基于偏好的列车运行过程多目标鲨鱼优化算法

Multi-objective shark smell optimization algorithm for train operation process based on preference

针对列车运行过程优化问题,以节能、准时、停靠精确和舒适作为优化目标建立了多目标的列车运行过程优化模型,并提出了一种基于偏好的列车运行过程多目标鲨鱼优化算法。将解向量与目标需求向量的夹角余弦作为评价指标,以增强评价的合理性;并将决策者的偏好融入解优劣的评价策略,以自适应的调整各个鲨鱼个体的交叉、选择和变异的概率,从而引导个体快速奔向决策者期望的偏好区域,以增强评价的实用性。与此同时,为提升全局优化性能,结合偏好引导、融合距离、遗传进化等有效机制,设计了一种优化性能更强的更新鲨鱼种群的计算流程。以大连地铁13号线九里至十九局的列车运行场景作为实验对象,其硬件在环实验环境下得到的实验结果表明,相比于传统的多目标优化算法,所提出的优化算法具有较快的寻优速度和较高的寻优精度,且能够寻优得到更符合决策者预期的优化解。

Aiming at solving the optimization problem of train operation process,a multi-objective optimization model for train operation process is established with energy conservation,punctuality,accurate parking and comfort as indexes,and a multi-objective shark smell optimization algorithm based on preference for train operation process optimization is proposed.The angle cosine between solution target vector and target demand vector as the evaluation index is proposed to enhance the evaluation reasonable.The preference of decision maker is integrated into the evaluation strategy of solution,and the probability of crossover,selection and mutation for each individual shark is adaptively adjusted,thus guiding individuals to quickly run to the desired preference region of decision-makers,which enhances the evaluation applicability.Meanwhile,in order to improve the global convergence performance,combined with the effective mechanisms such as preference guidance,fusion distance,genetic evolution,a computational process for updating shark population with better optimization performance is designed.Taking train operation scenario of rail transit line 13 from Jiuli to 19 th Bureau in Dalian as the test object,the results of hardware-in-the-loop test show that compared with the traditional multi-objective optimization algorithm,the proposed optimization algorithm has faster optimization speed and higher optimization accuracy,and can obtain the optimal solution which is more accordant with decision-makers′expectation.