Computational Many-Body QM

A numerical output is not yet a physical conclusion. Even an exactly diagonalized matrix answers only the declared finite model, basis, sector, boundary condition, and arithmetic problem. A defensible many-body claim also needs an observable-specific method, separated error sources, independent tests, and control of the physical limit being inferred.

This chapter is the problem-to-algorithm selection and numerical-evidence audit gateway. Computational Many-Body Overview owns the detailed method comparison and evidence framework. The specialist leaves own the algorithms and diagnostics. This gateway identifies the shortest sound route and the checks required before a result can support a claim.

Required background. Scaling of Hilbert Space supplies the basis-growth problem. Math Needed for Computational QM supplies numerical linear algebra, floating-point, integration, and probability preparation.

Helpful background. Many-Body Hilbert Spaces and Operators prepares sectors, bases, and operator construction. Finite-Size Effects separates physical size mechanisms from numerical error. Boundaries with Neighboring Subjects distinguishes the physics developed here from reusable algorithms in the Computational QM volume.

Start with a numerical-claim ledger

Use this contract:

finite problem + state or ensemble + target observable + representation and sector + structural obstruction + physical limit and resolution + algorithm and estimator + resource model + convergence controls + uncertainty decomposition + independent benchmark + provenance → bounded physical claim.

Record these entries before choosing software.

1. Problem. State the Hamiltonian or generator, parameters and units, degrees of freedom, statistics, constraints, geometry, dimension, boundaries, and finite family.
2. State. Specify the ground state, targeted excited state, thermal ensemble, initial density matrix, driven protocol, or other object to be computed.
3. Observable. Name the energy, gap, correlator, response, order parameter, entanglement quantity, real-time signal, or spectrum, including normalization and requested resolution.
4. Representation. Declare basis, mode ordering, symmetry labels, sector, truncation, real- or momentum-space convention, and any transformation of the weights or Hamiltonian.
5. Obstruction. Identify what limits the calculation: dimension, memory, sparsity, near degeneracy, sign or phase cancellations, entanglement growth, autocorrelation, inverse reconstruction, long time, or finite size.
6. Physical limit. State whether the claim is finite-system, continuum, zero-temperature, long-time, thermodynamic, critical, or some ordered combination of limits.
7. Method. Name the algorithm, estimator, initialization, update or optimization scheme, stopping rule, and why its assumptions match the obstruction and observable.
8. Resources. Estimate memory, operation count, sample count, wall time, precision, storage, and the largest controlled system before production work begins.
9. Controls. Vary basis cutoff, Krylov dimension, residual, bond dimension, sweep count, time step, evolution time, sample count, equilibration, bin size, broadening, size, and boundaries as applicable.
10. Uncertainty. Keep model, discretization, truncation, solver, floating-point, statistical, autocorrelation, finite-size, fitting, and continuation errors separate. Not every method has every error.
11. Validation. Use exact limits, identities, Hermiticity and symmetry checks, residuals, variances, sum rules, independent implementations, cross-method overlap, or a registered benchmark.
12. Provenance. Record code and data versions, environment, parameters, seeds, sector and basis conventions, raw outputs, analysis scripts, and failed or excluded runs.

Read as a dependency graph

The sidebar is a catalog, not one mandatory chain.

1. Take the common first pass. Read Computational Many-Body Overview for representation choice, method families, error classes, validation, and finite-to-bulk inference.
2. Build a transparent finite baseline. Exact Diagonalization Preview constructs the basis and finite operator. Symmetry Sectors in Many-Body Numerics then reduces compatible blocks and explains what must still be reconstructed across sectors.
3. Target part of a sparse spectrum. After the finite operator is controlled, Lanczos Method Preview develops Krylov projection, Ritz values, residuals, reconstruction, and continued fractions. Symmetry reduction is useful but not universal.
4. Infer a large-system law only through a controlled family. Finite-Size Scaling in Numerics is a cross-cutting inference branch after any solver that supplies comparable sizes; it is not evidence supplied automatically by that solver.
5. Sample an equilibrium representation. Quantum Monte Carlo Preview develops configurations, estimators, equilibration, autocorrelation, and systematic limits. Pair it with Sign Problem Preview when weights cancel.
6. Compress low-entanglement structure. Enter the entanglement chapter’s conceptual tensor-network pages, then use Tensor Networks: Computational Guide for family selection and convergence design and DMRG Preview for finite-system MPS optimization.
7. Compute frequency-resolved observables. First declare the canonical correlator in the Correlations chapter. Then Dynamical Correlation Functions Numerically compares Lehmann sums, continued fractions, and time evolution with matched broadening and resolution.
8. Validate every branch. Benchmark Problems supplies registered reference contracts. Reproducible Notebooks owns promotion gates, execution records, and the status-aware notebook inventory. They are validation spines, not optional end matter.

