On relaxations of the max k-cut problem formulations

A tight continuous relaxation is a crucial factor in solving mixed integer formulations of many NP-hard combinatorial optimization problems. The (weighted) max k-cut problem is a fundamental combinatorial optimization problem with multiple notorious mixed integer optimization formulations. In this paper, we explore four existing mixed integer optimization formulations of the max k-cut problem. Specifically, we show that the continuous relaxation of a binary quadratic optimization formulation of the problem is: (i) stronger than the continuous relaxation of two mixed integer linear optimization formulations and (ii) at least as strong as the continuous relaxation of a mixed integer semidefinite optimization formulation. We also conduct a set of experiments on multiple sets of instances of the max k-cut problem using state-of-the-art solvers that empirically confirm the theoretical results in item (i). Furthermore, these numerical results illustrate the advances in the efficiency of global non-convex quadratic optimization solvers and more general mixed integer nonlinear optimization solvers. As a result, these solvers provide a promising option to solve combinatorial optimization problems. Our codes and data are available on GitHub.     

An annotated bibliography of combinatorial optimization problems with fixed cardinality constraints

In this paper, we consider combinatorial optimization problems with additional cardinality constraints. In k-cardinality combinatorial optimization problems, a cardinality constraint requires feasible solutions to contain exactly k elements of a finite set E. Problems of this type have applications in many areas, e.g. in the mining and oil industry, telecommunications, circuit layout, and location planning. We formally define the problem, mention some examples and summarize general results. We provide an annotated bibliography of combinatorial optimization problems of which versions with cardinality constraint have been considered in the literature.     

Revisiting <Emphasis Type="Italic">k</Emphasis>-sum optimization

In this paper, we address continuous, integer and combinatorial k-sum optimization problems. We analyze different formulations of this problem that allow to solve it through the minimization of a relatively small number of minisum optimization problems. This approach provides a general tool for solving a variety of k-sum optimization problems and at the same time, improves the complexity bounds of many ad-hoc algorithms previously reported in the literature for particular versions of this problem. Moreover, the results developed for k-sum optimization have been extended to the more general case of the convex ordered median problem, improving upon existing solution approaches.     

Heuristically guided algorithm for k-parity matroid problems

It is known that a large class of “hard” combinatorial optimization problems can be put in the form of a k-parity (weighted) matroid problem. In this paper we describe a heuristically guided algorithm for solving the above class of problems, which utilizes the information obtainable from the problem domain by computing, at each step, a possibly tight lower bound to the solution.     

Combinatorial flexibility problems and their computational complexity

The concept of flexibility—originated in the context of heat exchanger networks—is associated with a substructure which guarantees the performance of the original structure, in a given range of possible states. We extend this concept to combinatorial optimization problems, and prove several computational complexity results in this new framework.Under some monotonicity conditions, we prove that a combinatorial optimization problem polynomially transforms to its associated flexibility problem, but that the converse need not be true.In order to obtain polynomial flexibility problems, we have to restrict ourselves to combinatorial optimization problems on matroids. We also prove that, when relaxing in different ways the matroid structure, the flexibility problems become NP-complete. This fact is shown by proving the NP-completeness of the flexibility problems associated with the Shortest Path, Minimum Cut and Weighted Matching problems.     

On a classification of independence systems

By generalizing matroid axiomatics we provide a framework in which independence systems may be classified. The concept is applied to independence systems arising from well known combinatorial optimization problems such as k-matroid intersection, matchoid, vertex packing in finite graphs and travelling salesman problems.     

Integer programming formulation of combinatorial optimization problems

This paper considers in a somewhat general setting when a combinatorial optimization problem can be formulated as an (all-integer) integer programming (IP) problem. The main result is that any combinatorial optimization problem can be formulated as an IP problem if its feasible region S is finite but there are many rather sample problems that have no IP formulation if their S is infinite. The approach used for finite S usually gives a formulation with a relatively small number of additional variables for example, an integer polynomial of n 0?1 variables requires at most n + 1 additional variables by our approach, whereas 2ⁿ - (n + 1) additional variables at maximum are required by other existing methods. Finally, the decision problem of deciding whether an arbitrarily given combinatorial optimization problem has an IP formulation is considered and it is shown by an argument closely related to Hilbert's tenth problem (drophantine equations) that no such algorithm exists.     

On p-norm linear discrimination

We consider a p-norm linear discrimination model that generalizes the model of Bennett and Mangasarian (1992) and reduces to a linear programming problem with p-order cone constraints. The proposed approach for handling linear programming problems with p-order cone constraints is based on reformulation of p-order cone optimization problems as second order cone programming (SOCP) problems when p is rational. Since such reformulations typically lead to SOCP problems with large numbers of second order cones, an “economical” representation that minimizes the number of second order cones is proposed. A case study illustrating the developed model on several popular data sets is conducted.     

