Révision 267 CSL17/tech-report/soundness.tex

soundness.tex (revision 267)
47 47
In the presence of a compatible sorting, we may easily define functions that \emph{evaluate} safe formulae in $\mubc$:
48 48

  
49 49

  
50
\begin{proposition}
51
Given a $\Sigma^\safe_i$ formula $A$ and compatible sorting $(\vec u; \vec x)$ of its variables, there is a $\mubci{i}$ program $\charfn{\vec u ;\vec x}{A} (l, \vec u ; \vec x)$ computing the characteristic function of $A (\vec u \mode l ; \vec x \mode l)$.
52
\end{proposition}
53

  
54 50
\begin{definition}
55 51
	[Length bounded characteristic functions]
56 52
	We define $\mubci{}$ programs $\charfn{}{A} (l , \vec u; \vec x)$, parametrised by a formula $A$ and a compatible typing $(\vec u ; \vec x)$ of its varables, as follows.
......
85 81
\end{definition}
86 82

  
87 83

  
84
\begin{proposition}
85
Given a $\Sigma^\safe_i$ formula $A$ and compatible sorting $(\vec u; \vec x)$ of its variables, 
86
$\charfn{\vec u ;\vec x}{A} (l, \vec u ; \vec x)$ is in $\mubci{i}$ and computes the characteristic function of $A (\vec u \mode l ; \vec x \mode l)$.
87
\end{proposition}
88

  
89

  
90

  
88 91
We will use the programs $\charfn{}{}$ in the witness functions we define below.
89 92
Let us write $\charfn{}{i}$ to denote the class of functions $\charfn{}{A}$ for $A \in \Sigma^\safe_{i}$.
90 93
For the notion of bounding polynomial below we are a little informal with bounds, using `big-oh' notation, since it will suffice just to be `sufficiently large'.
......
96 99
	For a $\Sigma^\safe_{i}$ formula $A$ with a compatible sorting $(\vec u ; \vec x)$, we define the \emph{length-bounded witness function} $\wit{\vec u ; \vec x}{A} (l, \vec u ; \vec x , w)$ in $\bc (\charfn{}{i-1})$ and its \emph{bounding polynomial} $b_A (l)$ as follows:
97 100
	\begin{itemize}
98 101
		\item If $A$ is $\Pi^\safe_{i-1}$ then $\wit{\vec u ; \vec x}{A} (l, \vec u ; \vec x , w) \dfn \charfn{\vec u ; \vec x}{A} (l, \vec u ; \vec x )$ and we set $b_A (l) = 1$.
99
		\item If $A$ is $B \cor C$ then we may set $b_A = O(b_B + b_C)$ and define $		\wit{\vec u ; \vec x}{A} (l, \vec u ;\vec x , w) \dfn		\orfn ( ; \wit{\vec u ; \vec x}{B} (l, \vec u ; \vec x , \beta (b_B (l) , 0 ; w ) ) , \wit{\vec u ; \vec x}{C} (l, \vec u ; \vec x , \beta (b_C (l) , 0 ; w ) )  )$.
102
		\item If $A$ is $B \cor C$ then we may set $|b_A| = O(|b_B| + |b_C|)$ and define $		\wit{\vec u ; \vec x}{A} (l, \vec u ;\vec x , w) \dfn		\orfn ( ; \wit{\vec u ; \vec x}{B} (l, \vec u ; \vec x , \beta (b_B (l) , 0 ; w ) ) , \wit{\vec u ; \vec x}{C} (l, \vec u ; \vec x , \beta (b_C (l) , 0 ; w ) )  )$.
100 103
%		\[
101 104
%		\wit{\vec u ; \vec x}{A} (l, \vec u ;\vec x , w) 
102 105
%		\quad \dfn \quad
......
121 124
		%appealing to Lemma~\ref{lem:sharply-bounded-recursion}, 
122 125
		and 
123 126
		we set 
124
		$b_A = O(b_{B(t)}^2 )$.
127
		$|b_A| = O(|b_{B(t)}|^2 )$.
125 128
		\item Similarly if $A$ is $\exists u^\normal \leq |t(\vec u;)|. A'(u)$, but with $\exists u \leq |t|$ in place of $\forall u \leq |t|$.
126 129
		\item If $A$ is $\exists x^\safe . B(x) $ then
127 130
		\(
......
129 132
		\dfn 
130 133
		\wit{\vec u ; \vec x , x}{B(x)} ( l, \vec u ; \vec x , \beta( b_{B} (l) , 0;w ) , \beta (q(l) , 1 ;w ))
131 134
		\)
132
		where $q$ is obtained by the polychecking and bounded minimisation properties, Lemmas~\ref{lem:polychecking} and \ref{lem:bounded-minimisation}, for $\wit{\vec u ; \vec x , x}{B(x)}$.
133
		We may set $b_A = O(b_B + q )$.
135
		where $q$ is obtained by the polychecking and bounded minimisation properties,\footnote{Here let us assume that $q$ is formulated as a corresponding quasipolynomial in $l$ as opposed to a polynomial in $|l|$, as in Lemma~\ref{lem:bounded-minimisation}.} Lemmas~\ref{lem:polychecking} and \ref{lem:bounded-minimisation}, for $\wit{\vec u ; \vec x , x}{B(x)}$.
136
		We may set $|b_A| = O(|b_B | + |q|)$.
134 137
	\end{itemize}
135 138
%	\[
136 139
%	\begin{array}{rcl}
......
151 154
\end{definition}
152 155

