Abstract
A quantum thermodynamic system is described by a Hamiltonian and a list ofconserved, non-commuting charges, and a fundamental goal is to determine theminimum energy of the system subject to constraints on the charges. Recently,[Liu et al., arXiv:2505.04514] proposed first- and second-order classical andhybrid quantum-classical algorithms for solving a dual chemical potentialmaximization problem, and they proved that these algorithms converge to globaloptima by means of gradient-ascent approaches. In this paper, we benchmarkthese algorithms on several problems of interest in thermodynamics, includingone- and two-dimensional quantum Heisenberg models with nearest andnext-to-nearest neighbor interactions and with the charges set to the total$x$, $y$, and $z$ magnetizations. We also offer an alternative compellinginterpretation of these algorithms as methods for designing ground and thermalstates of controllable Hamiltonians, with potential applications in molecularand material design. Furthermore, we introduce stabilizer thermodynamic systemsas thermodynamic systems based on stabilizer codes, with the Hamiltonianconstructed from a given code's stabilizer operators and the chargesconstructed from the code's logical operators. We benchmark the aforementionedalgorithms on several examples of stabilizer thermodynamic systems, includingthose constructed from the one-to-three-qubit repetition code, the perfectone-to-five-qubit code, and the two-to-four-qubit error-detecting code.Finally, we observe that the aforementioned hybrid quantum-classicalalgorithms, when applied to stabilizer thermodynamic systems, can serve asalternative methods for encoding qubits into stabilizer codes at a fixedtemperature, and we provide an effective method for warm-starting theseencoding algorithms whenever a single qubit is encoded into multiple physicalqubits.