Bayesian Approach to Determining the Orientation and Position of Sound-soft Obstacles with Partial Phaseless Data via Quadratic Wasserstein Metric      

主讲人：杨志鹏（兰州大学）

摘要：In this talk, we consider to reconstruct the orientation and position of an acoustically sound-soft obstacle from few phaseless far-field observation directions with respect to one single incident wave in two dimensions. The Bayesian approach is adopted to deal with the small number of observation data and to design robust inversion algorithms against different initial guesses and wavenumbers. The quadratic Wasserstein metric is utilized to measure the distance between two observation data to improve the Bayesian approach. The first-order convergence rate for computing the quadratic Wasserstein metric and the well-posedness of the posterior distribution are verified. The advantages of using the Wasserstein metric over the L2-norm and the effectiveness of the proposed method are illustrated.

主讲人简介：杨志鹏，兰州大学数学与统计学院副教授。主要研究方向为不确定性量化、反问题的数值方法等。主持完成中国博士后科学基金面上项目一项。主持在研国家自然科学基金青年项目（C类）一项。目前已在IP、CMAME、JSC、SIIMS、AMO、IPI、CAMWA、SISC、Math. Ann.等杂志发表多篇文章。   

邀请人：明炬

时间：2026年10月27日上午9:00-10:30

地点： 腾讯会议室：962-114-352