Kevin Ross
A new approach to solving singular stochastic control problems
with applications to investment and stochastic networks
Beginning with a class of singular stochastic control problems
that can be transformed to optimal stopping problems, we use the
equivalence with optimal stopping to develop a convergent and
computationally efficient backward induction algorithm for
approximating the value function and an optimal control policy. We then
use the method of finite differences to modify the backward induction
algorithm for much more general singular stochastic control problems,
including those that arise in applications in finite-horizon optimal
investment and consumption and in control of stochastic networks in two
dimensions. Joint work with Tze Leung Lai and Tiong Wee Lim.