On optimal forest management: a dynamic programming approach

Abstract

This paper studies optimal any-aged forest management, including uneven-aged and even-aged management as special cases. The approach is based on applying a dynamic programming approach. Dynamic optimisation is applied to study optimal profit-maximising harvesting policies, based on decomposing the forest management problem to smaller subproblems. The optimal harvesting policy implies a modification of the optimal rotation period of the Faustman model to a discrete state model allowing for uneven-aged stands. When maximising the present value of revenue from harvesting, it is optimal to periodically harvest trees from a group of efficient size classes, determined by an optimal stopping problem, while maintaining sufficient ingrowth. Numerical examples with constant or non-linear transition functions suggest that dynamically optimal harvesting policies are robust to varying the initial conditions.