Event-Triggered Adaptive Optimal Control of Switched Nonlinear Systems With Arbitrary Assigned-Time Performance Constraints

Abstract

This paper investigates the event-triggered adaptive optimal control problem for switched nonlinear systems with arbitrary assigned-time performance constraints. An identifier-critic-actor-based reinforcement learning framework approximates the solution of the Hamilton-Jacobi-Bellman equation online without requiring precise knowledge of system dynamics. A mode-dependent event-triggered control strategy is developed to reduce communication load and handle the mismatch between the controller and system modes. A piecewise constant variable is introduced to compensate for triggering error jumps and exclude Zeno behavior. Furthermore, a novel performance index is constructed based on a transformation function to achieve optimal control under performance constraints. Finally, simulation results demonstrate that the proposed scheme guarantees boundedness of all closed-loop signals and achieves the desired tracking performance within an arbitrarily assigned time. Note to Practitioners—This work is motivated by performance-constrained control problems in switched nonlinear systems subject to communication constraints, which commonly arise in industrial processes such as continuous stirred tank reactors. In practical applications, operating-condition variations may induce mode transitions, while stringent transient tracking performance must still be maintained under limited communication resources. To address these challenges, this paper develops a reinforcement learning-based event-triggered adaptive optimal control scheme for switched nonlinear systems. The proposed method learns the optimal control policy online without requiring precise system models and reduces the communication burden. A key practical advantage is that the proposed method removes the restrictive initial-condition requirement in conventional prescribed performance control, guaranteeing that the tracking error enters and remains within a predefined region in an arbitrarily assigned time. Furthermore, a mode-dependent event-triggered mechanism is developed to resolve controller-subsystem mismatches caused by asynchronous switching and triggering processes. The proposed framework can also be extended to other networked control systems with switching behaviors and communication constraints, including smart grids and autonomous vehicles.