In artificial intelligence and operations research, hierarchical constraint satisfaction (HCS) is a method of handling constraint satisfaction problems where the variables have large domains by exploiting their internal structure.
[1] For many real-world problems the domain elements cluster together into sets with common properties and relations.
This structure can be represented as a hierarchy and is partially ordered on the subset of a relation.
The expectation is that the domains are structured so that the elements of a set frequently share consistency properties permitting them to be retained or eliminated as a unit.
Thus, structuring the domain helps in considering sets of elements all at a time and hence helps in pruning the search space more quickly.