Traffic Signal Timing Optimization Based On Ant Colony Algorithm

With the rapid development of society, traffic congestion has become one ofthe most serious problems in many cities at present, which brings great challengeto the development of cities. However，the primary reason for traffic congestionis the irrational cycle time of traffic lights. Traffic signal timing problem is amulti-objective optimization problem and has more conflicting objectives，whichhas being obtained comprehensive attention from domestic and alien scholars.Therefore, to design an effective algorithm is obviously important for multi-objective optimization. Ant colony algorithm (ACA) is a new simulatedevolutionary optimization algorithm with the characteristics of positive feedback,distributed computing and strong robustness, which provides some new ideas forsolving optimization problems. Nowadays, with the development of artificialintelligence technology, ACA has been applied to signal timing optimizationproblems.Based on traffic signal timing optimization theory, aiming at theirrationality of traffic light time, the work in this paper is expanded fromfollowing aspects.Firstly, we study the normally running vehicles when the yellow light is on,and analyze two decisions of drivers such as parking at a reduced speed andpassing the intersection at an accelerated speed, then we can easily obtain theminimum of yellow light time.Secondly, single signal timing optimization based on ACA is mainly studiedin this paper. After three important parameters in single signal control are defined,its optimization model is established based on three weighting coefficients oftime delay, number of stops and traffic capacity, then it is solved by means ofthree optimization algorithms—Webster algorithm, genetic algorithm (GA) andACA. In a word, we can achieve better performance by ACA, which can well meet the actual traffic demand.Thirdly, signal timing optimization for the trunk road based on ACA isstudied in this paper. On the basis of three parameters such as time delay, numberof stops and traffic capacity, its optimization model is established and solved bymeans of two optimization algorithms—GA and ACA. The results show thatACA is a simple and feasible method for signal timing optimization problems.