Skip to main content
U.S. flag

An official website of the United States government

Monitoring as a partially observable decision problem

Formally Refereed

Abstract

Monitoring is an important and costly activity in resource man-agement problems such as containing invasive species, protectingendangered species, preventing soil erosion, and regulating con-tracts for environmental services. Recent studies have viewedoptimal monitoring as a Partially Observable Markov Decision Pro-cess (POMDP), which provides a framework for sequential decisionmaking under stochastic resource dynamics and uncertainty aboutthe resource state. We present an overview of the POMDP frame-work and its applications to resource monitoring. We discuss theconcept of the information content provided by monitoring systemsand illustrate how information content affects optimal monitoringstrategies. Finally, we demonstrate how the timing of monitoring inrelation to resource treatment and transition can have substantialeffects on optimal monitoring strategies.

Keywords

Environmental monitoring, Dynamic programming, Partial observability, Value of information, POMDP

Citation

Fackler, Paul L.; Haight, Robert G. 2014. Monitoring as a partially observable decision problem. Resource and Energy Economics. 37: 226-241.
Citations
https://www.fs.usda.gov/research/treesearch/46099