On the cost of approximating all roots of a complex polynomial

This work examines the computational complexity of a homotopy algorithm in approximating all roots of a complex polynomialf. It is shown that, probabilistically, monotonic convergence to each of the roots occurs after a determined number of steps. Moreover, in all subsequent steps, each rootz is approximated by a complex numberx, where ifx ₀ =x, x _j =x _j–1 –f(x _j–1)/f(x _j–1),j = 1, 2,, then |x _j –z| < (1/|x ₀ –z|)|x _j–1–z|².     

Finding all real roots of a polynomial by matrix algebra and the Adomian decomposition method

In this paper, we put forth a combined method for calculation of all real zeroes of a polynomial equation through the Adomian decomposition method equipped with a number of developed theorems from matrix algebra. These auxiliary theorems are associated with eigenvalues of matrices and enable convergence of the Adomian decomposition method toward different real roots of the target polynomial equation. To further improve the computational speed of our technique, a nonlinear convergence accelerator known as the Shanks transform has optionally been employed. For the sake of illustration, a number of numerical examples are given.     

A method for computing all the zeros of a polynomial with real coefficients

The problem of finding all the zeros of a polynomialP _n (x)=x ⁿ +a _n–1 x ^n–1+...+a ₁ x+a ₀, where the coefficientsa _i are real, can be posed as a system ofn nonlinear equations. The structure of this system allows an efficient numerical solution using a damped Newton method; in particular it is possible to generate the triangular factors of the associated Jacobian matrix directly. This approach provides a natural generalisation of the well-known method of Durand and Kerner.     

A PL homotopy for finding all the roots of a polynomial

This paper presents a constructive method which gives, for any polynomialF(Z) of the degreen, approximate values of all the roots ofF(Z).. The point of the method is on the use of a piecewise linear function (Z, t) which approximates a homotopyH(Z, t) betweenF(Z) and a polynomialG(Z) of the degreen withn known simple roots. It is shown that the set of solutions to (Z, t) = 0 includesn distinct paths,m of which converges to a root ofF(Z) if and only if the root has the multiplicitym. Starting from givenn roots ofG(Z), a complementary pivot algorithm generates thosen paths.This work was supported by grants from Management Science Development Foundation and Takeda Science Foundation.     

On arrangements of roots for a real hyperbolic polynomial and its derivatives

In this paper we count the number ?_n^(0,k), k?n−1, of connected components in the space Δ_n^(0,k) of all real degree n polynomials which a) have all their roots real and simple; and b) have no common root with their kth derivatives. In this case, we show that the only restriction on the arrangement of the roots of such a polynomial together with the roots of its kth derivative comes from the standard Rolle's theorem. On the other hand, we pose the general question of counting all possible root arrangements for a polynomial p(x) together with all its nonvanishing derivatives under the assumption that the roots of p(x) are real. Already the first nontrivial case n=4 shows that the obvious restrictions coming from the standard Rolle's theorem are insufficient. We prove a generalized Rolle's theorem which gives an additional restriction on root arrangements for polynomials.     

Computing the real roots of a polynomial by the exclusion algorithm

We describe a new algorithm for localizing the real roots of a polynomialP(x). This algorithm determines intervals on whichP(x) does not possess any root. The remainder set contains the real roots ofP(x) and can be arbitrarily small.     

An approximating polynomial algorithm for a sequence partitioning problem

We consider a strongly NP-hard problem of partitioning a finite sequence of vectors in Euclidean space into two clusters using the criterion of minimum sum-of-squares of distances from the elements of clusters to their centers. We assume that the cardinalities of the clusters are fixed. The center of one cluster has to be optimized and is defined as the average value over all vectors in this cluster. The center of the other cluster lies at the origin. The partition satisfies the condition: the difference of the indices of the next and previous vectors in the first cluster is bounded above and below by two given constants. We propose a 2-approximation polynomial algorithm to solve this problem.     

On the minimum real roots of the adjoint polynomial of a graph

For a graph G, we denote by h(G,x) the adjoint polynomial of G. Let β(G) denote the minimum real root of h(G,x). In this paper, we characterize all the connected graphs G with .     

Finding roots of a real polynomial simultaneously by means of Bairstow's method

Aberth's method for finding the roots of a polynomial was shown to be robust. However, complex arithmetic is needed in this method even if the polynomial is real, because it starts with complex initial approximations. A novel method is proposed for real polynomials that does not require any complex arithmetic within iterations. It is based on the observation that Aberth's method is a systematic use of Newton's method. The analogous technique is then applied to Bairstow's procedure in the proposed method. As a result, the method needs half the computations per iteration than Aberth's method. Numerical experiments showed that the new method exhibited a competitive overall performance for the test polynomials.     

A note about the average number of real roots of a Bernstein polynomial system

We prove that the real roots of normal random homogeneous polynomial systems with n+1$n+1$ variables and given degrees are, in some sense, equidistributed in the projective space P(Rⁿ⁺¹)$\mathbb{P}\left({\mathbb{R}}^{n+1}\right)$. From this fact we compute the average number of real roots of normal random polynomial systems given in the Bernstein basis.     

