Distributed CSPs by graph partitioning
作者:
Highlights:
•
摘要
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.
论文关键词:Constraint satisfaction problems,Distributed CSPs,Artificial intelligence
论文评审过程:Available online 31 July 2006.
论文官网地址:https://doi.org/10.1016/j.amc.2006.05.090