The feasibility pump

In this paper we consider the NP-hard problem of finding a feasible solution (if any exists) for a generic MIP problem of the form min{c^Tx:Ax≥b,x_j integer ∀j ∈ }. Trivially, a feasible solution can be defined as a point x* ∈ P:={x:Ax≥b} that is equal to its rounding , where the rounded point is defined by := x*_j if j ∈ and := x*_j otherwise, and [·] represents scalar rounding to the nearest integer. Replacing “equal” with “as close as possible” relative to a suitable distance function Δ(x*, ), suggests the following Feasibility Pump (FP) heuristic for finding a feasible solution of a given MIP.We start from any x* ∈ P, and define its rounding . At each FP iteration we look for a point x* ∈ P that is as close as possible to the current by solving the problem min {Δ(x, ): x ∈ P}. Assuming Δ(x, ) is chosen appropriately, this is an easily solvable LP problem. If Δ(x*, )=0, then x* is a feasible MIP solution and we are done. Otherwise, we replace by the rounding of x*, and repeat.We report computational results on a set of 83 difficult 0-1 MIPs, using the commercial software ILOG-Cplex 8.1 as a benchmark. The outcome is that FP, in spite of its simple foundation, proves competitive with ILOG-Cplex both in terms of speed and quality of the first solution delivered. Interestingly, ILOG-Cplex could not find any feasible solution at the root node for 19 problems in our test-bed, whereas FP was unsuccessful in just 3 cases.     

The volumetric barrier for convex quadratic constraints

Let where and _i is an n×n positive semidefinite matrix. We prove that the volumetric and combined volumetric-logarithmic barriers for are and self-concordant, respectively. Our analysis uses the semidefinite programming (SDP) representation for the convex quadratic constraints defining , and our earlier results on the volumetric barrier for SDP. The self-concordance results actually hold for a class of SDP problems more general than those corresponding to the SDP representation of .Mathematics Subject Classification (1991):90C25, 90C30     

A branch-and-cut algorithm for the stochastic uncapacitated lot-sizing problem

This paper addresses a multi-stage stochastic integer programming formulation of the uncapacitated lot-sizing problem under uncertainty. We show that the classical (ℓ,S) inequalities for the deterministic lot-sizing polytope are also valid for the stochastic lot-sizing polytope. We then extend the (ℓ,S) inequalities to a general class of valid inequalities, called the inequalities, and we establish necessary and sufficient conditions which guarantee that the inequalities are facet-defining. A separation heuristic for inequalities is developed and incorporated into a branch-and-cut algorithm. A computational study verifies the usefulness of the inequalities as cuts. This research has been supported in part by the National Science Foundation under Award number DMII-0121495.     

Optimal 3-terminal cuts and linear programming

Given an undirected graph G=(V,E) and three specified terminal nodes t ₁,t ₂,t ₃, a 3-cut is a subset A of E such that no two terminals are in the same component of G\A. If a non-negative edge weight c _e is specified for each e∈E, the optimal 3-cut problem is to find a 3-cut of minimum total weight. This problem is -hard, and in fact, is max- -hard. An approximation algorithm having performance guarantee has recently been given by Călinescu, Karloff, and Rabani. It is based on a certain linear-programming relaxation, for which it is shown that the optimal 3-cut has weight at most times the optimal LP value. It is proved here that can be improved to , and that this is best possible. As a consequence, we obtain an approximation algorithm for the optimal 3-cut problem having performance guarantee . In addition, we show that is best possible for this algorithm. Research of this author was supported by NSERC PGSB. Research supported by a grant from NSERC of Canada.     

On the gap between the quadratic integer programming problem and its semidefinite relaxation

Consider the semidefinite relaxation (SDR) of the quadratic integer program (QIP): where Q is a given symmetric matrix and D is diagonal. We consider the SDR gap We establish the uniqueness of the SDR solution and prove that if and only if γ_r:=n⁻¹max{x^TVV^Tx:x ∈ {-1, 1}ⁿ}=1 where V is an orthogonal matrix whose columns span the (r–dimensional) null space of D−Q and where D is the unique SDR solution. We also give a test for establishing whether that involves 2^r−1 function evaluations. In the case that γ_r<1 we derive an upper bound on γ which is tighter than Thus we show that `breaching' the SDR gap for the QIP problem is as difficult as the solution of a QIP with the rank of the cost function matrix equal to the dimension of the null space of D−Q. This reduced rank QIP problem has been recently shown to be solvable in polynomial time for fixed r.     

Ambiguous chance constrained problems and robust optimization

In this paper we study ambiguous chance constrained problems where the distributions of the random parameters in the problem are themselves uncertain. We focus primarily on the special case where the uncertainty set of the distributions is of the form where ρ_p denotes the Prohorov metric. The ambiguous chance constrained problem is approximated by a robust sampled problem where each constraint is a robust constraint centered at a sample drawn according to the central measure The main contribution of this paper is to show that the robust sampled problem is a good approximation for the ambiguous chance constrained problem with a high probability. This result is established using the Strassen-Dudley Representation Theorem that states that when the distributions of two random variables are close in the Prohorov metric one can construct a coupling of the random variables such that the samples are close with a high probability. We also show that the robust sampled problem can be solved efficiently both in theory and in practice. Research partially supported by NSF grant CCR-00-09972. Research partially supported by NSF grants CCR-00-09972, DMS-01-04282, and ONR grant N000140310514.     

