Skip to main content Skip to secondary navigation
Tyler Hall
Main content start
Ph.D. Student

Tyler Hall

Ph.D. Student in Geological Sciences

TomKat Graduate Fellow for Translational Research

Research Lab: Jef Caers

Year Awarded: 2021

Tyler Hall is a PhD Candidate in the Geological Sciences Department, advised by Dr. Jef Caers in the Stanford Center for Earth Resources Forecasting. His research focuses on sequential decision-making algorithms for subsurface exploration, as part of a collaboration with the Stanford Intelligent Systems Laboratory. Tyler has five years of work experience with major mining companies Freeport-McMoRan and Glencore in production and exploration geology as well as strategic business development. Tyler earned his bachelor’s degree in Geological Sciences from Hartwick College.

Google Scholar

LinkedIn

Optimal sequential decision-making for subsurface exploration

Global development of renewable energy technology and infrastructure necessitates exploration for geologically-controlled subsurface resources and storage reservoirs. In this research, we study data scientific techniques, decision theory, and algorithms for sequential decision-making commonly applied in autonomous navigation to improve exploration and uncertainty reduction in the subsurface. The very nature of exploration is sequential – an explorer updates their belief of the subsurface as they collect information. Therefore, it is reasonable to investigate quantitative methods that emulate sequential exploration.

For a given dataset, we determine an optimal sequence of exploration decisions: where to collect information, what kind of information to collect, and its extent. We demonstrate the effect of various parameters on the optimal exploration sequence, such as risk preference, cost of data collection, and target resource volume. The systematic framework reduces bias and illuminates exploration decisions or sensitivities that otherwise may not be obvious to decision-makers.

The initial application for mineral exploration extends to other subsurface resources and storage reservoirs due to similarities in geological controls, data types, and exploration methodologies. Improving the efficacy of exploration and uncertainty reduction allows us to understand and engineer the subsurface in a more sustainable and economic manner.