Completely positive reformulations for polynomial optimization

Polynomial optimization encompasses a very rich class of problems in which both the objective and constraints can be written in terms of polynomials on the decision variables. There is a well established body of research on quadratic polynomial optimization problems based on reformulations of the original problem as a conic program over the cone of completely positive matrices, or its conic dual, the cone of copositive matrices. As a result of this reformulation approach, novel solution schemes for quadratic polynomial optimization problems have been designed by drawing on conic programming tools, and the extensively studied cones of completely positive and of copositive matrices. In particular, this approach has been applied to solve key combinatorial optimization problems. Along this line of research, we consider polynomial optimization problems that are not necessarily quadratic. For this purpose, we use a natural extension of the cone of completely positive matrices; namely, the cone of completely positive tensors. We provide a general characterization of the class of polynomial optimization problems that can be formulated as a conic program over the cone of completely positive tensors. As a consequence of this characterization, it follows that recent related results for quadratic problems can be further strengthened and generalized to higher order polynomial optimization problems. Also, we show that the conditions underlying the characterization are conceptually the same, regardless of the degree of the polynomials defining the problem. To illustrate our results, we discuss in further detail special and relevant instances of polynomial optimization problems.     

Cardinality constrained combinatorial optimization: Complexity and polyhedra

Given a combinatorial optimization problem and a subset $N$ of nonnegative integer numbers, we obtain a cardinality constrained version of this problem by permitting only those feasible solutions whose cardinalities are elements of $N$. In this paper we briefly touch on questions that address common grounds and differences of the complexity of a combinatorial optimization problem and its cardinality constrained version. Afterwards we focus on the polyhedral aspects of the cardinality constrained combinatorial optimization problems. Maurras (1977) [5] introduced a class of inequalities, called forbidden cardinality inequalities in this paper, that can be added to a given integer programming formulation for a combinatorial optimization problem to obtain one for the cardinality restricted versions of this problem. Since the forbidden cardinality inequalities in their original form are mostly not facet defining for the associated polyhedron, we discuss some possibilities to strengthen them, based on the experiments made in Kaibel and Stephan (2007) and Maurras and Stephan (2009) [2], [3].     

Combinatorial Stokes formulas via minimal resolutions

We describe an explicit chain map from the standard resolution to the minimal resolution for the finite cyclic group Zk of order k. We then demonstrate how such a chain map induces a “Zk-combinatorial Stokes theorem,” which in turn implies “Dold's theorem” that there is no equivariant map from an n-connected to an n-dimensional free Zk-complex. Thus we build a combinatorial access road to problems in combinatorics and discrete geometry that have previously been treated with methods from equivariant topology. The special case k=2 for this is classical; it involves Tucker's (1949) combinatorial lemma which implies the Borsuk-Ulam theorem, its proof via chain complexes by Lefschetz (1949), the combinatorial Stokes formula of Fan (1967), and Meunier's work (2006).     

On a classification of independence systems

By generalizing matroid axiomatics we provide a framework in which independence systems may be classified. The concept is applied to independence systems arising from well-known combinatorial optimization problems such ask-matroid-intersection-, matchoid-, vertex packingor travelling salesman-problems.     

Linear spaces and transversal designs: k-anonymous combinatorial configurations for anonymous database search notes

Anonymous database search protocols allow users to query a database anonymously. This can be achieved by letting the users form a peer-to-peer community and post queries on behalf of each other. In this article we discuss an application of combinatorial configurations (also known as regular and uniform partial linear spaces) to a protocol for anonymous database search, as defining the key-distribution within the user community that implements the protocol. The degree of anonymity that can be provided by the protocol is determined by properties of the neighborhoods and the closed neighborhoods of the points in the combinatorial configuration that is used. Combinatorial configurations with unique neighborhoods or unique closed neighborhoods are described and we show how to attack the protocol if such configurations are used. We apply k-anonymity arguments and present the combinatorial configurations with k-anonymous neighborhoods and with k-anonymous closed neighborhoods. The transversal designs and the linear spaces are presented as optimal configurations among the configurations with k-anonymous neighborhoods and k-anonymous closed neighborhoods, respectively.     

Covering moving points with anchored disks

Consider a set of mobile clients represented by n points in the plane moving at constant speed along n different straight lines. We study the problem of covering all mobile clients using a set of k disks centered at k fixed centers. Each disk exists only at one instant and while it does, covers any client within its coverage radius. The task is to select an activation time and a radius for each disk such that every mobile client is covered by at least one disk. In particular, we study the optimization problem of minimizing the maximum coverage radius. First we prove that, although the static version of the problem is polynomial, the kinetic version is NP-hard. Moreover, we show that the problem is not approximable by a constant factor (unless P = NP). We then present a generic framework to solve it for fixed values of k, which in turn allows us to solve more general optimization problems. Our algorithms are efficient for constant values of k.     

