Which predictive uncertainty to resolve? Value of information sensitivity analysis for environmental decision models
Fridolin Haag
Sara Miñarro
Arjun Chennu
Abstract
Uncertainties in environmental decisions are large, but resolving them is costly. We provide a framework for value of information (VoI) analysis to identify key predictive uncertainties in a decision model. The approach addresses characteristics that complicate this analysis in environmental management: dependencies in the probability distributions of predictions, trade-offs between multiple objectives, and divergent stakeholder perspectives. For a coral reef fisheries case, we predict ecosystem and fisheries trajectories given different management alternatives with an agent-based model. We evaluate the uncertain predictions with preference models based on utility theory to find optimal alternatives for stakeholders. Using the expected value of partially perfect information (EVPPI), we measure how relevant resolving uncertainty for various decision attributes is. The VoI depends on the stakeholder preferences, but not directly on the width of an attribute’s probability distribution. Our approach helps reduce costs in structured decision-making processes by prioritizing data collection efforts.
Haag F., Miñarro S. and Chennu A. (2022) Which predictive uncertainty to resolve? Value of information sensitivity analysis for environmental decision models. Environmental Modelling & Software 158: 105552. 10.1016/j.envsoft.2022.105552
@article{Haag2022,
Title = {Which predictive uncertainty to resolve? Value of information sensitivity analysis for environmental decision models},
Author = {Haag, Fridolin and Miñarro, Sara and Chennu, Arjun},
Editor = {},
Journal = {Environmental Modelling & Software},
Year = {2022},
Pages = {105552},
Volume = {158},
Doi = {10.1016/j.envsoft.2022.105552},
Abstract = {Uncertainties in environmental decisions are large, but resolving them is costly. We provide a framework for value of information (VoI) analysis to identify key predictive uncertainties in a decision model. The approach addresses characteristics that complicate this analysis in environmental management: dependencies in the probability distributions of predictions, trade-offs between multiple objectives, and divergent stakeholder perspectives. For a coral reef fisheries case, we predict ecosystem and fisheries trajectories given different management alternatives with an agent-based model. We evaluate the uncertain predictions with preference models based on utility theory to find optimal alternatives for stakeholders. Using the expected value of partially perfect information (EVPPI), we measure how relevant resolving uncertainty for various decision attributes is. The VoI depends on the stakeholder preferences, but not directly on the width of an attribute’s probability distribution. Our approach helps reduce costs in structured decision-making processes by prioritizing data collection efforts.},
}
TY - JOUR
AU - Haag, Fridolin
AU - Miñarro, Sara
AU - Chennu, Arjun
TI - Which predictive uncertainty to resolve? Value of information sensitivity analysis for environmental decision models
T2 - Environmental Modelling & Software
PY - 2022
SP - 105552
VL - 158
DO - 10.1016/j.envsoft.2022.105552
AB - Uncertainties in environmental decisions are large, but resolving them is costly. We provide a framework for value of information (VoI) analysis to identify key predictive uncertainties in a decision model. The approach addresses characteristics that complicate this analysis in environmental management: dependencies in the probability distributions of predictions, trade-offs between multiple objectives, and divergent stakeholder perspectives. For a coral reef fisheries case, we predict ecosystem and fisheries trajectories given different management alternatives with an agent-based model. We evaluate the uncertain predictions with preference models based on utility theory to find optimal alternatives for stakeholders. Using the expected value of partially perfect information (EVPPI), we measure how relevant resolving uncertainty for various decision attributes is. The VoI depends on the stakeholder preferences, but not directly on the width of an attribute’s probability distribution. Our approach helps reduce costs in structured decision-making processes by prioritizing data collection efforts.
ER -
