Event-Driven Optimal Control of Integral Continuous-Time Hybrid Automata

This paper proposes an event-driven optimal control strategy for hybrid systems composed of continuous-time integral dynamics with inputs and of Anite state machines triggered by endgenous and exogenous events. Endogenous events are caused by continuous states or continuous inputs crossing certain linear thresholds or by the elapse of time intervals, while exogenous events are forced by changes of binary and continuous inputs. The advantages of the proposed strategy are the reduction of the amount of computation required to solve optimal control problems, and the reduction of approximation errors typical of discrete-time approaches. We examine several performance objectives and constraints that lead to mixed-integerlinear or quadratic optimization problems and we exemplify the approach on a simple example.