Closed arcondello closed 6 years ago
For a csp with only the above constraint
In [15]: bqm = dwavebinarycsp.stitch(csp, max_graph_size=9)
In [16]: bqm
Out[16]: BinaryQuadraticModel({'BC3': 0.0, 'BC2': 0.0, 'BC0': 0.0, 'BC1': 0.0, 'AB3': 0.0, 'AB1': 0.0, 'AB2': 0.0, 'AB0': 0.0}, {('AB3', 'AB0'): 0.0, ('BC2', 'AB2'): 4.0, ('BC3', 'AB0'): 0.0, ('AB3', 'BC0'): 0.0, ('BC2', 'AB1'): 0.0, ('BC0', 'AB0'): 4.0, ('BC3', 'BC0'): 0.0, ('BC3', 'AB3'): 4.0, ('BC3', 'BC2'): 0.0, ('BC1', 'AB2'): 0.0, ('BC0', 'AB2'): 0.0, ('AB3', 'AB2'): 0.0, ('BC1', 'AB1'): 4.0, ('BC0', 'AB1'): 0.0, ('AB3', 'AB1'): 0.0, ('BC0', 'BC1'): 0.0, ('AB3', 'BC1'): 0.0, ('BC3', 'BC1'): 0.0, ('AB2', 'AB0'): 0.0, ('AB1', 'AB2'): 0.0, ('BC2', 'AB0'): 0.0, ('AB1', 'AB0'): 0.0, ('BC1', 'BC2'): 0.0, ('BC0', 'BC2'): 0.0, ('BC3', 'AB2'): 0.0, ('AB3', 'BC2'): 0.0, ('BC1', 'AB0'): 0.0, ('BC3', 'AB1'): 0.0}, -4.0, Vartype.BINARY)
This is a reducible constraint, but each of the four components can be mapped to two variables. So this should be able to be stitched to a bqm over 8 variables.