Performance-Driven Operating Regime Selection in Proactive Management of Cyber-Physical Systems

Proactive management of cyber-physical systems (CPS) aims to extend equipment lifetime by adjusting operating parameters based on remaining useful life (RUL) predictions. Existing methods optimize toward a target RUL increment, often driving the system toward conservative, low-speed regimes that reduce operational throughput. This paper addresses the complementary problem: given a minimum acceptable RUL threshold, find operating parameters that maximize a performance function reflecting equipment productivity. A genetic algorithm with a penalty-based fitness function carries out this constrained regime selection using a Weibull-based RUL model adapted to varying operating conditions. The performance function is a configurable component of the framework: any application-specific throughput model can be substituted without modifying the optimisation procedure, making the approach applicable across different classes of equipment. Validation on the XJTU-SY bearing dataset confirms the effectiveness of the proposed method.