Computer Algebra and Bifurcations

In this paper we will present the family of Newton algorithms. From the computer algebra point of view, the most basic of them is well known for the local analysis of plane algebraic curves f(x,y)=0 and consists in expanding y as Puiseux series in the variable x. A similar algorithm has been developped for multi-variate algebraic equations and for linear differential equations, using the same basic tools: a “regular” case, associated with a “simple” class of solutions, and a “simple” method of calculus of these solutions; a Newton polygon; changes of variable of type ramification; changes of unknown function of two types y=ct ^μ+? or y=exp?(c/t ^μ)?. Our purpose is first to define a “regular” case for nonlinear implicit differential equations f(t,y,y′)=0. We will then apply the result to an explicit differential equation with a parameter y′=f(y,α) in order to make a link between the expansions of the solutions obtained by our local analysis and the classical theory of bifurcations.     

Rational and algebraic approximations of algebraic numbers and their application

Effective rational and algebraic approximations of a large class of algebraic numbers are obtained by Thue-Siegel’s method. As an application of this result, it is proved that: if D>0 is not a square, and ε =x ₀ denotes the fundamental solution ofx ²?Dy ²=?1, thenx ²+1=Dy ⁴ is solvable if and only ify ₀=A ², where A is an integer. Moreover, if ε>64, thenx ²+1=Dy ⁴ has at most one positive integral solution (x, y).     

Pillai's conjecture revisited

We prove a generalization of an old conjecture of Pillai (now a theorem of Stroeker and Tijdeman) to the effect that the Diophantine equation 3^x−2^y=c has, for |c|>13, at most one solution in positive integers x and y. In fact, we show that if N and c are positive integers with N?2, then the equation |(N+1)^x−N^y|=c has at most one solution in positive integers x and y, unless (N,c)∈{(2,1),(2,5),(2,7),(2,13),(2,23),(3,13)}. Our proof uses the hypergeometric method of Thue and Siegel and avoids application of lower bounds for linear forms in logarithms of algebraic numbers.     

Exponentially fitted one-step methods for the numerical solution of the scalar Riccati equation

Using stability analysis and information from the constant coefficient problem, we motivate an explicit exponentially fitted one-step method to approximate the solution of a scalar Riccati equation ϵy′ = c(x)y² + d(x)y + e(x), 0 < x ⩽ x, y(0) = y₀, where ϵ > 0 is a small parameter and the coefficients c, d and e are assumed to be real valued and continuous. An explicit Euler-type scheme is presented which, when applied to the numerical integration of the continuous problem, give solutions satisfying a uniform (in ϵ) error estimate with order one (where suitable restrictions are imposed on the coefficients c, d and e together with the choice of y(0)). Using a counterexample, we show that, for a particular class of problems, the solutions of the fitted scheme do not converge uniformly (in ϵ) to the corresponding solutions of the continuous problems. Numerical results are presented which compare the fitted scheme with a number of implicit schemes when applied to the numerical integration of some sample problems.     

Functions with Nonzero Wronskian form a fundamental solution set

This self-contained note could find classroom use in a course on differential equations. It is proved that if y₁(x) and y₂(x) are C ² -functions whose Wronskian is never zero for α < x < β, then y₁ and y₂ form a fundamental solution set for a uniquely determined second-order linear homogeneous ordinary differential equation, y″ + p(x)y′ + q(x)y = 0, whose coefficients, p(x) and q(x), are continuous on (α, β).     

On the Solutions of a Second Order Differential Equation Arising in the Theory of Resonant Oscillations in a Tank

The properties of the solutions of the differential equation y″ = y² ? x² ? c subject to the condition that y is bounded for all finite x discussed. The arguments of Holmes and Spence have been used by Ockendon, Ockendon, and Johnson to show that there are no solutions if c is large and negative. Numerically we find that solutions exist provided c is greater than a critical value c* and estimate this value to be c* = ?…. As x tends to + ∞ the solutions are asymptotic to . The relation between A₊ and ?₊ are found analytically as A₊ → ∞. This problem arises as a connection problem in the theory of resonant oscillations of water waves.     

