基于模拟退火精英协同进化算法的交通信号协调优化控制

发布时间：2018-05-27 22:20

本文选题：单交叉口 + 多交叉口　；参考：《南京邮电大学》2014年硕士论文

【摘要】：随着城市化水平的提高，机动车数量急剧增加，，城市交通拥堵问题日益严重，传统的交通信号控制方式已经不能获得良好的控制效果。因此，智能交通信号控制成为了主要控制手段。利用进化算法进行交通信号优化控制是一种有效途径，但其在收敛精度和收敛速度方面还有待改进。协同进化算法作为一种新型智能算法，由于其具有高度的协调性，在智能交通信号控制领域已崭露头角。本文对协同进化算法进行了改进，将其用于交通信号控制，取得了良好的效果。本文结合精英策略、协同进化思想和模拟退火机制，提出了一种基于模拟退火机制的精英协同进化算法（SACEA），将其应用到交通信号控制领域。算法维持三个种群：精英种群、普通种群和随机种群，精英个体组团，并和其他组员个体协作或对其引导来达到进化目的。算法在精英组团过程中引入随机种群以增加种群多样性，同时随机个体和精英个体的合作采用快速模拟退火机制来实现，使算法获得了更好的全局寻优性和更快的收敛速度。本文通过实验验证了算法的有效性。本文在观察研究单交叉口车流量分布特性的基础上，分别建立了以单交叉口的平均延误最小和平均停车率最小为目标的两个单交叉口信号控制模型。在此基础上利用SACEA算法对两个模型各自进行优化配时仿真，并和已有优化算法进行对比，结果表明:利用SACEA算法进行优化更能减少单交叉口平均延误时间和平均停车率。在分析了城市干线协调控制理论的基础上，本文以一个绿波系统为例，以车辆通过干线所需平均总共时间（平均行程时间和总平均延误时间之和）为控制目标，进一步利用SACEA算法进行了多交叉口配时优化，并与已有算法进行对比，结果表明：利用SACEA算法进行优化更能减少车辆通过干线所需时间。
[Abstract]:With the improvement of urbanization level, the number of motor vehicles has increased rapidly, traffic congestion in cities is becoming more and more serious, and the traditional traffic signal control methods have been unable to obtain good control effect. Therefore, intelligent traffic signal control has become the main control means. But it still needs to be improved in the convergence precision and convergence speed. As a new intelligent algorithm, coevolution algorithm has come to the fore in the field of intelligent traffic signal control because of its high coordination. This paper has improved the coevolution algorithm and applied it to the communication signal control, which has achieved good results.
In this paper, combining elite strategy, coevolutionary thought and simulated annealing mechanism, an elite cooperative evolution algorithm (SACEA) based on simulated annealing mechanism is proposed, which is applied to the field of traffic signal control. The algorithm maintains three populations: Elite population, ordinary population and random species group, elite individual groups, and collaborating with other group individuals. The algorithm introduces the random population to increase the population diversity in the elite group process, and the cooperation of the random and elite individuals adopts the fast simulated annealing mechanism to achieve better global optimality and faster convergence speed. This paper proves the effectiveness of the algorithm by experiments. Sex.
On the basis of observing the flow distribution characteristics of the single intersection vehicle, this paper sets up two single intersection signal control models with the goal of the minimum average delay and the minimum average parking rate of the single intersection. On this basis, the SACEA algorithm is used to optimize the timing simulation of the two models respectively, and the optimization algorithm is carried out with the existing optimization algorithms. The results show that the optimization of SACEA algorithm can reduce the average delay time and average parking rate of single intersection.
On the basis of the analysis of the coordinated control theory of urban trunk lines, this paper takes a green wave system as an example, taking the average total time (average travel time and the sum of the average delay time) as the control target by the vehicle through the trunk line, and further optimizes the timing of the multiple intersection using the SACEA algorithm, and compares the results with the existing algorithms. It shows that using SACEA algorithm to optimize can reduce the time required for vehicles to pass through the main road.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.54;TP18

【参考文献】