论文部分内容阅读
随着工业化进程的加剧,雾霾已严重影响到人类的日常生活,分析天气因素进而得出影响雾霾天气的关键因子尤为重要.预测雾霾天气形成的关键因子是一个不断剔除冗余因素保留关键要素的过程,每一个天气因素都有两种状态,被选中为关键因子与否,文章根据该特点,从一维细胞自动机入手,提出了一种以二元蚁群算法作为搜索策略,分形理论作为子集评估度量准则的混合方法.因二元蚁群算法前期信息素匮乏需要较长搜索时间,引入二元粒子群算法对其进行优化,将粒子经过多次迭代之后得到的最优位置通过模糊函数映射成蚂蚁所需的信息素,在较短的时间内形成一条信息素落差明显的路径,缩短算法前期运行时间.最后将所用方法应用于北京,广州和上海三地雾霾天气关键影响因子的预测中,并结合10-交叉验证和SVM算法对预测结果分类准确率进行分析,通过与其它算法进行对比,结果表明文章算法预测结果具有较高可信度,为后期的雾霾治理工作提供了重要的参考依据.
As the process of industrialization aggravates, haze has a serious impact on the daily life of human beings, and it is particularly important to analyze the weather factors and then draw the key factors that affect the hazy weather.The key factor in forecasting the hazy weather is a continuous elimination of redundant factors Key elements of the process, each weather has two states, was selected as the key factor or not, according to the characteristics of the article, starting from one-dimensional cellular automata, a binary ant colony algorithm is proposed as a search strategy, The fractal theory as a hybrid method of evaluating metrics for subsets.Because of the long search time of binary ant colony algorithm, the binary particle swarm optimization algorithm is adopted to optimize the algorithm, and the best one obtained after multiple iterations The position is mapped into a pheromone required by ants through ambiguity function, and a clear path of pheromone gap is formed in a short period of time, which shortens the pre-operation time of the algorithm.Finally, the method is applied to haze weather in Beijing, Guangzhou and Shanghai Key impact factors in the prediction, combined with 10-cross-validation and SVM algorithm to predict the accuracy of classification results were analyzed by other Method comparison, the results indicate that the article has a high reliability prediction algorithm, provides an important reference for the governance haze of late.