摘要
为解决当前道路安全评价模型中存在评价指标繁多,数据维数冗杂、极易陷入局部最优等问题,将能够将高维数据投影到低维子空间并进行分析的投影寻踪法与通过模拟鱼类在觅食时相互尾随到达食物点,依靠集体智慧聚集成群,并最终寻找最佳生存环境的人工鱼群算法相结合,构建道路安全评价模型,并选用合理评价指标进行实例验证。结果表明,该模型能够有效避免最优值局部收敛,计算精度高,抗干扰能力强,并且评价结果符合实际情况。
In order to solve the problems in the current evaluation model of road safety,including excessive evaluation index and data dimension,as well as occurrence of the partial optimum,an evaluation model of road safety was established which combines projection which analyzes the obtained data through projecting the high-dimensional data to the low-dimensional subspace,and artificial fish-swarm algorithm which is to imitate the fish＇s actions of clustering for safety and food and finding the best living environment.Some reasonable evaluation indexes were selected to test the availability of the model.The results showed that the model can effectively avoid the partial convergence of optimal value.In addition,it has high accuracy and strong anti-interference capability,as well as accordance with the actual situation of the results.
出处
《森林工程》
2013年第4期96-99,共4页
Forest Engineering
基金
哈尔滨市科技创新人才研究专项资金项目(2012RFXXS089)
关键词
道路安全
评价
投影寻踪
人工鱼群
road safety
evaluation
projection pursuit
artificial fish-swarm
