pylibmgm.ParallelGenerator

class pylibmgm.ParallelGenerator

Parallel solution generator for multi-graph matching.

Constructs a solution by matching graphs in parallel in a tree structure, starting with pairs of graphs as the leaves. Can be faster for large problems, may produce different results than sequential generation. Showed better results in experiments than sequential generator.

Methods

__init__(model)	Initialize parallel generator.
generate()	Generate a complete solution using parallel matching.
init(order)	Initialize the generator with a matching order.

__init__(model: pylibmgm.MgmModel) → None

Initialize parallel generator.

Parameters:

model (MgmModel) – The MGM model to solve.

generate() → pylibmgm.MgmSolution

Generate a complete solution using parallel matching.

Returns:

Generated cycle-consistent solution.

Return type:

MgmSolution

init(order: pylibmgm.MgmGenerator.matching_order) → List[int]

Initialize the generator with a matching order.

Parameters:

order (MgmGenerator.matching_order) – Order to match graphs (sequential or random).

Returns:

Sequence of graph IDs to be matched.

Return type:

list[int]