochastic variable, a robust optimization model for feeder bus timetable design was proposed to minimize expected value and variance of scenario objective. For combinatorial optimization and stochastic programming features of the proposed model, genetic algorithm was provided to solve the proposed model. a numerical example was used to verify the proposed mathematical model and the solution algorithm. The result shows that with variation weight coefficient increase, robust model tends to reduce the disturbance brought by random demand.