粒子滤波算法在评标中的应用-河北科技大学学报.pdfVIP

29 4 Vol. 29, No .4 2008 12 Journal of Hebei University of Science and T ec nology Dec. 2008 : 1008-1542( 2008) 04-0341- 06 1 1, 2 1 张晓宇, 胡士强 , 梁国壮 ( 1. 河北科技大学电气信息学院, 河北石家庄 050018; 2. 上海交通大学信息与控制所, 上海 200240) : 针对现行评标方法中 次分析法易受主观意见影响以及模糊综合评价法隶属函数的选取 带有主观随意性的不足, 提出了一种基于粒子滤波算法的评标方法该方法依据粒子滤波思想, 选 取最优值作为高斯抽样变量, 进行粒子滤波迭代, 求取某指标下某单位的中标概率密度, 解决了其 他方法在评定各指标时无法客观准确反映各投标单位真实差距的缺陷, 使评标在合理性公平性等 方面都得到了改进通过对同一投标案例的计算分析, 说明了粒子滤波算法的有效性, 同时该算法 也可以推广到其他类似的评价中 : 评标; 粒子滤波; 次分析法; 模糊综合评价法 : C934; T P391 : A Evaluating projects by particle filter 1 1, 2 1 ZHANG Xiao-yu , HU S -i qiang , LIANG Guo-z uang ( 1. College of Electrical Engineering and Information Science, H ebei University of Science and Tec nology, S ijiaz uang H ebei 050018, C ina; 2. Information and Control, S ang ai Jiaotong Universit y, S ang ai 200240, C ina) Abstract: In view of t e fact t at present evaluat ion met ods suc as AH P(t e analytic ierarc y process) are vulnerable to t e impact of subjective view s and t at fuzzy compre ensive evaluations members ip selection tends to be subjective and arbi- t rary , a evaluation met od based on particle filter algorit m is proposed. T is met od, based on t e principle of particle filter, select s opt imal value as a Gaussian random variable and uses particle f ilter iterat ive to strike a units success probabilit y density under a indicat or. T is met od resolves t e problem t at ot er met ods cannot resolve, t e unability accurately and objectively reflect all tenders real gap under various indicators.T is met od makes evaluation get imp