Model-checking agent refinement

We present a proof-technique for reducing the nondeterminism of abstract agent specifications in a BDI framework by means of refinement. We implement the operational semantics of agent specifications in rewrite systems such that we can automatically check if refinement between (fair) executions of agents holds.