Multi-Armed Bandit Regularized Expected Improvement for Efficient Global Optimization of Expensive Computer Experiments With Low Noise
Computer experiments are widely used to mimic expensive physical processes as black-box functions. A typical challenge of expensive computer experiments is to find the set of inputs that produce the desired response. This study proposes a multi-armed bandit regularized expected improvement (BREI) me...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2021-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9477602/ |
| _version_ | 1818669701066129408 |
|---|---|
| author | Rajitha Meka Adel Alaeddini Chinonso Ovuegbe Pranav A. Bhounsule Peyman Najafirad Kai Yang |
| author_facet | Rajitha Meka Adel Alaeddini Chinonso Ovuegbe Pranav A. Bhounsule Peyman Najafirad Kai Yang |
| author_sort | Rajitha Meka |
| collection | DOAJ |
| description | Computer experiments are widely used to mimic expensive physical processes as black-box functions. A typical challenge of expensive computer experiments is to find the set of inputs that produce the desired response. This study proposes a multi-armed bandit regularized expected improvement (BREI) method to adaptively adjust the balance between exploration and exploitation for efficient global optimization of long-running computer experiments with low noise. The BREI adds a stochastic regularization term to the objective function of the expected improvement to integrate the information of additional exploration and exploitation into the optimization process. The proposed study also develops a multi-armed bandit strategy based on Thompson sampling for adaptive optimization of the tuning parameter of the BREI based on the preexisting and newly tested points. The performance of the proposed method is validated against some of the existing methods in the literature under different levels of noise using a case study on optimization of the collision avoidance algorithm in mobile robot motion planning as well as extensive simulation studies. |
| first_indexed | 2024-12-17T06:56:23Z |
| format | Article |
| id | doaj.art-035d727ec15d4111bfd8ee28eda90e6d |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-17T06:56:23Z |
| publishDate | 2021-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-035d727ec15d4111bfd8ee28eda90e6d2022-12-21T21:59:24ZengIEEEIEEE Access2169-35362021-01-01910012510014010.1109/ACCESS.2021.30957559477602Multi-Armed Bandit Regularized Expected Improvement for Efficient Global Optimization of Expensive Computer Experiments With Low NoiseRajitha Meka0https://orcid.org/0000-0002-2622-8412Adel Alaeddini1https://orcid.org/0000-0003-4451-3150Chinonso Ovuegbe2https://orcid.org/0000-0002-7300-5713Pranav A. Bhounsule3https://orcid.org/0000-0002-7504-6009Peyman Najafirad4https://orcid.org/0000-0001-9671-577XKai Yang5Department of Mechanical Engineering, The University of Texas at San Antonio, San Antonio, TX, USADepartment of Mechanical Engineering, The University of Texas at San Antonio, San Antonio, TX, USADepartment of Mechanical Engineering, The University of Texas at San Antonio, San Antonio, TX, USADepartment of Mechanical and Industrial Engineering, University of Illinois, Chicago, IL, USADepartment of Information Systems and Cyber Security, The University of Texas at San Antonio, San Antonio, TX, USADepartment of Industrial and Systems Engineering, Wayne State University, Detroit, MI, USAComputer experiments are widely used to mimic expensive physical processes as black-box functions. A typical challenge of expensive computer experiments is to find the set of inputs that produce the desired response. This study proposes a multi-armed bandit regularized expected improvement (BREI) method to adaptively adjust the balance between exploration and exploitation for efficient global optimization of long-running computer experiments with low noise. The BREI adds a stochastic regularization term to the objective function of the expected improvement to integrate the information of additional exploration and exploitation into the optimization process. The proposed study also develops a multi-armed bandit strategy based on Thompson sampling for adaptive optimization of the tuning parameter of the BREI based on the preexisting and newly tested points. The performance of the proposed method is validated against some of the existing methods in the literature under different levels of noise using a case study on optimization of the collision avoidance algorithm in mobile robot motion planning as well as extensive simulation studies.https://ieeexplore.ieee.org/document/9477602/Computer experimentsGaussian process regressionexpected improvementmulti-armed banditThompson sampling |
| spellingShingle | Rajitha Meka Adel Alaeddini Chinonso Ovuegbe Pranav A. Bhounsule Peyman Najafirad Kai Yang Multi-Armed Bandit Regularized Expected Improvement for Efficient Global Optimization of Expensive Computer Experiments With Low Noise IEEE Access Computer experiments Gaussian process regression expected improvement multi-armed bandit Thompson sampling |
| title | Multi-Armed Bandit Regularized Expected Improvement for Efficient Global Optimization of Expensive Computer Experiments With Low Noise |
| title_full | Multi-Armed Bandit Regularized Expected Improvement for Efficient Global Optimization of Expensive Computer Experiments With Low Noise |
| title_fullStr | Multi-Armed Bandit Regularized Expected Improvement for Efficient Global Optimization of Expensive Computer Experiments With Low Noise |
| title_full_unstemmed | Multi-Armed Bandit Regularized Expected Improvement for Efficient Global Optimization of Expensive Computer Experiments With Low Noise |
| title_short | Multi-Armed Bandit Regularized Expected Improvement for Efficient Global Optimization of Expensive Computer Experiments With Low Noise |
| title_sort | multi armed bandit regularized expected improvement for efficient global optimization of expensive computer experiments with low noise |
| topic | Computer experiments Gaussian process regression expected improvement multi-armed bandit Thompson sampling |
| url | https://ieeexplore.ieee.org/document/9477602/ |
| work_keys_str_mv | AT rajithameka multiarmedbanditregularizedexpectedimprovementforefficientglobaloptimizationofexpensivecomputerexperimentswithlownoise AT adelalaeddini multiarmedbanditregularizedexpectedimprovementforefficientglobaloptimizationofexpensivecomputerexperimentswithlownoise AT chinonsoovuegbe multiarmedbanditregularizedexpectedimprovementforefficientglobaloptimizationofexpensivecomputerexperimentswithlownoise AT pranavabhounsule multiarmedbanditregularizedexpectedimprovementforefficientglobaloptimizationofexpensivecomputerexperimentswithlownoise AT peymannajafirad multiarmedbanditregularizedexpectedimprovementforefficientglobaloptimizationofexpensivecomputerexperimentswithlownoise AT kaiyang multiarmedbanditregularizedexpectedimprovementforefficientglobaloptimizationofexpensivecomputerexperimentswithlownoise |