Abstract:   In order to relieve the impact of the uncertainty of traffic flow on the actualization of signal control scheme under a unsaturated condition，this paper initially proposes a new optimization model of cycle length for single intersection by using the timed control． Secondly，through the analysis and comparison of the degree of mutual influence of the uncertainty of traffic flow，design of cycle length and effectiveness of control plan，the interval number multiple attribute decision model of signal cycle length at intersection is proposed on the basis of deviation and possibility degrees of interval number when attribute weights are completely unknown，attribute values are interval numbers，and the method shows no preference． Finally，the intersection in Beijing is simulated for analysis by using VISSIM simulation software． The results show that when the traffic flow ratio is between 0. 804 and 0. 909， the vehicle average delay decreases by 17. 4% and capacity increases by 6%，compared with the Webster method． The method is feasible and effective． 

Key words: urban traffic, cycle length, interval number multiple attribute decision making, signal intersection, uncertainty