edge_driver

edge_driver(graph, reward, *, n_qubits=None)

Edge-driver Hamiltonian rewarding edges whose endpoints match reward.

Mirrors the formulations in pennylane.qaoa.cost.edge_driver:

• reward ⊆ {"00","01","10","11"}; "01" and "10" are treated as the undirected pair (must appear together or not at all).

• len(reward) ∈ {0, 4} reduces to a constant identity term.

• Otherwise the difference between rewarded and penalised energies is always 1.

The graph nodes must be integer qubit indices. n_qubits may be provided to preserve isolated trailing qubits.

Return type:

SparsePauliOp