Showing papers in "Acta Cybernetica in 1989"

On the performance of on-line algorithms for partition problems

Abstract: We consider the performance of the greedy algorithm and of on-line algorithms for partition problems in combinatorial optimization. After surveying known resuls we give bounds for matroid and graph partitioning, and discuss the power of non-adaptive adversaries for proving lower bounds.

Nonlinear parameter estimation by global optimization—efficiency and reliability

Abstract: In this paper we first show that the objective function of a least squares type nonlinear parameter estimation problem can be any non- negative real function, and therefore this class of problems corre- sponds to global optimization. Two non-derivative implementations of a global optimization method are presented; with nine standard test functions applied to measure their efficiency. A new nonlinear test problem is then presented for testing the reliability of global op- timization algorithms. This test function has a countable infinity of local minima and only one global minimizer. The region of attraction of the global minimum is of zero measure. The results of efficiency and reliability tests are given.

On the worst-case performance of the NKF bin-packing heuristic

Abstract: of items (elements) with a weight function on items and a sequence of unit-capacity bins Bt, B2, .... In this paper, we assume that the item weights are real numbers in the range (0, 1] and that the list is given by the weights. The problem is to find a packing of the items in the bins such that the sum of the items in each bin is not greater than 1, and the number of bins used is minimized. This problem is NP-hard [GJ] and therefore heuristic algorithms which give "good" solutions in an acceptable computing time are investigated [J], [JDUGG]. We are interested in the worst-case behaviour of the Next-k Fit' (NkF) algorithm. For this, an upper and a lower bound were given in Johnson's paper. We shall improve both bounds.

Two transformations of attribute grammars improving the complexity of their evaluation

Abstract: Several papers have been written recently on designing efficient evaluators for attribute grammars (AGs). Some of these papers (e.g. [6], [7]) provide techniques to optimize the time complexity of the evaluators for certain classes of AGs (the class of absolutely noncircular AGs in the referenced papers), other ones (e.g. [5], [9]) try to reduce the storage requirement of the evaluators. The same goal of these papers is, however, to optimize evaluation by improving the evaluator itself in some respect. Our aim is to improve the AG to be evaluated — by the application of a suitable transformation — not the evaluator (of a fixed type) by which the evaluation is actually performed. Of course, this approach cannot provide general optimization results as the previous one, but in some cases it can be quite powerful. In this paper we present two transformation techniques and show how they work in restricted classes of AGs.