A Deep Reinforcement Learning Bidding Algorithm on Electricity Market

Jia, Shuai; Gan, Zhongxue; Yang, Xi; Li, Dewei; Xue, Shibei; Wang, Limin

doi:10.1007/s11630-020-1308-0

Cited by 12 publications

(6 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CNNs are expensive in terms of computations and memories. Deeper networks are preferred for accuracy, while smaller networks are widely used due to their efficiency [30]- [32]. So model compression becomes a focus, which intends to speed up running times while maintaining accuracies.…”

Section: Knowledge Distillationmentioning

confidence: 99%

HRPose: Real-Time High-Resolution 6D Pose Estimation Network Using Knowledge Distillation

2023

Self Cite

View full text Add to dashboard Cite

Real-time six degrees-of-freedom (6D) object pose estimation is essential for many real-world applications, such as robotic grasping and augmented reality. To achieve an accurate object pose estimation from RGB images in real-time, we propose an effective and lightweight model, namely high-resolution 6D pose estimation network (HRPose). We adopt the efficient and small HRNetV2-W18 as a feature extractor to reduce computational burdens while generating accurate 6D poses. With only 33% of the model size and lower computational costs, our HRPose achieves comparable performance compared with state-of-the-art models. Moreover, by transferring knowledge from a large model to our proposed HRPose through output and feature-similarity distillations, the performance of our HRPose is improved in effectiveness and efficiency. Numerical experiments on the widely-used benchmark LINEMOD demonstrate the superiority of our proposed HRPose against state-of-the-art methods.

show abstract

Section: Knowledge Distillationmentioning

confidence: 99%

HRPose: Real-Time High-Resolution 6D Pose Estimation Network Using Knowledge Distillation

2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…With the rapid development of edge devices, e.g., mobile phones and wearable devices, unprecedented amounts of data are generated. With these data, machine learning has made breakthrough progresses and provided many successful applications, e.g., next word prediction [1], cardiac health monitoring [2], load forecasting [3,4], pose estimation [5,6] and autonomous driving [7][8][9], etc. In these applications, traditional methods for machine learning are centralized where corresponding data is aggregated and calculated on a server [10].…”

Section: Introductionmentioning

confidence: 99%

Global Update Guided Federated Learning

Wu¹,

Liu²,

Xue³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Federated learning protects data privacy and security by exchanging models instead of data. However, unbalanced data distributions among participating clients compromise the accuracy and convergence speed of federated learning algorithms. To alleviate this problem, unlike previous studies that limit the distance of updates for local models, we propose global-update-guided federated learning (FedGG), which introduces a model-cosine loss into local objective functions, so that local models can fit local data distributions under the guidance of update directions of global models. Furthermore, considering that the update direction of a global model is informative in the early stage of training, we propose adaptive loss weights based on the update distances of local models. Numerical simulations show that, compared with other advanced algorithms, FedGG has a significant improvement on model convergence accuracies and speeds. Additionally, compared with traditional fixed loss weights, adaptive loss weights enable our algorithm to be more stable and easier to implement in practice.

show abstract

“…Jia et al. [6] used deep reinforcement learning methods to dynamically learn incomplete information in the electricity market and predict the strategies of competitors. Wang et al.…”

Section: Introductionmentioning

confidence: 99%

“…In real-time market bidding, He and Zhang [5] used system dynamics simulation to predict dynamic real-time electricity price levels. Jia et al [6] used deep reinforcement learning methods to dynamically learn incomplete information in the electricity market and predict the strategies of competitors. Wang et al [7] combined reinforcement learning algorithms with system dynamics simulation, and obtained information and adapted to the environment through continuous interaction with the environment.…”

Section: Introductionmentioning

confidence: 99%

Bidding strategy of thermal power compound differential evolution game under the market mechanism of peak regulation auxiliary service

Dong

et al. 2021

IET Generation Trans & Dist

View full text Add to dashboard Cite

As the main provider of peak regulation auxiliary services, thermal power units are particularly important to ensure their revenue in an uncertain environment. To obtain the optimal bidding strategy for thermal power units, a thermal power peak regulation bidding model based on the Northeast Power Grid's auxiliary service market bidding mechanism is established. Secondly, the evolutionary game theory is introduced into the bidding strategy of thermal power units. Meanwhile, in order to solve the multi-party game problem, a compound differential evolution game algorithm is constructed to solve the two-tiers bidding. Finally, based on the actual operating data of a typical day, the efficiency of the compound differential evolution game algorithm bidding strategy is verified, and the bidding strategy of thermal power units in different situations is discussed. The results show that when all units participate in peak regulation, the first-tier quotation is an optimal strategy at 0.19 yuan kWh −1 and below, and the second-tier quotation is distributed between 0.6-0.8 yuan kWh −1. When only some units are involved in peak regulation, the first-tier quotation of thermal power units adopts a high price strategy, mostly at 0.3 yuan kWh −1 , and the second-tier quotation is still based on a high price strategy.

show abstract

A Deep Reinforcement Learning Bidding Algorithm on Electricity Market

Cited by 12 publications

References 14 publications

HRPose: Real-Time High-Resolution 6D Pose Estimation Network Using Knowledge Distillation

HRPose: Real-Time High-Resolution 6D Pose Estimation Network Using Knowledge Distillation

Global Update Guided Federated Learning

Bidding strategy of thermal power compound differential evolution game under the market mechanism of peak regulation auxiliary service

Contact Info

Product

Resources

About