Abstract:
Let X be a one-dimensional regular diffusion, $A$ a positive continuous additive functional of $X$, and h a measurable real-valued function. A method is proposed to determine a stopping rule $T^*$ that maximizes $\mathbf{E}\{e^{-A_T} h(X_T) 1_{\{T < \infty\}}\}$ over all stopping times $T$ of $X$. Several examples are discussed.
Keywords:diffusions, generalized parking problems, optimal stopping, random regret.