On modified hybrid steepest-descent methods for general variational inequalities

We consider the general variational inequality GVI(F,g,C), where F and g are mappings from a Hilbert space into itself and C is intersection of the fixed point sets of a finite family of nonexpansive mappings. We suggest and analyze an iterative algorithm with variable parameters as follows:

     

Convergence of Hybrid Steepest-Descent Methods for Generalized Variational Inequalities

In this paper, we consider the generalized variational inequality GVI（F, g, C）, where F and g are mappings from a Hilbert space into itself and C is the fixed point set of a nonexpansive mapping. We propose two iterative algorithms to find approximate solutions of the GVI（F,g, C）. Strong convergence results are established and applications to constrained generalized pseudo-inverse are included.     

A globally and superlinearly convergent quasi-Newton method for general box constrained variational inequalities without smoothing approximation

A new quasi-Newton algorithm for the solution of general box constrained variational inequality problem (GVI(l, u, F, f)) is proposed in this paper. It is based on a reformulation of the variational inequality problem as a nonsmooth system of equations by using the median operator. Without smoothing approximation, the proposed quasi-Newton algorithm is directly applied to solve this class of nonsmooth equations. Under appropriate assumptions, it is proved that the algorithmic sequence globally and superlinearly converges to a solution of the equation reformulation and also of GVI(l, u, F, f). Numerical results show that our new algorithm works quite well.     

Über Funktional-Differentialgleichungen mit voreilendem Argument

We consider a functional differential equation (1)u′(t)=F(t,u) for )≤t≤∞ together with a generalized initial condition (2)u(t)=?(t) forr≤t≤0 or a generalized Nicoletti condition (3)N u=η. Here,N is a linear operator; in the case of a system ofn equations the classical Nicoletti operator is given byN u=(u ₁(t ₁),...,u _n(t _n)), with givent _i. The functionsu, F ? are Banach space valued, the functionF(t, z) is defined fort≥0 andz∈C ⁰[r,∞). The main point is that the value ofF(t, z) may depend on the values ofz(s) forr≤s≤t+σ(t), where σ(t)>0. Simple examples show that without a restriction on the magnitude of the advancement σ(t) there is neither existence nor uniqueness. Our results show that when σ(t) is properly bounded and when the solution is to satisfy a certain growth condition which depends on σ(t), then there exists exactly one solution, and it depends continuously on the given data. In the case of the Nicoletti problem (1), (3) there is convergence to the solution satisfyingu(0)=η if 0≤t _i≤T andT→0 (this holds in infinite-dimensional spaces, too). These results are true ifF satisfies a Lipschitz condition of the form $$\left| {F(t,z) - F(t,y)} \right| \leqslant h(t)\max \left\{ {\left| {z(s) - y(s)} \right|:r \leqslant s \leqslant t + \delta (t)} \right\}.$$ . In the case where (1) is a finite system andF is only continuous, an existence theory is developped based onSchauder's fixed point theorem. Again, growth conditions play an essential role here.     

Viscosity approximation methods for nonexpansive mappings and monotone mappings

Viscosity approximation methods for nonexpansive mappings are studied. Consider the iteration process {xn}, where x₀∈C is arbitrary and xn+1=αnf(xn)+(1−αn)SPC(xn−λnAxn), f is a contraction on C, S is a nonexpansive self-mapping of a closed convex subset C of a Hilbert space H. It is shown that {xn} converges strongly to a common element of the set of fixed points of nonexpansive mapping and the set of solutions of the variational inequality for an inverse strongly-monotone mapping which solves some variational inequality.     

Asymptotic behavior of eigenvalues of two-parameter nonlinear Sturm-Liouville problems

The nonlinear two-parameter Sturm-Liouville problemu ^"+μg(u)=λf(u) is studied for μ, λ>0. By using Ljusternik-Schnirelman theory on the general level set developed by Zeidler, we shall show the existence of ann-th variational eigenvalue λ=λ_n(μ). Furthermore, for specialf andg, the asymptotic formula of λ₁(μ)) as μ→∞ is established.     

