高速公路可变限速控制多目标优化与决策方法研究

发布时间：2018-05-08 21:46

本文选题：高速公路 + 可变限速　；参考：《吉林大学》2015年硕士论文

【摘要】：高速公路一直是我国交通运输业的重点建设对象。然而，随着不断增长的交通需求，各种各样的交通问题开始突显，相关实践表明，作为智能控制子系统之一的可变限速控制技术，对于解决高速公路所存在问题具有突出优势，该技术不仅能通过降低行车速度的离散性显著减少限速区域的事故率，还可以使得限速区域的交通流更加平滑，从而有效提高高速公路的通行能力。然而，目前我国关于高速公路可变限速控制的研究还停留在以交通流量最大或者行程时间最低为目标的单目标优化层次上，关于可变限速控制多目标问题的研究少之又少，且鲜见对于环境污染的考虑。日益严重的空气污染，导致了大范围的持续雾霾天气，这迫切要求交通管理者转变思路，不再一味地仅以提高通行能力或者减少行程时间为目标进行高速公路的管理与控制，而应该将多目标引入到可变限速控制中去。针对上述问题，本文依托于吉林省青年科研基金计划项目《高速公路可变限速控制算法研究》，在对高速公路可变限速控制研究背景、研究现状进行介绍、分析的基础上，从多目标优化与多目标决策两大方面对包括尾气排放量、交通流量与行程时间等多个控制目标的多目标问题进行了深入研究。其中，，多目标优化是多目标决策的首要前提，优化完成后可以得到各个目标函数相互协调的非劣解集；多目标决策是解决多目标问题的最终步骤，通过多目标决策能够从非劣解集中决策出单个解进行执行。因此，本论文的研究内容大致包括了以下方面： 1）对高速公路可变限速控制多目标优化进行研究，以多目标优化数学模型的构建理论为基础，结合高速公路可变限速控制的具体特性，构建包括尾气排放量优化函数、交通流量优化函数、行程时间优化函数及约束条件的多目标优化模型。然后，通过对相关算法的分析，明确采用遗传算法进行模型求解的优势与可行性，在对多目标优化模型与遗传算法进行具体匹配的基础上，设定求解参数，明确求解过程，从而完成可变限速多目标优化部分的研究，得到各个目标函数互相协调的非劣解集。 2）对高速公路可变限速控制多目标决策进行研究，在对现有多目标决策技术进行总结分析的基础上，明确采用强化学习算法进行决策的优势与可行性，然后，以强化学习原理为基础，构建包括基本组成元素、回报函数与准则函数的多目标决策模型，最后通过对动作选择机制的设定，明确多目标决策模型求解的具体步骤，完成可变限速控制多目标决策部分的研究，基于实时环境状态，从非劣解集中自适应地决策出单个解进行执行。 3）高速公路可变限速控制多目标优化与决策仿真，通过对各仿真软件的比较，选用Paramics进行仿真实验。然后，对仿真平台进行构建，具体包括仿真路段的搭建与路段参数设置、布设可变信息板与检测器、确定多目标优化结果及嵌入多目标决策算法等。最后，对输出结果进行分析与对比，完成对可变限速控制多目标优化与决策模型可行性与有效性的验证。 4）全文总结与展望，通过对论文的主要研究成果进行归纳，提出论文尚存的缺点与后期研究的重点。本论文针对高速公路可变限速控制进行的多目标优化与决策研究，一方面可以实现尾气排放、交通流量与行程时间等控制目标的有效协调，另一方面还可以根据高速公路实时交通状态，智能决策出限速值进行执行。论文是对高速公路智能控制的有益探讨，所取得的研究成果可以为高速公路可变限速控制提供一定的技术支持。
[Abstract]:The highway has always been the key object of the construction of the transportation industry in China. However, with the increasing traffic demand, all kinds of traffic problems have begun to highlight. The relevant practice shows that the variable speed limit control technology, one of the intelligent control subsystems, has a prominent advantage to solve the problems existing in the high speed highway. By reducing the discreteness of driving speed, it can significantly reduce the rate of accident in the speed limiting region, and can also make the traffic flow more smooth in the speed limit area, thus effectively improving the traffic capacity of the expressway.
However, at present, the research on the variable speed control of expressway is still in the single objective optimization level with the goal of maximum traffic flow or the lowest travel time. There are few studies on the multi target of variable speed limit control, and few of them are considered for environmental pollution. The continuous fog and haze weather, which urgently requires traffic managers to change their thinking, no longer only to improve the traffic capacity or reduce travel time as the goal of the management and control of the highway, and should be introduced into the variable speed limit control.
In view of the above problems, this paper, based on the study of the variable speed control algorithm of expressway, based on the project of Jilin youth scientific research fund, introduces the research background and current situation of the variable speed control of the expressway. On the basis of the analysis, from the two aspects of multi-objective optimization and multi target decision, the exhaust emissions, traffic flow and traffic flow are included. Multiple objective problems, such as travel time, are studied in depth. Among them, multi-objective optimization is the primary premise of multi-objective decision making. After the optimization is completed, the non inferior set of each target function can be coordinated, and multi objective decision is the ultimate step for solving the multi-objective problem, and the multi objective decision can be from the non inferior solution by multi target decision. A single solution is executed in a centralized decision.
Therefore, the research contents of this thesis include the following aspects:
1) multi objective optimization of variable speed control of expressway is studied. Based on the construction theory of multi-objective optimization mathematical model, combined with the specific characteristics of variable speed control of expressway, a multi-objective optimization model, including exhaust emission optimization function, traffic flow optimization function, travel time optimization function and constraint conditions, is constructed. Then, through the analysis of the related algorithms, the advantages and feasibility of the genetic algorithm for the model solution are clearly defined. On the basis of the specific matching of the multi-objective optimization model and the genetic algorithm, the solution parameters are set and the solution process is clearly defined, thus the study of the variable speed limit multi target optimization part is completed, and each target function is obtained. A coordinated non inferior set of solutions.
2) study on multi-objective decision making of variable speed control of freeway. On the basis of summarizing and analyzing the existing multi-objective decision technology, the advantages and feasibility of adopting the reinforcement learning algorithm to make decision are made clear. Then, based on strengthening the learning principle, the multi-objective of the basic composition elements, the return function and the criterion function is constructed. In the end, by setting the mechanism of action selection, the specific steps of the multi-objective decision model are defined and the multi target decision part of variable speed limit control is studied. Based on the state of the real-time environment, a single solution is automatically executed from the non inferior solution.
3) multi target optimization and decision simulation of variable speed limit control of expressway. Through comparison of various simulation software, Paramics is selected to carry out simulation experiment. Then, the simulation platform is constructed, including the construction of the simulation section and the setting of section parameters, setting of variable information board and detector, determining the results of multi-objective optimization and embedding multiple targets. Finally, the output results are analyzed and compared, and the feasibility and effectiveness of variable speed limit control multi-objective optimization and decision making model are verified.
4) the summary and Prospect of the thesis. By summarizing the main research results of the paper, the shortcomings of the paper and the emphasis of the later research are proposed.
On the one hand, the paper can realize the effective coordination of the exhaust emission, the traffic flow and the travel time and other control targets. On the other hand, the speed limit of the expressway can be implemented according to the real-time traffic state of the expressway, and the speed limit value is implemented by the intelligent decision. The research results can provide some technical support for variable speed limit control of freeway.

【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491

【参考文献】