The diophantine equation x3 + 3y3 = 2n

It is proved that the equation of the title has a finite number of integral solutions (x, y, n) and necessary conditions are given for (x, y, n) in order that it can be a solution (Theorem 2). It is also proved that for a given odd x₀ there is at most one integral solution (y, n), n ≥ 3, to x₀³ + 3y³ = 2ⁿ and for a given odd y₀ there is at most one integral solution (x, n), n ≥ 3, to x³ + 3y₀³ = 2ⁿ.     

Asymptotic behaviour of solutions near a turning point: The example of the Brusselator equation

The Brusselator equation is an example of a singularly perturbed differential equation with an additional parameter. It has two turning points: at x=0 and x=-1. We study some properties of so-called canard solutions, that remain bounded in a full neighbourhood of 0 and in the largest possible domain; the main goal is the complete asymptotic expansion of the difference between two values of the additional parameter corresponding to such solutions. For this purpose we need a study of behaviour of the solutions near a turning point; here we prove that, for a large class of equations, if 0 is a turning point of order p, any solution y not exponentially large has, in some sector centred at 0, an asymptotic behaviour (when ε→0) of the form ∑Y_n(x/ε^′)ε^′n, where ε^′p+1=ε, for x=ε^′X with X large enough, but independent of ε. In the Brusselator case, we moreover compute a Stokes constant for a particular nonlinear differential equation.     

Extending Elliptic Curve Chabauty to higher genus curves

We give a generalization of the method of “Elliptic Curve Chabauty” to higher genus curves and their Jacobians. This method can sometimes be used in conjunction with covering techniques and a modified version of the Mordell–Weil sieve to provide a complete solution to the problem of determining the set of rational points on an algebraic curve Y. We show how to apply these explicitly by using them to prove that the equation y ² = (x ³ + x ² ? 1) Φ₁₁(x) has no rational solutions.     

On the oscillation of Hill's equations under periodic forcing

We use the Floquet theory of the Hill's equation to prove the conjecture that all solutions of the second order forced linear differential equation y^″+c(sint)y=cost, are oscillatory on [0,∞) for all c≠0.     

An approximation property of exponential functions

We solve the inhomogeneous linear first order differential equations of the form y′(x) ? λy(x) = Σ _m=0 ^∞ a _m (x ? c) ^m , and prove an approximation property of exponential functions. More precisely, we prove the local Hyers-Ulam stability of linear first order differential equations of the form y′(x) = λy(x) in a special class of analytic functions.     

On a linear transcendence measure for the solutions of a universal differential equation at algebraic points

In this paper the author continues his work on arithmetic properties of the solutions of a universal differential equation at algebraic points. Every real continuous function on the real line can be uniformly approximated by C^∞-solutions of a universal differential equation. An algebraic universal differential equation of order five and degree 11 is explicitly given, such that every finite set of nonvanishing derivatives y^(k₁)(τ),…,y^(k_r)(τ) (1?k₁<?<k_r) at an algebraic point τ is linearly independent over the field of algebraic numbers. A linear transcendence measure for these values is effectively computed.     

Exact number of local extreme points of curvature function of solutions of second-order linear differential equations

We study local properties of the curvature ?? _y (x) of every nontrivial solution y=y(x) of the second-order linear differential equation?(P): (p(x)y??)??+q(x)y=0, x??(a,b)=I, where p(x) and q(x) are smooth enough functions. It especially includes the Euler, Bessel and other important types of second-order linear differential equations. Some sufficient conditions on the coefficients p(x) and q(x) are given such that the curvature ?? _y (x) of every nontrivial solution y of (P) has exactly one extreme point between each two its consecutive simple zeros. The problem of three local extreme points of ?? _y (x) is also considered but only as an open problem. It seems it is the first paper dealing with this kind of problems. Finally in Appendix, we pay attention to an application of the main results to a study of non-regular points (the cusps) of the ??-parallels of graph ??(y) of?y (the offset curves of???(y)).     

