On optimal control of reflected diffusions

Abstract

We study a simple singular control problem for a Brownian motion with constant drift and variance reflected at the origin. Exerting control pushes the process towards the origin and generates a concave increasing state-dependent yield which is discounted at a fixed rate. The most interesting feature of the problem is that its solution can be more complicated than anticipated. Indeed, for some parameter values, the optimal policy involves two reflecting barriers and one repelling boundary, the action region being the union of two disjoint intervals. We also show that the apparent anomaly can be understood as involving a switch between two strategies with different risk profiles: The risk-neutral decision maker initially gambles on the more risky strategy, but lowers risk if this strategy underperforms.