The curse of dimensionality is an obstacle to the computation of numerical solutions for PDE in high dimensions. In the setting of free boundary problems, the high number of degrees of freedom prevents us from representing a general boundary in the computer. Therefore, it becomes natural to specify a parametric form for the boundary and design algorithms to find the optimal parameter.To carry out this optimization one must be able to estimate the sensitivity of the solution with respect to the boundary parameters. In this talk I present an efficient Monte Carlo estimator for this sensitivity, based on simulating both the stopping time and the delta at the boundary accurately.
Speaker: Jose Alcala, New York University
Slides: (TBA)
Event Details
Optimizing the exercise boundary for the holder of an American
- Event Date: March 23, 2010
- Event End Date: March 23, 2010
- Event Start Time: 1:45 PM
- Event End Time: 2:45 PM
- Event Location: Hill 525
- Event Type: Mathematical Finance and Probability Seminars
- Extra Info: Speaker: Jose Alcala, New York University