Critical extreme points of the 2-edge connected spanning subgraph polytope

In this paper we study the extreme points of the polytope P(G), the linear relaxation of the 2-edge connected spanning subgraph polytope of a graph G. We introduce a partial ordering on the extreme points of P(G) and give necessary conditions for a non-integer extreme point of P(G) to be minimal with respect to that ordering. We show that, if is a non-integer minimal extreme point of P(G), then G and can be reduced, by means of some reduction operations, to a graph G' and an extreme point of P(G') where G' and satisfy some simple properties. As a consequence we obtain a characterization of the perfectly 2-edge connected graphs, the graphs for which the polytope P(G) is integral. Received: May, 2004 Part of this work has been done while the first author was visiting the Research Institute for Discrete Mathematics, University of Bonn, Germany.     

Improved complexity results on solving real-number linear feasibility problems

We present complexity results on solving real-number standard linear programs LP(A,b,c), where the constraint matrix the right-hand-side vector and the objective coefficient vector are real. In particular, we present a two-layered interior-point method and show that LP(A,b,0), i.e., the linear feasibility problem A x = b and x ≥ 0, can be solved in in O(n ^2.5 c(A)) interior-point method iterations. Here 0 is the vector of all zeros and c(A) is the condition measure of matrix A defined in [25]. This complexity iteration bound is reduced by a factor n from that for general LP(A, b, c) in [25]. We also prove that the iteration bound will be further reduced to O(n ^1.5 c(A)) for LP(A, 0, 0), i.e., for the homogeneous linear feasibility problem. These results are surprising since the classical view has been that linear feasibility would be as hard as linear programming. This author was supported in part by NSF Grants DMS-9703490 and DMS-0306611     

A -algorithm for convex minimization

For convex minimization we introduce an algorithm based on -space decomposition. The method uses a bundle subroutine to generate a sequence of approximate proximal points. When a primal-dual track leading to a solution and zero subgradient pair exists, these points approximate the primal track points and give the algorithm's , or corrector, steps. The subroutine also approximates dual track points that are -gradients needed for the method's -Newton predictor steps. With the inclusion of a simple line search the resulting algorithm is proved to be globally convergent. The convergence is superlinear if the primal-dual track points and the objective's -Hessian are approximated well enough. Dedicated to Terry Rockafellar who has had a great influence on our work via strong support for proximal points and for structural definitions that involve tangential convergence. On leave from INRIA Rocquencourt Research of the first author supported by the National Science Foundation under Grant No. DMS-0071459 and by CNPq (Brazil) under Grant No. 452966/2003-5. Research of the second author supported by FAPERJ (Brazil) under Grant No.E26/150.581/00 and by CNPq (Brazil) under Grant No. 383066/2004-2.     

Second-order conditions in C 1,1 constrained vector optimization

We consider the constrained vector optimization problem min _C f(x), g(x) ∈ ?K, where f:? ⁿ →? ^m and g:? ⁿ →? ^p are C ^1,1 functions, and C ? ^m and K ? ^p are closed convex cones with nonempty interiors. Two type of solutions are important for our considerations, namely w-minimizers (weakly efficient points) and i-minimizers (isolated minimizers). We formulate and prove in terms of the Dini directional derivative second-order necessary conditions for a point x ⁰ to be a w-minimizer and second-order sufficient conditions for x ⁰ to be an i-minimizer of order two. We discuss the reversal of the sufficient conditions under suitable constraint qualifications of Kuhn-Tucker type. The obtained results improve the ones in Liu, Neittaanmäki, K?í?ek [21].     

Analysis of nonsmooth vector-valued functions associated with second-order cones

Let be the Lorentz/second-order cone in . For any function f from to , one can define a corresponding function f^soc(x) on by applying f to the spectral values of the spectral decomposition of x with respect to . We show that this vector-valued function inherits from f the properties of continuity, (local) Lipschitz continuity, directional differentiability, Fréchet differentiability, continuous differentiability, as well as (-order) semismoothness. These results are useful for designing and analyzing smoothing methods and nonsmooth methods for solving second-order cone programs and complementarity problems.Mathematics Subject Classification (1991): 26A27, 26B05, 26B35, 49J52, 90C33, 65K05     

Newton methods for nonsmooth convex minimization: connections among -Lagrangian, Riemannian Newton and SQP methods

This paper studies Newton-type methods for minimization of partly smooth convex functions. Sequential Newton methods are provided using local parameterizations obtained from -Lagrangian theory and from Riemannian geometry. The Hessian based on the -Lagrangian depends on the selection of a dual parameter g; by revealing the connection to Riemannian geometry, a natural choice of g emerges for which the two Newton directions coincide. This choice of g is also shown to be related to the least-squares multiplier estimate from a sequential quadratic programming (SQP) approach, and with this multiplier, SQP gives the same search direction as the Newton methods. This paper is dedicated to R.T. Rockafellar, on the occasion of his 70th birthday.     