A unified semilocal convergence analysis of a family of iterative algorithms for computing all zeros of a polynomial simultaneously

In this paper, we first present a family of iterative algorithms for simultaneous determination of all zeros of a polynomial. This family contains two well-known algorithms: Dochev-Byrnev’s method and Ehrlich’s method. Second, using Proinov’s approach to studying convergence of iterative methods for polynomial zeros, we provide a semilocal convergence theorem that unifies the results of Proinov (Appl. Math. Comput. 284: 102–114, 2016) for Dochev-Byrnev’s and Ehrlich’s methods.     

High order iterative methods for approximating square roots

Given any natural numberm 2, we describe an iteration functiong _m (x) having the property that for any initial iterate \sqrt \alpha $$ " align="middle" border="0"> , the sequence of fixed-point iterationx _k +1 =g _m (x _k ) converges monotonically to having anm-th order rate of convergence. Form = 2 and 3,g _m (x) coincides with Newton's and Halley's iteration functions, respectively, as applied top(x) =x ² – .This research is supported in part by the National Science Foundation under Grant No. CCR-9208371.     

Greedy splitting algorithms for approximating multiway partition problems

Given a system (V,T,f,k), where V is a finite set, is a submodular function and k2 is an integer, the general multiway partition problem (MPP) asks to find a k-partition ={V₁,V₂,...,V_k} of V that satisfiesfor all i and minimizes f(V₁)+f(V₂)+···+f(V_k), where is a k-partition of hold. MPP formulation captures a generalization in submodular systems of many NP-hard problems such as k-way cut, multiterminal cut, target split and their generalizations in hypergraphs. This paper presents a simple and unified framework for developing and analyzing approximation algorithms for various MPPs.Mathematics Subject Classification (1991): 20E28, 20G40, 20C20Acknowledgement This research is partially supported by the Scientific Grant-in-Aid from Ministry of Education, Science, Sports and Culture of Japan. The authors would like to thank the anonymous referees for their valuable comments and suggestions.     

A method for finding the roots of a polynomial

Summary Two variants of a method for finding the roots of a polynomial are described. A proof of the method for a general polynomial with complex coefficients is not based on neighbourhoods of saddle points which are sufficiently small. The speed of convergence is guaranteed since the method is a modification of the downhill method and since it can be used in combination with an arbitrary method which quickly converges in practice, but whose convergence cannot be proved.     

A symplectic method for approximating all the eigenvalues of a Hamiltonian matrix

A fast method for computing all the eigenvalues of a Hamiltonian matrix M is given. The method relies on orthogonal symplectic similarity transformations which preserve structure and have desirable numerical properties. The algorithm requires about one-fourth the number of floating-point operations and one-half the space of the standard QR algorithm. The computed eigenvalues are shown to be the exact eigenvalues of a matrix M + E where ∥E∥ depends on the square root of the machine precision. The accuracy of a computed eigenvalue depends on both its condition and its magnitude, larger eigenvalues typically being more accurate.     

Detection of positive roots of a polynomial with five parameters

We give a complete set of discriminatory criteria for a polynomial with five parameters to have a positive root. This polynomial arises from the characteristic equation of a difference equation Δⁿ(x_k+px_k−τ)+qx_k−σ=0, which is used to model population dynamics. Our investigations are self-contained and based on the theory of envelopes.     

Approximation algorithms for homogeneous polynomial optimization with quadratic constraints

In this paper, we consider approximation algorithms for optimizing a generic multi-variate homogeneous polynomial function, subject to homogeneous quadratic constraints. Such optimization models have wide applications, e.g., in signal processing, magnetic resonance imaging (MRI), data training, approximation theory, and portfolio selection. Since polynomial functions are non-convex, the problems under consideration are all NP-hard in general. In this paper we shall focus on polynomial-time approximation algorithms. In particular, we first study optimization of a multi-linear tensor function over the Cartesian product of spheres. We shall propose approximation algorithms for such problem and derive worst-case performance ratios, which are shown to be dependent only on the dimensions of the model. The methods are then extended to optimize a generic multi-variate homogeneous polynomial function with spherical constraint. Likewise, approximation algorithms are proposed with provable approximation performance ratios. Furthermore, the constraint set is relaxed to be an intersection of co-centered ellipsoids; namely, we consider maximization of a homogeneous polynomial over the intersection of ellipsoids centered at the origin, and propose polynomial-time approximation algorithms with provable worst-case performance ratios. Numerical results are reported, illustrating the effectiveness of the approximation algorithms studied.     

Fast algorithms for discrete polynomial transforms

Consider the Vandermonde-like matrix , where the polynomials satisfy a three-term recurrence relation. If are the Chebyshev polynomials , then coincides with . This paper presents a new fast algorithm for the computation of the matrix-vector product in arithmetical operations. The algorithm divides into a fast transform which replaces with and a subsequent fast cosine transform. The first and central part of the algorithm is realized by a straightforward cascade summation based on properties of associated polynomials and by fast polynomial multiplications. Numerical tests demonstrate that our fast polynomial transform realizes with almost the same precision as the Clenshaw algorithm, but is much faster for .