Title
Parameter Estimation Algorithms for Computationally Intensive Spatial Problems
Speaker
Christopher H. Holloman, Statistical Consulting Service, The Ohio State University
Abstract
We propose a multi-resolution genetic algorithm that allows efficient estimation of parameters in large-dimensional models. Such models typically arise in modeling spatial phenomena. Fitting these models often requires the use of complex numerical methods and large amounts of computing power. Unfortunately, the numerical maximization and sampling techniques used to fit such complex models often explore the parameter space slowly resulting in unreliable estimates. Our algorithm improves this exploration by incorporating elements of simulated tempering into a genetic algorithm framework for maximization. Our algorithm can also be adapted to perform Markov chain Monte Carlo sampling from a posterior distribution in a Bayesian setting, which can greatly improve mixing and exploration of the posterior compared to ordinary MCMC methods. The proposed algorithm can be used to estimate parameters in any model where the solution can be solved on different scales, even if the data are not inherently multi-scale. We address parallel implementation of the algorithms and demonstrate their use on examples from groundwater hydrology.
Meet the speaker in Room 212 Cockins Hall at 4:30 p.m. Refreshments will be served.
