基于属性偏好学习的配电网综合评价方法

发布时间：2018-01-06 19:41

本文关键词：基于属性偏好学习的配电网综合评价方法　出处：《计算机应用研究》2017年03期 　论文类型：期刊论文

【摘要】：为了摆脱在传统地区配电网评价方法中对参评人员个人评价偏好的过度依赖,实现合理、精准的属性权重确定,提出了一种基于属性偏好学习的配电网多指标智能综合评价方法。依据属性测度理论,在置信度准则与评分准则下完成对配电网综合评价模型的构造;进而提出数值绝对偏移率指标以实现对中间值指标的数据预处理;最后,应用随机权神经学习方法,通过对配电网历史训练样本进行有监督学习,计算得到指标属性偏好权重,并依据配电网综合评价模型以及计算所得属性偏好权重完成对配电网待测样本的智能综合评价。与传统的AHP、PSO-SVM以及RWN算法的对比仿真实验验证了该方法的精确性与稳定性。该方法实现了合理、客观的配电网综合评价,对地区配电网评价具有一定的实际应用价值。
[Abstract]:In order to get rid of the excessive dependence on the individual evaluation preference of the participants in the traditional regional distribution network evaluation method, to achieve reasonable and accurate attribute weight determination. A multi-index intelligent comprehensive evaluation method based on attribute preference learning is proposed in this paper. According to the theory of attribute measure, the comprehensive evaluation model of distribution network is constructed under the confidence criterion and scoring criterion. Then, the numerical absolute deviation rate index is proposed to realize the data preprocessing of the intermediate value index. Finally, by using the stochastic weight neural learning method, the index attribute preference weight is calculated through supervised learning of the distribution network history training samples. And according to the distribution network comprehensive evaluation model and calculated attribute preference weight to complete the intelligent comprehensive evaluation of distribution network samples to be tested. And the traditional AHP. The accuracy and stability of the method are verified by the comparison of PSO-SVM and RWN algorithm. The method realizes reasonable and objective comprehensive evaluation of distribution network. It has certain practical application value to the regional distribution network evaluation.
【作者单位】：华北电力大学电气与电子工程学院;
【基金】：国家“863”计划资助项目(2015AA050203)
【分类号】：TM727;TP18
【正文快照】： 0引言配电网作为我国电力系统的重要组成部分,对供电质量、电网运行效率以及用电质量等多方面有着重要的作用。对配电网进行科学、综合的评价,在分析配电网薄弱环节、促进配电网智能化转变、提高配电网供电可靠性以及确定配电网发展方向上都具有重要的意义[1,2]。近年来,智能

【相似文献】