Data-Driven Optimization-Yinyu Ye英文学习资料 .pdfVIP

Distributionally Robust Optimization (DRO) Online Linear Programming (OLP) Least Squares with Nonconvex Regularization (LSNR) Alternating Direction Method of Multipliers (ADMM) Data-Driven Optimization Yinyu Ye K.T. Li Chair Professor of Engineering Department of Management Science and Engineering Stanford University June, 2014 Yinyu Ye June 2014 Distributionally Robust Optimization (DRO) Online Linear Programming (OLP) Least Squares with Nonconvex Regularization (LSNR) Alternating Direction Method of Multipliers (ADMM) Outline We present several optimization models and/or computational algorithms dealing with uncertain, dynamic/online, structured and/or massively distributed data: Distributionally Robust Optimization (data uncertainty) Online Linear Programming (data dynamics) Least Squares with Nonconvex Regularization (data structure) The ADMM Method with Multiple Blocks (data size) Yinyu Ye June 2014 Distributionally Robust Optimization (DRO) Online Linear Programming (OLP) Least Squares with Nonconvex Regularization (LSNR) Alternating Direction Method of Multipliers (ADMM) Mathematical Optimization Classic mathematical optimization considers: maximizex∈D h(x) Since h(x) may be partially decided by other input data, say ξ , we actually maximizex∈D h(x, E [ξ]) Yinyu Ye June 2