Choose a shorter route

Establish a finite ground-state baseline. Read Computational Overview → Exact Diagonalization → optional Symmetry Sectors. Add Lanczos when only a few extremal states are needed. Stop when basis closure, sector coverage, residuals, degeneracies, and finite-system observables are independently checked.

Test a thermodynamic or critical claim. Choose a solver appropriate to the model, obtain several comparable sizes and boundaries, then use Finite-Size Scaling. Stop only after numerical error is smaller than the resolved size drift and competing scaling forms or crossover explanations have been tested.

Study finite-temperature equilibrium by sampling. Read Quantum Monte Carlo and diagnose the Sign Problem when applicable; add Finite-Size Scaling for a bulk claim. Stop when equilibration, integrated autocorrelation, estimator bias, representation dependence, average sign, and systematic discretization limits are reported.

Study a one-dimensional low-entanglement state. Read the conceptual Tensor Networks and Matrix Product States pages → Tensor Networks: Computational Guide → DMRG. Add finite-size or finite-entanglement scaling as required. Stop when energy, variance or residual, bond dimension, discarded information, sweep history, boundaries, initialization dependence, and target observables agree within the declared budget.

Compute a dynamical spectrum. Read Correlation Functions and Linear Response → Dynamical Correlation Functions Numerically. Choose direct Lehmann, Lanczos, or time evolution according to size, state, and resolution. Stop when conventions, sum rules, finite-time or recurrence effects, broadening kernels, and cross-route checks are explicit.

Release a result. Run the relevant Benchmark Problem before the production model, then satisfy the Reproducible Notebooks contract. A plot without its model record, raw data, environment, and validation trail is not a reproducible result.

Worked method-selection audit

Suppose the target is the lowest gap of a periodic spin chain and its large-size behavior. First declare the Hamiltonian sign convention, chain lengths, periodic boundary, spin sector, momentum and parity labels, and whether the target is the lowest gap overall or within one sector. Exact diagonalization on the smallest chains tests basis construction, symmetry labels, degeneracies, and operator conventions. Sector-resolved Lanczos can extend the targeted eigenvalues, but a small Ritz residual supports a nearby eigenvalue—not every eigenvector-derived observable. Near degeneracy requires an invariant-subspace audit. The thermodynamic claim then requires a consistent sequence of sizes and sectors, boundary and commensurability audits, competing scaling forms, and overlap with exact limits or another method. Solver convergence and large-size inference remain separate arguments.

Exit checkpoint

You are ready to leave this gateway when you can:

• translate a physics question into a finite computational contract with a declared observable and limit;
• estimate the state-space and resource bottlenecks before selecting a method;
• choose ED, Lanczos, QMC, tensor networks, DMRG, or a dynamical route for structural reasons rather than familiarity;
• distinguish a complete finite solution, a targeted approximation, a stochastic estimator, a variational result, and an extrapolated bulk claim;
• design observable-specific convergence tests and keep numerical, statistical, finite-size, and modeling uncertainties separate;
• require an independent benchmark, cross-method or exact-limit check, and a reproducible record;
• route reusable algorithm and implementation work onward without moving the many-body physics claim out of its canonical home.

Canonical boundaries

• Computational Many-Body Overview owns the detailed method comparison, evidence pipeline, and error taxonomy. This gateway owns entry routing, dependency selection, readiness and exit checks, and the claim ledger.
• Many-Body Hilbert Spaces and Operators owns sectors, bases, and operator forms. Lattice Models owns model declarations; Correlations owns the exact observables; Entanglement owns tensor-network and MPS concepts.
• Exact Diagonalization, Lanczos, Quantum Monte Carlo, the Sign Problem, Tensor Networks, DMRG, numerical dynamics, finite-size scaling, benchmarks, and notebook governance retain their specialist derivations and protocols in the leaves.
• Finite-Size Effects owns physical size mechanisms, while Finite-Size Scaling in Numerics owns extrapolation practice. Phases owns phase and critical claims; Nonequilibrium Dynamics owns dynamical interpretation.
• Computational Quantum Matter owns material-specific method and evidence routing across electronic-structure, localized-representation, correlated, response, pairing, and disorder questions; this chapter retains generic many-body method selection and derivations.
• This chapter owns many-body method selection and evidence. The Computational QM volume owns reusable algorithms, software architecture, performance engineering, and implementation details; the Reference owns shared validation and environment records.

