摘要
为充分反映城市道路交叉口的综合交通效益需求,以定周期信号为约束条件,建立以机动车延误、行人延误和停车率联合最小为目标的配时优化模型。针对传统求解算法收敛速度较慢且易受约束条件限制问题,采用一种约束粒子群算法。通过对非可行解添加惩罚函数,有效消除其对最优解搜索的干扰;针对传统粒子群算法易陷入局部最优解的缺点,引入改进的杂交算子,提高解空间的多样性,加快算法的收敛速度。对给出算法进行性能仿真,分析周期时长对优化目标的灵敏性,结果证明给出算法的有效性。
To fully respond the need of comprehensive traffic benefits at urban intersection, the timing optimization model is built, which takes fixed - cycle signal as constraint and minimum vehicle delay, pedestrian delay and stopping rate as ob- jectives. To solve traditional algorithm' s problems of low convergence speed and liability to constraint, a CPS algorithm is employed. Through adding penalty function to the infeasible solution, the interference with optima searching is eliminated. Aiming at the disadvantage of traditional PSO algorithm, which is easy to trap in local optima, the improved crossover oper- ator is introduced, which improves the diversity of solution space and convergence speed. In the paper, the performance of the given algorithm is simulated and the sensibility of cycle time to optimization objectives is analyzed. The simulation result proves the effectiveness of the given algorithm.
出处
《军事交通学院学报》
2014年第7期86-90,共5页
Journal of Military Transportation University
关键词
约束条件
粒子群算法
周期时长
信号配时
constraint
PSO ( particle swarm optimization) algorithm
cycle time
signal timing