Research on Deep Reinforcement Learning-Driven Enterprise Risk Dynamic Prevention and Control System

- Empirical Evidence of Synergistic Governance of Market Risk and Internal Control Based on Decision Gradient Optimization

DOI

10.2991/978-94-6463-845-5_113How to use a DOI?

Keywords

Deep reinforcement learning; Risk Transfer Networks; Bi-objective optimization; Decision Gradient Optimization; Enterprise Risk Dynamics Prevention and Control

Abstract

Aiming at the problems of response lag and strategy fragmentation of traditional enterprise risk prevention and control systems in dynamic and complex environments, this study proposes a deep reinforcement learning-driven risk dynamic collaborative governance framework. By constructing a risk conduction network (RCN) model to quantify the interaction between market risk and internal control failure, designing a dual-objective optimization function to achieve a dynamic equilibrium between risk loss minimization and internal control cost constraints, and introducing decision gradient optimization (TRPO) and adaptive entropy regularization to enhance the stability of the strategy. Taking a branch office of a large state-owned enterprise as an empirical object, the model shows significant performance in offline environment: the response delay of cross-module collaboration is shortened by 72.9%. The theoretical contribution lies in the breakthrough of the traditional “prediction-response” paradigm, and the construction of a closed loop of intelligent prevention and control of “environment perception-strategy game-dynamic evolution”; the practical value is reflected in the quantifiable benefits such as 35% reduction of value at risk (VaR). Future research will focus on risk characterization intelligence (dynamic knowledge graph fusion), multi-intelligence game architecture optimization, and lightweight model deployment, to promote the transformation of enterprise risk management from “passive defense” to “active evolution”. This research provides both theoretical innovation and engineering feasibility solutions for risk management in the digital economy.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Jie Sun
PY  - 2025
DA  - 2025/09/16
TI  - Research on Deep Reinforcement Learning-Driven Enterprise Risk Dynamic Prevention and Control System
BT  - Proceedings of the 2025 6th International Conference on Management Science and Engineering Management (ICMSEM 2025)
PB  - Atlantis Press
SP  - 1157
EP  - 1163
SN  - 2667-1271
UR  - https://doi.org/10.2991/978-94-6463-845-5_113
DO  - 10.2991/978-94-6463-845-5_113
ID  - Sun2025
ER  -