Fault-Tolerant Information Sharing Networks

Agents are goal-driven entities, and the goal achievement is affected by the robustness of information supply required to achieve the goal. This research addresses fault tolerance in information sharing networks, and shows that the construction of information sharing networks, in the way that an agent can tolerate k concurrent information source failures, can be expressed as a constraint satisfaction problem. By solving the resulting constraint satisfaction problem, an agent can efficiently decide on the information sources from which to request the required information, so that the agent can achieve the goals even if a subset of the information sources fail.