Three-Stage Bidding Strategy of Generation Company Based on Double Deep Q-Network under Incomplete Information Condition

In power market with incomplete information, a generation company only knows its own relevant information, while biddings of other market members and market environment may affect the market clearing result, which impacts the generation company’s revenue, so its bidding strategy should consider mult...

Full description

Saved in:
Bibliographic Details
Main Authors: Pengpeng YANG, Beibei WANG, Peng XU, Gaoqin WANG, Yaxian ZHENG
Format: Article
Language:zho
Published: State Grid Energy Research Institute 2021-11-01
Series:Zhongguo dianli
Subjects:
Online Access:https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202103163
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850053179375353856
author Pengpeng YANG
Beibei WANG
Peng XU
Gaoqin WANG
Yaxian ZHENG
author_facet Pengpeng YANG
Beibei WANG
Peng XU
Gaoqin WANG
Yaxian ZHENG
author_sort Pengpeng YANG
collection DOAJ
description In power market with incomplete information, a generation company only knows its own relevant information, while biddings of other market members and market environment may affect the market clearing result, which impacts the generation company’s revenue, so its bidding strategy should consider multi-dimensional market information. On the basis of deep learning reinforcement method, this paper proposes a framework based on the multi-agent DDQN (Double Deep Q-Network) algorithm to simulate the bidding strategy of generation company in the spot market. Firstly, the elements of the Markov Decision Process and action-value function in the model is defined. Secondly, the framework of the generator’s double deep Q network is established and the ε-greedy algorithm and Experience Replay Memory is adopted to train the neural network. The proposed model can make decisions based on multi-dimensional continuous states such as the market clearing price and load levels. Finally, a PJM 5-bus test case is used to compare the rewards obtained by DDQN and traditional Q-learning algorithm. The results shows that the DDQN algorithm can make appropriate decisions according to the complex state while the Q-learning algorithm has poor performance. This paper also analyzes the effectiveness of the generation company’s adoption of the DDQN algorithm for generating market strategy in terms of selection of different state vector, network generalization ability and adaptability to larger-scale calculation examples.
format Article
id doaj-art-0a6ac5e41bef427b8bd3cb023d83b2e7
institution DOAJ
issn 1004-9649
language zho
publishDate 2021-11-01
publisher State Grid Energy Research Institute
record_format Article
series Zhongguo dianli
spelling doaj-art-0a6ac5e41bef427b8bd3cb023d83b2e72025-08-20T02:52:37ZzhoState Grid Energy Research InstituteZhongguo dianli1004-96492021-11-015411475810.11930/j.issn.1004-9649.202103163zgdl-54-11-yangpengpengThree-Stage Bidding Strategy of Generation Company Based on Double Deep Q-Network under Incomplete Information ConditionPengpeng YANG0Beibei WANG1Peng XU2Gaoqin WANG3Yaxian ZHENG4School of Electrical Engineering, Southeast University, Nanjing 210096, ChinaSchool of Electrical Engineering, Southeast University, Nanjing 210096, ChinaSchool of Electrical Engineering, Southeast University, Nanjing 210096, ChinaChina Electric Power Research Institute, Nanjing 210003, ChinaChina Electric Power Research Institute, Nanjing 210003, ChinaIn power market with incomplete information, a generation company only knows its own relevant information, while biddings of other market members and market environment may affect the market clearing result, which impacts the generation company’s revenue, so its bidding strategy should consider multi-dimensional market information. On the basis of deep learning reinforcement method, this paper proposes a framework based on the multi-agent DDQN (Double Deep Q-Network) algorithm to simulate the bidding strategy of generation company in the spot market. Firstly, the elements of the Markov Decision Process and action-value function in the model is defined. Secondly, the framework of the generator’s double deep Q network is established and the ε-greedy algorithm and Experience Replay Memory is adopted to train the neural network. The proposed model can make decisions based on multi-dimensional continuous states such as the market clearing price and load levels. Finally, a PJM 5-bus test case is used to compare the rewards obtained by DDQN and traditional Q-learning algorithm. The results shows that the DDQN algorithm can make appropriate decisions according to the complex state while the Q-learning algorithm has poor performance. This paper also analyzes the effectiveness of the generation company’s adoption of the DDQN algorithm for generating market strategy in terms of selection of different state vector, network generalization ability and adaptability to larger-scale calculation examples.https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202103163deep reinforcement learningbidding strategy of generatorthree stage quotation rulesddqn
spellingShingle Pengpeng YANG
Beibei WANG
Peng XU
Gaoqin WANG
Yaxian ZHENG
Three-Stage Bidding Strategy of Generation Company Based on Double Deep Q-Network under Incomplete Information Condition
Zhongguo dianli
deep reinforcement learning
bidding strategy of generator
three stage quotation rules
ddqn
title Three-Stage Bidding Strategy of Generation Company Based on Double Deep Q-Network under Incomplete Information Condition
title_full Three-Stage Bidding Strategy of Generation Company Based on Double Deep Q-Network under Incomplete Information Condition
title_fullStr Three-Stage Bidding Strategy of Generation Company Based on Double Deep Q-Network under Incomplete Information Condition
title_full_unstemmed Three-Stage Bidding Strategy of Generation Company Based on Double Deep Q-Network under Incomplete Information Condition
title_short Three-Stage Bidding Strategy of Generation Company Based on Double Deep Q-Network under Incomplete Information Condition
title_sort three stage bidding strategy of generation company based on double deep q network under incomplete information condition
topic deep reinforcement learning
bidding strategy of generator
three stage quotation rules
ddqn
url https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202103163
work_keys_str_mv AT pengpengyang threestagebiddingstrategyofgenerationcompanybasedondoubledeepqnetworkunderincompleteinformationcondition
AT beibeiwang threestagebiddingstrategyofgenerationcompanybasedondoubledeepqnetworkunderincompleteinformationcondition
AT pengxu threestagebiddingstrategyofgenerationcompanybasedondoubledeepqnetworkunderincompleteinformationcondition
AT gaoqinwang threestagebiddingstrategyofgenerationcompanybasedondoubledeepqnetworkunderincompleteinformationcondition
AT yaxianzheng threestagebiddingstrategyofgenerationcompanybasedondoubledeepqnetworkunderincompleteinformationcondition