Re: Petri Nets and FSM

Indeed, I agree with this point; FSM are fine as long as they are small,
but they lack a little structure when things become complex.
An approach maybe more pragmatic than PNs is synchronous languages,
like e.g. Esterel (www.esterel.org), where structured programming is
offered, for reactive systems, with a compilation into a FSM (or
equivalent formalism), so you don't have to make the transformation
yourself, and modification is easy. 
And ORCCAD, as Daniel Simon mentioned, is making a particular use of
this synchronous technology in a framework specialized in robot and
control system programming.

Eric

> Cezary Zielinski wrote:
> 
> Finite State Machines are very good at modelling sequential processes
> (single threads). They are equivalent to flow charts. Petri nets model
> concurrent processes. Safe Petri nets can be transformed into FSM (I
> think that also holds for bounded Petri nets, but I am not sure).
> Although this transformation is possible it is not advisable, because
> in this process clarity is lost. Usually a fairly simple Petri net
> results in quite a complex FSM. It is possible to use both models,
> i.e. to model single threads as FSM and to model the overall system
> functioning by PN. PN then can also be responsible for modelling (and
> control) of synchronisation of threads.
> 
> If you specify exactly what you want to do I can try and work out some
> means of its formal
> description.
> Cheers,
> Cezary
> 
> > In our software framework design, we want to include the capability
> > for dynamic reconfiguration of software components. Therefore, some
> > sort of state modelling is necessary, and I was wondering whether
> > somebody has good insights into the differences between Finite State
> > Machines on the one hand, and (Coloured) Petri Nets on the other
> hand?
> > Or maybe other techniques (CSP model?) are even more appropriate?
> 
> > Herman

-- 
Eric Rutten                  INRIA Rhône-Alpes                projet BIP 
ZIRST, 655 avenue de l'Europe, MONTBONNOT, 38334 ST ISMIER Cedex  FRANCE 
tel:+33 (0)4 76 61 54 02                        fax:+33 (0)4 76 61 54 77 
e-mail: [email protected]       web: http://www.inrialpes.fr/bip/