Proceedings of the Web Conference 2021 2021
DOI: 10.1145/3442381.3449934
|View full text |Cite
|
Sign up to set email alerts
|

Intelligent Electric Vehicle Charging Recommendation Based on Multi-Agent Reinforcement Learning

Abstract: Electric Vehicle (EV) has become a preferable choice in the modern transportation system due to its environmental and energy sustainability. However, in many large cities, EV drivers often fail to find the proper spots for charging, because of the limited charging infrastructures and the spatiotemporally unbalanced charging demands. Indeed, the recent emergence of deep reinforcement learning provides great potential to improve the charging experience from various aspects over a long-term horizon. In this paper… Show more

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
15
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
4
2
2

Relationship

0
8

Authors

Journals

citations
Cited by 60 publications
(15 citation statements)
references
References 43 publications
0
15
0
Order By: Relevance
“…Figure 2 plots ∇J(ω t ) 2 v.s. communication complexity (t(T c + T c + T ) = 65t, t(T c + T c + T + T z ) = 70t and 2t for Algorithms 1 & 3, and both DAC-RP algorithms, respectively)7 and sample complexity (t(T c N c + N ), 2t and 110t for both of our AC-type algorithms, DAC-RP1 and DAC-RP100, respectively) 8. For each curve, its upper and lower envelopes denote the 95% and 5% percentiles of the 10 repetitions, respectively.…”
mentioning
confidence: 87%
See 2 more Smart Citations
“…Figure 2 plots ∇J(ω t ) 2 v.s. communication complexity (t(T c + T c + T ) = 65t, t(T c + T c + T + T z ) = 70t and 2t for Algorithms 1 & 3, and both DAC-RP algorithms, respectively)7 and sample complexity (t(T c N c + N ), 2t and 110t for both of our AC-type algorithms, DAC-RP1 and DAC-RP100, respectively) 8. For each curve, its upper and lower envelopes denote the 95% and 5% percentiles of the 10 repetitions, respectively.…”
mentioning
confidence: 87%
“…(ηλ F ) 2 implied by the item 2 of this Lemma, (iv) uses the inequality that x + y + z 2 ≤ 3 x 2 + 3 y 2 + 3 z 2 , ∀x, y, z ∈ R d , and (v) uses the items 4 of this Lemma. Taking unconditional expectation of the above inequality and iterating it over t yield that 7 and sample complexity (t(T c N c + N ), 2t and 110t for both of our AC-type algorithms, DAC-RP1 and DAC-RP100, respectively). 8 For each curve, its upper and lower envelopes denote the 95% and 5% percentiles of the 10 repetitions, respectively.…”
Section: For Anymentioning
confidence: 99%
See 1 more Smart Citation
“…Recently, reinforcement learning-based technique has been widely applied to solve the recommendation issues, such as the noisy items in sequential recommendation [34], dialogbased interactive recommendation [35], news recommendation [36] and social recommendation [37], [38]. For example, Wu et al [35]…”
Section: Reinforcement Learning-based Recommendationmentioning
confidence: 99%
“…We can avoid this by combining different greedy strategies to a more refined policy by adopting reinforcement learning. Reinforcement learning has evolved as an essential method to solve sequential decision problems in a dynamic environment, and was recently adopted for recommending publicly accessible charging stations [24], [20]. In reinforcement learning, the agent seeks the best policy to solve a specific problem in the environment.…”
Section: Related Workmentioning
confidence: 99%