Dynamic multiobjective optimisation has gained increasing attention in recent years. Test problems are of great importance in order to facilitate the development of advanced algorithms that can handle dynamic environments well. However, many of existing dynamic multiobjective test problems have not been rigorously constructed and analysed, which may induce some unexpected bias when they are used for algorithmic analysis. In this paper, some of these biases are identified after a review of widely used test problems. These include poor scalability of objectives and, more importantly, problematic overemphasis of static properties rather than dynamics making it difficult to draw accurate conclusion about the strengths and weaknesses of the algorithms studied. A diverse set of dynamics and features is then highlighted that a good test suite should have. We further develop a scalable continuous test suite, which includes a number of dynamics or features that have been rarely considered in literature but frequently occur in real life. It is demonstrated with empirical studies that the proposed test suite is more challenging to the dynamic multiobjective optimisation algorithms found in the literature. The test suite can also test algorithms in ways that existing test suites can not.
RDEx-CMOP: Feasibility-Aware Indicator-Guided Differential Evolution for Fixed-Budget Constrained Multiobjective Optimization
University of Toyama(富山大学);Tohoku University(东北大学);Hirosaki University(弘前大学);Tongji University(同济大学);University of Electronic Science and Technology of China(电子科技大学);Chongqing Polytechnic University of Electronic Technology(重庆电子科技职业大学);Yantai Vocational College(烟台职业学院)
RDEx-MOP: Indicator-Guided Reconstructed Differential Evolution for Fixed-Budget Multiobjective Optimization
Department of Engineering, University of Toyama(富山大学工学部);Cyberscience Center, Tohoku University(东北大学网络科学中心);Faculty of Science and Technology, Hirosaki University(弘前大学理工学部);Sino-German College of Applied Sciences, Tongji University(同济大学中德工程学院);School of Mechanical Engineering, Tongji University(同济大学机械工程学院);Department of Information Engineering, Yantai Vocational College(烟台职业学院信息工程系)
A Detection Region Method-Based Evolutionary Algorithm for Binary Constrained Multiobjective Optimization
National University of Defense Technology(国防科技大学);Hunan Key Laboratory of Multi-energy System Intelligent Interconnection Technology(湖南省多能系统智能互联技术重点实验室);Xiangjiang Laboratory(湘江实验室);Tsinghua University(清华大学)
Clustering-based Transfer Learning for Dynamic Multimodal MultiObjective Evolutionary Algorithm
School of Automation and Electrical Engineering, Zhongyuan University of Technology(中原文理大学自动化与电气工程学院);School of Electrical Engineering, Zhengzhou University(郑州大学电气工程学院)
Benchmark for CEC 2024 Competition on Multiparty Multiobjective Optimization
School of Computer Science and Technology, Harbin Institute of Technology(哈尔滨工业大学计算机科学与技术学院);School of Artificial Intelligence, Nanjing University of Information Science and Technology(南京信息工程大学人工智能学院);Department of Computer Science and Engineering, Southern University of Science and Technology(南方科技大学计算机科学与工程系)
Semantics-Guided Automatic Tensorization for Multiobjective Evolutionary Algorithms: A Multi-Agent Framework
The Hong Kong Polytechnic University(香港理工大学);The Hong Kong Polytechnic University Shenzhen Research Institute(香港理工大学深圳研究院);The Hong Kong Polytechnic University-Daya Bay Technology and Innovation Research Institute(香港理工大学-大亚湾技术与创新研究院)
Neural Architecture Search as Multiobjective Optimization Benchmarks: Problem Formulation and Performance Assessment
Southern University of Science and Technology(南方科技大学);Bielefeld University(比勒费尔德大学);University of Surrey(萨里大学);The Hong Kong Polytechnic University(香港理工大学);Michigan State University(密歇根州立大学)
Harbin Institute of Technology, Shenzhen(哈尔滨工业大学(深圳));University of Science and Technology of China(中国科学技术大学);Southern University of Science and Technology(南方科技大学)
A New Approach to Characterising Optimisation Problems Using Programmatic Representation and Complexity Measures
School of Electrical Engineering and Computer Science, University of Queensland(昆士兰大学电气工程与计算机科学学院);Department of Decision Sciences, University of South Africa(南非大学决策科学系)
Machine Learning Research Group, University of Oxford(牛津大学机器学习研究组);Lattice Lab, Toyota Motor Corporation / Machine Learning Research Group, University of Oxford(丰田汽车公司晶格实验室 / 牛津大学机器学习研究组)
神经网络在现代机器学习领域占据主导地位,但它们的训练和成功仍然受到对模型架构、损失函数和优化算法等超参数的经验选择的敏感性的影响。在这项工作中,我们提出了\emph{Population Based Training (PBT)},一个简单的异步优化算法,有效地利用固定的计算预算,共同优化人口的模型和他们的超参数,以最大限度地提高性能。重要的是,PBT发现了一个超参数设置的时间表,而不是遵循通常的次优策略,即试图找到一个用于整个训练过程的固定集合。只需对典型的分布式超参数训练框架进行一个小的修改,我们的方法就可以对模型进行鲁棒和可靠的训练。我们证明了PBT在深度强化学习问题上的有效性,通过优化一套超参数,显示出更快的挂钟收敛和更高的代理最终性能。此外,我们还展示了同样的方法可以应用于机器翻译的监督学习,其中PBT用于直接最大化BLEU分数,以及生成对抗网络的训练,以最大化生成图像的初始分数。在所有情况下,PBT都会自动发现超参数计划和模型选择,从而实现稳定的训练和更好的最终性能。
LLM Driven Design of Continuous Optimization Problems with Controllable High-level Properties
Machine Learning and Optimization, Paderborn University, Germany(机器学习与优化,帕德博恩大学,德国);Leiden Institute of Advanced Computer Science, Leiden University, The Netherlands(莱顿先进计算机科学研究所,莱顿大学,荷兰);Data Management and Biometrics, University of Twente, The Netherlands(数据管理与生物特征,埃因霍温大学,荷兰)