A computational comparison of the dinic and network simplex methods for maximum flow

We study the implementation of two fundamentally different algorithms for solving the maximum flow problem: Dinic's method and the network simplex method. For the former, we present the design of a storage-efficient implementation. For the latter, we develop a "steepest-edge" pivot selection criterion that is easy to include in an existing network simplex implementation. We compare the computational efficiency of these two methods on a personal computer with a set of generated problems of up to 4 600 nodes and 27 000 arcs.This research was supported in part by the National Science Foundation under Grant Nos. MCS-8113503 and DMS-8512277.     

A largest-distance pivot rule for the simplex algorithm

We propose a new pivot rule for the simplex algorithm, which is demonstrative in the dual space intuitively. Although it is based on normalized reduced costs, like the steepest-edge rule and its variants, the rule is much simpler and cheaper than the latter. We report computational results obtained with the 47 largest Netlib problems in terms of the number of rows and columns, all of the 16 Kennington problems, and the 17 largest BPMPD problems. Over the total 80 problems, a variant of the rule outperformed the Devex rule with iterations and time ratio 1.43 and 3.24, respectively.     

Pivot rules for linear programming: A survey on recent theoretical developments

The purpose of this paper is to discuss the various pivot rules of the simplex method and its variants that have been developed in the last two decades, starting from the appearance of the minimal index rule of Bland. We are mainly concerned with finiteness properties of simplex type pivot rules. Well known classical results concerning the simplex method are not considered in this survey, but the connection between the new pivot methods and the classical ones, if there is any, is discussed.In this paper we discuss three classes of recently developed pivot rules for linear programming. The first and largest class is the class of essentially combinatorial pivot rules including minimal index type rules and recursive rules. These rules only use labeling and signs of the variables. The second class contains those pivot rules which can actually be considered as variants or generalizations or specializations of Lemke's method, and so they are closely related to parametric programming. The last class has the common feature that the rules all have close connections to certain interior point methods. Finally, we mention some open problems for future research.On leave from the Eötvös University, Budapest, and partially supported by OTKA No. 2115.     

Computational aspects of simplex and MBU-simplex algorithms using different anti-cycling pivot rules

Abstract

Several variations of index selection rules for simplex-type algorithms for linear programming, like the Last-In-First-Out or the Most-Often-Selected-Variable are rules not only theoretically finite, but also provide significant flexibility in choosing a pivot element. Based on an implementation of the primal simplex and the monotonic build-up (MBU) simplex method, the practical benefit of the flexibility of these anti-cycling pivot rules is evaluated using public benchmark LP test sets. Our results also provide numerical evidence that the MBU-simplex algorithm is a viable alternative to the traditional simplex algorithm.     

On Gradient Simplex Methods for Linear Programs

A variety of pivot column selection rules based upon the gradient criteria (including the steepest edge) have been explored to improve the efficiency of the primal simplex method. Simplex-like algorithms have been proposed imbedding the gradient direction (GD) which includes all variables whose increase or decrease leads to an improvement in the objective function. Recently a frame work has been developed in the simplex method to incorporate the reduced-gradient direction (RGD) consisting of only variables whose increase leads to an improvement in the objective function. In this paper, the results are extended to embed GD in the simplex method based on the concept of combining directions. Also mathematical properties related to combining directions as well as deleting a variable from all basic directions are presented.     

The s-monotone index selection rules for pivot algorithms of linear programming

In this paper we introduce the concept of s-monotone index selection rule for linear programming problems. We show that several known anti-cycling pivot rules like the minimal index, Last-In–First-Out and the most-often-selected-variable pivot rules are s-monotone index selection rules. Furthermore, we show a possible way to define new s-monotone pivot rules. We prove that several known algorithms like the primal (dual) simplex, MBU-simplex algorithms and criss-cross algorithm with s-monotone pivot rules are finite methods.     

