A Predictive Operating-System Control Layer for Quantum Computing Hardware and Execution Stacks

Quantum processors exhibit time-varying hardwaredrift, environmental fluctuations, and correlated control errorsthat challenge reliable execution and increase the physical errorburden presented to higher-level error-correction procedures.Existing approaches address important aspects of quantumcalibration, feedback control, scheduling, and error correction,but these functions are typically studied and implemented asdistinct components. We introduce QID-OS (Quantum Intelligent Dynamic Operating System), an operating-system-levelcontrol-plane architecture for quantum execution that mediatesbetween physical hardware telemetry and higher-level compilation, runtime, and quantum error-correction services. QID-OSintegrates causal recursive least-squares (RLS) online parameterestimation, predictive pulse-level compensation, and adaptiveexecution scheduling within a closed-loop runtime architecture.The proposed framework is evaluated using a distance-3 surfacecode simulation under a specified non-Markovian noise model, afull-factorial 23 module ablation, and experiments on a 5-qubitconnected subset of an IBM Quantum superconducting processor.In simulation, QID-OS improves logical success probability overthe evaluated noise conditions, and the full configuration achievesthe highest success probability among the evaluated modulecombinations. On superconducting hardware, the experimentallyvalidated online RLS estimation and active pulse-compensationcomponents increase computational-basis success probabilityfrom 68.5% ± 0.4% to 88.4% ± 0.3% across three independent8-hour calibration windows, comprising 300 randomized testcircuits and 2,457,600 total measurement shots. These resultsdemonstrate the feasibility of using a predictive control-planearchitecture to mitigate hardware-level errors prior to higherlevel quantum error-correction processing.