FINTECH THRUST SEMINAR | Integrated Estimation and Optimization: Theory and Application
Supporting the below United Nations Sustainable Development Goals:支持以下聯合國可持續發展目標：支持以下联合国可持续发展目标：
Integrated Estimation and Optimization: Theory and Application
Abstract:
As data-driven decision-making becomes increasingly central to operations and analytics, the integration of estimation and optimization has emerged as one of the attractive research frontiers. This talk presents a unified perspective on Integrated Estimation and Optimization, drawing on several strands of my recent research. I will begin with decision-driven regularization, a framework that embeds downstream decision objectives directly into predictive modeling to enhance decision quality. I will then introduce a robust satisficing approach that incorporates parameter uncertainty and residual ambiguity into a unified robust optimization formulation. Next, I will discuss contextual robust optimization, which explicitly accounts for endogeneity in estimation. These frameworks are illustrated through applications in shortest-path planning, wine investment, advertising, inventory management, and pricing optimization.
随着数据驱动决策在运营与分析领域的重要性日益提升，如何有效地协同“预测”与“优化”，已成为备受关注的研究前沿之一。本次讲座将介绍我近期在协同预测与优化的研究工作。首先，我将介绍决策驱动正则化框架，通过将决策目标直接嵌入预测模型来提升决策质量。随后，我将展示鲁棒预测与优化方法，聚焦如何在鲁棒优化模型中同时处理参数不确定性与残余模糊性。进一步地，当预测存在内生性时，我将讨论如何用情境鲁棒优化的框架进行处理。上述方法将通过多个应用场景进行展示，包括最短路径规划、红酒投资、广告投放、库存管理与定价优化。
Dr TANG Qinshen is currently an assistant professor at the Division of Information Technology and Operations Management, Nanyang Business School, Nanyang Technological University, Singapore. He joined ITOM in 2020. Prior to that, he received his PhD in Analytics and Operations from the National University of Singapore, M.S. in Management Science and Engineering, and B.S. in Industrial Engineering from the South China University of Technology. His primary research interests lie in data-driven and target-based decision-making under uncertainty, with operations and supply chain management applications. He is also interested in applying cooperative and non-cooperative game theory to solve problems at the interface of operations management and marketing/economics. His paper has been published in flagship journals such as Management Science, Operations Research, Manufacturing & Service Operations Management, etc. He is a recipient of the APORS Young Researcher Best Paper Award, NBS Research Excellent Award and nominee of the NBS Teaching Excellence Award and NBS Teacher of the Year Award.
汤勤深，本科和硕士就读于华南理工大学，于2019年获得新加坡国立大学商业分析与运营系博士学位，现就职于新加坡南洋理工大学南洋商学院，担任信息技术和运营管理系助理教授。他目前主要研究兴趣包括数据驱动和目标导向的决策模型、鲁棒和鲁棒性优化，以及他们在运营和供应链管理中的应用。其研究结果发表在Management Science, Operations Research, Manufacturing & Service Operations Management等期刊。曾获APORS青年研究最佳论文奖，南洋商学院科研卓越奖，并多次入围南洋商学院教学卓越奖与年度最佳教师提名。