Environmental and Ecological Statistics

, Volume 8, Issue 3, pp 269-288

First online:

Uncertainty, learning, and the optimal management of wildlife

  • Byron K. WilliamsAffiliated withBiological Resources Division, United States Geological Survey

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access


Wildlife management is limited by uncontrolled and often unrecognized environmental variation, by limited capabilities to observe and control animal populations, and by a lack of understanding about the biological processes driving population dynamics. In this paper I describe a comprehensive framework for management that includes multiple models and likelihood values to account for structural uncertainty, along with stochastic factors to account for environmental variation, random sampling, and partial controllability. Adaptive optimization is developed in terms of the optimal control of incompletely understood populations, with the expected value of perfect information measuring the potential for improving control through learning. The framework for optimal adaptive control is generalized by including partial observability and non-adaptive, sample-based updating of model likelihoods. Passive adaptive management is derived as a special case of constrained adaptive optimization, representing a potentially efficient suboptimal alternative that nonetheless accounts for structural uncertainty.

adaptive resource management Bayes’ theorem dynamic model optimal control wildlife populations