基于PSO-RBF神经网络的项目招标评标应用研究  

Application Research on Construction Bidding Based on PSO-RBF

作　　者：汪学清[1] 黄嘉林[1] 胡德俊[1] 

机构地区：[1]中国矿业大学(北京)力学与建筑工程学院,北京100083

出　　处：《工程管理学报》2014年第6期96-101,共6页Journal of Engineering Management

基　　金：煤炭联合基金项目(A1605102)

摘　　要：根据我国目前在招标投标过程中评标方式存在的缺陷,运用组合优化思想,将经过粒子群算法优化过的径向基神经网络模型(PSO-RBF)运用到招标评标的实践中。该模型作为一种新型的评标方法,先是利用粒子群优化算法对单纯的径向基神经网络参数进行优化,进而训练和测试优化后的RBF神经网络。通过对比分析单纯的RBF神经网络和经过参数优化后的RBF神经网络模型,结果表明后者比前者在性能和效率上更加优越,在招标投标系统中更加合理、科学。最后指出了该模型还存在的缺陷有待进一步的研究。For the shortcoming of evaluation methods of bidding system in china,the radical basis function based on particle swarm optimization（PSO-RBF）algorithm is applied to biding practice. This model is a new bid evaluation method. firstly,radical basis function（RBF）is optimized by particle swarm optimization（PSO）. Secondly,the optimized RBF neural network is trained and tested,and then used to predict a sample. The paper compares the two models of RBF and PSO-RBF,and finds that the PSO-RBF model has better performance than RBF in the aspect of function and efficiency. The PSO-RBF model is more suitable for bidding system due to its reasonability. At last,the paper points out the limitation of the PSO-RBF model and directs path for further study.

关 键 词：径向基神经网络 粒子群算法 评标系统 项目招标 

分 类 号：TU723.2[建筑科学—建筑技术科学]