Article ID Journal Published Year Pages File Type
4635917 Applied Mathematics and Computation 2006 8 Pages PDF
Abstract

Nowadays, many real problems in artificial intelligence can be modelled as constraint satisfaction problems (CSPs). A general CSP is known to be NP-complete. Nevertheless, distributed models may reduce the exponential complexity by partitioning the problem into a set of subproblems. In this paper, we present a preprocess technique to break a single large problem into a set of smaller loosely connected ones. These semi-independent CSPs can be efficiently solved and, furthermore, they can be solved concurrently.

Related Topics
Physical Sciences and Engineering Mathematics Applied Mathematics
Authors
, ,