Fuzzy particle swarm optimization algorithm (NFPSO) for reachability analysis of complex software systems
Nowadays, model checking is applied as an accuratetechnique to verify software systems. The main problem of modelchecking techniques is the state space explosion. This problemoccurs due to the exponential memory usage by the model checker.In this situation, using meta-heuristic and evolutionary algorithmsto search for a state in which a property is satisfied/violated is apromising solution. Recently, different evolutionary algorithmslike GA, PSO, etc. are applied to find deadlock state. Even thoughuseful, most of them are concentrated on finding deadlock. Thispaper proposes a fuzzy algorithm in order to analyze reachabilityproperties in systems specified through GTS with enormous statespace. To do so, we first extend the existing PSO algorithm (forchecking deadlocks) to analyze reachability properties. Then, toincrease the accuracy, we employ a Fuzzy adaptive PSO algorithmto determine which state and path should be explored in each stepto find the corresponding reachable state. These two approachesare implemented in an open-source toolset for designing andmodel checking GTS called GROOVE. Moreover, theexperimental results indicate that the hybrid fuzzy approachimproves speed and accuracy in comparison with other techniquesbased on meta-heuristic algorithms such as GA and the hybrid ofPSO-GSA in analyzing reachability properties.