The soft-capacitated facility location problem, where each facility is composed of a variable number of fixed-capacity production units, has been recently studied in several papers, especially in the metric case. In this paper, we only consider the general problem where connection costs do not systematically satisfy the triangle inequality property. We show that an adaptation of the set covering greedy heuristic, where the subproblem is approximately solved by a fully polynomial-time approximation scheme based on cost scaling and dynamic programming, achieves a logaritmic approximation ratio of (1 + ε)H(n) for the problem, where n is the number of customers to be served and H is the harmonic series. This improves the previous bound of 2H(n) for this problem.

A co-biclique of a simple undirected graph G = (V,E) is the edge-set of two disjoint complete subgraphs of G.(A co-biclique is the complement of a biclique.)A subset F ⊆ E is an independent of G if there is a co-biclique B such that F ⊆ B, otherwise F is a dependent of G. This paper describes the minimal dependents of G. (A minimal dependent is a dependent C such that any proper subset of C is an independent.) It is showed that a minimum-cost dependent set of G can be determined in polynomial time for any nonnegative cost vector $x\in \mathbb Q_+^E$ . Based on this, we obtain a branch-and-cut algorithm for the maximum co-biclique problem which is, given a weight vector $w\in \mathbb Q_+^E$ , to find a co-biclique B of G maximizing w(B) = ∑e∈B we .

The line balancing problem consits in assigning tasks to stations in orderto respect precedence constraints and cycle time constraints. In thispaper, the cycle time is fixed and the objective is to minimize the numberof stations. We propose to use metaheuristics based on simulated annealingby exploiting the link between the line balancing problem and the binpacking problem. The principle of the method lies in the combinationbetween a metaheuristic and a bin packing heuristic. Two representationsof a solution and two neighboring systems are proposed and the methods arecompared with results from the literature. They are better or similar totabu search based algorithm.


In telecommunications network design, one of the most frequentproblems is to adjust the capacity on the links of the network in order to satisfy a set of requirements. In the past, theserequirements were demands based on historical data and/or demographic predictions. Nowadays, because of new technologydevelopment and customer movement due to competitiveness, thedemands present considerable variability. Thus, network robustness w.r.t demand uncertainty is now regarded as a majorconsideration. In this work, we propose a min-max-min formulation and a methodology to cope with this uncertainty. We model the uncertainty as the convex hull of certain scenarios and show thatcutting plane methods can be applied to solve the underlying problems. We will compare Kelley, Elzinga-Moore and bundle methods.

We discuss the use of operations research methods for computer-aided design of mechanical transmission systems. We consider how to choose simultaneously transmission ratios and basic design parameters of transmission elements (diameters, widths, modules and tooth number for gears, diameters of shafts). The objectives, by the order of importance, are: to minimize the deviation of the obtained speeds from desired; to maximize the transmission life; to minimize the total mass. To solve this problem, we propose a multi-level decomposition scheme in combination with methods of quadratic and dynamic programming. Some industrial cases are solved. For these cases, the developed software tool improves the design decisions by decreasing total metal consumption of the transmission as much as 7–10% and considerable simplifies the work of the designer.

Given a weighted undirected graph G = (V,E),a tree (respectively tour) cover of an edge-weighted graph is a set of edges which forms a tree (resp. closed walk) and covers every other edge in the graph. The tree (resp. tour) cover problem is of finding a minimum weight tree (resp. tour) cover of G. Arkin, Halldórsson and Hassin (1993) give approximation algorithms with factors respectively 3.5 and 5.5. Later Könemann, Konjevod, Parekh, and Sinha (2003) study the linear programming relaxations and improve both factors to 3. We describe in the first part of the paper a 2-approximation algorithm for the metric case of tree cover.In the second part, we will consider a generalized version of tree (resp. tour) covers problem which is to find a minimum tree (resp. tours) which covers a subset D ⊆ E of G.We show that the algorithms of Könemann et al.can be adapted for the generalized tree and tours covers problem with the same factors.

We prove that MAX-3SAT can be approximated in polynomial time within a factor 1.0957 on random instances.

