A calculus of variations approach to the optimal control of discrete event dynamic systems

We present some early results on the control of discrete event dynamic systems (DEDS) using calculus of variations techniques. The idea is motivated by the observation that DEDS can be described by recursive equations of the same form as those used to describe conventional discrete-time continuous-variable dynamic systems. The calculus of variations is one of the few techniques that can handle nonlinearities such as the "max" and "min" operations commonly encountered in DEDS models. We apply the idea to a DEDS control problem in transportation systems and obtain a simple expression for the optimal control policy.