Efficiently Checking Actual Causality with SAT Solving
This addresses the computational challenge of causality reasoning for multi-agent and distributed socio-technical systems, representing an incremental advance in algorithmic efficiency.
The paper tackled the problem of efficiently checking actual causality in complex systems by developing a novel algorithmic approach using SAT solving, achieving computation times under 5 seconds for models with over 4000 variables.
Recent formal approaches towards causality have made the concept ready for incorporation into the technical world. However, causality reasoning is computationally hard; and no general algorithmic approach exists that efficiently infers the causes for effects. Thus, checking causality in the context of complex, multi-agent, and distributed socio-technical systems is a significant challenge. Therefore, we conceptualize an intelligent and novel algorithmic approach towards checking causality in acyclic causal models with binary variables, utilizing the optimization power in the solvers of the Boolean Satisfiability Problem (SAT). We present two SAT encodings, and an empirical evaluation of their efficiency and scalability. We show that causality is computed efficiently in less than 5 seconds for models that consist of more than 4000 variables.