TY - GEN
T1 - On Updating Probabilistic Graphical Models in Bayesian Optimisation Algorithm
AU - El Yafrani, Mohamed
AU - Scoczynski, Marcella
AU - Delgado, Myriam
AU - Luders, Ricardo
AU - Sung, Inkyung
AU - Wagner, Markus
AU - Oliva, Diego
PY - 2019/10
Y1 - 2019/10
N2 - The Bayesian Optimisation Algorithm (BOA) is an Estimation of Distribution Algorithm (EDA) that uses a Bayesian network as probabilistic graphical model (PGM). During the evolutionary process, determining the optimal Bayesian network structure by a given solution sample is an NP-hard problem resulting in a very time-consuming process. However, we show in this paper that significant changes in PGM structure do not occur so frequently, and can be particularly sparse at the end of evolution. A statistical study of BOA is thus presented to characterise a pattern of PGM adjustments that can be used as a guide to reduce the frequency of PGM updates. This is accomplished by proposing a new BOA-based optimisation approach (FBOA) whose PGM is not updated at each iteration. This new approach avoids the computational burden usually found in the standard BOA. Inspired by fitness landscape analysis concepts, we perform an investigation in the search space of an NK-landscape optimisation problem and compare the performances of both algorithms by using the correlation between the landscape ruggedness of the problem and the expected runtime of the algorithms. The experiments show that FBOA presents competitive results with significant saving of computational time.
AB - The Bayesian Optimisation Algorithm (BOA) is an Estimation of Distribution Algorithm (EDA) that uses a Bayesian network as probabilistic graphical model (PGM). During the evolutionary process, determining the optimal Bayesian network structure by a given solution sample is an NP-hard problem resulting in a very time-consuming process. However, we show in this paper that significant changes in PGM structure do not occur so frequently, and can be particularly sparse at the end of evolution. A statistical study of BOA is thus presented to characterise a pattern of PGM adjustments that can be used as a guide to reduce the frequency of PGM updates. This is accomplished by proposing a new BOA-based optimisation approach (FBOA) whose PGM is not updated at each iteration. This new approach avoids the computational burden usually found in the standard BOA. Inspired by fitness landscape analysis concepts, we perform an investigation in the search space of an NK-landscape optimisation problem and compare the performances of both algorithms by using the correlation between the landscape ruggedness of the problem and the expected runtime of the algorithms. The experiments show that FBOA presents competitive results with significant saving of computational time.
KW - Bayesian Networks
KW - Estimation of Distribution Algorithms
KW - Model-based Metaheuristics
KW - Probabilistic Graphical Models
UR - http://www.scopus.com/inward/record.url?scp=85077050027&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85077050027&partnerID=8YFLogxK
U2 - 10.1109/BRACIS.2019.00062
DO - 10.1109/BRACIS.2019.00062
M3 - Conference contribution
AN - SCOPUS:85077050027
T3 - Proceedings - 2019 Brazilian Conference on Intelligent Systems, BRACIS 2019
SP - 311
EP - 316
BT - Proceedings - 2019 Brazilian Conference on Intelligent Systems, BRACIS 2019
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 8th Brazilian Conference on Intelligent Systems, BRACIS 2019
Y2 - 15 October 2019 through 18 October 2019
ER -