模糊环境下的组合预测模型及其在电力负荷中的应用  

Application of Combination Forecasting Model in Load Forecasting under the Fuzzy Environment

作　　者：朱家明[1] 丁珍妮 陈华友[1] 周礼刚[1] 刘金培[2] ZHU Jia-ming;DING Zhen-ni;CHEN Hua-you;ZHOU Li-gang;LIU Jin-pei(School of Mathematical Sciences, Anhui University, Hefei 230601,China;School of Business, Anhui University, Hefei 230601, China)

机构地区：[1]安徽大学数学科学学院,安徽合肥230601 [2]安徽大学商学院,安徽合肥230601

出　　处：《模糊系统与数学》2018年第2期147-158,共12页Fuzzy Systems and Mathematics

基　　金：国家自然科学基金资助项目(71371011;71501002);教育部人文社科青年项目(13YJC630092)

摘　　要：传统的组合预测中,各单项预测方法的权重系数均为固定实数,并且预测信息往往是实数或区间数,而实际系统的不确定性又表现为模糊性的特征。针对权重系数和预测信息均为三角模糊数的情形,提出三角模糊数的相对重合度和重心距离测度的概念,构造新的精确度指数,并将其作为模糊优化组合预测模型的准则。考虑到各单项预测方法在不同时刻的预测精度有所不同,引入诱导有序三角模糊加权平均(IOTFWA)算子,构建基于IOTFWA算子的单目标模糊优化组合预测模型。最后,对电力负荷预测的实例说明该模糊组合预测方法的有效性,并对参数做了灵敏度分析。The weighting coefficients of all single forecasting methods are determined by positive real numbers and the prediction information is always expressed by real numbers or interval numbers in traditional combination forecasting methods. However, uncertainty of real systems has some fuzziness. The concepts of relative overlap degree and center distance measure of any two triangular fuzzy numbers are proposed in the circumstances that the weighting coefficients and prediction information are all triangular fuzzy numbers. We construct the new precision index as the optimal criteria of fuzzy combination forecasting model. Considering that the forecasting accuracy of each single method is different in different time, we present the induced ordered triangular fuzzy weighted averaging （IOTFWA） operator. The single-object fuzzy optimal combination forecasting model is developed based on IOTFWA operator. An example about the power load forecasting is illustrated to show that the effectiveness of the proposed method is guaranteed, and sensitivity analysis is given to some parameters.

关 键 词：组合预测 三角模糊数 模糊优化 IOTFWA算子 有效性 

分 类 号：O159[理学—数学]