An approximation algorithm for the general max-min resource sharing problem

We propose an approximation algorithm for, the general max-min resource sharing problem with M nonnegative concave constraints on a convex set B. The algorithm is based on a Lagrangian decomposition method and it uses a c-approximation algorithm (called approximate block solver) for a simpler maximization problem over the convex set B. We show that our algorithm achieves within iterations or calls to the approximate block solver a solution for the general max-min resource sharing problem with approximation ratio The algorithm is faster and simpler than the previous known approximation algorithms for the problem [12, 13] Research of the author was supported in part by EU Thematic Network APPOL, Approximation and Online Algorithms, IST-2001-30012, by EU Project CRESCCO, Critical Resource Sharing for Cooperation in Complex Systems, IST-2001-33135 and by DFG Project, Entwicklung und Analyse von Approximativen Algorithmen für Gemischte und Verallgemeinerte Packungs- und überdeckungsprobleme, JA 612/10-1. Part of this work was done while visiting the Department of Computer Science at ETH Zürich. An extended abstract of this paper appeared in SWAT 2004, Scandinavian Workshop on Algorithm Theory, LNCS 3111, 311–322.     

Global Solutions to the Two Dimensional Quasi-Geostrophic Equation with Critical or Super-Critical Dissipation

The two dimensional quasi-geostrophic (2D QG) equation with critical and super-critical dissipation is studied in Sobolev space H^s(ℝ²). For critical case (α=), existence of global (large) solutions in H^s is proved for s≥ when is small. This generalizes and improves the results of Constantin, D. Cordoba and Wu [4] for s = 1, 2 and the result of A. Cordoba and D. Cordoba [8] for s=. For s≥1, these solutions are also unique. The improvement for pushing s down from 1 to is somewhat surprising and unexpected. For super-critical case (α ∈ (0,)), existence and uniqueness of global (large) solution in H^s is proved when the product is small for suitable s≥2−2α, p ∈ [1,∞] and β ∈ (0,1].     

Asymptotic laws for regenerative compositions: gamma subordinators and the like

For a random closed set obtained by exponential transformation of the closed range of a subordinator, a regenerative composition of generic positive integer n is defined by recording the sizes of clusters of n uniform random points as they are separated by the points of . We focus on the number of parts K_n of the composition when is derived from a gamma subordinator. We prove logarithmic asymptotics of the moments and central limit theorems for K_n and other functionals of the composition such as the number of singletons, doubletons, etc. This study complements our previous work on asymptotics of these functionals when the tail of the Lévy measure is regularly varying at 0+. Research supported in part by N.S.F. Grant DMS-0071448     

On pairs of vectors achieving the maximal angle of a convex cone

In this paper we explore the concept of antipodality relative to a closed convex cone . The problem under consideration is that of finding a pair of unit vectors in K achieving the maximal angle of the cone. We mention also a few words on the attainability of critical angles. By way of application of the general theory, we briefly discuss the problem of estimating the radius of pointedness of a cone.     

Tannaka duality on quotient stacks

Let = [X/G] be the quotient stack of a scheme X by an affine group scheme G over a field k. Assume that there is a line bundle on whose underlying line bundle on X is very ample. Let VB() be the category of vector bundles on .We show that is canonically isomorphic to the stack of fiber functors on VB(). This is an analogue of the Tannaka duality for affine groups. Partially supported by CNCSIS contract no. 33079/2004     

All double coverings of cyclotomic fields

A double covering of a Galois extension K/F in the sense of [3] is an extension /K of degree ≤2 such that /F is Galois. In this paper we determine explicitly all double coverings of any cyclotomic extension over the rational number field in the complex number field. We get the results mainly by Galois theory and by using and modifying the results and the methods in [2] and [3]. Project 10571097 supported by NSFC     

Partial regularity for stationary harmonic maps at a free boundary

Let and be smooth Riemannian manifolds, of the dimension n≥2 with nonempty boundary, and compact without boundary. We consider stationary harmonic maps u ∈ H¹(, ) with a free boundary condition of the type u(∂) ⊂ Γ, given a submanifold Γ⊂. We prove partial boundary regularity, namely (sing(u))=0, a result that was until now only known in the interior of the domain (see [B]). The key of the proof is a new lemma that allows an extension of u by a reflection construction. Once the partial regularity theorem is known, it is possible to reduce the dimension of the singular set further under additional assumptions on the target manifold and the submanifold Γ.     

On topological Kadec norms

We present a topological analogue of the classic Kadec Renorming Theorem, as follows. Let be two separable metric topologies on the same set X. We prove that every point in X has an -neighbourhood basis consisting of sets that are -closed if and only if there exists a function φ: X→ℝ that is -lower semi-continuous and such that is the weakest topology on X that contains and that makes φ continuous. An immediate corollary is that the class of almost n-dimensional spaces consists precisely of the graphs of lower semi-continuous functions with at most n-dimensional domains.