An approach to constrained global optimization based on exact penalty functions

In the field of global optimization many efforts have been devoted to solve unconstrained global optimization problems. The aim of this paper is to show that unconstrained global optimization methods can be used also for solving constrained optimization problems, by resorting to an exact penalty approach. In particular, we make use of a non-differentiable exact penalty function ${P_q(x;\varepsilon)}$ . We show that, under weak assumptions, there exists a threshold value ${\bar \varepsilon >0 }$ of the penalty parameter ${\varepsilon}$ such that, for any ${\varepsilon \in (0, \bar \varepsilon]}$ , any global minimizer of P _q is a global solution of the related constrained problem and conversely. On these bases, we describe an algorithm that, by combining an unconstrained global minimization technique for minimizing P _q for given values of the penalty parameter ${\varepsilon}$ and an automatic updating of ${\varepsilon}$ that occurs only a finite number of times, produces a sequence {x ^k } such that any limit point of the sequence is a global solution of the related constrained problem. In the algorithm any efficient unconstrained global minimization technique can be used. In particular, we adopt an improved version of the DIRECT algorithm. Some numerical experimentation confirms the effectiveness of the approach.     

Primal-dual interior-point algorithm for semidefinite optimization based on a new kernel function with trigonometric barrier term

In this paper we propose primal-dual interior-point algorithms for semidefinite optimization problems based on a new kernel function with a trigonometric barrier term. We show that the iteration bounds are $O(\sqrt{n}\log(\frac{n}{\epsilon}))$ for small-update methods and $O(n^{\frac{3}{4}}\log(\frac{n}{\epsilon}))$ for large-update, respectively. The resulting bound is better than the classical kernel function. For small-update, the iteration complexity is the best known bound for such methods.     

Complexity analysis and numerical implementation of primal-dual interior-point methods for convex quadratic optimization based on a finite barrier

In this paper, we present primal-dual interior-point methods for convex quadratic optimization based on a finite barrier, which has been investigated earlier for the case of linear optimization by Bai et al. (SIAM J Optim 13(3):766–782, 2003). By means of the feature of the finite kernel function, we study the complexity analysis of primal-dual interior-point methods based on the finite barrier and derive the iteration bounds that match the currently best known iteration bounds for large- and small-update methods, namely, $O(\sqrt{n}\log{n}\log{\frac{n}{\varepsilon}})$ and $O(\sqrt{n}\log{\frac{n}{\varepsilon}})$ , respectively, which are as good as the linear optimization analogue. Numerical tests demonstrate the behavior of the algorithms with different parameters.     

Dimension Reduction for Finite Trees in \varvec{\ell _1}

We show that every $n$ -point tree metric admits a $(1+\varepsilon )$ -embedding into $\ell _1^{C(\varepsilon ) \log n}$ , for every $\varepsilon > 0$ , where $C(\varepsilon ) \le O\big ((\frac{1}{\varepsilon })^4 \log \frac{1}{\varepsilon })\big )$ . This matches the natural volume lower bound up to a factor depending only on $\varepsilon $ . Previously, it was unknown whether even complete binary trees on $n$ nodes could be embedded in $\ell _1^{O(\log n)}$ with $O(1)$ distortion. For complete $d$ -ary trees, our construction achieves $C(\varepsilon ) \le O\big (\frac{1}{\varepsilon ^2}\big )$ .     

The third moment of quadratic Dirichlet L-functions

We study the third moment of quadratic Dirichlet $L$ -functions, obtaining an error term of size $O(X^{3/4 + \varepsilon })$ .     

A Predictor-corrector algorithm with multiple corrections for convex quadratic programming

Recently an infeasible interior-point algorithm for linear programming (LP) was presented by Liu and Sun. By using similar predictor steps, we give a (feasible) predictor-corrector algorithm for convex quadratic programming (QP). We introduce a (scaled) proximity measure and a dynamical forcing factor (centering parameter). The latter is used to force the duality gap to decrease. The algorithm can decrease the duality gap monotonically. Polynomial complexity can be proved and the result coincides with the best one for LP, namely, $O(\sqrt{n}\log n\mu^{0}/\varepsilon)$ .     

On the complexity of finding first-order critical points in constrained nonlinear optimization

The complexity of finding $\epsilon $ -approximate first-order critical points for the general smooth constrained optimization problem is shown to be no worse that $O(\epsilon ^{-2})$ in terms of function and constraints evaluations. This result is obtained by analyzing the worst-case behaviour of a first-order short-step homotopy algorithm consisting of a feasibility phase followed by an optimization phase, and requires minimal assumptions on the objective function. Since a bound of the same order is known to be valid for the unconstrained case, this leads to the conclusion that the presence of possibly nonlinear/nonconvex inequality/equality constraints is irrelevant for this bound to apply.     

