What is the relationship between physical bit consumption and problem size when using EmbeddingComposite tools
I read the paper “Vicky Choi. Minor-embedding in adiabatic quantum computation: II. minor-universal graph design.“.
It says that you can use (n-3)/(d-2) physical qubits to achieve a n-bit problem's minor-embedding.
But when I try to use EmbeddingComposite tools to embed a 80 bits random MIS problem , sometimes it coasts more than that number of qubits.
ps:I achieve a 80 bits random MIS problem by the codes below
problem_node_count = 80
G = nx.random_geometric_graph(problem_node_count,radius=0.0063*problem_node_count,seed=1)
bqm = dimod.BQM.from_qubo(qubo)
And I use EmbeddingComposite(DWaveSampler()) as sampler