In real algebraic geometry, Krivine–Stengle (German for "positive-locus-theorem") characterizes polynomials that are positive on a semialgebraic set, which is defined by systems of inequalities of polynomials with real coefficients, or more generally, coefficients from any real closed field.
It can be thought of as a real analogue of Hilbert's Nullstellensatz (which concern complex zeros of polynomial ideals), and this analogy is at the origin of its name. It was proved by French mathematician and then rediscovered by the Canadian .
Let be a real closed field, and = and = finite sets of polynomials over in variables. Let be the semialgebraic set
W=\{x\inRn\mid\forallf\inF,f(x)\ge0;\forallg\inG,g(x)=0\},
and define the preordering associated with as the set
P(F,G)=\left\{
m} | |
\sum | |
\alpha\in\{0,1\ |
\sigma\alpha
\alpha1 | |
f | |
1 |
…
\alpham | |
f | |
m |
+
r | |
\sum | |
\ell=1 |
\varphi\ellg\ell:\sigma\alpha\in
2[X | |
\Sigma | |
1,\ldots,X |
n]; \varphi\ell\inR[X1,\ldots,Xn]\right\}
where Σ2[{{Var|X}}<sub>1</sub>,...,{{Var|X}}<sub>{{Var|n}}</sub>] is the set of sum-of-squares polynomials. In other words, = +, where is the cone generated by (i.e., the subsemiring of [{{Var|X}}<sub>1</sub>,...,{{Var|X}}<sub>{{Var|n}}</sub>] generated by and arbitrary squares) and is the ideal generated by .
Let ∈ [{{Var|X}}<sub>1</sub>,...,{{Var|X}}<sub>{{Var|n}}</sub>] be a polynomial. Krivine–Stengle Positivstellensatz states that
(i)
\forallx\inW p(x)\ge0
\existsq1,q2\inP(F,G)
s\inZ
q1p=p2s+q2
(ii)
\forallx\inW p(x)>0
\existsq1,q2\inP(F,G)
q1p=1+q2
The weak is the following variant of the . Let be a real closed field, and,, and finite subsets of [{{Var|X}}<sub>1</sub>,...,{{Var|X}}<sub>{{Var|n}}</sub>]. Let be the cone generated by, and the ideal generated by . Then
\{x\inRn\mid\forallf\inFf(x)\ge0\land\forallg\inGg(x)=0\land\forallh\inHh(x)\ne0\}=\emptyset
\existsf\inC,g\inI,n\inN f+g+\left(\prodH\right)2n=0.
(Unlike, the "weak" form actually includes the "strong" form as a special case, so the terminology is a misnomer.)
The Krivine–Stengle Positivstellensatz also has the following refinements under additional assumptions. It should be remarked that Schmüdgen's Positivstellensatz has a weaker assumption than Putinar's Positivstellensatz, but the conclusion is also weaker.
Suppose that
R=R
W=\{x\inRn\mid\forallf\inF,f(x)\ge0\}
p\inR[X1,...,Xn]
W
W
p\inP(F,\emptyset)
W
p(x)>0
x\inW
R=R
Define the quadratic module associated with as the set
Q(F,G)=\left\{\sigma0+
m | |
\sum | |
j=1 |
\sigmajfj+
r | |
\sum | |
\ell=1 |
\varphi\ellg\ell:\sigmaj\in\Sigma2[X1,\ldots,Xn]; \varphi\ell\inR[X1,\ldots,Xn]\right\}
Assume there exists L > 0 such that the polynomial
L-
n | |
\sum | |
i=1 |
2 | |
x | |
i |
\inQ(F,G).
p(x)>0
x\inW
de:Konrad Schmüdgen
. 1991. The K-moment problem for compact semi-algebraic sets. Mathematische Annalen. 289. 1. 203–206. 10.1007/bf01446568. 0025-5831.