A simple expected running time analysis for randomized “divide and conquer” algorithms

There are many randomized “divide and conquer” algorithms, such as randomized Quicksort, whose operation involves partitioning a problem of size n uniformly at random into two subproblems of size k and n-k that are solved recursively. We present a simple combinatorial method for analyzing the expected running time of such algorithms, and prove that under very weak assumptions this expected running time will be asymptotically equivalent to the running time obtained when problems are always split evenly into two subproblems of size n/2.     

Approximation algorithms for group prize-collecting and location-routing problems

In this paper we develop approximation algorithms for generalizations of the following three known combinatorial optimization problems, the Prize-Collecting Steiner Tree problem, the Prize-Collecting Travelling Salesman Problem and a Location-Routing problem.Given a graph G=(V,E) on n vertices and a length function on its edges, in the grouped versions of the above mentioned problems we assume that V is partitioned into k+1 groups, {V₀,V₁,…,V_k}, with a penalty function on the groups. In the Group Prize-Collecting Steiner Tree problem the aim is to find S, a collection of groups of V and a tree spanning the rest of the groups not in S, so as to minimize the sum of the costs of the edges in the tree and the costs of the groups in S. The Group Prize-Collecting Travelling Salesman Problem, is defined analogously. In the Group Location-Routing problem the customer vertices are partitioned into groups and one has to select simultaneously a subset of depots to be opened and a collection of tours that covers the customer groups. The goal is to minimize the costs of the tours plus the fixed costs of the opened depots. We give a -approximation algorithm for each of the three problems, where I is the cardinality of the largest group.     

Semiconductor final test scheduling with Sarsa(λ, k) algorithm

Semiconductor test scheduling problem is a variation of reentrant unrelated parallel machine problems considering multiple resource constraints, intricate {product, tester, kit, enabler assembly} eligibility constraints, sequence-dependant setup times, etc. A multi-step reinforcement learning (RL) algorithm called Sarsa(λ, k) is proposed and applied to deal with the scheduling problem with throughput related objective. Allowing enabler reconfiguration, the production capacity of the test facility is expanded and scheduling optimization is performed at the bottom level. Two forms of Sarsa(λ, k), i.e. forward view Sarsa(λ, k) and backward view Sarsa(λ, k), are constructed and proved equivalent in off-line updating. The upper bound of the error of the action-value function in tabular Sarsa(λ, k) is provided when solving deterministic problems. In order to apply Sarsa(λ, k), the scheduling problem is transformed into an RL problem by representing states, constructing actions, the reward function and the function approximator. Sarsa(λ, k) achieves smaller mean scheduling objective value than the Industrial Method (IM) by 68.59% and 76.89%, respectively for real industrial problems and randomly generated test problems. Computational experiments show that Sarsa(λ, k) outperforms IM and any individual action constructed with the heuristics derived from the existing heuristics or scheduling rules.     

Edge-Coloring Partialk-Trees

Many combinatorial problems can be efficiently solved for partialk-trees (graphs of treewidth bounded byk). The edge-coloring problem is one of the well-known combinatorial problems for which no efficient algorithms were previously known, except a polynomial-time algorithm of very high complexity. This paper gives a linear-time sequential algorithm and an optimal parallel algorithm which find an edge-coloring of a given partialk-tree with the minimum number of colors for fixedk.     

Global minimization using an Augmented Lagrangian method with variable lower-level constraints

A novel global optimization method based on an Augmented Lagrangian framework is introduced for continuous constrained nonlinear optimization problems. At each outer iteration k the method requires the ${\varepsilon_{k}}A novel global optimization method based on an Augmented Lagrangian framework is introduced for continuous constrained nonlinear optimization problems. At each outer iteration k the method requires the e_k{\varepsilon_{k}} -global minimization of the Augmented Lagrangian with simple constraints, where e_k ? e{\varepsilon_k \to \varepsilon} . Global convergence to an e{\varepsilon} -global minimizer of the original problem is proved. The subproblems are solved using the αBB method. Numerical experiments are presented.     

Estimating L ∞ Norms by L 2k Norms for Functions on Orbits

Abstract. Let G be a compact group acting in a real vector space V . We obtain a number of inequalities relating the L ^∞ norm of a matrix element of the representation of G with its L ^2k norm for a positive integer k . As an application, we obtain approximation algorithms to find the maximum absolute value of a given multivariate polynomial over the unit sphere (in which case G is the orthogonal group) and for the assignment problem of degree d , a hard problem of combinatorial optimization generalizing the quadratic assignment problem (in which case G is the symmetric group).