Results 1 
7 of
7
A simplex algorithm whose average number of steps is bounded between two quadratic functions of the smaller dimension
 JOURNAL OF THE ACM
, 1985
"... It has been a challenge for mathematicians to confirm theoretically the extremely good performance of simplextype algorithms for linear programming. In this paper the average number of steps performed by a simplex algorithm, the socalled selfdual method, is analyzed. The algorithm is not started ..."
Abstract

Cited by 30 (2 self)
 Add to MetaCart
It has been a challenge for mathematicians to confirm theoretically the extremely good performance of simplextype algorithms for linear programming. In this paper the average number of steps performed by a simplex algorithm, the socalled selfdual method, is analyzed. The algorithm is not started at the traditional point (1,..., but points of the form (1, e, e2,...)T, with t sufficiently small, are used. The result is better, in two respects, than those of the previous analyses. First, it is shown that the expected number of steps is bounded between two quadratic functions cl(min(m, n))' and cz(min(m, n)) ' of the smaller dimension of the problem. This should be compared with the previous two major results in the field. Borgwardt proves an upper bound of 0(n4m1'(n1') under a model that implies that the zero vector satisfies all the constraints, and also the algorithm under his consideration solves only problems from that particular subclass. Smale analyzes the selfdual algorithm starting at (1,..., He shows that for any fixed m there is a constant c(m) such the expected number of steps is less than ~(m)(lnn)"'("+~); Megiddo has shown that, under Smale's model, an upper bound C(m) exists. Thus, for the first time, a polynomial upper bound with no restrictions (except for nondegeneracy) on the problem is proved, and, for the first time, a nontrivial lower bound of precisely the same order of magnitude is established. Both Borgwardt and Smale require the input vectors to be drawn from
Further Development on the Interior Algorithm for Convex Quadratic Programming
 Dept. of EngineeringEconomic Systems, Stanford University
, 1987
"... The interior trust region algorithm for convex quadratic programming is further developed. This development is motivated by the barrier function and the "center" pathfollowing methods, which create a sequence of primal and dual interior feasible points converging to the optimal solution. At each it ..."
Abstract

Cited by 8 (1 self)
 Add to MetaCart
The interior trust region algorithm for convex quadratic programming is further developed. This development is motivated by the barrier function and the "center" pathfollowing methods, which create a sequence of primal and dual interior feasible points converging to the optimal solution. At each iteration, the gap between the primal and dual objective values (or the complementary slackness value) is reduced at a global convergence ratio (1 \Gamma 1 4 p n ), where n is the number of variables in the convex QP problem. A safeguard line search technique is also developed to relax the smallstepsize restriction in the original path following algorithm. Key words: Convex Quadratic Programming, Primal and Dual, Complementarity Slackness, Polynomial Interior Algorithm. Abbreviated title: Interior Algorithm for Convex Quadratic Programming Since Karmarkar proposed the new polynomial algorithm (Karmarkar [19]), several developments have been made to the growing literature on interior a...
The Linear Complementarity Problem, Sufficient Matrices and the CrissCross Method
, 1990
"... Specially structured Linear Complementarity Problems (LCP's) and their solution by the crisscross method are examined in this paper. The crisscross method is known to be finite for LCP's with positive semidefinite bisymmetric matrices and with Pmatrices. It is also a simple finite algorithm for o ..."
Abstract

Cited by 6 (4 self)
 Add to MetaCart
Specially structured Linear Complementarity Problems (LCP's) and their solution by the crisscross method are examined in this paper. The crisscross method is known to be finite for LCP's with positive semidefinite bisymmetric matrices and with Pmatrices. It is also a simple finite algorithm for oriented matroid programming problems. Recently Cottle, Pang and Venkateswaran identified the class of (column, row) sufficient matrices. They showed that sufficient matrices are a common generalization of P and PSDmatrices. Cottle also showed that the principal pivoting method (with a clever modification) can be applied to row sufficient LCP's. In this paper the finiteness of the crisscross method for sufficient LCP's is proved. Further it is shown that a matrix is sufficient if and only if the crisscross method processes all the LCP's defined by this matrix and all the LCP's defined by the transpose of this matrix and any parameter vector.
A Polynomial Method of Weighted Centers for Convex Quadratic Programming
 Journal of Information & Optimization Sciences
, 1991
"... A generalization of the weighted central pathfollowing method for convex quadratic programming is presented. This is done by uniting and modifying the main ideas of the weighted central pathfollowing method for linear programming and the interior point methods for convex quadratic programming. B ..."
Abstract

Cited by 3 (2 self)
 Add to MetaCart
A generalization of the weighted central pathfollowing method for convex quadratic programming is presented. This is done by uniting and modifying the main ideas of the weighted central pathfollowing method for linear programming and the interior point methods for convex quadratic programming. By means of the linear approximation of the weighted logarithmic barrier function and weighted inscribed ellipsoids, `weighted' trajectories are defined. Each strictly feasible primal dual point pair define such a weighted trajectory. The algorithm can start in any strictly feasible primaldual point pair that defines a weighted trajectory, which is followed through the algorithm. This algorithm has the nice feature, that it is not necessary to start the algorithm close to the central path and so additional transformations are not needed. In return, the theoretical complexity of our algorithm is dependent on the position of the starting point. Polynomiality is proved under the usual mild cond...
Edmonds Fukuda Rule And A General Recursion For Quadratic Programming
"... A general framework of nite algorithms is presented here for quadratic programming. This algorithm is a direct generalization of Van der Heyden's algorithm for the linear complementarity problem and Jensen's `relaxed recursive algorithm', which was proposed for solution of Oriented Matroid programmi ..."
Abstract
 Add to MetaCart
A general framework of nite algorithms is presented here for quadratic programming. This algorithm is a direct generalization of Van der Heyden's algorithm for the linear complementarity problem and Jensen's `relaxed recursive algorithm', which was proposed for solution of Oriented Matroid programming problems. The validity of this algorithm is proved the same way as the finiteness of the crisscross method is proved. The second part of this paper contains a generalization of EdmondsFukuda pivoting rule for quadratic programming. This generalization can be considered as a finite version of Van de Panne  Whinston algorithm and so it is a simplex method for quadratic programming. These algorithms uses general combinatorial type ideas, so the same methods can be applied for oriented matroids as well. The generalization of these methods for oriented matroids is a subject of another paper.
Finding Nash Equilibria of Bimatrix Games
"... This thesis concerns the computational problem of finding one Nash equilibrium of a bimatrix game, a twoplayer game in strategic form. Bimatrix games are among the most basic models in noncooperative game theory, and finding a Nash equilibrium is important for their analysis. The Lemke–Howson algo ..."
Abstract
 Add to MetaCart
This thesis concerns the computational problem of finding one Nash equilibrium of a bimatrix game, a twoplayer game in strategic form. Bimatrix games are among the most basic models in noncooperative game theory, and finding a Nash equilibrium is important for their analysis. The Lemke–Howson algorithm is the classical method for finding one Nash equilibrium of a bimatrix game. In this thesis, we present a class of square bimatrix games for which this algorithm takes, even in the best case, an exponential number of steps in the dimension d of the game. Using polytope theory, the games are constructed using pairs of dual cyclic polytopes with 2d suitably labelled facets in dspace. The construction is extended to two classes of nonsquare games where, in addition to exponentially long Lemke–Howson computations, finding an equilibrium by support enumeration takes exponential time on average. The Lemke–Howson algorithm, which is a complementary pivoting algorithm, finds at least one solution to the linear complementarity problem (LCP) derived from a bimatrix