报告题目：Queueing Causal Models: Comparative Analytics in Queueing Systems

报 告 人：徐正航

报告时间：2025年12月10日（星期三），10:00-11:30

报告地点：明哲楼517

主办单位：东北财经大学现代供应链管理研究院

【报告人简介】

Zhenghang Xu is a fifth-year PhD candidate in Operations Management and Statistics at the Rotman School of Management, University of Toronto, advised by Professors Opher Baron, Dmitry Krass, Philipp Afèche, Arik Senderovich, and Mark van der Laan. He received his bachelor’s degree in Statistical Science from The Chinese University of Hong Kong, Shenzhen (CUHKSZ) in 2021. Zhenghang develops data-driven frameworks that integrate causal inference, machine learning, and stochastic modeling to support decision-making in complex service environments. His recent work introduces causal models for queueing systems that recover system dynamics directly from data and enable counterfactual analysis without restrictive analytical assumptions, and he also studies Bayesian pricing strategies that adaptively learn customer valuations and optimize revenue under operational constraints.

【摘要】

Problem Definition Much of the focus of queueing theory (QT) is on performance evaluation that supports comparative analytics, i.e., comparing performance measures under different interventions. However, closed-form queueing models are very sensitive to assumptions. We develop a data-driven Structural Causal Queueing Model (SCQM)--a form of structural causal models that automatically adapts to the data generating process of queueing systems, finds causal relations, and supports comparative analytics. Numerical experiments show that the accuracy of SCQM is competitive with QT even for examples where analytical queueing solutions are available. Methodology We employ structural causal modeling methodology that uses queueing-relevant features to develop a simulator that replicates the system’s data-generating process without requiring prior knowledge of its dynamics. We apply Machine Learning (ML) models for identifying the parent sets and causal relations. We then provide intervention analysis using Monte Carlo simulation. Managerial Implications We use queueing knowledge to develop an accurate self-adapting data-driven performance evaluator for congested systems that requires no prior knowledge of the system dynamics. Using this method, companies can perform comparative analytics of interventions for queueing systems that may not be analytically solvable.