基于智能体仿真的可持续交通综合策略优化研究

The contradiction between China's urban travel demand and transportation supply has become increasingly prominent. Many cities are trying different traffic management strategies. How to rationally analyze the impact of these strategies on various travelers and their behavior changes in response to such strategies, how to theoretically simulate the interaction process and performance at the system level, how to manage and optimize strategies especially under abnormal conditions. This project aims to the methodological development and applications from the following four perspectives of agent-based dynamic travel demand model (Agent-Based Modeling, ABM), multi-objective system optimization and algorithm based on traffic simulation (Simulation-Based Optimization, SBO), travel behavior adjustment under abnormal conditions, and integrated simulation platform development (Dynamic Transportation Simulation Platform, DTSP).The key issues that will be addressed include travel behavior data collection, behavioral parameter calibration and validation, agent-based travel demand modeling, behavioral traffic assignment, comprehensive evaluation indices of sustainable transportation systems, simulated based multi-objective optimization and efficient algorithm, integration of macroscopic and microscopic simulation platform. This project combines behavioral science, system optimization with transportation modeling and simulation. The main features of the project are the agent-based simulation and optimization of sustainable transportation systems. The modeling and simulation process reflects the variation of individual traveler. The simulation platform will be feasible in both normal and abnormal transportation situations to achieve the integration of the different levels of transportation strategies.

当前我国城市交通需求和供给矛盾越来越突出，多个城市都在尝试不同的交通管理策略。如何理性分析这些策略对于不同交通出行者的影响,以及出行者如何做出调整，理论上应该如何模拟这个互动过程以及在系统层面的表现，系统管理者又应该如何管理和优化，特别是在情况复杂的非常态条件下。本项目从基于智能体的动态交通需求模型、基于交通仿真的多目标交通系统优化模型和算法、非常态条件下的交通行为调整模型、交通系统仿真优化平台的构建四个方面开展研究工作。重点突破交通行为数据采集、参数标定和校验、基于智能体的交通需求建模、可持续交通系统的综合评价指标、基于仿真的多目标优化模型及算法、宏观与微观的整合仿真平台等关键问题。结合行为科学、系统优化与交通建模与仿真，主要特色是对可持续交通系统进行基于智能体仿真的优化建模，仿真过程体现了出行者个体差异，仿真平台能够对常态和非常态交通进行模拟，实现不同层次的交通策略的整合优化 。

当前我国城市交通需求和供给矛盾越来越突出，本项目针对如何理性分析不同交通出行者的行为与交通管理策略之间的相互影响，如何从理论上模拟这个互动过程以及在系统层面的表现，如何为交通管理者提供优化的交通策略，如何应对常态和非常态的复杂交通情况等角度进行研究。本项目的主要研究内容和重要成果包括：研究了基于智能体出行行为模型建模理论，提出了基于智能体的交通方式选择和出发时间选择联合模型；研究了可持续交通系统的复合评价指标体系，建立能够全面综合地反映城市交通的人、车、路、环境及管理各子系统的城市道路交通系统评价指标体系；研究了交通系统运行的多目标函数和基于仿真的元模型优化方法，提出的高效元模型随机优化算法；研究了交通规划和交通管理相结合的交通仿真技术，开发了交通仿真-综合策略优化平台。本项目的特色是对可持续交通系统进行基于智能体仿真的优化建模，仿真过程体现了出行者个体差异，仿真平台能够对常态和非常态交通进行模拟，实现不同层次的交通策略的整合优化。