节点文献
基于CM、GCA、PCA和GA-BP的公共工程分包伙伴施工能力评价研究
A Research on Public Works Subcontracting Partners Construction Ability Evaluation Based on CM,GCA,PCA and GA-BP
【摘要】 基于公共工程分包伙伴施工能力评价的相关文献分析,确定了公共工程分包伙伴施工能力评价的决定因素,构建了公共工程分包伙伴施工能力评价的指标体系。运用云模型及云的不确定性推理,将定性评价指标量化为分值,并运用灰关联度理论,确定了各评价指标的权重和各评价对象的评价结果。通过主成分分析将众多指标进行综合,消除样本间的信息重叠,降低了BP神经网络的输入维数。将遗传算法引入BP神经网络训练,优化神经网络的权值和阈值,充分发挥遗传算法的全局寻优能力和BP算法的局部搜索优势,形成了一种新的GA-BP神经网络。针对分包伙伴施工能力评价系统的非线性特征和专家评价结果具有较强的主观性的特点,采用GA-BP神经网络高度非线性映射能力,对某高速公路建设工程的分包伙伴的施工能力进行了评价。实证结果表明CM、GCA、PCA与GA-BP神经网络相结合的方法,提高了评价结果的科学性和客观性。
【Abstract】 Based on the analysis of literature regarding public works subcontracting partners’ construction ability evaluation,this paper identifies the determinants of public works subcontracting partners’ construction ability evaluation and establishes its evaluation criteria system.A cloud model and the uncertain inference are employed to translate the qualitative evaluation criteria into quantitative scores,and the weight of every criterion and the evaluation result of every evaluation object are ascertained using gray correlation analysis.Through principal component analysis,numerous criteria are synthesized,information overlapping of the sample is eliminated,and the input dimension of BP network is reduced.The genetic algorithms is introduced to optimize the right values and thresholds of BP ncural network,and then a new GA- BP neural network is presented,which takes full use of the global optimization of GA and local accurate searching of BP.Against the nonlinear feature of public works subcontracting partners’ construction ability evaluation system and the subjectivity characteristic of the evaluation result by experts,using GA-BP network’ highly nonlinear mapping ability,the construction ability of subcontracting partners of a highway construction project is evaluated.The empirical results show that the conjoint methods-CM,GCA,PCA and GA-BP neural network improve the scientific and objective of the evaluation results.
【Key words】 public works; subcontracting partners; construction ability; cloud models; gray correlation analysis; Principal component analysis; BP neural network; Genetic algorithms;
- 【文献出处】 管理评论 ,Management Review , 编辑部邮箱 ,2013年10期
- 【分类号】F284
- 【被引频次】5
- 【下载频次】297