本文關(guān)鍵詞： 交叉口 信號配時 斗雞博弈 Nash bargainingg Shapley估值　出處：《北京交通大學(xué)》2015年碩士論文　論文類型：學(xué)位論文

【摘要】：平面交叉口是城市交通的關(guān)鍵,對交叉口信號配時的研究具有重大實際意義。本文將博弈論的思想應(yīng)用于解決交叉口信號配時問題,在分析交叉口信號配時體現(xiàn)的博弈特性的基礎(chǔ)上,針對交叉口信號配時中不同的問題和目標(biāo),建立幾種交叉口信號配時的博弈模型,通過博弈求解,提出基于博弈論的信號配時方案。主要研究內(nèi)容包括以下幾個方面： 首先,針對兩相位非固定周期的單交叉口和相鄰兩交叉口,提出基于斗雞博弈的交叉口信號配時方案。對兩相位交叉口,以各相位作為博弈參與者、各相位信號燈狀態(tài)(紅燈、綠燈)作為策略集、利用排隊長度相反數(shù)作為贏得值建立博弈模型；對相鄰兩交叉口,以每個路口為參與者,每個路口的信號燈狀態(tài)為策略集,每個路口整體排隊長度相反數(shù)為贏得值建立博弈模型。針對上述博弈模型,以固定時長為博弈周期,基于非合作博弈中經(jīng)典的斗雞博弈模型,求得上述博弈混合策略Nash均衡,給出下一個博弈周期內(nèi)的信號配時方案,實現(xiàn)信號燈狀態(tài)的合理分配,提高交叉口的通行效率。 其次,針對四相位固定周期的單交叉口,提出基于合作博弈的交叉口信號配時方案。以各相位作為合作博弈的參與者,各相位綠燈時間作為策略集,各相位車輛排隊長度作為贏得值,建立相應(yīng)的合作博弈模型,針對上述博弈模型分別采用Nash bargaining解法和Shapley估值解法進(jìn)行求解。由于Nash bargaining解法針對初始bargaining能力相同的二人博弈,因此適用于不區(qū)分干路、支路的交叉口,并需要進(jìn)行兩次二人博弈；Shapley估值解法則可直接應(yīng)用于區(qū)分干路、支路的交叉口進(jìn)行博弈。根據(jù)上一周期結(jié)束時各相位排隊車輛的長度,實時地調(diào)整下一周期各相位綠燈時間。通過博弈求解達(dá)到參與者贏得值的最大,即下一周期結(jié)束時路口車輛排隊長度的最小化。 最后,利用MATLAB對所提的各個方案進(jìn)行仿真,驗證了所提配時方案的有效性,并與固定配時進(jìn)行比較說明所提方案可以有效減少路口各相位的車輛排隊長度,對提高路口通行效率有一定實際指導(dǎo)意義。
[Abstract]:Intersection is the key of urban traffic, which is of great practical significance to the study of intersection signal timing. In this paper, the game theory is applied to solve the intersection signal timing problem. On the basis of analyzing the game characteristics of intersection signal timing, aiming at different problems and objectives of intersection signal timing, several game models of intersection signal timing are established and solved by game. A signal timing scheme based on game theory is proposed. The main research contents include the following aspects:. First of all, a signal timing scheme based on game game is proposed for single intersection with two phases and adjacent intersections. For two-phase intersections, each phase is used as a game participant. Each phase signal state (red light, green light) is used as the strategy set, and the game model is established by using the opposite number of queue length as the winning value, and for the adjacent two intersections, each intersection is regarded as the participant, and the signal light state of each intersection is the strategy set. Based on the above game model, based on the classic game model of the non-cooperative game, the Nash equilibrium of the above game mixed strategy is obtained by taking the fixed time as the game period, and the total queue length of each intersection is opposite to the winning value. The signal timing scheme in the next game cycle is given to realize the reasonable distribution of signal lights and improve the traffic efficiency of intersection. Secondly, a signal timing scheme based on cooperative game is proposed for single intersection with four-phase fixed period. Each phase is used as the participant of the cooperative game, and the green light time of each phase is used as the strategy set. The corresponding cooperative game model is established for the queue length of each phase vehicle as the winning value. The Nash bargaining solution and the Shapley estimation method are used to solve the above game model, respectively. Because the Nash bargaining solution is aimed at the two-person game with the same initial bargaining capability, the corresponding cooperative game model is established. Therefore, it can be applied to the intersection without differentiating the trunk road and the branch road, and it needs two times of two-player game and Shapley estimation method, which can be directly applied to distinguish the trunk road from the branch road intersection and play the game. According to the length of each phase queue vehicle at the end of the last cycle, The green time of each phase of the next cycle is adjusted in real time and the maximum of the participants' winning value is obtained by the game solution which is the minimization of the queue length at the end of the next cycle. Finally, the proposed scheme is simulated by MATLAB, and the validity of the proposed scheme is verified. The comparison with the fixed timing scheme shows that the proposed scheme can effectively reduce the queue length of the vehicle at each phase of the intersection. It has certain practical guiding significance to improve the traffic efficiency of intersection.
【學(xué)位授予單位】：北京交通大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：U491.54

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 夏新海;許倫輝;;交叉口Agent間的博弈學(xué)習(xí)協(xié)調(diào)方法[J];重慶交通大學(xué)學(xué)報(自然科學(xué)版);2010年02期

2 郭繼孚;劉瑩;余柳;;對中國大城市交通擁堵問題的認(rèn)識[J];城市交通;2011年02期

3 管志忠;;合作對策問題的MATLAB程序?qū)崿F(xiàn)及應(yīng)用[J];池州學(xué)院學(xué)報;2008年03期

4 鄭長江,王煒;混合交通流條件下信號交叉口配時優(yōu)化設(shè)計[J];公路交通科技;2005年04期

5 王秋平;譚學(xué)龍;張生瑞;;城市單點交叉口信號配時優(yōu)化[J];交通運輸工程學(xué)報;2006年02期

6 李立源,曹大鑄;道路交通流最優(yōu)預(yù)測與交叉口最優(yōu)控制[J];控制理論與應(yīng)用;1993年01期

7 齊馳;侯忠生;賈琰;;基于排隊長度均衡的交叉口信號配時優(yōu)化策略[J];控制與決策;2012年08期

8 蔣亞磊;;基于博弈論的城市交通擁擠成因及對策研究[J];蘭州交通大學(xué)學(xué)報;2007年01期

9 郝建根;侯忠生;柳向斌;;基于多參數(shù)規(guī)劃的單交叉口排隊長度均衡控制[J];控制與決策;2013年04期

10 朱銘琳;陳陽舟;;基于博弈的交叉口交通信號配時系統(tǒng)設(shè)計[J];計算機仿真;2013年07期

，

本文編號：1515893