Welfare-maximizing correlated equilibria using Kantorovich polynomials with sparsity

We provide motivations for the correlated equilibrium solution concept from the game-theoretic and optimization perspectives. We then propose an algorithm that computes ${\varepsilon}$ -correlated equilibria with global-optimal (i.e., maximum) expected social welfare for normal form polynomial games. We derive an infinite dimensional formulation of ${\varepsilon}$ -correlated equilibria using Kantorovich polynomials, and re-express it as a polynomial positivity constraint. We exploit polynomial sparsity to achieve a leaner problem formulation involving sum-of-squares constraints. By solving a sequence of semidefinite programming relaxations of the problem, our algorithm converges to a global-optimal ${\varepsilon}$ -correlated equilibrium. The paper ends with two numerical examples involving a two-player polynomial game, and a wireless game with two mutually-interfering communication links.     

Singular perturbation method for inhomogeneous nonlinear free boundary problems

In this paper, we study solutions of one phase inhomogeneous singular perturbation problems of the type: $ F(D^2u,x)=\beta _{\varepsilon }(u) + f_{\varepsilon }(x) $ and $ \Delta _{p}u=\beta _{\varepsilon }(u) + f_{\varepsilon }(x)$ , where $\beta _{\varepsilon }$ approaches Dirac $\delta _{0}$ as $\varepsilon \rightarrow 0$ and $f_{\varepsilon }$ has a uniform control in $L^{q}, q>N.$ Uniform local Lipschitz regularity is obtained for these solutions. The existence theory for variational (minimizers) and non variational (least supersolutions) solutions for these problems is developed. Uniform linear growth rate with respect to the distance from the $\varepsilon -$ level surfaces are established for these variational and nonvaritional solutions. Finally, letting $\varepsilon \rightarrow 0$ basic properties such as local Lipschitz regularity and non-degeneracy property are proven for the limit and a Hausdorff measure estimate for its free boundary is obtained.     

An Approximation Algorithm for Computing Shortest Paths in Weighted 3-d Domains

We present an approximation algorithm for computing shortest paths in weighted three-dimensional domains. Given a polyhedral domain $\mathcal D $ D , consisting of $n$ n tetrahedra with positive weights, and a real number $\varepsilon \in (0,1)$ ε ∈ ( 0 , 1 ) , our algorithm constructs paths in $\mathcal D $ D from a fixed source vertex to all vertices of $\mathcal D $ D , the costs of which are at most $1+\varepsilon $ 1 + ε times the costs of (weighted) shortest paths, in $O(\mathcal{C }(\mathcal D )\frac{n}{\varepsilon ^{2.5}}\log \frac{n}{\varepsilon }\log ^3\frac{1}{\varepsilon })$ O ( C ( D ) n ε 2.5 log n ε log 3 1 ε ) time, where $\mathcal{C }(\mathcal D )$ C ( D ) is a geometric parameter related to the aspect ratios of tetrahedra. The efficiency of the proposed algorithm is based on an in-depth study of the local behavior of geodesic paths and additive Voronoi diagrams in weighted three-dimensional domains, which are of independent interest. The paper extends the results of Aleksandrov et al. (J ACM 52(1):25–53, 2005), to three dimensions.     

A full-Newton step feasible interior-point algorithm for P_*(\kappa )-linear complementarity problems

In this paper, a full-Newton step feasible interior-point algorithm is proposed for solving $P_*(\kappa )$ -linear complementarity problems. We prove that the full-Newton step to the central path is local quadratically convergent and the proposed algorithm has polynomial iteration complexity, namely, $O\left( (1+4\kappa )\sqrt{n}\log {\frac{n}{\varepsilon }}\right) $ , which matches the currently best known iteration bound for $P_*(\kappa )$ -linear complementarity problems. Some preliminary numerical results are provided to demonstrate the computational performance of the proposed algorithm.     

Iteration complexity of randomized block-coordinate descent methods for minimizing a composite function

In this paper we develop a randomized block-coordinate descent method for minimizing the sum of a smooth and a simple nonsmooth block-separable convex function and prove that it obtains an $\varepsilon $ -accurate solution with probability at least $1-\rho $ in at most $O((n/\varepsilon ) \log (1/\rho ))$ iterations, where $n$ is the number of blocks. This extends recent results of Nesterov (SIAM J Optim 22(2): 341–362, 2012), which cover the smooth case, to composite minimization, while at the same time improving the complexity by the factor of 4 and removing $\varepsilon $ from the logarithmic term. More importantly, in contrast with the aforementioned work in which the author achieves the results by applying the method to a regularized version of the objective function with an unknown scaling factor, we show that this is not necessary, thus achieving first true iteration complexity bounds. For strongly convex functions the method converges linearly. In the smooth case we also allow for arbitrary probability vectors and non-Euclidean norms. Finally, we demonstrate numerically that the algorithm is able to solve huge-scale $\ell _1$ -regularized least squares problems with a billion variables.     

Multilevel correction for collocation solutions of Volterra integral equations with proportional delays

In this paper, we propose a convergence acceleration method for collocation solutions of the linear second-kind Volterra integral equations with proportional delay qt $(0<q<1)$ . This convergence acceleration method called multilevel correction method is based on a kind of hybrid mesh, which can be viewed as a combination between the geometric meshes and the uniform meshes. It will be shown that, when the collocation solutions are continuous piecewise polynomials whose degrees are less than or equal to ${m} (m \leqslant 2)$ , the global accuracy of k level corrected approximation is $O(N^{-(2m(k+1)-\varepsilon)})$ , where N is the number of the nodes, and $\varepsilon$ is an arbitrary small positive number.     

