Performance Evaluation of Variable Stiffness Vibration Control for a Single-Degree-of-Freedom System Using DDPG Reinforcement Learning

Wonsuk Park

doi:10.7734/COSEIK.2026.39.1.49

All Issue

2026 Vol.39, Issue 1 Preview Page Next Page

Research Paper

Performance Evaluation of Variable Stiffness Vibration Control for a Single-Degree-of-Freedom System Using DDPG Reinforcement Learning DDPG 강화학습을 이용한 단자유도 진동계의 가변강성 진동제어 성능평가

28 February 2026. pp. 49-57

PDF XML

Abstract

This study aims to clarify the performance characteristics and limitations of reinforcement learning (RL)-based controllers for structural vibration control through an objective comparison with conventional control methods under identical conditions. A single-degree-of-freedom variable stiffness system is considered, and an RL-based controller is designed using the deep deterministic policy gradient (DDPG) algorithm. Its performance is compared with that of bang-bang control and clipped-optimal control. Numerical simulations are conducted for both free vibration and forced vibration under the El Centro earthquake ground acceleration, and the nominal performance and robustness against sensor noise are evaluated. The results show that the DDPG-based controller exhibits competitive robustness in free vibration control but does not consistently outperform conventional controllers in forced vibration control. These findings offer fundamental insights into the practical capabilities and inherent limitations of RL-based vibration control under equivalent objective functions and system conditions.

Keywords

structural vibration control

reinforcement learning

variable stiffness control

deep deterministic policy gradient (DDPG)

robust control

이 논문에서는 강화학습 기반 제어기와 전통적인 제어기를 동일한 조건에서 비교함으로써 구조 진동 제어 문제에서 강화학습 제어기의 성능 특성과 한계를 규명하는 것을 목적으로 한다. 가장 단순한 비선형 제어로서 단자유도 가변 강성 시스템을 대상으로 심층 결정적 정책 경사(DDPG) 기반의 강화 학습 제어기를 설계하고, bang-bang 제어 및 제한 최적 제어와의 성능 비교를 수행하였다. 자유 진동 및 El Centro 지진 가속도에 의한 강제 진동 조건에서 공칭 성능과 센서 잡음이 존재하는 경우의 강인 성능을 분석하였다. 그 결과, 강화학습 제어기는 자유 진동 조건에서 우수한 강인 성능을 보였으나, 강제 진동 제어에서는 기존 제어기를 일관되게 상회하지는 못하였다. 이 연구는 동일한 보상 함수와 시스템 조건 하에서 강화학습 기반 진동 제어의 실질적 기여와 적용상의 한계를 기초적으로 제시하였다.

키워드

구조 진동 제어

강화학습

가변 강성 제어

심층 결정적 정책 경사(DDPG)

강인제어

References

Anderson, B.D., Moore, J.B. (2007) Optimal Control: Linear Quadratic Methods, Courier Corporation, p.448.

Bradtke, S. (1992) Reinforcement Learning Applied to Linear Quadratic Regulation, Adv. Neural Inf. Process. Syst., 5.

Djajakesukma, S.L., Samali, B., Nguyen, H. (2002) Study of a Semi-Active Stiffness Damper under Various Earthquake Inputs, Earthq. Eng. & Struct. Dyn., 31(10), pp.1757~1776.

10.1002/eqe.181

Dyke, S.J., Spencer, B.F., Sain, M.K., Carlson, J.D. (1996) Modeling and Control of Magnetorheological Dampers, J. Eng. Mech., 122(8), pp.770~780.

Housner, G.W., Bergman, L.A., Caughey, T.K., Chassiakos, A.G., Claus, R.O., Masri, S.F., Skelton, R.E., Soong, T.T., Spencer, B.F., Yao, J.T.P. (1997) Structural Control: Past, Present, and Future, J. Eng. Mech., 123(9), pp.897~971.

10.1061/(ASCE)0733-9399(1997)123:9(897)

Ikeda, Y., Saito, K. (2000) Seismic Response Control of Structures using Variable Stiffness Systems, Earthq. Eng. & Struct. Dyn., 29(12), pp. 1825~1844.

Karnopp, D. (1984) Active Damping in Road Vehicle Suspension Systems, Veh. Syst. Dyn., 12(6), pp. 291~311.

10.1080/00423118308968758

Kim, T.-H., Lee, J.-H. (2015) Vibration Control of Structures using Variable Stiffness Systems, J. Comput. Struct. Eng. Inst. Korea, 28(6), pp. 673~681.

Kirk, D.E. (2004) Optimal Control Theory: An Introduction, Courier Corporation, p.452.

Li, H., Liu, M., Li, X. (2021) Deep Reinforcement Learning-based Semi-Active Vibration Control, Mech. Syst. & Signal Process., 152, 107457.

Lillicrap, T.P., Hunt, J.J., Pritzel, A., Pritzel, A., Heess, N., Erez, T., Tassa, Y., Silver, S., Wierstra, D. (2015) Continuous Control with Deep Reinforcement Learning, arXiv preprint arXiv:1509.02971.

Nagarajaiah, S., Varadarajan, N. (2005) Short Time Fourier Transform Algorithm for Wind Response Control, Eng. Struct., 27(3), pp.431~441.

10.1016/j.engstruct.2004.10.015

Recht, B. (2019) A Tour of Reinforcement Learning, Annu. Rev. Control, Robotics, & Auton. Syst., 2, pp.253~279.

10.1146/annurev-control-053018-023825

Slotine, J.J.E., Li, W. (1991) Applied Nonlinear Control, 199(1), Englewood Cliffs, NJ: Prentice Hall.

Soong, T.T. (1990) Active Structural Control, Longman Scientific & Technical.

Spencer, B.F., Dyke, S.J., Sain, M.K., Carlson, J.D. (1997) Phenomenological Model for Magnetorheological Dampers, J. Eng. Mech., 123(3), pp.230~238.

10.1061/(ASCE)0733-9399(1997)123:3(230)

Sutton, R.S., Barto, A.G. (2018) Reinforcement Learning: An Introduction (2nd ed.). MIT Press.

Symans, M.D., Constantinou, M.C. (1999) Semi-Active Control Systems for Seismic Protection of Structures, Eng. Struct., 21(6), pp.469~487.

10.1016/S0141-0296(97)00225-3

Utkin, V.I. (1992) Sliding Modes in Control and Optimization, Springer Berlin Heidelberg, p.286.

10.1007/978-3-642-84379-2

Wang, H., Chen, G. (2020) Reinforcement Learning-based Control of Nonlinear Structures, Struct. Control & Health Monit., 27(1), e2471.

10.1002/stc.2608

Yang, T., Sun, Y., Nagarajaiah, S. (2018) Deep Reinforcement Learning for Structural Control, Eng. Struct., 172, pp.772~785.

Information

Publisher :Computational Structural Engineering Institute of Korea
Publisher(Ko) :한국전산구조공학회
Journal Title :Journal of the Computational Structural Engineering Institute of Korea
Journal Title(Ko) :한국전산구조공학회 논문집
Volume : 39
No :1
Pages :49-57
Received Date : 2025-12-15
Revised Date : 2025-12-17
Accepted Date : 2025-12-17
DOI :https://doi.org/10.7734/COSEIK.2026.39.1.49

Journal of the Computational Structural Engineering Institute of Korea ISSN:1229-3059(Print) 2287-2302(Online) 한국전산구조공학회 논문집

All Issue