POLIZA DE FIANZA COLECTIVA GRUPAL DE FIDELIDAD
AMPLIACION Y DISMINUCION DE LA SUMA AFIANZADA
Definition 3.4.3 (Satisfaction). Given the model M, the satisfaction of a CTLcc,αformula
ϕin a social state s denoted by (M, s)|= ϕ is defined in Tables 4, 5, 6, 7, 8, and 9.
The formal semantics of the CTL formulae, a temporal fragment of CTLcc,α, is intro-
duced in Chapter 2. In Table 4, the state formula W CC(i, j, ψ, ϕ) is satisfied in the model M at s iff the consequence ϕ holds in every state satisfying ψ and accessible from s via
∼i→j. The semantics of the strong commitment SCC(i, j, ψ, ϕ) is similar, but we add
Condition 1 to stress that the antecedent ψ and the consequent ϕ of commitments should be achieved at least in one state accessible from s, which is necessitated in business com- mitments in a strong manner. The proposed semantics is close to the semantics introduced
Table 4: The semantics of the weak and strong commitment formulae (M, s) |= W CC(i, j, ψ, ϕ) iff ∀s ∈ S s.t. s ∼
i→j s and (M, s) |= ψ, we have
(M, s) |= ϕ
(M, s) |= SCC(i, j, ψ, ϕ) iff 1) ∃s∈ S s.t. s∼i→j sand (M, s) |= ψ, and 2) (M, s) |= W CC(i, j, ψ, ϕ)
by Singh [81] for practical commitments. Singh’s semantics is defined by first comput- ing the set of states where the consequence holds and then testing whether these states are among the set of sets of states satisfying the antecedent, which in turn explicitly means this set of states should satisfy the consequence and antecedent of commitment together. Moreover, our semantics guarantee that any conflict between two or more strong commit- ments (enforceable commitments) will not exist as if one commitment has been activated, the other conflicting commitments will be not active in the same state (cf. R11 in Section 3.6). Also, two conflicting commitments can be individually enforceable in different states in the future.
In Table 5, the state formula F uW (i, W CC(i, j, ψ, ϕ)) is satisfied in the model M at s iff s satisfies the consequence ϕ and the negation of the weak commitment W CC(i, j, ψ, ϕ) as well as there exists a state s at which i performs an action to fulfill her commitment
holding at s and s is “seen” and reached from this state via ∼
i→j and Rci. The idea
behind this semantics is to say that a weak commitment is fulfilled when we reach an Table 5: The semantics of the fulfillment action formulae
(M, s) |= F uW (i, W CC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼
i→j s and (li(s), Fulfilli,
li(s)) ∈ Rciand (M, s) |= W CC(i, j, ψ, ϕ) and (M, s)|= ϕ ∧ ¬W CC(i, j, ψ, ϕ)
(M, s) |= F uS(i, SCC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼i→j s and (li(s), Fulfilli, li(s)) ∈ Rciand (M, s) |= SCC(i, j, ψ, ϕ) and (M, s)|= ψ ∧ ¬SCC(i, j, ψ, ϕ)
accessible state from the weak commitment state by performing the fulfillment action in which the consequence holds and the weak commitment becomes no longer active. The semantics of the strong fulfillment F uS(i, SCC(i, j, ψ, ϕ)) is similar, but the focus is on checking the satisfiability of the antecedent ψ. This is because—from the semantics of
the strong conditional commitment—we guarantee that whenever ψ holds in an accessible state, then the consequence ϕ holds as well. By stressing that the active commitment should be terminated in the fulfillment state, we address the fulfillment paradox appeared in [15] (Proposition 2) and discussed in Chapter 1. Recall that this paradox results from the assumption: unconditional commitment should be active when it comes time to its fulfillment, formally: F u(C(i, j, ϕ)) → C(i, j, ϕ). Terminating (or deleting) commitment being fulfilled is stated explicitly in the operational semantics introduced in [106, 97, 81, 23]. Singh, for instance, uses the following postulate: ϕ → ¬C(i, j, ψ, ϕ) where C could be a practical or dialectical conditional commitment [81].
In Table 6, the formula CaS(i, SCC(i, j, ψ, ϕ)) (respectively, CaW (i, W CC(i, j, ψ, ϕ))) is satisfied in M at s iff there exists a state s at which i performs an action to cancel her
strong (respectively, weak) commitment holding at s and s is seen and reached from the
state s via ∼i→j and Rci. And the current state s satisfies the negation of both the an-
tecedent ψ (respectively, consequence ϕ) and strong commitment SCC(i, j, ψ, ϕ) (respec- tively, weak commitment W CC(i, j, ψ, ϕ)).
