【正文】 解決這一問題，本文提出一種基于多代理聚類算法控制交通信號燈。在我們的方法中，聚類算法與道路使用者的價值函數是用來確定每個交通燈的最優(yōu)決策的，這項決定是基于所有道路使用者站在交通路口累積投票，通過估計每輛車的好處(或收益)來確定綠燈時間增益值與總時間是有差異的，它希望在它往返的時候等待，如果燈是紅色，或者燈是綠色。等待，直到車輛到達目的地，通過有聚類算法的基礎設施，最后經過監(jiān)測車的監(jiān)測。我們對自己的聚類算法模型和其它使用綠燈模擬器的系統做了比較。綠燈模擬器是一個交通模擬器，監(jiān)控交通流量統計，如平均等待時間，并測試不同的交通燈控制器。結果表明，在擁擠的交通條件下，聚類控制器性能優(yōu)于其它所有測試的非自適應控制器，我們也測試理論上的平均等待時間，用以選擇車輛通過市區(qū)的道路，并表明，道路使用者采用合作學習的方法可避免交通瓶頸。本文安排如下：第2部分敘述如何建立交通模型，預測交通情況和控制交通。第3部分是就相關問題得出結論。第4部分說明了現在正在進一步研究的事實，并介紹了我們的新思想。The times is a automation times nowadays,traffic light waits for much the industey equipment to go hand in hand with the puter under the control ,a good traffic light controls system,will give road aspect such as being crowded,controlling against rules to give a technical the fact that the largescale integrated circuit and the puter art promptness develop,as well as artificial intelligence broad in the field of control technique applies,intelligence equipment has had very big development,the main current being that modern science and technology develops main body of a book is designed having introduccd a intelligence traffic light function being intelligence traffic light navar’s turn to be able to e true has：The crossing carries out supervisory control on four main traffic of some downtown area。Every crossing has the fixed duty period,charges centrefor being able to change it’s period and in depending on a road when being crowded。The motro vehicle breaking rules and regulations to the crossing is able to take a photo immediately,abstracts and the vehicle shop world range ,one uses the microelectronics technology,the puter and the technology municating by letter are a guide’s,centering on IT and IT industry information revolution is in the ,how,puter art applies more effective union and there is an effect’s brought it’s effect into play with reality is the most popular topic of scientific munity,is also that puter applications is hit by the unparalleled active field main body of a book is applied up mainly from slicing machine’s only realizing intellectualized administration of crossroads traffic light,use operation in controlling the vehicular traffic research has the goal to optimize transportation flow of people and the number of road users constantly increases, and resources provided by current infrastructures are limited, intelligent control of traffic will bee a very important issue in , some limitations to the usage of intelligent tra?c control jams for example is thought to be beneficial to both environment and economy, butimproved trafficflow may also lead to an increase in demand [Levinson, 2003].There are several models for traffic our research we focus on microscopicmodels that model the behavior of individual vehicles, and thereby can simulate dynamics of groups of has shown that such models yield realistic behavior[Nagel and Schreckenberg, 1992, Wahle and Schreckenberg, 2001].Cars in urban traffic can experience long travel times due to inefficient traffic light control of traffic lights using sophisticated sensors and intelligent optimizationalgorithms might therefore bevery of traffic light switching increasesroad capacity and traffic flow, and can prevent tra?c light control is aplex optimization problem and several intelligent algorithms, such as fuzzy logic, evolutionary algorithms, and reinforcement learning(RL)have already been used in attemptsto solve this paper we describe a modelbased, multiagent reinforcement learningalgorithm for controlling traffic our approach, reinforcement learning [Sutton and Barto, 1998, Kaelbling et al., 1996]with roaduserbased value functions [Wiering, 2000] is used to determine optimal decisionsfor each traffic decision is based on a cumulative vote of all road users standingfor a traffic junction, where each car votes using its estimated advantage(or gain)of settingits light to gainvalue is the difference between the total time it expects to waitduring the rest of its trip if the light for which it is currently standing is red, and if it is waiting time until cars arrive at their destination is estimated by monitoring cars flowingthrough the infrastructure and using reinforcement learning(RL) pare the performance of our modelbased RL method to that of other controllersusing the Green Light District simulator(GLD).GLD is a traffic simulator that allows usto design arbitrary infrastructures and traffic patterns, monitor traffic flow statistics such asaverage waiting times, and test different traffic light experimental resultsshow that in crowded traffic, the RL controllers outperform all other tested also test the use of the learned average waiting times for choosing routes of cars through the city(colearning), and show that by using colearning road users can avoidbottlenecks.