中文摘要：

为合理控制交叉口交通流,优化相位方案和信号配时,基于元胞传播模型和双层规划方法进行信号优化设计。以交叉口入口引道交通流为研究对象,改进元胞划分及其状态描述方法,建立了交叉口元胞传播模型;以相位优化问题为上层规划,以配时优化问题为下层规划,同时优化信号相位和配时。应用精确罚函数法转化下层规划后,集成遗传算法和混沌优化方法,设计了信号优化求解算法。实例计算结果表明：该算法克服了混沌优化在大范围内失效的缺点,提高了遗传算法的局部搜索能力和搜索精度,与相位固定的感应控制相比,车流总延误为270.2pcu·h,总延误减少了5.6%。

英文摘要：

In order to effectively control traffic flow,and synchronously optimize phase,period and green time at intersection,cell transmission model（CTM）and bi-level programming method were studied,and a signal optimization method was proposed.A CTM of entrance traffic flow at intersection was built by modifying cell partition and its state expression,and a bi-level programming method was designed,whose upper programming was to optimize phase scheme,while lower programming was to optimize period and green time.After lower programming was translated by using precision penalty method,an solving algorithm integrated genetic algorithm and chaos optimization algorithm was put forward.Applied result shows that the algorithm not only averts the invalidation problem of chaos optimization algorithm in a large scale,but also increases the local search ability and search precision of genetic algorithm;the method is feasible,the total delay is 270.2 pcu·h,and it is reduced by 5.6% compared with induction control with fixed phases.6 figs,11 refs.