An End-to-end Flow Control Method Based on DQN

Gao, Guichao; Jin, Rong

doi:10.1109/bdicn55575.2022.00098

Cited by 3 publications

(2 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…DDPG uses the experience replay of Deep Q-learning (Gao and Jin, 2022), and adds two target networks, namely, the targetactor network and the target-critic network. The loss function L of the critic network is defined as…”

Section: Td3 Algorithmmentioning

confidence: 99%

Multi-objective optimal scheduling of charging stations based on deep reinforcement learning

Cui

Lin²,

Zhang³

et al. 2023

Front. Energy Res.

View full text Add to dashboard Cite

With the green-oriented transition of energy, electric vehicles (EVs) are being developed rapidly to replace fuel vehicles. In the face of large-scale EV access to the grid, real-time and effective charging management has become a key problem. Considering the charging characteristics of different EVs, we propose a real-time scheduling framework for charging stations with an electric vehicle aggregator (EVA) as the decision-making body. However, with multiple optimization objectives, it is challenging to formulate a real-time strategy to ensure each participant’s interests. Moreover, the uncertainty of renewable energy generation and user demand makes it difficult to establish the optimization model. In this paper, we model charging scheduling as a Markov decision process (MDP) based on deep reinforcement learning (DRL) to avoid the afore-mentioned problems. With a continuous action space, the MDP model is solved by the twin delayed deep deterministic policy gradient algorithm (TD3). While ensuring the maximum benefit of the EVA, we also ensure minimal fluctuation in the microgrid exchange power. To verify the effectiveness of the proposed method, we set up two comparative experiments, using the disorder charging method and deep deterministic policy gradient (DDPG) method, respectively. The results show that the strategy obtained by TD3 is optimal, which can reduce power purchase cost by 10.9% and reduce power fluctuations by 69.4%.

show abstract

Section: Td3 Algorithmmentioning

confidence: 99%

Multi-objective optimal scheduling of charging stations based on deep reinforcement learning

Cui

Lin²,

Zhang³

et al. 2023

Front. Energy Res.

View full text Add to dashboard Cite

show abstract

“…DRL's foray into the domain of path planning is particularly noteworthy, where it demonstrates remarkable competence in negotiating complex environments and executing challenging tasks with notable alacrity. The Deep Q-Network (DQN) paradigm [9]- [11], an amalgamation of Q-learning and deep learning, epitomizes this synergy by transforming the state-action value function into a construct amenable to neural network interpretation, thereby enabling the calculation of action values pertinent to the current state and facilitating the identification of an optimal navigational strategy.…”

Section: Introductionmentioning

confidence: 99%

An Improved Deep Reinforcement Learning Algorithm for Path Planning in Unmanned Driving

Yang,

Liu

2024

IEEE Access

View full text Add to dashboard Cite

In the domain of intelligent transportation systems, the advent of autonomous driving technology represents a critical milestone, profoundly shaping the automotive industry's evolutionary path. This technology's core, particularly the algorithms facilitating driverless path planning, has attracted significant scholarly interest. This paper presents an advanced Deep Reinforcement Learning algorithm for Path Planning (DRL-PP), designed to rectify the shortcomings inherent in existing path planning techniques. Considering the complex nature of the environment, the DRL-PP algorithm is meticulously crafted to ascertain optimal actions, thereby effectively reducing the propensity for overfitting. The algorithm harnesses the capabilities of deep reinforcement learning, utilizing neural networks to identify the most advantageous action corresponding to a specific state. It then constructs an optimal action sequence, extending from the vehicle's initial position to its designated target. Additionally, the algorithm enhances the reward function by incorporating data pertinent to the objective. This refinement enables the nuanced differentiation of action values based on dynamically adjusted reward metrics, thereby augmenting the efficiency of the action selection process and yielding improved results in path planning. Empirical results validate the algorithm's proficiency in stabilizing the reward metric while minimizing exploratory steps, consistently surpassing comparative models in pathfinding effectiveness. INDEX TERMSDeep reinforcement learning, Path planning, Autonomous driving, Deep Q-learning network. Deep Reinforcement Learning (DRL) marries the perceptual acuity of deep learning with the strategic prowess of

show abstract

Deep Reinforcement Learning-Based Joint Task Offloading in Cloud-Edge-End Cooperation Environments

Yang

Zhang

et al. 2022

2022 2nd International Conference on Frontiers of Electronics, Information and Computation Technologies (ICFEICT)

View full text Add to dashboard Cite

An End-to-end Flow Control Method Based on DQN

Cited by 3 publications

References 9 publications

Multi-objective optimal scheduling of charging stations based on deep reinforcement learning

Multi-objective optimal scheduling of charging stations based on deep reinforcement learning

An Improved Deep Reinforcement Learning Algorithm for Path Planning in Unmanned Driving

Deep Reinforcement Learning-Based Joint Task Offloading in Cloud-Edge-End Cooperation Environments

Contact Info

Product

Resources

About