Spectrum-Based Fault Localisation for Multi-Agent Systems

However, generation of MAS models that SFL is a well-suited technique for MASs. is both error-prone and time intense, as it exponentially Literature has shown that there is no standard similarity increases with the number of agents coefficient that yields the best result for SFL [Yoo et al., 2014; and their interactions. In this paper, we propose Hofer et al., 2015; Le et al., 2013]. Empirical evaluation is a lightweight, automatic debugging-based technique, therefore essential to establish which set of heuristics excels coined ESFL-MAS, which shortens the diagnostic for the specific context to which SFL is being applied. To the process, while only relying on minimal best of our knowledge, SFL has not as yet been applied to information about the system. ESFL-MAS uses a diagnose behavioural faults in MASs; there is hence the need heuristic that quantifies the suspiciousness of an to empirically evaluate different formulae using known faults agent to be faulty; therefore, different heuristics to compare the performance yielded by several coefficients.