Multi-Agent Reinforcement Learning for Decentralized Resilient Secondary Control of Energy Storage Systems Against DoS Attacks
Bibliographic record
Abstract
While distributed secondary controllers have been studied for multiple energy storage systems in islanded microgrids, information infrastructure has to be added for the extensive information transmission among these secondary controllers and the additional communication among distributed controllers is costly and increases the vulnerability surface to cyberattacks. In this work, a data-driven decentralized secondary control scheme is proposed for multiple heterogeneous battery energy storage systems (BESSs). The proposed secondary control scheme can achieve frequency regulation and the state-of-charge (SoC) balancing simultaneously for BESSs without requiring accurate BESS models. This scheme leverages an asynchronous advantage actor-critic (A3C) based multi-agent deep reinforcement learning (MA-DRL) algorithm where the centralized off-line learning with shared convolutional neural networks (CNN) is designed to maximize global rewards and ensure the performance of the entire system and a decentralized online execution mechanism is applied to each BESS. Furthermore, in view of possible denial-of-service (DoS) attack on local communication networks used for signal transfer between secondary controllers and remote sensors, a signal-to-interference-plus-noise ratio (SINR)-based dynamic and proactive event-triggered communication mechanism is proposed to alleviate the impact of DoS attacks and reduce the occupation of communication resources. Simulation results on a four-bus multiple BESS system show that the proposed decentralized secondary controller can achieve simultaneous frequency regulation and SoC balancing. Comparison results with other event-triggered mechanisms and MA-DRL algorithms show the A3C based MA-DRL algorithm with CNN can obtain a comparatively optimal policy through training and the designed event-triggered strategy can dynamically adapt the release frequency based on real-time SINR and significantly reduce the occupied network bandwidth and packet loss rate (PER) induced by DoS attacks.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".