  
153 156

  
154

  
157
From Lemmas~\ref{lem:polychecking} and \ref{lem:bounded-minimisation} we have:
155 158
\begin{proposition}
156 159
	\label{prop:wit-rfn}
157 160
	If, for some $w$, $\wit{\vec u ; \vec x}{A} (l, \vec u ; \vec x,w) =1$, then $A (\vec u \mode l ; \vec x \mode l)$ is true.
......
177 180
	\begin{array}{rl}
178 181
&	\bigwedge\limits_{A \in \Gamma} \wit{\vec u ; \vec x}{ A} (l, \vec u ; \vec x , w_A \mode b_A(l)) =1 \\
179 182
\noalign{\medskip}
180
\implies & 	\bigvee\limits_{B\in \Delta} \wit{\vec u ; \vec x}{B} (l, \vec u ; \vec x , f^\pi_B((\vec u ; \vec x )\mode l, \vec w \mode p(l))) = 1
183
\implies & 	\bigvee\limits_{B\in \Delta} \wit{\vec u ; \vec x}{B} (l, \vec u ; \vec x , f^\pi_B((\vec u ; \vec x )\mode l, \vec w \mode p^\pi(l))) = 1
181 184
	\end{array}
182 185
	\]
183
	for some polynomial $p$.
186
	for some polynomial $p^\pi$.
184 187
%	\anupam{Need $\vec w \mode p(l)$ for some $p$.}
185 188
%	\anupam{$l$ may occur freely in the programs $f^\pi_B$}
186 189
\end{lemma}
187 190

  
188 191

  
189 192

  
190
For the implication above, let us simply refer to the LHS as $\Wit{\vec u ; \vec x}{\Gamma} (l , \vec u ; \vec x , \vec w)$ and the RHS as $\Wit{\vec u ; \vec x}{ \Delta} (l, \vec u ; \vec x , \vec w')$, with $\vec w'$ in place of $\vec f( \cdots )$, which is a slight abuse of notation: we assume that LHS and RHS are clear from context.
193
For the implication above, let us simply refer to the LHS as $\Wit{\vec u ; \vec x}{\Gamma} (l , \vec u ; \vec x , \vec w)$ and the RHS as $\Wit{\vec u ; \vec x}{ \Delta} (l, \vec u ; \vec x , \vec f^\pi ((\vec u ; \vec x )\mode l, \vec w \mode p^\pi(l))  )$, which is a slight abuse of notation: we assume that LHS and RHS are clear from context.
194
Also let us call $p^\pi$ the \emph{modulus} of $f^\pi$ with respect to $l$.
191 195

