Advanced techniques with Aqarios Constrained Quantum Optimizer

Use the following techniques to get the most out of the Aqarios Constrained Quantum Optimizer. For every option and its default value, see the API reference.

---

Trade solution quality against runtime

The number of chains (num_parallel), the total shot budget per chain (total_shots), and the number of QAOA repetitions (reps) are the primary controls:

• Increase num_parallel to improve the probability of finding high-quality solutions. This raises the total shot budget.
• Increase total_shots to run more iterations for each chain. Use shots to set the number of shots for each iteration.
• Increase reps for higher solution quality, at the cost of deeper circuits that can add noise.
• Lower approximation_degree to reduce circuit depth through approximations, which can affect solution quality.

---

Control postprocessing and penalties

• Set postprocessing_level to an integer from 0 (off) to 3 (greedy local search, the default). Higher levels improve solution quality at the cost of classical compute time.
• The optimizer derives penalty strengths automatically. If the results are infeasible, set penalty_override to a larger value.
• Use beta and epsilon to control how strongly each new warm-start state concentrates on low-energy samples.

---

Reuse a session and tag jobs

• Set use_session to True to keep a dedicated connection to the QPU open across iterations, which can lower wall-clock time.
• Set seed_transpiler to make circuit synthesis and routing reproducible across runs.
• Use job_tags to attach your own tags to every Qiskit Runtime job that the function submits.

---

Inspect the result

• feasible reports whether the solutions satisfy all constraints. Compare raw_energy with objective_value to diagnose constraint violations.
• metadata.circuit_metrics reports circuit sizes before and after transpilation.
• metadata.resource_usage breaks down QPU and CPU time by phase.