TY - JOUR
T1 - A novel learning-based binarization scheme selector for swarm algorithms solving combinatorial problems
AU - Lemus-Romani, José
AU - Becerra-Rozas, Marcelo
AU - Crawford, Broderick
AU - Soto, Ricardo
AU - Cisternas-Caneo, Felipe
AU - Vega, Emanuel
AU - Castillo, Mauricio
AU - Tapia, Diego
AU - Astorga, Gino
AU - Palma, Wenceslao
AU - Castro, Carlos
AU - García, José
N1 - Funding Information:
The APC was funded by Grant ANID/FONDECYT/REGULAR/1210810. Broderick Crawford, Wenceslao Palma and Gino Astorga are supported by Grant ANID/FONDECYT/REGULAR/1210810. Ricardo Soto is supported by grant CONICYT/FONDECYT/ REGULAR/1190129. Jos? Lemus-Romani is supported by National Agency for Research and Development (ANID)/ Scholarship Program/DOCTORADO NACIONAL/2019-21191692. Marcelo Becerra-Rozas is supported by National Agency for Research and Development (ANID)/ Scholarship Pro-gram/DOCTORADO NACIONAL/2021-21210740. Emanuel Vega is supported by National Agency for Research and Development ANID/Scholarship Program/DOCTORADO NACIONAL/2020-21202527. Jos? Garc?a was supported by the Grant CONICYT/FONDECYT/INICIACION/11180056. Jos? Garc?a acknowledge funding provided by DI Interdisciplinaria Pontificia Universidad Cat?lica de Valpara?so (PUCV). Valpara?so (PUCV), 039.414/2021. Broderick Crawford, Ricardo Soto and Marcelo Becerra-Rozas are supported by Grant Nucleo de Investigacion en Data Analytics/VRIEA/PUCV/ 039.432/2020. Marcelo Becerra-Rozas are supported by Grant DI Investigaci?n Interdisciplinaria del Pregrado/VRIEA/PUCV/039.421/2021.
Funding Information:
Data Availability Statement: You can find the code used and replicate the results in: https://github.com/joselemusr/BSS-QL Acknowledgments: Broderick Crawford, Wenceslao Palma and Gino Astorga are supported by Grant ANID/FONDECYT/REGULAR/1210810. Ricardo Soto is supported by grant CONICYT/FONDECYT/ REGULAR/1190129. José Lemus-Romani is supported by National Agency for Research and Development (ANID)/ Scholarship Program/DOCTORADO NACIONAL/2019-21191692. Marcelo Becerra-Rozas is supported by National Agency for Research and Development (ANID)/ Scholarship Program/DOCTORADO NACIONAL/2021-21210740. Emanuel Vega is supported by National Agency for Research and Development ANID/Scholarship Program/DOCTORADO NACIONAL/2020-21202527. José García was supported by the Grant CONICYT/FONDECYT/INICIACION/11180056. José García acknowledge funding provided by DI Interdisciplinaria Pontificia Universidad Católica de Valparaíso (PUCV). Valparaíso (PUCV), 039.414/2021. Broderick Crawford, Ricardo Soto and Marcelo Becerra-Rozas are supported by Grant Nucleo de Investigacion en Data Analytics/VRIEA/PUCV/ 039.432/2020. Marcelo Becerra-Rozas are supported by Grant DI Investigación Interdisciplinaria del Pregrado/VRIEA/PUCV/039.421/2021.
Publisher Copyright:
© 2021 by the authors. Licensee MDPI, Basel, Switzerland.
PY - 2021/11/1
Y1 - 2021/11/1
N2 - Currently, industry is undergoing an exponential increase in binary-based combinatorial problems. In this regard, metaheuristics have been a common trend in the field in order to design approaches to successfully solve them. Thus, a well-known strategy includes the employment of continuous swarm-based algorithms transformed to perform in binary environments. In this work, we propose a hybrid approach that contains discrete smartly adapted population-based strategies to efficiently tackle binary-based problems. The proposed approach employs a reinforcement learning technique, known as SARSA (State–Action–Reward–State–Action), in order to utilize knowledge based on the run time. In order to test the viability and competitiveness of our proposal, we compare discrete state-of-the-art algorithms smartly assisted by SARSA. Finally, we illustrate interesting results where the proposed hybrid outperforms other approaches, thus, providing a novel option to tackle these types of problems in industry.
AB - Currently, industry is undergoing an exponential increase in binary-based combinatorial problems. In this regard, metaheuristics have been a common trend in the field in order to design approaches to successfully solve them. Thus, a well-known strategy includes the employment of continuous swarm-based algorithms transformed to perform in binary environments. In this work, we propose a hybrid approach that contains discrete smartly adapted population-based strategies to efficiently tackle binary-based problems. The proposed approach employs a reinforcement learning technique, known as SARSA (State–Action–Reward–State–Action), in order to utilize knowledge based on the run time. In order to test the viability and competitiveness of our proposal, we compare discrete state-of-the-art algorithms smartly assisted by SARSA. Finally, we illustrate interesting results where the proposed hybrid outperforms other approaches, thus, providing a novel option to tackle these types of problems in industry.
KW - Binarization scheme
KW - Combinatorial problems
KW - Discretization methods
KW - Machine learning
KW - Metaheuristics
KW - Q-learning
KW - SARSA
UR - http://www.scopus.com/inward/record.url?scp=85119111942&partnerID=8YFLogxK
U2 - 10.3390/math9222887
DO - 10.3390/math9222887
M3 - Article
AN - SCOPUS:85119111942
VL - 9
JO - Mathematics
JF - Mathematics
SN - 2227-7390
IS - 22
M1 - 2887
ER -