大数据时代已经到来，它为经济、社会和科学带来了革命性的变化，创造了重要的经济价值。有效的数据分析及大数据优化决策成为重要的科学问题。本课题研究带约束的大数据优化方法及在管理中的应用。包括带约束大数据优化的原始对偶一阶方法，超大规模约束大数据优化的随机坐标下降方法；带稀疏约束的各类Lasso问题，Ivanov-SVM，多核学习（MKL）新方法。项目的理论研究成果将应用于一些能源管理问题，包括电力需求预测、能源规划和电力运行优化。

The era of Big Data has arrived. Big data challenges techniques include capturing data, data storage, data analysis in the area of economical life, social life and scientific research. Our purpose of this proposal is developing constrained big data optimization theory and discovery its applications in management. The main content include first-order primal-dual method for constrained big data optimization; stochastic primal dual coordinate descent method for large scale constrained big data optimization; Lasso problem with sparsity constraints; Ivanov SVM; Multi-Kernel Learning. We also study the application of big data optimization for energy management: electricity demand forecasting, energy planing and operational optimization.