arXiv:2609.39602
Warm-Start Iterative QITE for Distribution Network Reconfiguration via Branch-Exchange Encoding
Noah Crum, Pete Pritchard, Tanner Rase, Phillip C. Lotshaw, Shaked Regev, Claudio Girotto, Ananth Kaushik, Paul Smith
Abstract
We present a hybrid quantum-classical algorithm for distribution network reconfiguration, a combinatorial optimization problem on power distribution networks, that combines a radiality preserving branch-exchange encoding with iterative warm-start quantum imaginary-time evolution to minimize active line losses. The encoding uses a fixed-width binary representation of sequential branch-exchange actions, ensuring that every register outcome decodes to a radial configuration. The quantum subroutine is informed by a surrogate model fit to classically-solved alternating current power flow (ACPF) labels. After a set budget of ACPF solves and circuit samples is exhausted, the algorithm returns the lowest-loss configuration seen. We employ our algorithm on ten test cases, including seven commonly-used benchmark cases and three cases we modified from these and similar standard systems. We include two methods to reduce problem size, which enable extensions of the algorithm to larger networks. We successfully reach minimal-loss configurations for six cases and find configurations with losses from 0.2 percent to 2.18 percent above the reported reference minimal loss for the other four.
AI Summary
The paper addresses loss-minimizing distribution network reconfiguration using a hybrid quantum–classical search. A fixed-width binary encoding represents sequential branch exchanges: closing a tie creates a cycle, and opening a controlled edge restores radiality. Every bit string decodes to a radial configuration; coverage of all radial configurations is established under stated assumptions on available switches and register widths. Multiple strings can represent the same configuration. Classical alternating-current power-flow (ACPF) evaluations supply labels for a second-order surrogate over branch-exchange actions, with fitting and candidate ranking accounting for repeated representations. Each acquisition round prepares an incumbent-centered product state and applies a pairwise QITE-inspired circuit whose parameters are calculated classically by first-order expectation matching. Measurements propose configurations for further ACPF evaluation. Switch-fixing and rolling windows restrict register sizes for larger networks. Tests cover ten cases with 14–415 buses, using 100 ideal simulations, 25 noisy simulations, and one hardware run per case. Hardware reaches exhaustively verified minimum losses in Cases 1–5 using 11–26 qubits, first finding optima after 3–369 distinct ACPF evaluations. Rolling-window searches reproduce the published reference optimum for the 84-bus case; the other four hardware results are approximately $0.07\%$–$2.18\%$ above their reference losses. The study does not compare against classical optimization baselines or claim quantum advantage. Radiality is enforced, but additional operational constraints are not addressed. Register restrictions can exclude global optima, and finite QITE updates guarantee neither surrogate-energy nor evaluated-loss descent. Single hardware runs limit conclusions about repeatability.
Error found: The abstract incorrectly reports a 0.2%–2.18% excess-loss range: the reported hardware losses imply approximately 0.07%–2.18%, because Case 9 is only 0.0724% above its reference. The acquisition algorithm caps usable ACPF labels rather than all ACPF solve attempts, so its parameter N does not guarantee the hard solve budget described in the abstract when evaluations fail or return nonfinite losses.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 23 pgs, 11 figures