Preliminary Results on the Modeling of System Level Diagnosis Problems with Abstract Argumentation

Stefano Bistarelli, Antonio Caruso
2018 Italian Conference on Theoretical Computer Science  
System Level Diagnosis is a large subfield of fault-tolerance in parallel and distributed systems. Each node in the system can be 'faulty' or 'fault-free' and is able to perform a diagnostic test of it's neighbouring nodes. The collection of test results is called the syndrome of the system; the goal of the diagnosis is to provide algorithms and formal techniques to identify all faulty nodes correctly under various fault model. Abstract Argumentation is a framework used in AI to reason about
more » ... ief and valid conclusion among a network of conflicting arguments, it is used by Intelligent Agents to take decisions. Subset of arguments can be collectively accepted or refuted. In this paper, we study a connection between the two models, and show that the diagnosis problem can be represented using an argumentation system: arguments can be mapped to set of units such that refuted arguments can be correctly diagnosed as faulty.
dblp:conf/ictcs/BistarelliC18 fatcat:tjdmew36lbauppjetkqwtwak6e