A generalized weight for linear codes and a Witt-MacWilliams theorem

Let F be a finite field, H a subgroup of $\text{F}{}^1$ of index ν, and α₁,…, α_ν coset representatives. For each n-tuple u = (u₁,…, u_n) ?Fⁿ define W_H(u) = (w₁(u),…, w_ν(u)), where w_m(u) = #{u_i: u_i?α_mH}. An H-monomial map on Fⁿ is an automorphism of Fⁿ whose matrix with respect to the co-ordinate basis is of the form P · D, where P is a permutation matrix and D is a diagonal matrix with non-zero entries from H. Suppose C is an (n, k) code over F (that is, a k-dimensional subspace of Fⁿ) and that ?: C → Fⁿ is an injective homomorphism which preserves W_H in the sense that W_H(?(u)) = W_H(u) for all u ?C. We prove that ? may be extended to an H-monomial map on Fⁿ. This generalization of a theorem of MacWilliams on the (Hamming) equivalence of codes may be considered an analogue of the Witt theorem of metric vector space theory.     

On rational invariants of a set of symmetric matrices

Some identities resulting from the Cayley-Hamilton theorem are derived. Some applications include: (a) for k = 1,2,…,n ? 1 a condition is found for a pair (A,B) of symmetric operators acting in Euclidean n-space to have common invariant k-subspace (provided that A does not have multiple eigenvalues); (b) it is shown that the field of rational invariants of (A,B) is isomorphic to a subfield of a rational function field with n(n+3)/2 generators consisting of elements symmetric with respect to the permutaion group P_n; (c) it is shown that any rational invariant of (g+2) symmetric operators A,B,C₁,C₂,…, C_g can be expressed as a rational function of invariants of one or two operators that are taken for pairs (A,B), (A,C₂),…, (A,C_g, (A,B+C₁), (A,B+C₂),…,(A,B+C_g).     

Separation of variables of a generalized porous medium equation with nonlinear source

This paper considers a general form of the porous medium equation with nonlinear source term: u_t=(D(u)u_xⁿ)_x+F(u), n≠1. The functional separation of variables of this equation is studied by using the generalized conditional symmetry approach. We obtain a complete list of canonical forms for such equations which admit the functional separable solutions. As a consequence, some exact solutions to the resulting equations are constructed, and their behavior are also investigated.     

Uniform asymptotic expansion for a class of polynomials biorthogonal on the unit circle

An asymptotic expansion including error bounds is given for polynomials {P _n, Q_n} that are biorthogonal on the unit circle with respect to the weight function (1?e^iθ)^α+β(1?e^?iθ)^α?β. The asymptotic parameter isn; the expansion is uniform with respect toz in compact subsets ofC{0}. The pointz=1 is an interesting point, where the asymptotic behavior of the polynomials strongly changes. The approximants in the expansions are confluent hyper-geometric functions. The polynomials are special cases of the Gauss hyper-geometric functions. In fact, with the results of the paper it follows how (in a uniform way) the confluent hypergeometric function is obtained as the limit of the hypergeometric function₂ F ₁(a, b; c; z/b), asb→±∞,z≠b, withz=0 as “transition” point in the uniform expansion.     

非扩张映射和广义变分不等式的粘滞逼近法

应用已提出的非扩张映射的粘滞逼近方法,给定初值x_0∈C,考虑一般迭代过程{x_n},g(x_(n+1))=α_nf(x_n)+(1-α_n)SP_C(g(x_n)-λ_nAx_n),n≥0,其中{α_n}■(0,1),S:C→C是非扩张映射,C是实Hilbert空间H的非空闭凸子集.在{α_n}满足合适的条件下可证明,{x_n}强收敛到非扩张映射的不动点集和广义变分不等式解的公共元,且满足某变分不等式.     

