基于粒子群优化的多指标组合算子的大气污染预报模型

AIR POLLUTION PREDICTION MODEL WITH MULTIPLE INDEX COMBINATION OPERATORS BASED ON PARTICLE SWARM OPTIMIZATION (PSO)

根据多指标组合算子法建立了大气污染物浓度预报的参数模型,并采用一种新颖的粒子群优化算法对大气污染物浓度预报模型中的参数进行优化。通过实例计算,该模型同线性回归、模糊模式识别、参数化组合算子方法进行了结果比较,结果表明,所建立的模型比前三种方法平均误差率小,吻合度好,具有较好的预测效果。

With the parameters optimized by a new method of Particle Swarm Optimization （PSO）, the algorithm of multiple index combination operators was used to establish a air pollution prediction model. Comparison of the results of this model with those of linear regression, fuzzy pattern recognition and parameterized combination operator indicated that this model had better coherence and smaller mean error rate than the other three methods and thus had better prediction effect.