Table 6: The semantics of the cancelation action formulae
(M, s) |= CaS(i, SCC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼i→j s and (li(s), Canceli, li(s)) ∈ Rciand (M, s) |= SCC(i, j, ψ, ϕ) and (M, s)|= ¬ψ ∧ ¬SCC(i, j, ψ, ϕ) (M, s) |= CaW (i, W CC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼
i→j s and (li(s), Canceli,
li(s)) ∈ Rciand (M, s) |= W CC(i, j, ψ, ϕ) and (M, s)|= ¬ϕ ∧ ¬W CC(i, j, ψ, ϕ)
In Table 7, the formula ReS(i, SCC(i, j, ψ, ϕ)) (respectively, ReW (i, W CC(i, j, ψ, ϕ))) is satisfied in M at s iff there exists a state s at which j performs an action to release
the agent i from her strong (respectively, weak) commitment holding at s and s is seen
and reached from the state s via ∼
i→j and Rcj. And the current state s satisfies the
negation of both the antecedent ψ (respectively, consequence ϕ) and strong commitment SCC(i, j, ψ, ϕ) (respectively, weak commitment W CC(i, j, ψ, ϕ)). The only difference between the semantics of cancelation and release action formulae is the intelligent agent that performs these actions and the label of the local transition.
In Table 8, the formula DeS(i, k, SCC(i, j, ψ, ϕ)) (respectively, DeW (i, k, W CC(i, j, ψ, ϕ))) is satisfied in M at s iff there exists a state s at which i performs an action to del-
egate her strong (respectively, weak) commitment holding at s to another agent k and s
is seen and reached from the state s via ∼
i→j and Rci. And the current state s satis-
Table 7: The semantics of the release action formulae
(M, s) |= ReS(j, SCC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼i→j s and (lj(s), Releasej, lj(s)) ∈ Rcj and (M, s) |= SCC(i, j, ψ, ϕ) and (M, s)|= ¬ψ ∧ ¬SCC(i, j, ψ, ϕ) (M, s) |= ReW (j, W CC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼
i→j s and (lj(s), Releasej,
lj(s)) ∈ Rcj and (M, s) |= W CC(i, j, ψ, ϕ) and (M, s)|= ¬ψ ∧ ¬W CC(i, j, ψ, ϕ)
SCC(i, j, ψ, ϕ)(respectively, weak commitment W CC(i, j, ψ, ϕ)) as well as the new com- mitment created by k.
Table 8: The semantics of the delegation action formulae
(M, s) |= DeS(i, k, SCC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼i→j s and (li(s), Delegatei, li(s)) ∈ Rci and (M, s) |= SCC(i, j, ψ, ϕ) and (M, s) |= ¬ψ ∧ ¬SCC(i, j, ψ, ϕ)∧
SCC(k, j, ψ, ϕ) (M, s) |= DeW (i, k, W CC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼
i→j s and (li(s), Delegatei,
li(s)) ∈ Rci and (M, s) |= W CC(i, j, ψ, ϕ) and (M, s)|= ¬ϕ ∧ ¬W CC(i, j, ψ, ϕ) ∧W CC(k, j, ψ, ϕ)
In Table 9, the formula AsS(j, k, SCC(i, j, ψ, ϕ)) (respectively, AsW (j, k, W CC(i, j, ψ, ϕ))) is satisfied in M at s iff there exists a state sat which j performs an action to assign
her strong (respectively, weak) commitment holding at s to another agent k and s is seen
and reached from the state svia∼
i→j and Rcj. And the current state s satisfies the negation
of the antecedent ψ (respectively, consequence ϕ) and strong commitment SCC(i, j, ψ, ϕ) (respectively, weak commitment W CC(i, j, ψ, ϕ)) as well as the new commitment cre- ated by k. The differences between the semantics of delegation and assignment action
Table 9: The semantics of the assignment action formulae
(M, s) |= AsS(j, k, SCC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼i→j s and (lj(s), Assignj, lj(s)) ∈ Rcj and (M, s) |= SCC(i, j, ψ, ϕ) and (M, s)|= ¬ψ ∧ ¬SCC(i, j, ψ, ϕ)∧ SCC(i, k, ψ, ϕ)
(M, s) |= AsW (j, k, W CC(i, j, ψ, ϕ)) iff ∃s ∈ S s.t. s ∼
i→j s and (lj(s), Assignj,
lj(s)) ∈ Rcj and (M, s) |= W CC(i, j, ψ, ϕ) and (M, s)|= ¬ϕ ∧ ¬W CC(i, j, ψ, ϕ)∧ W CC(i, k, ψ, ϕ)
put, the idea of the proposed semantic rules is that a commitment is fulfilled (respectively cancelled, released, delegated and assigned) when we reach an accessible state from the commitment state by performing the fulfillment (respectively cancelation, release, delega- tion and assignment) action in which the antecedent holds (respectively not holds) and the commitment becomes no longer active as well as in the case of delegation and assignment action formulae, a new commitment is created.