A practicable steepest-edge simplex algorithm

It is shown that suitable recurrences may be used in order to implement in practice the steepest-edge simplex linear programming algorithm. In this algorithm each iteration is along an edge of the polytope of feasible solutions on which the objective function decreases most rapidly with respect to distance in the space of all the variables. Results of computer comparisons on medium-scale problems indicate that the resulting algorithm requires less iterations but about the same overall time as the algorithm of Harris [8], which may be regarded as approximating the steepest-edge algorithm. Both show a worthwhile advantage over the standard algorithm.     

On strongly polynomial dual simplex algorithms for the maximum flow problem

Several pivot rules for the dual network simplex algorithm that enable it to solve a maximum flow problem on ann-node,m-arc network in at most 2nm pivots and O(n ² m) time are presented. These rules are based on the concept of apreflow and depend upon the use of node labels which are either the lengths of a shortestpseudoaugmenting path from those nodes to the sink node orvalid underestimates of those lengths. Extended versions of our algorithms are shown to solve an important class of parametric maximum flow problems with no increase in the worst-case pivot and time bounds of these algorithms. This research was supported in part by NSF Grants DMS 91-06195, DMS 94-14438, and CDR 84-21402 and DOE Grant DE-FG02-92ER25126.     

A combinatorial abstraction of linear programming

In the context of oriented matroids we establish and elaborate upon an abstraction of linear programming duality foreseen by Rockafellar in his work on elementary vectors. We describe a pivoting operation for oriented matroids and a finite pivoting method, which elucidate the combinatorial nature of Dantzig's simplex method. The pivoting method specializes, when the oriented matroids arise from real vector spaces, to the simplex method with a new pivot selection rule. A very simple pivot selection rule for which finiteness has been established in the linear programming context, but not in the broader setting of oriented matroids, is also described.     

Criss-cross methods: A fresh view on pivot algorithms

Criss-cross methods are pivot algorithms that solve linear programming problems in one phase starting with any basic solution. The first finite criss-cross method was invented by Chang, Terlaky and Wang independently. Unlike the simplex method that follows a monotonic edge path on the feasible region, the trace of a criss-cross method is neither monotonic (with respect to the objective function) nor feasibility preserving. The main purpose of this paper is to present mathematical ideas and proof techniques behind finite criss-cross pivot methods. A recent result on the existence of a short admissible pivot path to an optimal basis is given, indicating shortest pivot paths from any basis might be indeed short for criss-cross type algorithms. The origins and the history of criss-cross methods are also touched upon.     

On the Bartels—Golub decomposition for linear programming bases

Many implementations of the simplex method require the row of the inverse of the basis matrixB corresponding to the pivot row at each iteration for updating either a pricing vector or the nonbasic reduced costs. In this note we show that if the Bartels—Golub algorithm [1, 2] or one of its variants is used to update theLU factorization ofB, then less computing is needed if one works with the factors of the updatedB than with those ofB. These results are discussed as they apply to the column selection algorithms recently proposed by Goldfarb and Reid [4, 5] and Harris [6].This research was supported in part by the National Science Foundation under Grant GJ 36472.     

Parallel search paths for the simplex algorithm

It is well known that the simplex method is inherently a sequential algorithm with little scope for parallelization. Even so, during the last decades several attempts were made to parallelize it since it is one of the most important algorithms for solving linear optimization problems. Such parallelization ideas mostly rely on iteration parallelism and overlapping. Since the simplex method goes through a series of basic solutions until it finds an optimal solution, each of them must be available before performing the next basis change. This phenomenon imposes a limit on the performance of the parallelized version of the simplex method which uses overlapping iterations. Another approach can be considered if we think about alternative paths on the n-dimensional simplex polyhedron. As the simplex method goes through the edges of this polyhedron it is generally true that the speed of convergence of the algorithm is not smooth. It depends on the actual part of the surface. If a parallel version of the simplex algorithm simultaneously goes on different paths on this surface a highly reliable algorithm can be constructed. There is no known dominating strategy for pivot selection. Therefore, one can try different pivot selection methods in parallel in order to guide the algorithm on different pathways. This approach can be used effectively with periodic synchronization on shared memory multi-core computing environments to speed up the solution algorithm and get around numerically and/or algorithmically difficult situations throughout the computations.     

