Article
Engineering, Electrical & Electronic
Haoxiang Chen, Hao Jia, Tao Wang, Jianhong Yang
Summary: The paper introduces a gradient-oriented binary search method for photonic device design, which accelerates convergence speed and eliminates the need for additional binarization steps compared to direct binary search methods. By designing a bimodal interferometer and other devices, the universality and fast optimization speed of the method are verified.
JOURNAL OF LIGHTWAVE TECHNOLOGY
(2021)
Article
Mechanics
Johann M. Bouwer, Schalk Kok, Daniel N. Wilke
Summary: This paper presents a shape optimization scheme for snap-through structures to match a target load-deflection curve. By using gradient-only optimization algorithms, correct solutions can be reliably located, while methods based on function values may wrongly terminate.
MECHANICS BASED DESIGN OF STRUCTURES AND MACHINES
(2023)
Article
Computer Science, Artificial Intelligence
Benson Shu Yan Lam, Alan Wee-Chung Liew
Summary: This paper proposes a BQP solver that alternates between deterministic search and stochastic neighborhood search to tackle large BQP problems. Experimental results demonstrate that the proposed solver outperforms other methods in terms of solution quality and computational complexity.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Computer Science, Software Engineering
A. L. Custodio, E. H. M. Krulikovski, M. Raydan
Summary: In this paper, we propose a new derivative-free optimization approach for solving convex constrained minimization problems. The proposed method combines different search and projection strategies, and utilizes simplified computations under specific constraints, overcoming the limitations of traditional methods.
OPTIMIZATION METHODS & SOFTWARE
(2023)
Article
Mathematics
Bushra Shakir Mahmood, Nazar K. Hussein, Mansourah Aljohani, Mohammed Qaraad
Summary: This study introduces the Multi-strategy Gradient-Based Algorithm (MAGBO) for precise parameter estimation of solar PV systems. MAGBO excelled in global optimization and demonstrated its accuracy in complex PV data analysis.
Article
Mathematics, Applied
Aliyu Muhammed Awwal, Lin Wang, Poom Kumam, Mohammed Ibrahim Sulaiman, Sani Salisu, Nasiru Salihu, Petcharaporn Yodjai
Summary: In this article, a generalized conjugate gradient parameter is proposed, which includes RMIL and its variant RMIL+ as special cases. The new method generates a sufficiently descent search direction. The convergence analysis of the proposed method is discussed under standard conditions. Experimental results demonstrate that the proposed method outperforms its main competitors.
MATHEMATICAL METHODS IN THE APPLIED SCIENCES
(2023)
Article
Engineering, Multidisciplinary
Chunxiao Liu, Xianliang Hu, Shengfeng Zhu
Summary: A binary level set method of two-grid type is proposed for structural topology optimization, which allows shape and topological changes without reinitialization. The efficiency of the gradient-type algorithm is improved and the computational efforts are balanced through a nested two-grid discretization strategy.
ENGINEERING OPTIMIZATION
(2023)
Article
Management
Said Hanafi, Yang Wang, Fred Glover, Wei Yang, Rick Hennig
Summary: This paper explores various strategies for overcoming local optimality in metaheuristic search. The characteristics of moves are analyzed to make informed decisions on steps that lead away from a local optimum and towards a new local optimum. The authors propose adaptive memory strategies based on exponential extrapolation to identify and take advantage of useful features of solution history. Experimental results on the Quadratic Unconstrained Binary Optimization (QUBO) problem show that the AA algorithm achieves a high solution quality with significantly shorter computation time compared to other algorithms.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2023)
Article
Computer Science, Artificial Intelligence
Zhi Zhao, Ke Xu, Yanxin Ma, Jianwei Wan
Summary: With the progress of deep learning on 3D point cloud, there is a growing demand for deploying point cloud neural network on edge devices. This paper presents a novel binarization framework for point cloud neural network. It proposes a gradient optimization method to overcome the shortcomings of Straight Through Estimator commonly used in binary network training, and introduces manifold preserving optimizations to restore manifold and improve feature fidelity. The proposed method outperforms state-of-the-art in accuracy and has good scalability.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Information Systems
Xiangjie Sun, Qiang Wei, Jiang Du, Yisen Wang
Summary: Binary code search is a technique used to find code similar to a given code within a code database. The proposed method, HEBCS, addresses the challenges of interpretability and efficiency in dealing with large-scale functions. It achieves high search efficiency while maintaining comparable accuracy, recall, and F1-score compared to existing methods.
Article
Computer Science, Artificial Intelligence
Bin She, Aime Fournier, Mengjie Yao, Yaojun Wang, Guangmin Hu
Summary: This paper introduces a hybrid self-adaptive gradient-based cuckoo search (HAGCS) method to address the low efficiency and high computational cost issues in using gradient information for stochastic global optimization (SGO) methods. HAGCS first employs a gradient-based local random walk to explore the search space, then uses gradient-based local optimization to find more precise solutions. Additionally, adaptive and diversity promotion strategies are introduced to avoid premature convergence and getting trapped in local minima or stationary points.
APPLIED SOFT COMPUTING
(2022)
Article
Computer Science, Artificial Intelligence
Laith Abualigah, Ali Diabat
Summary: This paper proposes a feature selection method called CGSO that combines chaotic maps and binary Group Search Optimizer. Experimental results demonstrate the superiority of this method over other published methods.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Quantum Science & Technology
Austin Gilliam, Stefan Woerner, Constantin Gonciulea
Summary: This paper discusses using Grover Adaptive Search (GAS) to solve Constrained Polynomial Binary Optimization (CPBO) problems, particularly Quadratic Unconstrained Binary Optimization (QUBO) problems, by developing efficient oracles, demonstrating potential speed-up in combinatorial optimization and QUBO, and applying to higher-degree polynomial objective functions and constrained optimization problems.
Article
Computer Science, Information Systems
Carlos O. Flor-Sanchez, Edgar O. Resendiz-Flores, Gerardo Altamirano-Guerrero
Summary: A new hybrid metaheuristic method called Kernel-based Gradient Evolution (KGE) is proposed, which introduces the concept of reproducing kernel and accurately estimates the numerical gradient for updating. The method shows superior convergence performance compared to the original method and achieves respectable results against other considered methods in numerical evaluations.
INFORMATION SCIENCES
(2022)
Article
Management
Markus Leitner, Ivana Ljubic, Michele Monaci, Markus Sinnl, Kuebra Tantnmts
Summary: This article proposes an exact solution algorithm for fortification games using a new mixed-integer linear programming reformulation. The algorithm generates fortification cuts in the natural space of fortification variables to measure the contribution of a given strategy to the objective function value. It includes branch-and-cut and speed-up techniques and outperforms existing methods in solving previously unsolved instances.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2023)