A Comparative Study of Energy Management Strategies for Battery-Ultracapacitor Electric Vehicles Based on Different Deep Reinforcement Learning Methods
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis paper analyzed the application method and effect of the DRL method as a way to efficiently manage energy in electric vehicles using a hybrid energy storage system (HESS) composed of a battery and a ultracapacitor. It is judged that the following supplements are needed to increase the completeness of this paper.
1. The DQN method and the DDPG method were reviewed to analyze the DRL effect, and it is necessary to explain the basis for selecting and comparing these two methods.
2. It is necessary to resize the characters included in the picture inserted in the paper.
3. Table 3 shows the parameters of the battery and capacitor, and specifically, the model information of the cells used in the experiment is required.
4. The purpose of this paper is to compare several DRL algorithms on the same line, but it is not convincing whether the design of the DDPG and DQN algorithms is designed with the same criteria.
Need to improve the quality of English language.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe article presents a comparative analysis of energy management strategies (EMS) for hybrid energy storage systems (HESS) in electric vehicles using deep reinforcement learning (DRL). It evaluates different DRL algorithms under the same benchmark and demonstrates their effects on energy efficiency, battery life, and power distribution.
Overall, the article provides a solid comparative study of DRL-based energy management strategies but suffers from grammar inconsistencies, ambiguous performance comparisons, and a lack of justification for benchmark selection. Strengthening the clarity of explanations, ensuring notation consistency, and addressing real-world implementation challenges would improve the paper's quality.
1. Lack of Clear Hypothesis and Research Question
The paper lacks a clearly stated research question in the introduction.
While it introduces the problem, it does not explicitly define what specific aspects of DRL-based EMSs are being compared.
Suggested state:
"This paper investigates how different DRL algorithms (DQN and DDPG) affect energy efficiency, battery longevity, and power distribution under identical training conditions."
2. The article compares only DQN and DDPG but does not explain why other DRL methods (e.g., PPO, SAC, TD3) were not considered. It briefly mentions "Actor-Critic methods" but does not justify why DDPG was chosen over TD3, which is a more stable variant.
So, the authors could explain why DQN and DDPG were chosen over alternatives like TD3 or PPO.
3. The study claims DDPG reduces energy loss by 28.3% compared to DQN, but the methodology does not explain if this was averaged across multiple driving cycles.
4. The term ‘economic efficiency’ (0.7% improvement) lacks a definition. Does it mean cost savings, battery lifespan extension, or reduced computational burden?
5. In Table 7, "Terminal SOC" values for the battery and ultracapacitor are given without explaining whether these values are averaged over multiple runs or if they are from a single scenario. Insert an explanation in the text.
6. All the labels and captions on all Figures are very hard to read. It is mandatory to increase the font size in all figures.
7. Table 6 lists hyperparameters but does not justify their values
8. The study does not mention how hyperparameters were optimized (grid search? Bayesian optimization?)
Grammar and Language Issues
9. "However, most scholars only study the impact of a single DRL algorithm on the performance of EMSs..."
Revise to: "However, most scholars study only the impact of a single DRL algorithm on EMS performance..."
10. "In addition, the rule-based EMS strongly relies on the engineering designers' experience." Suggestion: "Moreover, rule-based EMSs rely heavily on the experience of engineering designers."
11. "It is well known that the dynamic programming (DP) being a typical representative of global optimization strategies with an excellent control performance [15]."
Suggestion: "Dynamic programming (DP) is a well-known global optimization strategy with excellent control performance [15]."
12. “With the scarcity of oil resources and abnormal change of climate…” ‘abnormal change of climate’ is unnatural.
Suggestion: “With the depletion of oil resources and increasing climate variability...”
13. “The purpose of future research should focus on establishing standardized benchmarks to facilitate the comparison and evaluation of diverse EMSs.” Redundant phrase ‘The purpose should focus on’
Suggestion: "Future research should establish standardized benchmarks to facilitate the comparison and evaluation of diverse EMSs."
14. Unclear phrasing on: “The efficiency of the DC/DC converter is determined by the current and power, which is presented in Table 2 [32], and it can be represented by 𝜀dc = 𝐹(𝐼𝑑𝑐 , 𝑃𝑑𝑐).”
Suggestion: "Table 2 [32] presents the efficiency of the DC/DC converter as a function of current and power, expressed as 𝜀dc = 𝐹(𝐼𝑑𝑐 , 𝑃𝑑𝑐)."
15. Utilization of power supply potential’ is awkward and unclear: “As a consequence, the EMS plays a crucial role in optimizing the utilization of power supply potential.”
Suggestion: "As a result, EMS plays a crucial role in optimizing power distribution efficiency."
16. Unnecessary complexity on: "The DQN-based EMS fully utilizes the advantages of ultracapacitors to recover more regenerative energy when the driving cycle conditions undergo significant changes."
Suggestion: "The DQN-based EMS maximizes ultracapacitor efficiency in recovering regenerative energy under varying driving conditions."
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThanks to the authors. The issues have been addressed.