  
192 196
\begin{proof}
193 197
	Since the proof is in typed variable normal form we have that each line of the proof is of the same form, i.e.\ $\normal (\vec u), \safe (\vec x) , \Gamma \seqar \Delta$ over free variables $\vec u ; \vec x$.
......
202 206
Notice that, since $\pi$ is in De Morgan form, we have that $A$ is atomic ($s\leq t$) and so, in particular, $\Pi^\safe_{i-1}$.
203 207
So we can simply set the witness for both $A$ and $\cnot A$ to $0$.
204 208
Namely, if $\vec f (\vec u ; \vec x , \vec w , w )$ is obtained by the inductive hypothesis, then we may set $f^\pi_B ( \vec u ; \vec x , \vec w) \dfn f_B (\vec u ;\vec x , \vec w , 0)$ for $B\in \Delta$, and $f^\pi_A (\vec u ; \vec x , \vec w) \dfn 0$. 
205
The bounding polynomial remains the same. 
209
The modulus $p^\pi$ remains the same as that of the inductive hypothesis. 
210

  
211
Left negation is similar, relying on a dummy argument.
206 212
	
207 213
	\paragraph*{Logical rules}
208
	Pairing, depairing. Need length-boundedness.
209
	
210
	If we have a left conjunction step:
214
%	Pairing, depairing. Need length-boundedness.
215
	Suppose $\pi$ ends with a 
216
	left conjunction step:
211 217
	\[
212 218
	\vlinf{\lefrul{\cand}}{}{ \normal (\vec u ), \safe (\vec x) , A\cand B , \Gamma \seqar \Delta }{ \normal (\vec u ), \safe (\vec x) , A, B , \Gamma \seqar \Delta}
213 219
	\]
......
219 225
	\]
220 226
	for some polynomial $p$.
221 227
	%
222
	We define $\vec f^\pi (\vec u ; \vec x , w , \vec w) \dfn \vec f (\vec u ; \vec x , \beta (p(l),0; w) , \beta(p(l),1;w),\vec w )$ and, by the bounding polynomial for pairing, it suffices to set $p^\pi = O(p)$.
228
	We define $\vec f^\pi (\vec u ; \vec x , w , \vec w) \dfn \vec f (\vec u ; \vec x , \beta (p(l),0; w) , \beta(p(l),1;w),\vec w )$ and, by the bounding polynomial for pairing, it suffices to set $|p^\pi| = O(|p|)$.
223 229
	
230
		Right disjunction step:
231
		\[
232
		\vlinf{\rigrul{\cor}}{}{ \normal (\vec u ), \safe (\vec x) , \Gamma \seqar \Delta , A \cor B}{ \normal (\vec u ), \safe (\vec x) , \Gamma \seqar \Delta, A, B }
233
		\]
234
		$\vec f^\pi_\Delta$ remains the same as that of premiss.
235
		Let $f_A, f_B$ be the functions corresponding to $A$ and $B$ in the premiss, so that:
236
		\[
237
		\Wit{\vec u ; \vec x}{\Gamma} (l, \vec u ; \vec x , \vec w)
238
		\quad \implies \quad
239
		\Wit{\vec u ; \vec x}{\Delta} (l, \vec u ; \vec x , f_C ( (\vec u ; \vec x) \mode l , \vec w \mode p(l) ))		
240
		\]
241
		for $C = A,B$ and for some $p$, such that $f_A , f_B$ are bounded by $q(l)$ (again by IH).
242
		We define $f^\pi_{A\cor B} (\vec u ; \vec x, \vec w) \dfn \pair{q(l)}{f_A ((\vec u ; \vec x, \vec w))}{f_B ((\vec u ; \vec x, \vec w))}$.
224 243
	