Common routing errors

“Exact diagonalization gives the exact physical answer.” It is exact only for the declared finite model, basis or cutoff, sector coverage, arithmetic, and tolerance. The thermodynamic or continuum inference is separate.

“A tiny residual certifies the eigenvector and every observable.” A residual strongly constrains a nearby eigenvalue; individual eigenvectors require spectral isolation, while near degeneracy calls for an invariant-subspace or projector statement.

“A converged variational energy certifies the phase.” Energy convergence does not by itself certify fidelity, long-distance correlations, topology, excitation structure, or freedom from metastability.

“The Monte Carlo error bar is the total uncertainty.” It normally measures sampling uncertainty after an autocorrelation analysis. Equilibration, discretization, estimator, finite-size, continuation, and sign-related biases remain separate.

“The sign problem is an intrinsic label on a Hamiltonian.” Its severity depends on basis, representation, decomposition, parameters, and algorithm. A sign-free formulation in one setting is not a universal cure.

“A discarded weight or bond dimension is an error bar.” It is a convergence diagnostic. Observable errors, finite-size effects, initialization, symmetry enforcement, and entanglement structure still require direct tests.

“One successful fit proves the thermodynamic law.” Size range, boundaries, corrections, covariance, competing forms, and stability under omitted sizes must be audited.

“Numerical broadening is a physical linewidth.” Finite time, recurrence, kernel choice, level spacing, truncation, and inverse reconstruction can create width without a physical decay rate.

“A planned notebook is a reproduced result.” A filename or catalog entry is not a committed artifact, successful execution, validation record, or independent reproduction.

Exercises

Exercise 1: Route three targets for one Hamiltonian

For a spin-1/2 chain, route (a) the complete spectrum at ten sites, (b) the ground-state energy density at much larger size in a favorable one-dimensional regime, and (c) a finite-temperature structure factor when the sampling weights are nonnegative. State one mandatory validation test for each route.

Solution

For (a), use Exact Diagonalization with Symmetry Sectors and validate dimension, Hermiticity, trace or moment identities, and reconstruction across sectors. For (b), use the Tensor-Networks guide and DMRG, then a size and bond-dimension convergence study; compare the smallest systems with ED. For (c), use Quantum Monte Carlo, verify equilibration and integrated autocorrelation, test the estimator on a benchmark, and add finite-size scaling for a bulk claim. The same Hamiltonian does not imply the same method because the state, observable, size, and error structure differ.

Exercise 2: Repair an overclaim

Repair this statement: “The Lanczos residual is below $10^{-12}$, the DMRG discarded weight is small, and the Monte Carlo standard error is tiny, so all three methods prove the same thermodynamic phase.”

Solution

The three quantities diagnose different numerical layers. The Lanczos residual supports a finite eigenvalue or invariant-subspace claim, subject to sector and spectral-isolation checks. Discarded weight supports a bond-truncation convergence study but does not alone bound every observable or exclude metastability. A Monte Carlo standard error addresses correlated sampling after equilibration, not all biases. A phase claim additionally needs a common observable definition, cross-method agreement on overlapping finite systems, controlled size and boundary sequences, a phase diagnostic, uncertainty propagation, competing-explanation tests, and the declared thermodynamic limit.

References

• A. W. Sandvik, “Computational Studies of Quantum Spin Systems,” in AIP Conference Proceedings 1297, 135–338 (2010).
• U. Schollwöck, “The Density-Matrix Renormalization Group in the Age of Matrix Product States,” Annals of Physics 326, 96–192 (2011).
• H. Fehske, R. Schneider, and A. Weiße, eds., Computational Many-Particle Physics, Springer (2008).
• M. E. J. Newman and G. T. Barkema, Monte Carlo Methods in Statistical Physics, Oxford University Press (1999).
• Y. Saad, Numerical Methods for Large Eigenvalue Problems, 2nd ed., SIAM (2011).