1. MAFSM: From Trees to GraphsBn BB BA
BBBB
BABA
BABB
BBBA . . . . . . . . . . . .

2. MAFSM: From Trees to GraphsBn BB BA
BBBB
BABA
BABB
BBBA . . . . .
Front( Bn) . . . . . . .

3. MAFSM: From Trees to Graphs
We introduce a relation E included in Front(Bn )Bn
E(BBBA , BA) intuitively means that BBBA and BA are modelled by the same process (they are “behaviourally equivalent”).

4. MAFSM: From Trees to Graphs
We introduce a relation E included in Front(Bn )Bn
E(BBBA , BA) intuitively means that BBBA and BA are modelled by the same process (they are “behaviourally equivalent”).
Example:
E(BBBA , BA) as we assume that A “knows” that B behaves following the protocol.
E(BABB , BA) as we assume that B “knows” that A behaves following the protocol.

5. MAFSM: From Trees to GraphsBn BB BA
BBBB
BABA
BABB
BBBA . . . . .
Front( Bn) . . . . . . . . .

6. MAFSM: From Trees to Graphs
We introduce a relation E included in Front(Bn )Bn
E denotes the smallest equivalence relation on B* containing E
Example:
BA E BBBA E BABBBA E  E BBBABBBA  BBBA
BB E BABB E BBBABB E  E BABBBABB  BABB

7. MAFSM: From Trees to Graphs
We introduce a relation E included in Front(Bn )Bn
E denotes the smallest equivalence relation on B* containing E
Example:
BA E BBBA E BABBBA E  E BBBABBBA  BBBA
BB E BABB E BBBABB E  E BABBBABB  BABB
but
BBBB, BABA are not equivalent, via E , to any other view except BBBB, BABA, respectively.

8. MAFSM: From Trees to Graphs
BA E BBBA BB
BABA
BABB
BBBB . . . . . . . . .

9. MAFSM: From Trees to Graphs
BA E BBBA
BB E BABB
BABA
BBBB . . . . . .

10. MAFSM: From Trees to Graphs
BA E BBBA
BB E BABB
BBf ?
BABA
BBBB . . . . . .

11. MAFSM: From Trees to Graphs
BA E BBBA
BB E BABB
BBf ?
f ?
BABA
BBBB . . . . . .

12. MultiAgent Finite State Machine
Definition: Let {La } be a family of MATL languages on {Pa }. A MAFSM MF = <E, F> where
E  Front(Bn)×Bn is functional relation
F is as defined above