225
	Right disjunction step:
226
	\[
227
	\vlinf{\rigrul{\cor}}{}{ \normal (\vec u ), \safe (\vec x) , \Gamma \seqar \Delta , A \cor B}{ \normal (\vec u ), \safe (\vec x) , \Gamma \seqar \Delta, A, B }
228
	\]
229
	$\vec f^\pi_\Delta$ remains the same as that of premiss.
230
	Let $f_A, f_B$ be the functions corresponding to $A$ and $B$ in the premiss, so that:
231
	\[
232
	\Wit{\vec u ; \vec x}{\Gamma} (l, \vec u ; \vec x , \vec w)
233
	\quad \implies \quad
234
	\Wit{\vec u ; \vec x}{\Delta} (l, \vec u ; \vec x , f_C ( (\vec u ; \vec x) \mode l , \vec w \mode p(l) ))		
235
	\]
236
	for $C = A,B$ and for some $p$, such that $f_A , f_B$ are bounded by $q(l)$ (again by IH).
237
	We define $f^\pi_{A\cor B} (\vec u ; \vec x, \vec w) \dfn \pair{q(l)}{f_A ((\vec u ; \vec x, \vec w))}{f_B ((\vec u ; \vec x, \vec w))}$.
244
	The other logical cases use a similar argument.
245
	
246

  
238 247
	\paragraph*{Quantifiers}
239
	\anupam{Do $\exists$-right and $\forall$-right, left rules are symmetric.}
240
	
241
	
242
	
243
	Sharply bounded quantifiers are generalised versions of logical rules.
248
	If $\pi$ ends with a sharply bounded quantifier step, then we use a similar argument to that for logical rules. 
249
	For instance, suppose $\pi$ ends with a $\rigrul{|\forall|}$:
244 250
	\[
245
	\vlinf{|\forall|}{}{\normal (\vec u) , \safe (\vec x) , \Gamma \seqar  \Delta , \forall u^\normal \leq |t(\vec u;)| . A(u) }{ \normal(u) , \normal (\vec u) , \safe (\vec x) , u \leq |t(\vec u;)| , \Gamma \seqar  \Delta, A(u)  }
251
	\vlinf{\rigrul{|\forall|}}{}{\normal (\vec u) , \safe (\vec x) , \Gamma \seqar  \Delta , \forall u^\normal \leq |t(\vec u;)| . A(u) }{ \normal(u) , \normal (\vec u) , \safe (\vec x) , u \leq |t(\vec u;)| , \Gamma \seqar  \Delta, A(u)  }
246 252
	\]
247 253
	By the inductive hypothesis we have functions $\vec f(u , \vec u ; \vec x , w , \vec w)$ such that:
248 254
	\[
......
256 262
	
257 263
	We set $f^\pi_{\forall u^\normal \leq t . A} (\vec u ; \vec x , \vec w) \dfn F(q(|l|), t(\vec u;), \vec u ; \vec x , 0, \vec w  )$.
258 264
	
265
		\anupam{Do $\exists$-right and $\forall$-right, left rules are symmetric.}
259 266
	
260 267
	Right existential:
261 268
	\[
......
265 272
	
266 273
	
267 274
	\paragraph*{Contraction}
268
	Left contraction simply duplicates an argument, whereas right contraction requires a conditional on a $\Sigma^\safe_i$ formula.
275
	Left contraction simply duplicates an argument,\footnote{We ignore here the cases when contraction is on a $\normal (u) $ or $\safe (x)$ formula, treating these formulae form as forming a set rather than a multiset.} whereas right contraction requires a conditional on a $\Sigma^\safe_i$ formula.
276
	This is the reason why we need to the witness function encoded itself into $\mubc$ rather than simply using a predicate.
269 277
	
278
	For the sake of example, suppose $\pi$ ends with a right contraction step:
270 279
	\[
271 280
	\vlinf{\cntr}{}{\normal (\vec u) , \safe (\vec x) , \Gamma \seqar \Delta , A }{\normal (\vec u) , \safe (\vec x) , \Gamma \seqar \Delta , A , A}
272 281
	\]
273
	
274
	$\vec f^\pi_\Delta$ remains the same as that of premiss. Let $f_0 ,f_1$ correspond to the two copies of $A$ in the premiss.
275
	We define:
282
	By the inductive hypothesis we have functions $\vec f_\Delta , f_0, f_1$ corresponding to the RHS $\Delta , A ,A$.
283
	We may define $\vec f^\pi_\Delta = \vec f_\Delta$ and:
276 284
	\[
277 285
	f^\pi_A ( \vec u ; \vec x , \vec w  )
278 286
	\quad \dfn \quad

Formats disponibles : Unified diff