Modular Decomposition of Hierarchical Finite State Machines

11/09/2021
by   Oliver Biggar, et al.
0

In this paper we develop an analogue of the graph-theoretic `modular decomposition' in automata theory. This decomposition allows us to identify hierarchical finite state machines (HFSMs) equivalent to a given finite state machine (FSM). We provide a definition of a module in an FSM, which is a collection of nodes which can be treated as a nested FSM. We identify a well-behaved subset of FSM modules called thin modules, and represent these using a linear-space directed graph we call a decomposition tree. We prove that every FSM has a unique decomposition tree which uniquely stores each thin module. We provide an O(n^2k) algorithm for finding the decomposition tree of an n-state k-alphabet FSM. The decomposition tree allows us to extend FSMs to equivalent HFSMs. For thin HFSMs, which are those where each nested FSM is a thin module, we can construct an equivalent maximally-hierarchical HFSM in polynomial time.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset
Success!
Error Icon An error occurred

Sign in with Google

×

Use your Google Account to sign in to DeepAI

×

Consider DeepAI Pro