Efficient nested pricing in the simplex algorithm

We report a remarkable success of nested pricing rules over major pivot rules commonly used in practice, such as Dantzig’s original rule as well as the steepest-edge rule and Devex rule.     

New variants of finite criss-cross pivot algorithms for linear programming

In this paper we generalize the so-called first-in-last-out pivot rule and the most-often-selected-variable pivot rule for the simplex method, as proposed in S. Zhang (Operations Research Letters 10 (1991) 189–192) to the criss-cross pivot setting where neither the primal nor the dual feasibility is preserved. The finiteness of the new criss-cross pivot variants is proven.     

The pivot and probe algorithm for solving a linear program

In [7] we defined acandidate constraint as one which, for at least one pivot step, contains a potential pivot, discovered that most constraints are never candidate, and devised a modification of the simplex method in which only constraints which are currently candidates are updated. In this paper we present another way to take advantage of this fact. We begin by solving a relaxed linear program consisting of the constraints of the original problem which are initially candidates. Its solution gives an upper bound to the value of the original problem. We also introduce the idea of a probe, that is, a line segment joining two vectors for the primal problem, one of which is primal feasible, and use it to identify a most violated constraint; at the same time this gives a lower bound to the objective value of the original problem. This violated constraint is added to the relaxed problem which is solved again, which gives a new upper bound etc. We present computational experience indicating that time savings of 50–80% over the simplex method can be obtained by this method, which we call PAPA, the Pivot and Probe Algorithm. This report was prepared as part of the activities of the Management Science Research Group, Carnegie-Mellon University, under Contract No. N00014-75-C-0621 NR 047-048 with the U.S. Office of Naval Research. Reproduction in whole or part is permitted for any purpose of the U.S. Government.     

PRIMAL PERTURBATION SIMPLEX ALGORITHMS FOR LINEAR PROGRAMMING

1. IntroductionExtensive research in linear programming, such as [1,2,9,10,if, 12,13,14,19], hasbeen to improve pivot rules to reduce the number of iterations required. Relatively lesseffort was made on perturbing problem data with pivot rules unaltered (for instance, theself--dual parametric method [7] and perturbation--based methods [3,5]). And, becauseof the papametrization, the latter do not proceed as simply as the conventional simplexalgorithm itself.Recently, Pan [17] proposes new pert…     

Some fundamental issues of basic line search algorithm for linear programming problems

In this article we present the fundamental idea, concepts and theorems of a basic line search algorithm for solving linear programming problems which can be regarded as an extension of the simplex method. However, unlike the iteration of the simplex method from a basic point to an improved adjacent basic point via pivot operation, the basic line search algorithm, also by pivot operation, moves from a basic line which contains two basic feasible points to an improved basic line which also contains two basic feasible points whose objective values are no worse than that of the two basic feasible points on the previous basic line. The basic line search algorithm may skip some adjacent vertices so that it converges to an optimal solution faster than the simplex method. For example, for a 2-dimensional problem, the basic line search algorithm can find an optimal solution with only one iteration.     

Worst case behavior of the steepest edge simplex method

At each nondegenerate iteration of the steepest-edge simplex method one moves from a vertex of the polyhedron, P, of feasible points to an adjacent vertex along an edge that is steepest with respect to the linear objective function ψ. In this paper we show how to construct a sequence of linear programs (P_n,ψ_n) in n variables for which the number of iterations required by the steepest edge simplex method is 2ⁿ−1.     

An efficient algorithm for sparse null space basis problem using ABS methods

We propose a way to use the Markowitz pivot selection criterion for choosing the parameters of the extended ABS class of algorithms to present an effective algorithm for generating sparse null space bases. We explain in detail an efficient implementation of the algorithm, making use of the special MATLAB 7.0 functions for sparse matrix operations and the inherent efficiency of the ANSI C programming language. We then compare our proposed algorithm with an implementation of an efficient algorithm proposed by Coleman and Pothen with respect to the computing time and the accuracy and the sparsity of the generated null space bases. Our extensive numerical results, using coefficient matrices of linear programming problems from the NETLIB set of test problems show the competitiveness of our implemented algorithm.