Singular limits for Liouville-type equations on the flat two-torus

We consider the problem $$\begin{aligned} -\Delta u=\varepsilon ^{2}e^{u}- \frac{1}{|\Omega |}\int _\Omega \varepsilon ^{2} e^{u}+ {4\pi N\over |\Omega |} - 4 \pi N\delta _p, \quad \text{ in} {\Omega }, \quad \int _\Omega u=0 \end{aligned}$$ in a flat two-torus $\Omega $ with periodic boundary conditions, where $\varepsilon >0,\,|\Omega |$ is the area of the $\Omega $ , $N>0$ and $\delta _p$ is a Dirac mass at $p\in \Omega $ . We prove that if $1\le m<N+1$ then there exists a family of solutions $\{u_\varepsilon \}_{\varepsilon }$ such that $\varepsilon ^{2}e^{u_\varepsilon }\rightharpoonup 8\pi \sum _{i=1}^m\delta _{q_i}$ as $\varepsilon \rightarrow 0$ in measure sense for some different points $q_{1}, \ldots , q_{m}$ . Furthermore, points $q_i$ , $i=1,\dots ,m$ are different from $p$ .     

Properly discontinuous isometric group actions on inhomogeneous Lorentzian manifolds

If a homogeneous space $G/H$ is acted properly discontinuously upon by a subgroup $\varGamma $ of $G$ via the left action, the quotient space $\varGamma \backslash G/H$ is called a Clifford–Klein form. In Calabi and Markus (Ann Math (2) 75: 63–76, 1962) proved that there is no infinite subgroup of the Lorentz group $O(n+1,\,1)$ whose left action on the de Sitter space $O(n+1,\,1)/O(n,\,1)$ is properly discontinuous. It follows that a compact Clifford–Klein form of the de Sitter space never exists. In the present paper, we provide a new extension of the theorem of E. Calabi and L. Markus to a certain class of Lorentzian manifolds that are not necessarily homogeneous by using the techniques of differential geometry.     

Sufficient Criteria for \varepsilon Starlike Mappings in a Complex Banach Space

Several sufficient conditions for $\varepsilon $ starlike mappings on the unit ball $B$ in a complex Banach space are provided. From these, we may construct many concrete $\varepsilon $ starlike mappings on $B$ . Furthermore, several growth results associated with these sufficient conditions are also provided.     

Algorithms on Minimizing the Maximum Sensor Movement for Barrier Coverage of a Linear Domain

In this paper, we study the problem of moving $n$ n sensors on a line to form a barrier coverage of a specified segment of the line such that the maximum moving distance of the sensors is minimized. Previously, it was an open question whether this problem on sensors with arbitrary sensing ranges is solvable in polynomial time. We settle this open question positively by giving an $O(n^2\log n)$ O ( n 2 log n ) time algorithm. For the special case when all sensors have the same-size sensing range, the previously best solution takes $O(n^2)$ O ( n 2 ) time. We present an $O(n\log n)$ O ( n log n ) time algorithm for this case; further, if all sensors are initially located on the coverage segment, our algorithm takes $O(n)$ O ( n ) time. Also, we extend our techniques to the cycle version of the problem where the barrier coverage is for a simple cycle and the sensors are allowed to move only along the cycle. For sensors with the same-size sensing range, we solve the cycle version in $O(n)$ O ( n ) time, improving the previously best $O(n^2)$ O ( n 2 ) time solution.     

Tunneling for a Class of Difference Operators

We analyze a general class of difference operators ${H_\varepsilon = T_\varepsilon + V_\varepsilon}$ on ${\ell^2((\varepsilon \mathbb {Z})^d)}$ where ${V_\varepsilon}$ is a multi-well potential and ${\varepsilon}$ is a small parameter. We decouple the wells by introducing certain Dirichlet operators on regions containing only one potential well, and we shall treat the eigenvalue problem for ${H_\varepsilon}$ as a small perturbation of these comparison problems. We describe tunneling by a certain interaction matrix, similar to the analysis for the Schr?dinger operator [see Helffer and Sj?strand in Commun Partial Differ Equ 9:337–408, 1984], and estimate the remainder, which is exponentially small and roughly quadratic compared with the interaction matrix.     

Sharp geometric rigidity of isometries on Heisenberg groups

We prove sharp geometric rigidity estimates for isometries on Heisenberg groups. Our main result asserts that every $(1+\varepsilon )$ -quasi-isometry on a John domain of the Heisenberg group $\mathbb H ^n, n>1,$ is close to some isometry up to proximity order $\sqrt{\varepsilon }+\varepsilon $ in the uniform norm, and up to proximity order $\varepsilon $ in the $L_p^1$ -norm. We give examples showing the asymptotic sharpness of our results.