The recourse to operation research solutions has strongly increasedthe performances of scheduling task in the High-Level Synthesis(called hardware compilation). Scheduling a whole program is notpossible as too many constraints and objectives interact. We decomposehigh-level scheduling in three steps. Step 1: Coarse-grain schedulingtries to exploit parallelism and locality of the whole program (inparticular in loops, possibly imperfectly nested) with a rough view ofthe target architecture. This produces a sequence of logical steps,each of which contains a pool of macro-tasks. Step 2: Micro-schedulingmaps and schedules each macro-task independently taking into accountall peculiarities of the target architecture. This produces areservation table for each macro-task. Step 3: Fine-grain schedulingrefines each logical step by scheduling all its macro-tasks. Thispaper focuses on the third step.As tasks are modeled as reservation tables, we can express resourceconstraints using dis-equations (i.e., negations of equations). Asmost scheduling problems, scheduling tasks with reservation tables tominimize the total duration is NP-complete. Our goal here is todesign different strategies and to evaluate them, on practicalexamples, to see if it is possible to find optimal solution inreasonable time. The first algorithm is based on integer linearprogramming techniques for scheduling, which we adapt to our specificproblem. Our main algorithmic contribution is an exactbranch-and-bound algorithm, where each evaluation is accelerated byvariant of Dijkstra's algorithm. A simple greedy heuristic is alsoproposed for comparisons. The evaluation and comparison are done onpieces of scientific applications from the PerfectClub and theHLSynth95 benchmarks. The results demonstrate the suitability of thesesolutions for high-level synthesis scheduling.

In this paper, we propose a new class of adaptive trust region methods for unconstrained optimization problems and develop some convergence properties. In the new algorithms, we use the current iterative information to define a suitable initial trust region radius at each iteration. The initial trust region radius is more reasonable in the sense that the trust region model and theobjective function are more consistent at the current iterate. The global convergence, super-linear and quadratic convergence rate are analyzed under some mild conditions. Numericalresults show that some special adaptive trust region methods are available and efficient in practical computation.

The m-subspace polytope is defined as the convex hull of the characteristic vectors of all m-dimensional subspaces ofa finite affine space. The particular case of the hyperplane polytopehas been investigated by Maurras (1993) and Anglada and Maurras (2003), who gave a complete characterization of the facets. The general m-subspace polytope that we consider shows a much more involved structure, notably as regards facets. Nevertheless, several families of facets are established here. Then the group of automorphisms of the m-subspace polytope is completely described and the adjacency of vertices is fully characterized.

Most systems are characterized by uncertainties that cause throughput to be highly variable, for example, many modern production processes and services are substantially affected by random yields. When yield is random, not only is the usable quantity uncertain, but the random yield reduces usable capacity and throughput in the system. For these reasons, strategies are needed that incorporate random yield. This paper presents the analysis of the newsvendor model with a general random yield distribution, including the derivation of the optimal order quantity. Results are shown to converge to the basic newsvendor model for the case of perfect yield, and are further demonstrated using the case of general multiplicative random yield. Results have significant impact on both manufacturing and service sectors since the newsvendor model applies to many real-world situations.


The Progressive Second Price mechanism (PSP), recently introduced byLazar and Semret to share aninfinitely-divisible resource among users through pricing, has been shown to verifyvery interesting properties. Indeed, the incentive compatibilityproperty of that scheme, and the convergence toan efficient resource allocation where established, using the frameworkof Game Theory.Therefore, that auction-based allocation and pricing scheme seemsparticularly well-suited to solve congestion problems intelecommunication networks, where the resource to share is theavailable bandwidth on a link.This paper aims atsupplementing the existing results by highlighting some properties of thedifferent equilibria that can be reached. We precisely characterize the possible outcomes of thePSP auction game in terms of players bid price: when the bid fee (cost of a bid update) tends to zero then the bid price of all users at equilibrium gets close to the so-called market clearing price of the resource. Therefore, observing an equilibrium of the PSP auction game gives some accurate information about the market clearing price of the resource.