Generation System of Dispatching Order Sheet for Distribution Network with Self Checking Function Based on Generative Countermeasure Network. Issue 4 (October 2020)

• Record Type: Journal Article
• Authors: Jie, Huang; Qing, Ge; Kuanhong, Li; Yuxing, Chen; Ling, Ye
• Is Part Of: Journal of physics. Volume 1648:Issue 4(2020)
• Publication Date: 2020-10
• Subjects: Generative Countermeasure -- Self Calibration of Network -- Distribution Network -- Dispatching Instruction; Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1648/4/042066
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: Although the recent wave of artificial intelligence research led by deep learning has achieved very good results in the field of supervised learning. Unsupervised learning, as a method that can truly allow computers to learn from unlabeled real data from the real world, can avoid tedious and unavoidable data labeling work in supervised learning. If you want a computer to better understand the complex real world, the best way is to let the computer generate a representation of the real world in a certain way. The first thing needed to accomplish the above goals is the generative model. The most outstanding performance in generative models in recent years is the variational autoencoder and generative confrontation network introduced in this article. As an extension of the autoencoder, the former is a good combination of deep learning ideas and statistical learning. The high-dimensional distribution of the image can be reduced through the encoder network, and then the decoder network can be used to achieve low-dimensional Data distribution automatically generates an image similar to the original image. Therefore, in the subsequent improvement, the researchers took the advantages of combining the variational autoencoder and the generative confrontation network. The experimental results show that this paper has jointly trained VAE and GAN and achieved good results. However, due to the inherent shortcomings of the original generative adversarial network, the combination of the two still cannot achieve very good results.