摘要
城市快速公交线路选线决策规则的优劣直接关系到公交线路的客运能力能否满足城市客流的需求.本文根据现有公交线路的交通调查客运能力数据,结合BRT线路的'大运量、快速性'特点,发现并提取隐藏在其中的重要选线决策规则.针对选线决策问题中的多变量因素具有不确定性的特点,提出一种基于模糊神经网络实现决策规则获取的方法.给出了基于模糊推理的BP神经模糊系统结构及算法,通过神经网络提高其学习能力,采用GA改进的算法思想克服BP神经网络存在的局限.最后通过实例研究验证了模型和算法的有效性,结果可为公共交通线路选线、规划方案的制定提供参考.
Decision rule of route selection for rapid transit line determines the public transport effect.According to the current existing bus lines ability of passenger traffic survey data,combining with ＂rapidity,mass transit＂ characteristics of BRT lines,this paper is to find and extract important decision rules of route selection,For the problem of route selection,it is required to acquire evolution rule of system status from the changes of multiple environment variable factors,and all kinds of factors are of ＂fuzziness＂ uncertainty characteristics.FLS has the advantages of processing non-quantitative variable,but has the disadvantage of deficient learning capability.The BP neural fuzzy system structure and algorithm is given based on fuzzy inference,to improve its learning capability.Meanwhile,in order to overcome the existing limitations of BP neural network,it puts forward the thought of algorithm improved on the basis of GA.The model and effectiveness of the algorithm are clarified by computational experiments.The results can provide reference for the transit route selection and planning.
出处
《交通运输系统工程与信息》
EI
CSCD
北大核心
2014年第1期158-165,共8页
Journal of Transportation Systems Engineering and Information Technology
基金
国家自然科学基金项目(61263004)
甘肃省自然科学研究基金项目(0803RJZA020)
甘肃省科技支撑计划项目(090GKCA009
1304GKCA023)
关键词
智能交通
线路选线
遗传神经模糊网络
快速公交
决策规则
intelligent transportation
route selection
genetic neural fuzzy network
rapid transit
decision rule
