The methodology supported by DEEM relies upon Deterministic and Stochastic
Petri Nets (DSPN) as the modeling tool and on Markov Regenerative Processes
(MRGP) for the model solution. Due to their high expressiveness, DSPN models
are able to cope with the dynamic structure of MPS, and allow defining
very concise models. DEEM models are solved with a very simple and computationally
efficient analytical solution technique based on the separability of the
MRGP underlying the DSPN of a MPS.
A DEEM model may include immediate transitions, transitions with exponentially distributed firing times, and transitions with deterministic firing times. In addition, DEEM makes available a set of modeling features that significantly improve its expressiveness.
Through a Graphical User Interface (GUI), DEEM provides a general modeling scheme in which two logically separate parts are used to represent MPS models. One is the System Net (SN), which represents the failure/repair behavior of system components, and the other is the Phase Net (PhN), which represents the execution of the various phases.
The specific dependability measure of interest for the MPS evaluation is defined through the general mechanism of marking-dependent reward functions. Then results of the evaluation are returned in a file which can be further elaborated for producing plots or tables of the dependability measures.
DEEM is able to deal with all the scenarios of MPS that have been analytically
treated in the literature, at a cost which is comparable with that of the
cheapest ones. Compared to general purpose DSPN tools, DEEM offers a number
of advantages. On the modeling side, the tool GUI allows defining the PhN
and SN sub-models to neatly model the phase-dependent behaviors of MPS.
On the evaluation side, the specialized separate algorithm implemented
by DEEM results in a relevant reduction of the MPS model solution time.