报告时间、地点：
报告时间：2025年1月6日，星期一，下午14:00
报告地点：华南理工大学 大学城校区，B10中座101
报告摘要：
Cash Transfer Programs (CTPs) are widely used in different regions of the world as an effective approach to fighting extreme poverty. A key issue is how to utilize information from beneficiaries to design an allocation policy if potential beneficiaries cannot be accurately targeted. In this paper, we construct a robust allocation model to minimize the objective of the FGT index with parameter based on a scenario-wise ambiguity set. To deal with the nonlinearity and infinite number of constraints in the robust model, we employ the linear decision rule and duality theory, which does not lead to a loss of optimality. Specifically, we reduce the robust model into a compact second-order cone program for (Poverty Gap Index) and a compact mixed-integer second-order cone program for (Poverty Severity Index). Based on the China Family Panel Study (CFPS) dataset, we divide the beneficiaries into different clusters with selected covariates and estimate the parameters of the ambiguity set from the predicted demand by the regression tree. Our robust allocation policy turns out to have better performance than the robust benchmarks without covariate information and the robust model based on KL divergence in terms of efficiency and equity, indicating the advantage of scenario-based distributionally robust models in moderating conservativeness to balance inclusion and exclusion errors.
主讲人简介：
柯剑男，武汉大学经济与管理学院管理科学与工程系副教授、副系主任，博士毕业于上海交通大学，美国科罗拉多大学博尔德分校访问学者。主要研究方向为收益管理和供应链管理，研究成果发表在Manufacturing & Service Operations Management、Production and Operations Management、Naval Research Logistics、Omega等期刊中，主持国家自然科学基金青年项目。
