No label defined (Q4239088)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1277444
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 1277444 |
Statements
20 April 1999
0 references
ring of twice differentiable real functions
0 references
differential ring
0 references
differential ideal
0 references
0.8389626145362854
0 references
0.7552013397216797
0 references
0.7522289752960205
0 references
0.75107741355896
0 references
We consider a system of real algebraic differential equations with constant coefficients, i.e. polynomials in \(y,y',y'', \dots\) over the real numbers. Now it is possible to manipulate these equations by multipliying them with each other or differentiating them, and so on. - For example, we take the system \(S:=(y'{}^2-y'{}^3 = 0\), \(y'{}^2 +2y''y'+1 =0)\). Then differentiate the first equation and subtract it from the second equation we obtain the equation \(4y'{}^2 +1=0\). Clearly, this equation has no solution in the ring of twice differentiable real functions on any interval. We conclude that each twice differentiable function which solves one equation cannot solve the other equation. In particular it follows that \(S\) has no \(C^\infty\)-solution. At this point we come into contact with the problem that manipulating differential equations in this way may produce equations of the type 1+sum of squares=0. Hence if we are able to derive an equation \(f=0\) of this type, then there cannot exist a \(C^\infty\)-solution of the original system. Since, every solution of the original system must also be a solution of the equation \(f=0\), but this is impossible. -- It is not trivial to ask wether the reverse direction is also true. That means if a system \(S\) of algebraic differential equations has no \(C^\infty\)-solutions, then we ask the question: Is it possible to produce -- by applying to \(S\) manipulations of the type as illustrated above -- an equation of the type 1+sum of squares =0? It is the main result of the first chapter of this thesis that the answer is yes. So we have obtained an equivalent condition for the existence of solutions of algebraic differential equations, which is formulated in pure terms of algebraic operations and differentiation. Since we are dealing with algebraic operations plus differentiation, we are led to the notion of a differential ring. This is a commutative ring with unit together with an additive map \(\delta:A\to A\) such that the product rule holds, i.e. \(\delta (ab)=a \delta b+b \delta a\). This setting enables us to make the problem of the appearance of differential equations of the type 1+sums of squares =0 more concrete. Namely this problem is reflected in the property whether or not a differential ring, or equivalently a differential ideal, is semi-real. In fact we can prove the following result (corollary 6.1): Let \(\mathbb{R}[X]\{y_1,\dots,y_n\}\) be the ring of differential polynomials over \(\mathbb{R}[X]\). We consider the system \(f_1(X,y_1, \dots, y_n)=0, \dots, f_k(X,y_1, \dots, y_n)=0\) of algebraic differential equations. Let \({\mathfrak A}\) be the differential ideal generated by \(f_1,\dots,f_k\) in \(\mathbb{R}[X] \{y_1, \dots, y_n\}\). Then \({\mathfrak A}\) is a semi real ideal if and only if there exist an open interval \(I\) and an analytic map \(c:I\to \mathbb{R}^n\) such that \(f_1(t,c(t)) =0,\dots, f_k (t,c(t))=0\) for every \(t\in I\). Moreover, by combining methods of differential algebra and real algebra, we are able to generalize our above result to systems of equations and inequations as follows (corollary 6.2). We consider the system \(f_1 (X,y_1, \dots, y_n)=0, \dots, f_k(X,y_1, \dots, y_n=0\), \(g(X,y_1, \dots, y_n) \neq 0\). Then there exist an open interval \(I\) and an analytic map \(c:I\to \mathbb{R}^n\) such that \(f_1(t,c(t)) =0, \dots, f_k(t,c(t)) =0\), \(g(t,c(t)) \neq 0\) for every \(t\in I\) iff \(g\notin \root {re}\of {[f_1, \dots, f_k]}\) in \(\mathbb{R} [X]\{y_1, \dots, y_n\}\).
0 references