Informatica

Most efforts in business process reengineering to date are motivated by potential improvement in performance measures such as costs, quality of products or services, and cycle time for their delivery. While obviously important, these factors do not necessarily reflect any intrinsic improvement in how work is organized. This paper presents an approach to the structural analysis of business processes. The goal is to capture an inherent degree of synchronization of the various interdependent activities involved. This way, the process before and after reengineering, or alternative designs, can be compared without arbitrary scaling effects introduced by nonstructural variables. Our methodology for modeling and analysis incorporates established techniques in fuzzy logic and systems. Its application to a well-known example is used as illustration.

Efficiency of one automatic estimation and c1usterization procedure of one-dimensional Gaussian mixture which combines EM algorithm with non-parametric estimation is considered. The paper is based on mathematical methods of statistical estimation of a mixture of Gaussian distributions presented by R. Rudzkis and M. Radavičius (1995). The main result of the implementation of the mathematical methods is completely automatic procedure which can start from no information about unknown parameters and finish with final mixture model (tested for adequacy).

Statistical properties are examined for a class of pipelined-block linear time-varying (LTV) and linear time-invariant (LTI) discrete-time systems. Pipelined-block equations are derived, using the general solution of LTV discrete-time system in state space. Afterwards, we analysed the state covariance and output covariance matrices of pipelined-block LTV and LTI discrete-time systems in state space. For this class of pipelined-block realizations expressions are found for calculation of characteristics of the roundoff noise. Finally, scaling in the pipelined LTV discrete-time systems in state space is considered.

Queueing systems with a single device are well developed (see, for example, Borovkov, 1972; 1980). But there are only several works in the theory of multiphase queueing systems in heavy traffic (see Iglehart, Whitt, 1970b) and no proof of laws of the iterated logarithm for the probabilistic characteristics of multiphase queuing systems in heavy traffic. The law of the iterated logarithm for the waiting time of a customer is proved in the first part of the paper (see Minkevičius, 1995). In this work, theorems on laws of the iterated logarithm for the other main characteristics of multiphase queuing systems in heavy traffic (a summary queue length of customers, a queue length of customers, a waiting time of a customer) are proved.

In this paper we define a class of edge-weighted graphs having nonnegatively valued bisections. We show experimentally that complete such graphs with more than three vertices and also some special graphs with only positive edges can be applied to improve the existing lower bounds for a version of the quadratic assignment problem, namely with a matrix composed of rectilinear distances between points in the Euclidean space.

This paper deals with the design problem of generalized linear controllers for linear systems with after-effect so that the resulting closed-loop system is globally uniformly asymptotically stable in the Lyapunov's sense. The controllers are universal in the sense that they include the usual delays (namely, point, distributed and mixed point-distributed delays) which can be finite, infinite or even time-varying. In Part I of the paper, some preliminary concepts and results on stabilizability are given.

In this paper we are concerned with global optimization, which can be defined as the problem of finding points on a bounded subset of R^m, in which some real-valued function f(x) assumes its optimal value. We consider here a global optimization algorithm. We present a stochastic approach, which is based on the simulated annealing algorithm. The optimization function f(x) here is discrete and with noise.

Traditional population dynamics models, explaining separate aspects of observed ecological phenomena, are well understood mathematically, but their extension to a wider class of phenomena is difficult. Artificial life models represent another extreme, capable of simulating almost every biological phenomenon, but not accessible to mathematical analysis. A population model having a scope between the two extremes – but flexible, extensible, and within reach of mathematical analysis – is introduced in this paper. The model is based on piece-wise linear dynamical models of individual organisms. Parameter ranges in which the model meets natural ecological requirements is explored through analysis and simulations.