Behavioral Hierarchy with Hierarchical FSMs (HFSMs)

Abstract

We introduce the idea of behavioral hierarchy in this chapter using examples to illustrate the difference between traditional modeling and simulation and one that preserves the behavioral hierarchy. However, an intuitive understanding of behavioral hierarchy is insufficient in being able to implement it such that it is suitable for extension for heterogeneous behavioral hierarchy, which is another essential ingredient for SLDLS.We present the basic formal definitions allowing us to formalize the execution semantics for the hierarchical FSM MoC. This sets the foundation for extending the formalization of the HFSM with heterogeneity that we will see in the upcoming chapters. Another advantage of using behavioral hierarchy is abstraction of state space. To understand how behavioral hierarchy helps in the state space abstraction, let us look at a simple television control example in Figure 4.2 and Figure 4.3. Once the television is turned on, channels can be flipped either using the television itself (named with _tv_ ) or the remote device (named with _rmt_ ). This example also shows AND concurrency in FSMs [50]. In Figure 4.2, the tv on state encapsulates all behaviors of changing the channels via the remote or the television. Notice that only one transition to off tv is required from this state, whereas in Figure 4.3 every state must have a transition to turn the television off. Furthermore, the toplevel FSM in Figure 4.2 only sees two states, tv_on and off_tv and the hierarchical FSM sees the six other states. Figure 4.2 can use three instances of an FSM scheduler to simulate the model. One instance simulates the two toplevel states and the tv_on can use one instance for the remote controls and the other for the television controls. Hence, abstracting the state space effectively.