CMA-ES for Safe Optimization

被引:0
|
作者
Uchida, Kento [1 ]
Hamano, Ryoki [1 ,2 ]
Nomura, Masahiro [2 ]
Saito, Shota [1 ,3 ]
Shirakawa, Shinichi [1 ]
机构
[1] Yokohama Natl Univ, Yokohama, Kanagawa, Japan
[2] CyberAgent Inc, Shibuya Ku, Tokyo, Japan
[3] SKILLUP NeXt Ltd, Tokyo, Japan
来源
PROCEEDINGS OF THE 2024 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, GECCO 2024 | 2024年
关键词
safe optimization; covariance matrix adaptation evolution strategy; Gaussian process regression; Lipschitz constant;
D O I
10.1145/3638529.3654193
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In several real-world applications in medical and control engineering, there are unsafe solutions whose evaluations involve inherent risk. This optimization setting is known as safe optimization and formulated as a specialized type of constrained optimization problem with constraints for safety functions. Safe optimization requires performing efficient optimization without evaluating unsafe solutions. A few studies have proposed the optimization methods for safe optimization based on Bayesian optimization and the evolutionary algorithm. However, Bayesian optimization-based methods often struggle to achieve superior solutions, and the evolutionary algorithm-based method fails to effectively reduce unsafe evaluations. This study focuses on CMA-ES as an efficient evolutionary algorithm and proposes an optimization method termed safe CMA-ES. The safe CMA-ES is designed to achieve both safety and efficiency in safe optimization. The safe CMA-ES estimates the Lipschitz constants of safety functions transformed with the distribution parameters using the maximum norm of the gradient in Gaussian process regression. Subsequently, the safe CMA-ES projects the samples to the nearest point in the safe region constructed with the estimated Lipschitz constants. The numerical simulation using the benchmark functions shows that the safe CMA-ES successfully performs optimization, suppressing the unsafe evaluations, while the existing methods struggle to significantly reduce the unsafe evaluations.
引用
收藏
页码:722 / 730
页数:9
相关论文
共 50 条
  • [31] The lens design using the CMA-ES algorithm
    Nagata, Y
    GENETIC AND EVOLUTIONARY COMPUTATION GECCO 2004 , PT 2, PROCEEDINGS, 2004, 3103 : 1189 - 1200
  • [32] Acoustic topology optimisation using CMA-ES
    Ramamoorthy, V. T.
    Ozcan, E.
    Parkes, A. J.
    Sreekumar, A.
    Jaouen, L.
    Becot, F. X.
    PROCEEDINGS OF INTERNATIONAL CONFERENCE ON NOISE AND VIBRATION ENGINEERING (ISMA2020) / INTERNATIONAL CONFERENCE ON UNCERTAINTY IN STRUCTURAL DYNAMICS (USD2020), 2020, : 511 - 522
  • [33] Viability Principles for Constrained Optimization Using a (1+1)-CMA-ES
    Maesani, Andrea
    Floreano, Dario
    PARALLEL PROBLEM SOLVING FROM NATURE - PPSN XIII, 2014, 8672 : 272 - 281
  • [34] An Asynchronous Implementation of the Limited Memory CMA-ES
    Arkhipov, Viktor
    Buzdalov, Maxim
    Shalyto, Anatoly
    2015 IEEE 14TH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA), 2015, : 707 - 712
  • [35] Augmented Lagrangian, penalty techniques and surrogate modeling for constrained optimization with CMA-ES
    Dufosse, Paul
    Hansen, Nikolaus
    PROCEEDINGS OF THE 2021 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE (GECCO'21), 2021, : 519 - 527
  • [36] CMA-ES with exponential based multiplicative covariance matrix adaptation for global optimization
    Karmakar, Bishal
    Kumar, Abhishek
    Mallipeddi, Rammohan
    Lee, Dong-Gyu
    SWARM AND EVOLUTIONARY COMPUTATION, 2023, 79
  • [37] Electromagnetic Design Optimization using Mixed-Parameter and Multiobjective CMA-ES
    BouDaher, Elie
    Hoorfar, Ahmad
    2013 IEEE ANTENNAS AND PROPAGATION SOCIETY INTERNATIONAL SYMPOSIUM (APSURSI), 2013, : 406 - 407
  • [38] Topology Optimization of Plasmonic Device Using Function Expansion Method and CMA-ES
    Tsuji, Y.
    Morimoto, K.
    Iguchi, A.
    Kashiwa, T.
    2022 3RD URSI ATLANTIC AND ASIA PACIFIC RADIO SCIENCE MEETING (AT-AP-RASC), 2022,
  • [39] A CMA-ES with Multiplicative Covariance Matrix Updates
    Krause, Oswin
    Glasmachers, Tobias
    GECCO'15: PROCEEDINGS OF THE 2015 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, 2015, : 281 - 288
  • [40] Sequential Sampling for Noisy Optimisation with CMA-ES
    Groves, Matthew
    Branke, Juergen
    GECCO'18: PROCEEDINGS OF THE 2018 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, 2018, : 1023 - 1030