On improving efficiency of model checking through systematically ...

Chapter 4
Case studies
This chapter contains four systems which are of different complexity and
application domains chosen as case studies for our approach. We roughly
classify them according to physical size and characteristic of the state space.
We define classification standards as follows:
1. Physical size: In our context, physical size is represented by the
number of locations, transitions, Boolean operators, and variables.
Since input models are merely text files, the physical size may have
significant impact on the performance of tools applied. Our goal is to
see whether or not the physical size of a system affects the efficiency
of verification. We define two categories: small and large. A system
is categorized as small in physical size if the total number of locations,
transitions, Boolean operators, and variables is less than 10000 and
vice versa.
2. Characteristic of the state space: Two categories are defined including
Finite and Infinite. We found two reasons for which a state
space becomes infinite:
(a) The system involves unbounded numerical variables.
(b) The existence of real-time in the system.
We characterize the state space of models in the perspective of Uppaal
model checking approach. In an infinite system due to unbounded variables,
the iterative resolution of a fixed point equation defined by Uppaal involves
infinite iterations. Note that the fixed point equation defined by Nbac may
have a solution. It depends on the type of values on which the fixed point
equation is defined. For example, values x =(1..n) may be abstracted as a
unique value pos = x>0inNbac. In this case, if x always increases, the
37