大数据给数据分析和处理以及相关管理与决策带来了前所未有的机遇和挑战。本项目以几个典型运筹决策和生物医学影像中的重大国家和社会需求为驱动，从发展基础数学模型和高效共性最优化算法入手，加强重要基础科学问题研究，提高解决实际问题的能力：(1)针对数据量大,维度和复杂度高的特点，挖掘问题结构，探索模型与算法的结合点，发展有效的模型和数据表达格式，研究高效的一阶和二阶相结合的随机优化算法；研究求解线性规划以及整数规划高效算法；(2)基于异构众核并行平台，发展深度学习等典型机器学习的新方法和理论；(3)提供一套优化算法齐全，支持并行和分布式的数学优化软件平台；(4)将这些运筹优化基础算法和理论运用到生产实践，包括无人仓的路径规划，选址问题，冷冻电镜三维构象重建，基于影像的辅助诊断等等。本课题将为解决前沿科学研究和社会需要的瓶颈问题提供关键的理论和计算支撑，有力地促进产业、模型、理论和算法的协调发展。

Big data has brought unprecedented opportunities and challenges to data analysis and processing and related management and decision-making processes. This project is driven by the major national and social needs emerging in several typical operations research and biomedical imaging. Starting with the development of basic mathematical models and efficient common-based optimization algorithms, this project will strengthen the research on important basic scientific issues and enhance the capability to solve practical problems: (1) According to the characteristics of large amount of data, high dimensionality and high complexity, this project explores the structures of typical problems, explores the combination of model and algorithm, develops effective model and data expression format, studies the efficient first-order and second-order stochastic optimization algorithms, studies efficient algorithms for linear programming and integer programming; (2) We develop new machine learning theory and algorithms including deep learning based on heterogeneous parallel platforms; (3)We provide an optimization software package with a rich set of optimization algorithms which supports parallel and distributed computation; (4) Apply the fundamental theory and algorithms of operations research and optimization to production practice, including the path planning of unmanned warehouse, site selection, Cryo-EM 3D reconstruction, and imaging based computer aided diagnosis. This project will provide key theory and computational algorithms for solving the bottleneck problems in cutting-edge scientific research and social needs, and effectively promote the coordinated development of industries, models, theories and algorithms.