Model Checking of Possibilistic Linear-Time Properties Based on Generalized Possibilistic Decision Processes
Yongming Li,Wuniu Liu,Junmei Wang,Xianfeng Yu,Chao Li
DOI: https://doi.org/10.1109/tfuzz.2023.3260446
IF: 12.253
2023-01-01
IEEE Transactions on Fuzzy Systems
Abstract:Model checking possibilistic linear-time properties was investigated by Li in 2017. However, nondeterminism of the system is absent in previous studies. Therefore, in order to permit both possibilistic and nondeterministic choices, we use the generalized possibilistic decision process (GPDP) as a model of the system. First, the definition of GPDP describing the behavior of nondeterministic system is given in detail, the resolution of nondeterminism is performed by using the notion of schedulers, and the semantics of generalized possibilistic linear-temporal logic (GPoLTL) with schedulers are defined. Second, we study possibilistic model checking of some fuzzy linear-time properties under GPDP. Since there are many (infinite) schedulers satisfying a certain linear timing property in a given state of a GPDP, it is particularly critical to study the optimal strategy and its corresponding possible measure, which is called extremal possibility model checking. For some special fuzzy linear-time properties, such as constrained reachability, step-bounded constrained reachability, reachability, always reachability, repeated reachability, persistence reachability, we present complete solution to the optimal (including maximum and minimum cases) possibilistic model checking of the above reachability using the fixpoint techniques. We also introduce fuzzy $\omega$ -regular properties in GPDP and show that their model checking can be simplified by repeated reachability. The algorithms for model checking are also provided. Additionally, an example is presented to illustrate the methods described in the article.