Article
Computer Science, Artificial Intelligence
Mei Han, Linhan Ouyang
Summary: This study proposes a novel framework for Bayesian optimization of multiple stochastic responses, which improves the performance of multi-objective stochastic simulation optimization by constructing stochastic Kriging metamodels and integrating acquisition functions.
SWARM AND EVOLUTIONARY COMPUTATION
(2022)
Review
Energy & Fuels
Daniel J. Silva, Joao Ventura, Joao P. Araujo
Summary: The research focuses on modeling of heat effect device systems, optimizing key geometrical parameters, and utilizing a range of optimization strategies, including sensitivity analysis, brute force methods, statistical learning, and genetic algorithms. These models are crucial in designing the most efficient, cheap, and robust setups with reliable performances.
INTERNATIONAL JOURNAL OF ENERGY RESEARCH
(2021)
Article
Multidisciplinary Sciences
Libin Tan, Yuejin Yuan, Man Zhang
Summary: This research predicted and optimized the hydraulic performance of an engine cooling water pump using CFD analysis, with the optimized designs showing significant improvement in efficiency and head. The results of the prediction model and automatic optimization model were effective in providing a theoretical basis for the design, development, and optimization of engine cooling water pump.
Article
Engineering, Chemical
Pranav Kherdekar, Shantanu Roy, Divesh Bhatia
Summary: This study introduces a method for designing a specific fixed-bed reactor and analyzes the dynamic stability of the solutions obtained using the nondominated sorting genetic algorithm. The optimal values of decision variables are dependent on different objective functions and tolerable deviations, while the influence of steam to CO ratio on the H-2/CO ratio needs to be considered in terms of pressure drop and operating costs.
INDUSTRIAL & ENGINEERING CHEMISTRY RESEARCH
(2021)
Article
Automation & Control Systems
Anuj Pal, Ling Zhu, Yan Wang, Guoming G. Zhu
Summary: This article experimentally calibrates and optimizes the control parameters of an engine using a stochastic surrogate-assisted optimization method, and achieves good experimental results.
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2022)
Article
Operations Research & Management Science
P. B. Assuncao, O. P. Ferreira, L. F. Prudente
Summary: The paper analyzes the conditional gradient method, also known as the Frank-Wolfe method, for constrained multiobjective optimization. Different strategies for obtaining step sizes are considered, and asymptotic convergence properties and iteration-complexity bounds are established with and without convexity assumptions on the objective functions. Numerical experiments are provided to illustrate the effectiveness of the method and certify the obtained theoretical results.
COMPUTATIONAL OPTIMIZATION AND APPLICATIONS
(2021)
Article
Engineering, Electrical & Electronic
Shan Wang, Zijian Qiao, Pingjuan Niu
Summary: Most traditional stochastic resonance methods have limitations in extracting fault signature due to insufficient impulse signals, vulnerability to interference, and dependence on advanced requirements. This study investigates the dynamic model of rolling bearings and proposes the coupled hybrid stochastic resonance (CHSR) method to identify weak signal signatures. The proposed method offers advantages such as corresponding with real vibration signals, processing 2-D signals to enhance weak useful signals, and adapting parameter adjustment through multiobjective optimization. The CHSR method demonstrates superior performance in identifying weak fault signatures, as supported by dynamic simulation and applied machine learning models.
IEEE SENSORS JOURNAL
(2023)
Article
Operations Research & Management Science
Frank E. Curtis, Suyun Liu, Daniel P. Robinson
Summary: A strategy for fair supervised learning is proposed, which involves minimizing loss while satisfying a constraint on a measure of unfairness. This strategy can be incorporated into a multi-objective optimization method to generate a Pareto front for minimizing loss and unfairness. A scalable stochastic optimization algorithm is proposed for solving the resulting constrained optimization problems in large data settings. Numerical experiments on recidivism and income prediction problems demonstrate the effectiveness of this strategy in large-scale fair learning.
OPTIMIZATION LETTERS
(2023)
Article
Engineering, Chemical
Bruno Leite, Andrea Oliveira Souza da Costa, Esly Ferreira da Costa Junior
Summary: In this study, steady-state axial-flow and radial-flow multibed catalyst reactors for dehydrogenation of ethylbenzene into styrene were optimized, with radial-flow reactor proving to be more efficient at lower pressures. The obtained optimal results can be used for integrated process analysis and defining operational conditions.
CHEMICAL ENGINEERING SCIENCE
(2021)
Article
Automation & Control Systems
Seon Han Choi, Bong Gu Kang, Tag Gon Kim
Summary: This study presents an efficient ranking and selection procedure for simulation-based optimization, aiming to maximize selection accuracy under a limited simulation budget. By updating simulation data with a heuristic policy and considering the precision of sample mean ignored in previous studies, this procedure may be more efficient in simulation models with large stochastic noise. Experimental results show improved efficiency, and a case study on military network system design demonstrates its practical effectiveness.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2021)
Article
Geosciences, Multidisciplinary
Jichuan Zhao, Shuangquan Chen
Summary: This paper proposes a method for reconstructing complex reservoir structures using a combination of variational autoencoder and generative adversarial networks (VAE-GAN) with conditional image quilting algorithm. The method improves the stability and quality of traditional GAN-based simulation, enhances the diversity of facies patterns, and successfully reproduces complex geological structures in real training images.
JOURNAL OF APPLIED GEOPHYSICS
(2023)
Article
Green & Sustainable Science & Technology
Yingjie Xu, Songlin Huang, Jiafeng Wang, Mengjie Song, Jiaqi Yu, Xi Shen
Summary: Solar-driven refrigeration systems are useful for building cooling, but the assumption of unlimited and free solar heat is not always valid. This study proposes a new ejector-partially coupled enhanced compression refrigeration cycle that consumes less heat. Energy comparison and advanced exergy analysis are conducted to evaluate its performance. The results show that the new cycle achieves higher energy efficiency and lower annual cost compared to the traditional ejector-compressor cycle.
JOURNAL OF RENEWABLE AND SUSTAINABLE ENERGY
(2022)
Article
Engineering, Chemical
Wei-Ting Tang, Jeffrey D. Ward
Summary: This study conducts detailed energy and exergy analyses for a stacked complex sequence (SCS) and alternatives for a ternary zeotropic distillation problem. Multi-objective optimizations are performed, and the results suggest that SCSs are a viable alternative to thermally coupled or DWC configurations for the separation of ternary mixtures.
SEPARATION AND PURIFICATION TECHNOLOGY
(2023)
Article
Management
Kuo-Hao Chang, Robert Cuckler, Song-Lin Lee, Loo Hay Lee
Summary: This paper extends the concept of Conditional Expectation to the expected value of a loss function between the alpha- and beta-quantiles of the output distribution from a simulation model. A simulation optimization framework called APHS-CE is proposed to efficiently estimate and optimize this problem. The algorithm shows promising performance in exploring the feasible region and converging to the global optimum.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2022)
Article
Engineering, Civil
Xin Huang, Bin Xu, Ping-an Zhong, Hongyi Yao, Hao Yue, Feilin Zhu, Qingwen Lu, Yu Sun, Ran Mo, Zhen Li, Weifeng Liu
Summary: The study established a robust multiobjective operation and risk decision-making model for informing reservoir operations. The model utilizes copula function, robust multiobjective optimization, adaptive reference multiobjective evolutionary algorithm, and TOPSIS multicriteria decision-making method to achieve more widely distributed noninferior solutions than traditional models. The methodologies were verified by application to Xianghongdian reservoir in China and proved to be effective in reducing maximum release and improving system vulnerability.
JOURNAL OF HYDROLOGY
(2022)