Download Full Issue in PDF - Academy Publisher
Download Full Issue in PDF - Academy Publisher
Download Full Issue in PDF - Academy Publisher
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
JOURNAL OF COMPUTERS, VOL. 8, NO. 6, JUNE 2013 1439<br />
L<br />
L<br />
Let v = v = a , φ<br />
M, v M,<br />
w<br />
M ,<br />
= b<br />
w w<br />
, thus we must<br />
prove <strong>in</strong>f<br />
w( a⇒bw) ≤( a⇒ <strong>in</strong>f<br />
wbw)<br />
. On the one hand,<br />
<strong>in</strong>f w<br />
b w<br />
≤ b w<br />
, thus a⇒bw ≥( a⇒ <strong>in</strong>f<br />
wbw)<br />
for each w,<br />
thus <strong>in</strong>f<br />
w( a⇒bw) ≥( a⇒ <strong>in</strong>f<br />
wbw)<br />
. On the other hand if<br />
z ≤( a⇒ b w<br />
) for each w, then z∗a ≤ bw<br />
for each w,<br />
z∗a ≤ <strong>in</strong>f w<br />
b w<br />
, z ≤( a⇒ <strong>in</strong>f<br />
wbw)<br />
. Thus ( a⇒ <strong>in</strong>f<br />
wbw)<br />
is<br />
the <strong>in</strong>fimum of all ( a⇒<br />
b w<br />
). So (U19) holds.<br />
For (U20), we need to verify<br />
<strong>in</strong>f<br />
w( aw ⇒ bw) = (sup<br />
w<br />
aw<br />
⇒ b)<br />
. Indeed, sup w<br />
≥ a w<br />
, thus<br />
(sup<br />
w<br />
aw ⇒b) ≤( aw<br />
⇒ b)<br />
, hence<br />
(sup<br />
w<br />
aw ⇒b) ≤<strong>in</strong>f w( aw<br />
⇒ b)<br />
, If z ≤ aw<br />
⇒ b for all w ,<br />
then aw<br />
≤( z ⇒ b)<br />
for all w, then sup<br />
w<br />
aw<br />
≤( z ⇒ b)<br />
,<br />
z ≤(sup waw<br />
⇒ b)<br />
, so sup w<br />
a w<br />
⇒ b is the <strong>in</strong>fimum. So<br />
(U20) holds.<br />
F<strong>in</strong>ally we verify (U21), we need to verify<br />
<strong>in</strong>f<br />
w( a∨ bw) = a∨ <strong>in</strong>fwbw<br />
. Indeed, a ≤ a∨ bw<br />
, thus<br />
a ≤<strong>in</strong>f w( a∨ bw)<br />
; similarly, <strong>in</strong>fwbw ≤<strong>in</strong>f w( a∨ bw)<br />
, thus<br />
a∨<strong>in</strong>fwbw ≤<strong>in</strong>f w( a∨ b)<br />
. Conversely, let z ≤ a∨ bw<br />
for<br />
all w , we prove z ≤ a∨ <strong>in</strong>f w<br />
b w<br />
.<br />
Case 1: Let a ≤ <strong>in</strong>f w<br />
b w<br />
. Then z ≤ bw<br />
for each w ,<br />
z ≤ <strong>in</strong>f w<br />
b w<br />
and z ≤ a∨ <strong>in</strong>f w<br />
b w<br />
.<br />
Case 2: Let a ≥ <strong>in</strong>f w<br />
b w<br />
. Then for some w 0<br />
, a ≥ b w0<br />
,<br />
thus z ≤ a and z ≤ a∨ <strong>in</strong>f w<br />
b w<br />
.<br />
So we prove the soundness of axioms.<br />
Theorem 6 (Soundness of deduction rules) (1) For<br />
arbitrary formulas φ, ψ , safe-structure M and evaluation<br />
L L L<br />
v , ψ ≥ φ ∗ φ → ψ . In particular, if<br />
M, v M, v M,<br />
v<br />
L<br />
L<br />
L<br />
= → = 1<br />
M, v<br />
M,<br />
v L<br />
then ψ 1<br />
M ,<br />
=<br />
v L<br />
.<br />
L L L<br />
(2) Consequently, ψ φ φ ψ<br />
M M M<br />
φ φ ψ<br />
≥ ∗ → , thus if<br />
φ, φ → ψ are then ψ is 1 L<br />
-true <strong>in</strong> M.<br />
(3) If φ is 1 L<br />
-true <strong>in</strong> M then Δ φ is 1 L<br />
-true <strong>in</strong> M .<br />
(4) If φ is 1 L<br />
-true <strong>in</strong> M then ( ∀ x)<br />
φ is 1 L<br />
-true <strong>in</strong> M.<br />
Proof. (1) is just as <strong>in</strong> propositional calculus.<br />
To prove (2) put φ = aw, ψ = bw, <strong>in</strong>f<br />
w<br />
w<br />
w<br />
aw<br />
= a. We<br />
have to prove <strong>in</strong>f<br />
w( aw ⇒bw) ≤<strong>in</strong>fw aw ⇒ <strong>in</strong>fwbw<br />
.<br />
Observe the follow<strong>in</strong>g:<br />
<strong>in</strong>f( aw ⇒bw) ≤( aw ⇒bw) ≤( a⇒ bw)<br />
,<br />
thus <strong>in</strong>f<br />
w( aw ⇒bw) ≤<strong>in</strong>f w( a⇒ bw)<br />
. It rema<strong>in</strong>s to prove<br />
