基于集对分析的山区低等级公路路侧安全评价

为了提高山区低等级公路路侧安全,提出基于投影寻踪的集对分析评价方法。该方法运用投影寻踪技术从待评价路侧安全样本数据本身确定评价指标权重,避免了人为确定权重的主观性和局限性。考虑到评价指标确定性与不确定性信息并存的特点;引入集对分析理论,建立联系度评价矩阵。实现从同、异、反3方面描述指标与各评价等级之间的联系度。运用该方法对承德市滦平县部分路段路侧安全状况实施评价。并与基于灰色聚类理论的路侧安全评估法对比分析。该模型将样本看作信息的载体能够全面反映样本"携带"的信息,与灰色聚类白化函数仅将其作为数值相比;评价结果更可靠。尤其在对路段3的评价方面,灰色聚类评价法表现出了其仅将样本作为数值的局限性;其隶属度向量0.0332,0.1741,0.4171,0.3948],其中三级和四级的隶属度非常相近使得决策者难以抉择。而基于投影寻踪的集对分析模型的对应结果为0.180,0.970,12.382,0.2398],安全等级为三级;提高了信息利用率,评价方法更有优越性。 

Roadside Safety Assessment Based on Set Pair Analysis

In order to improve the roadside safety of low-grade highways in mountain areas,a set pair analysis based on projection pursuit model is proposed.This method using the projection pursuit technology that can objectively deter-mine the index weighs of sample data from candidate roads,which avoids the subjectivity and limitation of experts.Tak-ing the features of evaluation indices into account,the definitive information coexists with dubious information.The set pair analysis theory is applied to establish an evaluation matrix on degrees of connection,which describes the relationship between evaluation indices and risk ranks from three aspects:identity,difference,and antagonism.In order to verify this model,the low-grade roads in the City of Chengde,China are selected as a case study .Compared to the method based on the grey cluster theory which considers data as value,this proposed model takes data as an information carrier,which contains more information and can make evaluation results more reliable.When evaluating the third segment in the case study,the evaluation result of gray clustering is 0.0332 0.1741 0.4171 0.3948],which shows its limitation when regard-ing samples as data,makes it difficult to determine the safety level between III and IV.By contrast,the evaluation result of proposed model is 0.1800 0.9701 2.382 0.2398],which clearly determine the safety level of the road as III.The results show that roadside safety assessment based on set pair analysis is more feasible and reasonable.