On a generalization of Minkowski's convex body theorem

Given a lattice $\text{Λ ? R}{}^{\mathrm{n}}$ and a bounded function g(x), x ∈ Rⁿ, vanishing outside of a bounded set, the functions ?(x)$\text{g}\text{?}\text{(x)?}$max_u∈Λg(u +x), ?(x)?Σ_{u∈Λ g(u +x)}, and ?₊(x)?Σ_u∈Λ max_{v∈Λ min {g(v + x); g(u + v + x)}} are defined and periodic mod Λ on Rⁿ. In the paper we prove that ?(x) + ?₊(x) ? 2?(x) ≥ ?(x) + h?₊(x) ? 2?(x) holds for all x ∈ Rⁿ, where h(x) is any “truncation” of g by a constant c ≥ 0, i.e., any function of the form h(x)?g(x) if g(x) ≤ c and h(x)?c if g(x) > c. This inequality easily implies some known estimations in the geometry of numbers due to Rado [1] and Cassels [2]. Moreover, some sharper and more general results are also derived from it. In the paper another inequality of a similar type is also proved.     

Über komplementäre Ungleichungen mit drei Mittelwerten. Teil 2

In Part 1 we obtained lower and upper bounds of the expressionf(M _φ(x;α),M _ψ(y;α))?M _χ(f(x,y);α) by replacing the given sets(x)=(x ₁,...,x _n ),(y)=(y ₁,...,y _n ) by two suitably chosen sets ((u)=(u ₁,...,u _m ),(v)=(v ₁,...,v _m ), in general withm≥4. Now, in the case of upper bounds, the numberm will, under additional hypotheses, be reduced tom=3 (§ 4) and finally tom=2 (§ 5). Inequalities, complementary to the inequalities of Hölder and Minkowski and to another inequality are given as illustrations.     

Funktionalungleichungen und Iterationsverfahren

This paper is concerned with a class of iterative processes of the formu _k+1 =Tu _k (k = 0, 1, ?) for solving nonlinear operator equationsu = Tu orFu = 0. By studying the relationship between a linear functional inequality?(Ah) β(h) + γ(h) ? ?(h) and estimates for the iteration operatorT a general semilocal convergence theorem is obtained. The theorem contains as special cases theorems for various iterative methods. Numerical examples illustrate the accuracy of the error estimates for the approximationu _k .     

On the g-Ary Expansions of Apéry, Motzkin, Schröder and Other Combinatorial Numbers

Let g ≥ 2 be an integer and let (u _n ) _n≥1  be a sequence of integers which satisfies a relation u _n+1 = h(n)u _n for a rational function h(X). For example, various combinatorial numbers as well as their products satisfy relations of this type. Here, we show that under some mild technical assumptions the number of nonzero digits of u _n in base g is large on a set of n of asymptotic density 1. We also extend this result to a class of sequences satisfying relations of second order u _n+2 = h ₁(n)u _n+1 + h ₂(n)u _n with two nonconstant rational functions ${h_1(X), h_2(X) \in {\mathbb Q} [X]}Let g ≥ 2 be an integer and let (u _n ) _n≥1  be a sequence of integers which satisfies a relation u _n+1 = h(n)u _n for a rational function h(X). For example, various combinatorial numbers as well as their products satisfy relations of this type. Here, we show that under some mild technical assumptions the number of nonzero digits of u _n in base g is large on a set of n of asymptotic density 1. We also extend this result to a class of sequences satisfying relations of second order u _n+2 = h ₁(n)u _n+1 + h ₂(n)u _n with two nonconstant rational functions h₁(X), h₂(X) ? \mathbb Q [X]{h_1(X), h_2(X) \in {\mathbb Q} [X]}. This class includes the Apéry, Delannoy, Motzkin, and Schr?der numbers.     

A multipoint method of third order

LetF be a mapping of the Banach spaceX into itself. A convergence theorem for the iterative solution ofF(x)=0 is proved for the multipoint algorithmx _n+1=x _n ?ø(x _n ), where $$\phi (x) = F\prime_x^{ - 1} \left[ {F(x) + F\lgroup {x - F\prime_x^{ - 1} F(x)} \rgroup} \right]$$ andF′x is the Frechet derivative ofF. The theorem guarantees that, under appropriate conditions onF, the multipoint sequence {x _n } generated by ø converges cubically to a zero ofF. The algorithm is applied to the nonlinear Chandrasekhar integral equation $$\frac{1}{2}\omega _0 x(t)\int_0^1 {\frac{{tx(s)}}{{s + t}}ds - x(t) + 1 = 0}$$ where ω₀>0. A discretization of the equations of iteration is discussed, and some numerical results are given.     