The Hukuhara-Kneser property for parabolic systems with nonlinear boundary conditions

We analyze the asymptotic behavior as x → ∞ of the product integral Π_x0^xe^A(s)ds, where A(s) is a perturbation of a diagonal matrix function by an integrable function on [x₀,∞). Our results give information concerning the asymptotic behavior of solutions of certain linear ordinary differential equations, e.g., the second order equation y″ = a(x)y.     

Multiple positive solutions to a singular boundary value problem for a superlinear Emden-Fowler equation

A multiplicity result for the singular ordinary differential equation y^″+λx⁻²yσ=0, posed in the interval (0,1), with the boundary conditions y(0)=0 and y(1)=γ, where σ>1, λ>0 and γ?0 are real parameters, is presented. Using a logarithmic transformation and an integral equation method, we show that there exists Σ^?∈(0,σ/2] such that a solution to the above problem is possible if and only if λγσ−1?Σ^?. For 0<λγσ−1<Σ^?, there are multiple positive solutions, while if γ=(λ⁻¹Σ^?)^1/(σ−1) the problem has a unique positive solution which is monotonic increasing. The asymptotic behavior of y(x) as x→⁰⁺ is also given, which allows us to establish the absence of positive solution to the singular Dirichlet elliptic problem −Δu=d⁻²(x)uσ in Ω, where Ω⊂RN, N?2, is a smooth bounded domain and d(x)=dist(x,∂Ω).     

Sixth-order superstable two-step methods for second-order initial-value problems

For the numerical integration of general second-order initial-value problems y″ = f(x, y, y′), y(x₀) = y₀, y′(x₀) = y′₀, we report a family of two-step sixth-order methods which are superstable for the test equation y″ + 2αy′ + β²y = 0, α, β ⩾ 0, α + β\s>0, in the sense of Chawla [1].     

Lamé equations with algebraic solutions

In this paper we study Lamé equations L_n,By=0 in so-called algebraic form, having only algebraic functions as solution. In particular we provide a complete list of all finite groups that occur as the monodromy groups, together with a list of examples of such equations. We show that the set of such Lamé equations with is countable, up to scaling of the equation. This result follows from the general statement that the set of equivalent second-order equations, having algebraic solutions and all of whose integer local exponent differences are 1, is countable.     

Description of the algebraic structure of second-order differential operators that can be represented in divergence form

In the paper one proves that the second-order differential operators ?(x,u,u _x ,u _xx ), representable in divergence form, can be written as ?=c^AΔ_A, where Δ_A is the corresponding minor of order m of the Hessian \(\det (u_{xx} ) = \Delta _{\left( {\begin{array}{*{20}c} {1...n} \\ {1...n} \\ \end{array} } \right)} \) whose representation coefficients c^A=c^A(x,u,u _x ) satisfy certain algebraic relations. One introduces the concept of a second-order D-elliptic differential operator and one establishes that the totally elliptic operators of divergence form are linear with respect to the second derivatives of the function u.     

Differential transformations of parabolic second-order operators in the plane

Darboux’s classical results about transformations of second-order hyperbolic equations by means of differential substitutions are extended to the case of parabolic equations of the form Lu = (D _x ² + a(x, y)D _x + b(x, y)D _y + c(x, y))u = 0. We prove a general theorem that provides a way to determine admissible differential substitutions for such parabolic equations. It turns out that higher order transforming operators can always be represented as a composition of first-order operators that define a series of consecutive transformations. The existence of inverse transformations imposes some differential constrains on the coefficients of the initial operator. We show that these constraints may imply famous integrable equations, in particular, the Boussinesq equation.     

High-order one-step P-stable methods for the numerical integration of periodic initial value problems

Using Lobatto nodes, one-step methods of order six and eight have been obtained for the second-order differential equation y″ = f(x, y), y(x₀) = y₀, y′(x₀) = y′₀. The methods are shown to be P-stable. If

, then at each integration step a system of dimension 3s, 4s, respectively, has to be solved. The numerical results, for two problems, obtained by using these methods are given in the end.