On the location of the zeros of a polynomial (Q5906567): Difference between revisions
From MaRDI portal
Added link to MaRDI item. |
Set profile property. |
||
Property / MaRDI profile type | |||
Property / MaRDI profile type: MaRDI publication profile / rank | |||
Normal rank |
Revision as of 02:57, 5 March 2024
scientific article; zbMATH DE number 617509
Language | Label | Description | Also known as |
---|---|---|---|
English | On the location of the zeros of a polynomial |
scientific article; zbMATH DE number 617509 |
Statements
On the location of the zeros of a polynomial (English)
0 references
18 August 1994
0 references
It is known [see \textit{S. Kakeya}, Tôhoku Math. J. 2, 140-142 (1912- 1913)] that if \(p(z) = \sum^ n_{\nu = 0} a_ \nu z^ \nu\) is a polynomial of degree \(n\) with real coefficients satisfying \(0<a_ 0 \leq a_ 1 \leq \cdots \leq a_ n\), then all the zeros of \(p(z)\) lie in \(| z | \leq 1\). In this paper the authors give the generalization of the mentioned result: Theorem. Let \(p(z) = \sum^ n_{\nu = 0} a_ \nu z^ \nu\) be a polynomial of degree \(n\). If \(\text{Re} a_ j = \alpha_ j\) and \(\text{Im} a_ j = \beta_ j\) for \(j = 0,1,2, \dots,n\), \(a_ n \neq 0\) and for some \(k\) and \(r\) and for some \(t \geq 0\) \[ \alpha_ 0 \leq t \alpha_ 1 \leq t^ 2 \alpha_ 2 \leq \cdots \leq t^ k \alpha_ k \geq t^{k + 1} \alpha_{k + 1} \geq \cdots \geq t^ n \alpha_ n, \] \[ \beta_ 0 \leq t \beta_ 1 \leq t^ 2 \beta_ 2 \leq \cdots \leq t^ r \beta_ r \geq t^{r + 1} \beta_{r + 1} \geq \cdots \geq t^ n \beta_ n, \] then \(p(z)\) has all its zeros in \(R_ 1 \leq | z | \leq R_ 2\), where \[ R_ 1 = \min \Bigl \{t | a_ 0 |/ \biggl[ 2(t^ k \alpha_ k + t^ r \beta_ r)- (a_ 0 + \beta_ 0) - t^ n \bigl( \alpha_ n + \beta_ n - | a_ n | \bigr) \biggr], t \Bigr\}, \] \[ R_ 2 = \max \left\{ \left[ | a_ 0 | t^{n + 1} - t^{n - 1} (\alpha_ 0 + \beta_ 0) - t (\alpha_ n + \beta_ n) + \right. \right. \] \[ (t^ 2 + 1) (t^{n - k - 1} \alpha_ k + t^{n - r + 1} \beta_ r) + (t^ 2 - 1) \left( \sum^{k - 1}_{j = 1} t^{n - j - 1} \alpha_ j + \sum^{r - 1}_{j = 1} t^{n - j - 1} \beta_ j \right) + \] \[ \left. \left. (1 - t^ 2) \left( \sum^{n - 1}_{j = k + 1} t^{n - j - 1} \alpha_ j + \sum^{n - 1}_{j = k + 1} t^{n - j - 1} \alpha_ j + \sum^{n - 1}_{j = r + 1} t^{n - j - 1} \beta_ j \right) \right] \left/ | a_ n |, {1 \over t} \right. \right\}. \] If all \(\beta_ j = 0\) and \(0<a_ 0 \leq \alpha_ 1 \leq \cdots \leq \alpha_ n\) then \(p(z)\) has all its zeros in \({\alpha_ 0 \over 2 \alpha_ n - \alpha_ 0} \leq | z | \leq 1\).
0 references