The algebraic connectivity of lollipop graphs

Let C_n,g be the lollipop graph obtained by appending a g-cycle C_g to a pendant vertex of a path on n-g vertices. In 2002, Fallat, Kirkland and Pati proved that for and g?4, α(C_n,g)>α(C_n,g-1). In this paper, we prove that for g?4, α(C_n,g)>α(C_n,g-1) for all n, where α(C_n,g) is the algebraic connectivity of C_n,g.     

Approximation from Above of Systems of Differential Inclusions with Non-Lipschitzian Right-Hand Side

Suppose that ? ⁿ is the p-dimensional space with Euclidean norm ∥ ? ∥, K (? ^p ) is the set of nonempty compact sets in ? ^p , ?₊ = [0, +∞), D = ?₊ × ? ^m × ? ⁿ × [0, a], D ₀ = ?₊ × ? ^m , F ₀: D ₀ → K (? ^m ), and co F ₀ is the convex cover of the mapping F ₀. We consider the Cauchy problem for the system of differential inclusions $$\dot x \in \mu F(t,x,y,\mu ),\quad \dot y \in G(t,x,y,\mu ),\quad x(0) = x_0 ,\quad y(0) = y_0$$ with slow x and fast y variables; here F: D → K (? ^m ), G: D → K (? ⁿ ), and μ ∈ [0, a] is a small parameter. It is assumed that this problem has at least one solution on [0, 1/μ] for all sufficiently small μ ∈ [0, a]. Under certain conditions on F, G, and F ₀, comprising both the usual conditions for approximation problems and some new ones (which are weaker than the Lipschitz property), it is proved that, for any ε > 0, there is a μ₀ > 0 such that for any μ ∈ (0, μ₀] and any solution (x _μ(t), y _μ(t)) of the problem under consideration, there exists a solution u _μ(t) of the problem ${\dot u}$ ∈ μ co F ₀ (t, u), u(0) = x ₀ for which the inequality ∥x _μ(t) ? u _μ(t)∥ < ε holds for each t ∈ [0, 1/μ].     

Convergence of distributed optimal control problems governed by elliptic variational inequalities

First, let u _g be the unique solution of an elliptic variational inequality with source term g. We establish, in the general case, the error estimate between $u_{3}(\mu)=\mu u_{g_{1}}+ (1-\mu)u_{g_{2}}$ and $u_{4}(\mu)=u_{\mu g_{1}+ (1-\mu) g_{2}}$ for ????[0,1]. Secondly, we consider a family of distributed optimal control problems governed by elliptic variational inequalities over the internal energy g for each positive heat transfer coefficient h given on a part of the boundary of the domain. For a given cost functional and using some monotony property between u ₃(??) and u ₄(??) given in Mignot (J.?Funct. Anal. 22:130?C185, 1976), we prove the strong convergence of the optimal controls and states associated to this family of distributed optimal control problems governed by elliptic variational inequalities to a limit Dirichlet distributed optimal control problem, governed also by an elliptic variational inequality, when the parameter h goes to infinity. We obtain this convergence without using the adjoint state problem (or the Mignot??s conical differentiability) which is a great advantage with respect to the proof given in Gariboldi and Tarzia (Appl. Math. Optim. 47:213?C230, 2003), for optimal control problems governed by elliptic variational equalities.     

Vanishing exponential integrability for functions whose gradients belong to L(log(e+L))

If the gradient of u(x) is nth power locally integrable on Euclidean n-space, then the integral average over a ball B of the exponential of a constant multiple of |u(x)−u_B|^n/(n−1), u_B=average of u over B, tends to 1 as the radius of B shrinks to zero—for quasi almost all center points. This refines a result of N. Trudinger (1967). We prove here a similar result for the class of gradients in Lⁿ(log(e+L))^α, 0?α?n−1. The results depend on a capacitary strong-type inequality for these spaces.