Fuzzy Labeled Transition Refinement Tree: Application to Stepwise Designing Multi Agent Systems

Abstract

This paper aims to provide a formal framework that supports an incremental development of dynamic systems such as multi agents systems MAS. We propose a fuzzy labeled transition system model FLTS for short. FLTS allows a concise action refinement representation and deals with incomplete information through its fuzziness representation. Afterward, based on FLTS model, we propose a refinement model called fuzzy labeled transition refinement tree FLTRT for short. The FLTRT structure serves as a tree of potential concurrent design trajectories of the system. Also, we introduce bisimulation relations for both models in order to identify equivalent design trajectories, which could be assessed with respect to relevant design parameters.