基于多核并行计算技术的混合交通微观仿真

李珊珊^{1, 2}，钱大琳^{1, 2}

(1. 北京交通大学 交通运输学院，北京，100044；
2. 北京交通大学 城市交通复杂系统理论与技术教育部重点实验室，北京，100044)

摘 要：为提高混合交通微观仿真的运行速度，将多核并行计算技术应用于混合交通微观仿真中。提出混合交通微观仿真的并行化方法，包括基于对串行混合交通微观仿真中各个模块CPU运算时间的百分比分析，确定基于数据并行的任务分解方法，及结合多内核CPU架构特点确定Fork/Join的并行模式。提出基于多核技术的混合交通微观仿真的关键并行算法，包括初始路网分割算法、车辆穿越边界算法及动态负载平衡算法。通过仿真实验验证多核并行计算技术能够提高混合交通微观仿真的运行效率。

关键词：交通运输系统工程；多核；并行计算；混合交通；微观仿真

Mixed traffic microscopic simulation based on multi-core parallel computing technology

LI Shan-shan^{1, 2}, QIAN Da-lin^{1, 2}

(1. School of Traffic and Transportation, Beijing Jiaotong University, Beijing 100044, China;
2. MOE Key Laboratory for Urban Transportation Complex Systems Theory and Technology,
Beijing Jiaotong University, Beijing 100044, China)

Abstract:In order to speed up mixed traffic microscopic simulation, the multi-core parallel computing technology was applied to mixed traffic microscopic simulation. Firstly, a method for parallelizing mixed traffic microscopic simulation was proposed. Task decomposition based on data parallelism and Fork/Join parallel mode was included. Task decomposition was determined by analyzing the CPU time of each software module, and parallel mode was determined considering the characters of CPU architecture with multi-core. Secondly, key algorithms necessary for implementing parallel microscopic mixed traffic microscopic simulation based on multi-core technology were proposed. The key algorithms included initial traffic network segmentation algorithm， vehicles crossing border algorithm and dynamic load balancing algorithm. The simulation results show that the multi-core parallel computing technology can greatly improve the simulation efficiency.

Key words:engineering of communications and transportation system; multi-core; parallel computing; mixed traffic; microscopic simulation