Using stochastic dynamic programming for making decisions about invasive species

Invasive species present substantial ecological, economic, and public health issues. Deciding when, where, and how to intervene necessitates repeated decisions under uncertainty. While structured decision-making and adaptive management offer conceptual frameworks, their practical application is hindered by the perceived complexity, difficulty in application, and interpretability challenges of stochastic dynamic programming (SDP) methods.

This article offers a practical guide to employing SDP and its extensions for invasive species management. Coypu (*Myocastor coypus*) management serves as a case study, introducing the framework through a minimal working example. The guide then progressively expands the example to incorporate spatial dynamics, imperfect detection using partially observable Markov decision processes (POMDPs), adaptive management under model uncertainty, and value-of-information analyses to quantify the benefits of learning before taking action.

All examples are supported by reproducible R implementations, along with guidance on model formulation, parameterization, and interpretation. The authors also discuss the advantages and limitations of SDP. By reducing the technical barrier to SDP, this guide aims to enable ecologists and wildlife managers to integrate decision theory into invasive species management.

This will allow for the development of transparent, reproducible, and objective-driven management strategies under uncertainty. The methods discussed are not confined to invasive species management; they can be applied to various ecological decision problems where managing immediate actions against uncertain future outcomes is necessary.