Abstract:

In this study, we observe that, for several classes of lot sizing and production distribution problems, their optimal solution values have strong correlations with some other solution values associated with Dantzig-Wolfe decompositions, uncapacitated relaxation, linear programming relaxation, and Lagrangian relaxation, etc. Using these solution values, we build advanced analytics models that can yield information about how likely a solution pattern is the same as the optimum. We then develop an analytics branching and selection method where this information is utilized to generate numerous smaller restricted subproblems iteratively. These smaller subproblems are solved using some other combined heuristics to improve feasible solutions of the original problem progressively. To the best of our knowledge, this is the first research of using likelihood information to improve solution qualities for lot sizing and production distribution problems. Comparisons with other methods indicate that this optimization method is computationally tractable and can obtain better results.

Bio:

Tao Wu has been working as a Lead Data Scientist in the Advanced Analytics group at The Dow Company since 2014. Tao has expertise in data science, operations research, data-driven and analytics-driven optimization, and business analytics, with their applications in supply chain and operations management, production and inventory planning and distribution, transportation and logistics, and energy management. He is leading the research and development of scalable business analytics, operations research, and optimization techniques for supply chain and operations management processes at Dow. Tao has successfully implemented data science, operations research, and optimization algorithms which has enabled multi-million dollar value for Dow. For example, he has implemented Mixed Integer Linear Programs (MILPs) to optimize supply chain networks and aviation fuel tankering operations, developed econometric time series models to forecast demand and prices, used artificial intelligence techniques to optimize production scheduling processes, and built hybrid econometric time series and mathematical models for pricing optimization, etc. Prior to joining Dow, Tao had worked at Apollo Group Inc. and General Motors as Operations Research Scientist and Research Engineer, respectively, from 2011 to 2014. Tao received his Master degrees in computer science in 2009 and his Ph.D. degree in Industrial and Systems Engineering in 2010 from The University of Wisconsin-Madison.

Tao has made achievements and contributions to the industry and academia in the area of operations research, analytics-driven optimization, and supply chain and operations management by having published about 30 peer-reviewed papers in renowned international journals (e.g., INFORMS Journal on Computing (3), Transportation Science, Transportation Research, Part B, Part E, European Journal of Operational Research, Omega, and Annals of Operations Research, etc.) and 7 peer-reviewed international conference papers. His publications have garnered more than 675 citations from other scholars according to the Google Scholar search engine. He has gained recognition for achievements by professional organizations, has served as Associate Editor for the International Journal of Systems Science and the International Journal of Systems Science: Operations & Logistics, and has refereed more than 50 international journal papers and book proposals. He is the recipient of the best track paper award in the 2018 International Conference on Industrial Engineering and Operations Management held at Washington DC, USA, and the recipient of multiple silver and gold awards at Dow.

欢迎广大老师和同学们参加！