报告题目：An adaptive Lagrangian algorithm for optimal portfolio deleveraging with cross-impact

报告人：徐凤敏 教授 西安交通大学

报告时间：2016年11月4日15:40-16:30

报告地点：信远楼II206我院报告厅

报告人简介：徐凤敏，女，河南郑州人，计算数学博士，西安交通大学经济与金融学院教授、博士生导师，加拿大西蒙弗雷泽大学访问学者、香港理工大学访问学者、韩国首尔大学高级交换学者和客座研究员，西安交通大学经济与金融学院金融工程系系主任助理。中国双选法学会理事，中国双选法学会经济数学与管理数学分会副理事长兼秘书长,中国运筹学学会数学规划分会理事。

申请人长期致力于大数据所涉及的统计与稀疏优化理论算法的研究和典型金融问题微观研究。目前，已在国内外知名期刊上发表(含录用)论文23篇，其中被SCI检索20篇。参与编写专著1部。主持两项国家自然科学基金，参与一项自然科学基金重点项目。

报告摘要：With the frequent occurrence of financial crisis and much more severer devastating blows, more and more economists have been studying how to minimize the extent of the losses during the financial crisis. In this talk we analyze the problem of optimal portfolio deleveraging. we model the optimal portfolio liquidation problem to meet a specified liability/equity requirement, and prove that

it is NP-hard, which lies a theoretical foundation for algorithm design. Then without the restrictions on the relative magnitudes of permanent and temporary market impact, we elaborate the existence of rationality for non-convex assets through the intuition behind price factor, and also show the priority level of asset sales. Furthermore, taken the permanent and temporary price cross-impact into account, we establish a quadratic program with quadratic and box constrains. Under reasonable assumptions, we obtain an optimal trading strategy, and show that the optimal solution must be achieved when the quadratic constraint is active. We propose adaptive Lagrangian algorithm to solve these model, And the numerical results show that our algorithm is effective, and demonstrate the optimal trading strategy.