EPB and IonQ post three preprints on quantum-assisted power grid switching
Municipal utility EPB and IonQ published three preprints (not yet peer reviewed) on hybrid quantum-classical methods for distribution network reconfiguration, which selects switch states to cut line losses while keeping the grid radial. The work was funded by a $4 million NIST grant and run by EPB's Quantum Computing Fellows on an on-premises IonQ Forte Enterprise trapped-ion system at the EPB Quantum Center in Chattanooga, Tennessee. In the reinforcement learning approach, a quantum circuit cut parameters by 95.8% on the IEEE 33-bus system (9,156 vs. 220,052 classically), and hardware runs reached the exact optimum (0.1396 MW loss) for 33 buses and within 0.30% for 69 buses. A guided solver approach cut median Gurobi runtime from 1,415.47 s to 34.71 s on a 417-bus network using 29 physical qubits, while a third method, iterative quantum imaginary-time evolution, recovered global minimal-loss configurations on six benchmark networks and came within 0.2% to 2.18% of published reference losses on larger ones.