Iterative Methods in Combinatorial Optimization
نویسنده
چکیده
In these lectures, I will describe a simple iterative method that supplies new proofs of integrality of linear characterizations of various basic problems in combinatorial optimization, and also allows adaptations to design approximation algorithms for NP-hard variants of these problems involving extra “degree-like” budget constraints. It is inspired by Jain’s iterative rounding method for designing approximation algorithms for survivable network design problems, and augmented with a relaxation idea in the work of Lau, Naor, Salvatipour and Singh in their work on designing the approximation algorithm for its degree bounded version. Its application was further refined in recent work of Bansal, Khandekar and Nagarajan on degree-bounded directed network design. I will begin by reviewing the background material on LP relaxations and their solvability and properties of extreme point or vertex solutions to such problems. I will then introduce the basic framework of the method using the assignment problem, and show its application by re-deriving the approximation results of Shmoys and Tardos for the generalized assignment problem. I will then discuss linear characterizations for the spanning tree polyhedron in undirected graphs and give a new proof of integrality using an iterative method. I will then illustrate an application to approximating the degree-bounded version of the undirected problem, by proving the results of Goemans and Lau & Singh. I will continue with showing how these methods for spanning trees simplify and generalize to showing linear descriptions of maximum weight matroid bases and also maximum weight sets that are independent in two different matroids. This also leads to good additive approximation algorithms for a bounded degree version of the matroid basis problem. I will close with applications of the iterative method by revisiting Jain’s original proof for the SNDP and giving a new proof that unifies its treatment with that for the Symmetric TSP polyhedron (describing joint work with Nagarajan and Singh). I will also outline the versatility of the method by pointing out the other problems for which the method has been applied, summarizing the discussion in a recent monograph I have co-authored on this topic with Lap Chi Lau and Mohit Singh (published by Cambridge University Press, 2011). 1998 ACM Subject Classification F.2. Analysis of Algorithms and Problem Complexity
منابع مشابه
Distributed Combinatorial Optimization
This paper reports about research projects of the University of Paderborn in the eld of distributed combinatorial optimization. We give an introduction into combinatorial optimization and a brief deenition of some important applications. As a rst exact solution method we describe branch & bound and present the results of our work on its distributed implementation. Results of our distributed imp...
متن کاملA Real–time and Any–time Approach for a Dialog System
In this paper we present a knowledge based approach to a speech understanding and dialog system which possesses any–time and real–time capabilities. Knowledge is represented by means of a semantic–network formalism; the concept–centered knowledge base is automatically compiled into a fine–grained network which allows an efficient exploitation of parallelism. The interpretation problem is define...
متن کاملCombinatorial problems in solving linear systems
Numerical linear algebra and combinatorial optimization are vast subjects; as is their interaction. In virtually all cases there should be a notion of sparsity for a combinatorial problem to arise. Sparse matrices therefore form the basis of the interaction of these two seemingly disparate subjects. As the core of many of today’s numerical linear algebra computations consists of the solution of...
متن کاملDistribution Approximation , Combinatorial Optimization , and Lagrange - Barrier
In this paper, typical analog combinatorial optimization approaches, such as Hopfield net, Hopfield-Lagrange net, Maximum entropy approach, Lagrange-Barrier approach, are systematically examined from the perspective of learning distribution. The minimization of a combinatorial cost is turned into a procedure of learning a simple distribution to approximate the Gibbs distribution induced from th...
متن کاملA reusable iterative optimization software library to solve combinatorial problems with approximate reasoning
Real world combinatorial optimization problems such as scheduling are typically too complex to solve with exact methods. Additionally, the problems often have to observe vaguely speci ed constraints of di erent importance, the available data may be uncertain, and compromises between antagonistic criteria may be necessary. We present a combination of approximate reasoning based constraints and i...
متن کاملA L Use of a Validation Set
In this paper we derive novel algorithms for estimation of regularization parameters and for optimization of neural net architectures based on a validation set. Regularization parameters are estimated using an iterative gradient descent scheme. Architecture optimization is performed by approximative combinatorial search among the relevant subsets of an initial neural network architecture by emp...
متن کامل