Could architectural changes enforcing dimensional analysis or symmetry constraints reduce the gap on research-scale problems?

Your question asks whether building in dimensional analysis or symmetry constraints *at the architectural level* could help quantum solvers close the scaling gap on real research problems.

The preprints do address symmetry and structural constraints, but mostly as post-hoc optimizations rather than architectural first principles. Let me unpack what's available:

Symmetry-aware bounds, not architecture

The most direct symmetry result comes from quantum phase estimation. Researchers found that by respecting the symmetries already present in chemical Hamiltonians—rather than using generic 1-norm bounds—you can compute tighter spectral bounds [AG-2024.03-1561]. This matters because the spectral gap drives the cost of quantum algorithms. However, this is a *tuneable rescaling* applied after the algorithm is chosen, not a constraint baked into the hardware design itself.

Structural embedding as workaround

Several papers attack connectivity constraints (a topology problem) by embedding the logical problem onto sparse hardware:

• A scalable 2-local architecture uses chains of physical qubits to encode logical variables by decomposing the problem graph into triangles [AG-2024.03-1725]. This respects the hardware's geometric limits but requires extra qubits.
• Parity-constraint embeddings extend this idea modularly to map arbitrary Ising Hamiltonians onto existing annealers while preserving the original Hamiltonian's properties [AG-2024.05-2370].

These help, but they're *solutions to the connectivity problem*, not architectural changes that prevent the problem from arising.

The scaling gap itself

The real bottleneck is the divide-and-conquer approach: distributed quantum systems struggle because inter-device communication dominates cost [AG-2024.05-1606]. Even hybrid quantum–classical solvers on state-of-the-art hardware (thousands of qubits) show mixed results compared to classical optimizers like CPLEX [AG-2024.03-1860].

---

Bottom line: The archive shows how to exploit symmetries *after* problem formulation and how to work around hardware topology, but doesn't discuss designing quantum-annealer or gate-based architectures from first principles around dimensional-analysis or symmetry constraints to reduce the research-scale gap.