A stochastic dynamic programming framework for weed control decision making: an application to Avena fatua L.

This paper develops a stochastic multi-period decision model to analyse a continuous wheat cropping system infested by wild oats (Avena fatua L.), in southern Australia. The multi-period solutions is obtained by employing a dynamic programming model in conjunction with a bioeconomic simulation model. An empirically estimated dose response function is used to derive the optimal herbicide rate. Uncertainties due to environmental effects on the performance of herbicide and crop yields are modelled and optimal decision rules derived. The results indicate that substantial economic gains can be realised if herbicide dose decisions are taken by considering future profit effects of current decisions, as opposed to the more common approach of only considering the current-period effect.