上海市公路网安全宏观评价投影寻踪模型

Projection Pursuit Model for Shanghai Highway Network Safety Macroscopic Evaluation

从道路交通安全系统工程的角度,构建了上海市公路网安全宏观评价指标体系.基于投影寻踪算法的基本原理和求解过程,建立了上海市公路网安全宏观评价模型.该模型能从待评样本寻求各评价指标的客观权重,更重要的是能通过增加主观约束条件,来体现决策者对某些评价指标的偏好,实现了评价指标的主、客观权重的综合考虑.采用实码加速遗传算法优化投影指标函数,求解最优投影向量,有效地解决了高维数据全局寻优的问题.

Index system for Shanghai highway network safety macroscopic evaluation is established from the perspective of road safety system engineering. Shanghai highway network safety macroscopic evaluation model is built on projection pursuit technique which can determine the index objective weights resulting from sample data themselves, more importantly, can also take decision makers＇ bias into account by adding subjective restriction conditions to the model, which not only considers the subjective but also the objective evaluation index weight. Real coding based Accelerating Genetic Algorithm（RAGA） is proposed and applied in searching for the optimal projection directional vector, and the algorithm effectively can solve the global optimization problem with various restriction conditions.