07.01.2013 Views

Lecture Notes in Computer Science 3472

Lecture Notes in Computer Science 3472

Lecture Notes in Computer Science 3472

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

264 Verena Wolf<br />

(8) (⊑BC , ⊑CH ):Weshowherethat<br />

M1 ⊑BC M2 =⇒ φem(M1) ⊑CH φem(M2)<br />

holds which has not been considered by Bernardo and Cleaveland [BC00]<br />

or Christoff [Chr90]. Assume M1 ⊑BC M2. We have to show that for each<br />

T ∈T np,re , α ∈ Act ∗<br />

Pr trace<br />

trace<br />

φem (M1)�T (α) � Prφem (M2)�T (α).<br />

For a given T we set T ′ =(ST , →, sT ) ∈T pa<br />

τ<br />

t a −→T t ′ iff t (a,0)<br />

−−−→ t ′ for a ∈ Act.<br />

where → is such that<br />

For a given trace α <strong>in</strong> T there is only one s<strong>in</strong>gle prefix β1 of a path β =<br />

β1β2 ∈ Path(T )withtrace(β1) =α. Itiseasytoverifythat<br />

Pr trace<br />

M1�T ′(α) =Pr trace<br />

φem(M1)�T (α).<br />

So if the last state of the sequence β1 is the only success state <strong>in</strong> T ′ ,wecan<br />

derive<br />

Pr trace<br />

M1�T ′(α) =W �x<br />

M1�T ′ � W �x<br />

M2�T ′ = Pr trace<br />

M2�T ′(α).<br />

Hence we have φem(M1) ⊑CH φem(M2).<br />

⊑BC is strictly f<strong>in</strong>er than ⊑CH which can be seen by a simple counterexample<br />

similar to example 9.7.2 because we can apply test processes with<br />

τ-transitions.<br />

(9) (⊑BC , ⊑CL) : To relate the Markovian test<strong>in</strong>g relation ⊑BC and the fully<br />

probabilistic test<strong>in</strong>g relation ⊑CL, we have to consider the sets of applied test<br />

processes. For ⊑BC some k<strong>in</strong>d of ”passive” test processes are used because<br />

external actions must have rate zero. Only τ-transitions have nonzero rates.<br />

For ⊑CL all transitions <strong>in</strong> test processes are equipped with probabilities.<br />

S<strong>in</strong>ce probabilistic test processes have more dist<strong>in</strong>guish<strong>in</strong>g power than nonprobabilistic<br />

test processes (see relationship (6)), we have<br />

M1 ⊑BC M2 �=⇒ φem(M1) ⊑CL φem(M2)<br />

<strong>in</strong> general. It is easy to construct a counterexample (similar to Example 6).<br />

Of course, the converse of the statement is also wrong, because of the additional<br />

requirement on the expected duration of a successful path for ⊑BC .<br />

So ⊑BC and ⊑CL are <strong>in</strong>comparable. Bernardo and Cleaveland def<strong>in</strong>ed ⊑CL<br />

<strong>in</strong> a restrictive way such that the statement holds [BC00].<br />

Figure 9.14 shows a diagram of the relationships discussed above.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!