<strong>in</strong>f<br />
w( a⇒bw) ≤ a⇒ <strong>in</strong>fwbw, this is holds from Theorem<br />
5.<br />
L<br />
(3) If φ is 1 L<br />
-true <strong>in</strong> M then φ<br />
M<br />
= 1L<br />
, so<br />
L<br />
L<br />
Δ φ =Δ φ = . Then (3) holds.<br />
1<br />
M, v<br />
M,<br />
v L<br />
L<br />
M<br />
φ ′<br />
′<br />
,<br />
}<br />
L<br />
M<br />
L<br />
M<br />
(4) Be<strong>in</strong>g φ = <strong>in</strong>f{ φ ,v<br />
| v M − evaluation}<br />
≤ <strong>in</strong>f{<br />
Mv| v ≡ v = ( ∀ x) φ } , So (4) holds.<br />
So we can get the follow<strong>in</strong>g soundness theorem.<br />
L<br />
Theorem 7 (Soundness) Let L is l<strong>in</strong>early ordered<br />
ŁΠG − -algebra and φ is a formula <strong>in</strong> J, if φ , then φ is<br />
L<br />
L-tautology, i.e. φ = 1 . M L<br />
Similarly, we can get the follow<strong>in</strong>g strong soundness<br />
theorem.<br />
Def<strong>in</strong>ition 10 Let T be a theory, L be a l<strong>in</strong>early ordered<br />
ŁΠG − -algebra and M a safe L-structure for the language<br />
of T. M is an L-model of T if all axioms of T are 1 L<br />
-true<br />
<strong>in</strong> M, i.e. φ = 1 L<br />
<strong>in</strong> each φ ∈ T .<br />
Theorem 8 (Strong Soundness) Let T be a theory, L is<br />
l<strong>in</strong>early ordered ŁΠG − -algebra and φ is a formula <strong>in</strong> J,<br />
L<br />
if T φ ( φ is provable <strong>in</strong> T), then φ<br />
M<br />
= 1L<br />
for each<br />
l<strong>in</strong>early ordered ŁΠG − -algebra L and each L-model M<br />
of T.<br />
Proof. In fact, from the proof of Theorem 5, for each<br />
L-model M of T, the axioms are true, and the formulas <strong>in</strong><br />
T are true, from the proof of Theorem 6, the deduction<br />
rules preserve true. So the theorem holds.<br />
Theorem 9 (Deduction Theorem) Let T be a theory,<br />
φ, ψ are closed formulas. Then ( T ∪φ)<br />
ψ iff<br />
T Δφ →ψ<br />
.<br />
Proof. Sufficiency: Let T Δφ →ψ<br />
, from φ<br />
( φ∈ ( T ∪ φ)<br />
), then Δ φ by necessitation, so we can get<br />
ψ by MP rules.<br />
Necessity: Let m is the proof length from T ∪ φ}<br />
to ψ ,<br />
we prove by <strong>in</strong>duction for the length m.<br />
When m = 1 , ψ ∈T<br />
∪φ∪Axm(C ∀ ) , if ψ = φ , The<br />
result holds. If ψ ∈ T or ψ is axiom, from Lemma2 (2),<br />
we have ψ →( Δφ → ψ)<br />
, then by ψψ , →( Δφ→ ψ)<br />
, we<br />
get Δφ → ψ , thus T Δφ →ψ<br />
.<br />
Assume that the result holds when m≤<br />
k , i.e. we get<br />
γ at k step, then T Δφ → γ . Now Let m= k + 1.<br />
If ψ is obta<strong>in</strong>ed from MP rule by the above results<br />
γγ , → ψ <strong>in</strong> the proof sequence, then by <strong>in</strong>duction<br />
hypothesis, we get T Δφ → γ, T Δφ →( γ →ψ)<br />
.<br />
From Lemma 3, we can get<br />
T ( Δφ& Δφ) →( γ &( γ →ψ))<br />
. Be<strong>in</strong>g<br />
T ( Δφ) &( Δφ)<br />
≡ Δφ<br />
, so T Δφ →( γ &( γ →ψ))<br />
.<br />
From lemma 2 (4) we have ( γ & ( γ →ψ)<br />
→ ψ , so we<br />
get T Δφ →ψ<br />
by the hypothetical syllogism.<br />
If ψ is obta<strong>in</strong>ed from necessitation rule by the above<br />
step γ <strong>in</strong> the proof sequence, i.e. Δ γ = ψ , then by<br />
<strong>in</strong>duction hypothesis, we get T Δφ →γ<br />
.<br />
T Δ( Δφ →γ)<br />
, from (U16) we can get T ΔΔφ → Δγ<br />
,<br />
from (U15) we can get Δφ<br />
→ΔΔ φ , thus by the<br />
hypothetical syllogism we can get T Δφ →Δγ<br />
, i.e.<br />
T Δφ →ψ<br />
.<br />
If ψ is obta<strong>in</strong>ed from generalization rule by the above<br />
step γ <strong>in</strong> the proof sequence, i.e. ( ∀ x)<br />
γ = ψ , then by<br />
© 2013 ACADEMY PUBLISHER