Search

Zero-sum sequential games where both control variables are stopping times are considered. Two game problems are dealt with: (G[sub 1]) is a problem in which the players are allowed the possibility of not stopping the game, and in the other (G[sub 2]) they are obliged to stop the observed process at some finite (but not preassigned) time. The problem (G[sub 1]) is well known. In the present paper we mainly investigate (G[sub 2]) as compared with (G[sub 1]). We give sufficient conditions for the game problem (G[sub 2]) to have a value which is consequently equal to that of (G[sub 1]) and, in parallel with it, present the constructive algorithm of the value in the natural form. The saddle point in each problem is found under a certain condition. The monotone case and the Markov case are finally investigated as the special cases. [ABSTRACT FROM AUTHOR]

This paper presents two new combinatorial algorithms for the generalized circulation problem. After an initial step in which all flow-generating cycles are canceled and excesses are created, both algorithms bring these excesses to the sink via highest-gain augmenting paths. Scaling is applied to the fixed amount of flow that the algorithms attempt to send to the sink, and both node and arc excesses are used. The algorithms have worst-case complexities of O(m²(m + n log n) log B), where n is the number of nodes, in is the number of arcs, and B is the largest integer used to represent the gain factors and capacities in the network. This bound is better than the previous best bound for a combinatorial algorithm for the generalized circulation problem, and if m = O(n4/3-c), it is better than the previous best bound for any algorithm for this problem. [ABSTRACT FROM AUTHOR]

This paper studies approximation algorithms for the problem of nonpreemptively scheduling n jobs on a single machine to minimize total weighted late work, where the late work for a job is the amount of processing of this job that is performed after its due date. A family of approximation algorithms {DP€} is derived. For any € > 0, DP€ delivers a schedule having total weighted late work which does not exceed (1 + €) times that of an optimal schedule. Since DP€ requires O(n³ log n + n³/€) time, the family {DP€} is a fully polynomial approximation scheme. [ABSTRACT FROM AUTHOR]

We consider the n-player houseswapping game of Shapley and Scarf (1974), with indifferences in preferences allowed. It is well known that the strict core of such a game may be empty, single valued, or multi valued. We define a condition on such games called segmentability, which means that the set of players can be partitioned into a top trading segmentation. It generalizes Gale's well-known idea of the partition of players into top trading cycles (which is used to find the unique strict core allocation in the model with no indifference). We prove that a game has a nonempty strict core if and only if it is segmentable. We then use this result to devise an O n3 algorithm which takes as input any houseswapping game, and returns either a strict core allocation or a report that the strict core is empty. Finally, we are also able to construct a linear inequality system whose feasible region's extreme points precisely correspond to the allocations of the strict core. This last result parallels the results of Vande Vate (1989) and Rothblum (1992) for the marriage game of Gale and Shapley (1962). [ABSTRACT FROM AUTHOR]

We present a family of variants of the Simplex method, which are based on a Constraint-By-Constraint procedure: the solution to a linear program is obtained by solving a sequence of subproblems with an increasing number of constraints. We discuss several probabilistic models for generating linear programs. In all of them the underlying distribution is assumed to be invariant under changing the signs of rows or columns in the problem data. A weak regularity condition is also assumed. Under these models, for linear programs with d variables and m + d inequality constraints, the expected number of pivots required by these algorithms is bounded by a function of min(m, d) only. In particular this means that, for a fixed number of variables, the expected number of pivots is bounded by a constant when the number of constraints tends to infinity. Since Smale's original model [S1] satisfies our probabilistic assumptions, the same results apply to his model, although not to the particular algorithm he analyzes. We also present some results for models generating only feasible linear programs, and for Bland's pivoting rule. We conclude with a discussion of our probabilistic models, and show why they are inadequate for obtaining meaningful results unless d and m are of the same order of magnitude. [ABSTRACT FROM AUTHOR]