Resolving structural uncertainty in natural resources management using POMDP approaches

摘要: In recent years there has been a growing focus on the uncertainties of natural resources management, and importance accounting for uncertainty in assessing management effectiveness. This paper focuses resource terms discrete-state Markov decision processes (MDP) under structural partial observability. It describes treatment with approaches developed partially observable systems. particular, I show how value iteration MDPs (POMDP) can be extended to structurally uncertain MDPs. A key difference between these process classes is that require tracking system state as well probability structure uncertainty, whereas POMDPs only observation uncertainty. The added complexity optimization problem compensated by reduced dimensionality search optimal strategy. solution algorithm outlined simple example conservation biology. By building conceptual framework POMDPs, analysts makers who confront take advantage rapid growth POMDP methods approaches, thereby produce better strategies over larger class problems.