Skip to content

Quantum Cloud Services for Optimisation Problems

In this scenario, we want to show how our architecture behaves and interacts using a NP-hard combinatorial optimisation problem. Since the System and Physical Layer are almost identical to the first scenario, we only document the new building blocks.

It has the following specification:

  • Input: Max-Cut instance
  • Output: A cut consisting of a set of edges
  • Used in:
    • Flight-gate assignment in airports
    • Electronic Design Automation
    • Trajectory optimization in air traffic

Application Layer - Downwards

  1. Cloud Optimisation Application:

    • Processes the MaxCut problem instance and forwards the problem information to the QUBO Transformator

  2. QUBO Transformator

    • Creates a QUBO, using the information of the Cloud Optimisation Application

  3. Ising Transformator

    • Transforms the QUBO into an Ising Hamiltonian

  4. Penalty Encoder

    • Checks if the required optimization problem needs a penalty factor and adjusts it.

  5. QAOA

    • QAOA is a hybrid algorithm which uses tunable parameters.
    • It encodes the Ising Hamiltonian into a quantum circuit

  6. QAOA Program Generator

    • With the parametrised circuit the program generator creates a program for the transpilation pipeline in the system layer.

System Layer - Downwards

  1. Noise Suppression
    • Noise Supression reduces the error rate by actively canceling noise using adjusted pulse timing or add redundancy to detect and reduce errors.
    • It takes an optimized circuit and adds the above mentioned operations.

Physical Layer & System Layer - Upwards

Refer to the quantum simulation scenario for a full explanation.

Application Layer - Upwards

  1. SPSA
    • SPSA is classical optimizer used for optimizing the parameters of the QAOA algorithm.
    • It updates the parameters and sends them to the QAOA Program Generator, where the program is computed again until the optimizer terminates.