超大规模约束优化问题算法及其应用天元数学交流项目

项目名称： 超大规模约束优化问题算法及其应用天元数学交流项目

项目编号： No.11726505

项目类型： 专项基金项目

立项/批准年度： 2018

项目学科： 数理科学和化学

项目作者： 彭拯

作者单位： 福州大学

项目金额： 28万元

中文摘要： 由于现代科学技术的高速发展，大数据分析与应用、科学与工程计算、物理与工程设计、经济与金融分析等各领域提出一系列大规模（超大规模）面向实际应用的复杂数学优化问题。为了促进数据与应用驱动的数学优化理论、算法及其应用研究，本项目拟搭建一个由数学优化、数据科学、计算机应用、集成电路物理设计等若干领域专家学者共同参与的学术交流平台，通过学术报告和自由讨论等形式，一方面展现数学优化理论与算法研究的最新进展，使得应用领域专家充分了解数学优化算法研究的最新成果，加速学术成果的转化利用；另一方面，使得数学优化专家充分了解数据科学及其应用、机器学习、工业与工程设计等领域需要解决的数学优化问题，促进数据驱动与应用驱动的数学优化研究迈上新台阶。

中文关键词： 一阶优化方法；大样本随机优化；群Lasso模型；主成分分析；降维方法

英文摘要： With advancement of modern science and technology, many fields such as Big data analysis and applications, scientific and engineering computations, physics and engineering design, economics and finance analysis, pose a new set of large scale or very large scale complicated optimization problems. To promote the researches on theory、methods and applications of data-driven and application-driven optimization, this project will provide a platform for academic communication of the specialists from multi-fields including optimization, data science, computer science and VLSI physical design, etc. On the platform, some high level technical reports will be provided, and some sections of associative discussion will be organized. On the one hand, the newest advancement on mathematical optimization theory and methods will be presented such that, the new research findings of mathematical programming could be fully genned-up by the applied scientists and technologists, and the transformation of science and technology achievements could be accelerated. On the other hand, some real-world problems raised in data science, machine learning and engineering design will be also presented, such that these problems could be realized by optimization specialist, and the data-driven and application-driven optimization could be promoted.

英文关键词： The first order optimization methods;Stochastic optimization on Big Samples;Group Lasso Model;Principal component analysis;Dimension reduction methods