Multiagent Constraint Satisfaction with Multiply Sectioned Constraint Networks

Abstract

Abstract. Variables and constraints in problem domains are often distributed. These distributed constraint satisfaction problems (DCSPs) lend themselves to multiagent solutions. Most existing algorithms for DCSPs are extensions of centralized backtracking or iterative improvement with breakout. Their worst case complexity is exponential. On the other hand, directional consistency based algorithms solve centralized CSPs efficiently if primal graph density is bounded. No known multiagent algorithms solve DCSPs with the same efficiency. We propose the first such algorithm and show that it is sound and complete.

Privacy preserving existence recognition and construction of hypertree agent organization

“…Multiple sectioned constraint networks (MSCNs) [XZ07,XMZ14] solve distributed constraint satisfaction problems with complex local problems such as university timetabling [XZ08]. The core of each agent is a constraint subnet that encodes variables and constraints over its subenvironment.…”

“…In multiagent probabilistic reasoning, a message encodes the sender agent's belief over shared variables, e.g., in MSBNs [Xia96,Xia02]. In multiagent constraint reasoning, a message encodes partial solutions over shared variables, e.g., in MSCNs [XZ07,XMZ14]. In multiagent decision making, a message is either an expected utility function over shared variables, or a partial action profile over these variables, e.g., in CDNs [XCD04,XH11].…”