Quantum Modeler · Custom QUBO

Model optimization problems as QUBOs. Validate small instances with quantum simulation.

Define a Binary Quadratic Model, inspect its Ising Hamiltonian, execute bounded QAOA, verify the original energy exactly, project larger models and export the circuit.

A QUBO (Quadratic Unconstrained Binary Optimization) uses binary variables and linear or pairwise interactions: E(x) = c + Σ aᵢxᵢ + Σ bᵢⱼxᵢxⱼ. Substituting xᵢ = (1 − Zᵢ)/2 gives the QAOA cost Hamiltonian.

API execution limit: 8 variables. Minimization only. Negate objective coefficients to convert maximization. Business constraints must already be encoded as penalty terms. An inadequate penalty can make an infeasible assignment optimal. Validate penalties independently: exhaustive verification checks the supplied bounded QUBO, not an unstated business constraint.

Define your QUBO

Presets replace problem inputs. Your execution and projection settings stay as entered. Edit any input, then build to calculate new results.

Import compiled model

Import preserves original coefficients, offsets and ordering. It does not execute. Run submits coefficients to the configured Modeler API and uses the normal run allowance. Custom QUBO verifies penalized energy, not domain feasibility.

Source formulation metadata

Binary objectiveReducing this count removes coefficients and interactions outside the new range.

Linear coefficients

All coefficient magnitudes ≤ 1,000,000. Very small differences may fall below numerical/search tolerances.
Sparse interactions

Each distinct pair appears once; reversed duplicates are rejected. Use linear coefficients for xᵢ² = xᵢ. Each pair is counted once; this is not a symmetric matrix with doubled off-diagonal entries.

Execution and projectionExample: 500 variables and 1200 interactions. Change these values to match your target.Supply both or leave both blank. At most N(N−1)/2 distinct target pairs. Projection is analysis only — the larger model is not executed.
API JSON for current editor

Use Copy API JSON to validate and refresh this view. Includes this browser's session ID when copied; omit it where monetization is disabled.

Use Copy API JSON to generate a request.

Ready to validate a small binary objective.

NOT CLAIMED

Modelling evidence, explicit limits

Ideal/noiseless simulation on classical hardware; p=1 with at most 64 parameter evaluations. No quantum advantage, speedup, scalability proof, physical QPU execution, fault-tolerant resource estimates, QAOA global optimality or production optimization quality is demonstrated. Larger targets are symbolic projections only.

Understand and audit this model

Mathematical methodology · Worked examples · Full DSL specification → · OpenQASM compatibility