Uncertain measurements of renewable resources: approximations, harvesting policies and value of accuracy

This paper deals with the management of renewable resources subject to measurement error. The method, stochastic optimization in policy space (SOPS), allows for more general solutions to problems that are not LQ than earlier attempts. Measurement errors are found to be more important for quota policies than natural stochastic variation. With little error, filtering (in time or in policy space) is important. With considerable error caution becomes important. Various approximations are compared. Policies based simply on historical measurements are found to give higher expected values than policies based on advanced Kalman filter estimates. Values of measurement accuracy are estimated.