Yinyu Ye

Brief Bio

Yinyu Ye is currently the K.T. Li Professor of Engineering at Department of Management Science and Engineering and Institute of Computational and Mathematical Engineering, Stanford University. He is also the Director of the MS&E Industrial Affiliates Program. He received the B.S. degree in System Engineering from the Huazhong University of Science and Technology, China, and the M.S. and Ph.D. degrees in Engineering-Economic Systems and Operations Research from Stanford University. His current research interests include Continuous and Discrete Optimization, Algorithm Design and Analysis, Computational Game/Market Equilibrium, Metric Distance Geometry, Dynamic Resource Allocation, and Stochastic and Robust Decision Making, etc. He is an INFORMS (The Institute for Operations Research and The Management Science) Fellow since 2012, and has received several academic awards including the inaugural 2012 **ISMP Tseng Lectureship Prize** for outstanding contribution to continuous optimization, the 2009 **John von Neumann Theory Prize **for fundamental sustained contributions to theory in Operations Research and the Management Sciences, the inaugural 2006 **Farkas Prize **on Optimization, the 2009 IBM Faculty Award, etc.. He has supervised numerous doctoral students at Stanford who received the 2008 **Nicholson Prize** and the 2006 and 2010 **INFORMS Optimization Prizes** for Young Researchers. He is the Chairman of one of the major commercial international optimization software companies. His text book written with David Luenberger, “Linear and Nonlinear Programming,” has been popularly used in academic education. Ye demonstrated his leadership in managing a group of researchers on a broader range of government and industrial projects including Boeing, American Express, Oracle, and IBM, focusing on business analytics, sensor network, big data, risk management, electronic commerce, Internet economics, etc. He also manages a broader range of government and industry funded research projects. He has been the Director of the Stanford Management Science and Engineering Department Industrial Affiliates Program since 2002, where his role is to establish direct links between members of the faculty and industrial affiliates.

Current Research Field

Ye’s current research lies in a broader range of computational mathematics and engineering areas crossing complexity theory and numerical algorithm, continuous and combinatorial optimization, game and equilibrium analyses, etc. Linear Programming (LP) has been widely used to optimize communication systems, manage energy networks, control supply-chains, plan investments, and maximize productivities. Ye resolved a major open question in LP research by developing an O(n³L) potential reduction interior-point algorithm. He and his coauthors developed a predictor-corrector interior-point algorithm and proved the first quadratic convergence result for the algorithm. Furthermore, they developed a homogeneous and self-dual LP method, which became the *Default Solver* of the major optimization software *package CPLEX-Barrier of IBM* in 2011, and has been the foundation of the efficient commercial convex optimization software MOSEK and the most popular public-domain Conic Linear Optimization software such as SEDUMI (by Jos Sturm).

He and his co-authors resolved several other significant theoretical open questions in Operations Research and Mathematical Programming, such as developed “A primal-dual interior-point method whose running time depends only on the constraint matrix”, produced the first strongly polynomial-time algorithm for Markov Decision Processes (MDP) with a constant discount, proved that the Simplex method of Dantzig is strongly polynomial for the deterministic MDP regardless discount, showed that the Arrow-Debreu equilibrium computation is in PPAD when the utility is a Leontief function, constructed a tractable distributionally-robust optimization model under moment uncertainty, and established a unified convex optimization framework for dynamic and online prediction market design.

Yes was one of the pioneer researchers on developing efficient algorithms for semidefinite programming (SDP) and second-order cone programming (SOCP); both are generalized linear programming decision models. He and his students also proved a unified rank-reduction theorem for SDP which has a direct application in low-rank matrix completion. Also, they applied SDP for localizing sensor/target points onto Euclidean spaces with incomplete and noisy metric pair-wise distances information. The result from his research resolved an open problem in graph realization and universal rigidity theory, and their solution technologies are effectively adapted in industries.

He has also made significant contributions for discrete optimization, such as an approximation algorithm for un-capacitated metric facility location, a 70% efficient algorithm for Max-Bisection using a low dimension SDP relaxation, the current best approximation local-search algorithm for capacitated facility location, the best square-root of log(n) approximation algorithm for the radii of any given set of points on any flat of given dimensions (where n is the number of points), and best approximation efficiency results for a few non-convex quadratic optimization problems.

He not only work on computation theory and algorithm, but also actively do implementations and produces public-domain optimization software (such as DSDP and the initial CVX development) for industrial and academic applications. There are total 10 packages developed by his research group and they were widely used by both academics and industries.

Personal Homepage : http://www.stanford.edu/~yyye/