JAQM Volume 5, Issue 3 - September 30, 2010
International Symposium on Stochastic Models in Reliability Engineering, Life Sciences and Operations Management(SMRLO'10)
Upon Scheduling and Controlling Large-Scale Stochastic Network Projects (p 382)
The problem of controlling large-size stochastic network projects of PERT type is considered. A conclusion is drawn that the need of proper control models for PERT projects is very important. The authors suggest aggregating the initial model in order to modify the latter to an equivalent one, but of medium or small-size.
For those network models effective on-line control algorithms are already developed. After observing the project's output at a routine control point and introducing proper control actions the aggregated network is transformed to the initial one, and the project’s realization proceeds.
The developed control techniques are especially effective for those R&D projects, when an on-line control has to be undertaken under a chance constraint. The suggested control model can be regarded as an additional tool to help the project manager to realize the project in time.
Solving Nonlinear Optimization Problems by Means of the Network Programming Method (p 389)
Vladimir N. BURKOV,
Irina V. BURKOVA
We suggest a new approach to solve discrete optimization problems, based on the possibility of presenting a function as a superposition of simpler functions. Such a superposition can be easily represented in the form of a network for which the inputs correspond to variables, intermediate nodes – to functions entering the superposition, and in the final node the function is calculated. Due to such representation the method has been called the method of network programming (in particular, dichotomic). The network programming method is applied for solving nonlinear optimization problems. The concept of a dual problem is implemented. It is proved that the dual problem is a convex programming problem. Necessary and sufficient optimality conditions for a dual problem of integer linear programming are developed.
Structure Decision Making Based on Universal Generating Functions for Refrigeration System (p 397)
This paper presents a method for calculation of reliability measures for supermarket refrigeration system. The system and its components can have different performance levels ranging from perfect functioning to complete failure and, so it can be interpreted as a multi-state system. Calculated reliability measures are used for decision making of system structure. The suggested approach based on combined Universal Generating Functions and stochastic processes method for computation of availability, output performance and performance deficiency for multi-state system. Corresponding procedures are suggested. A numerical example is presented in order to illustrate the approach.
Production Planning under Uncertain Demands and Yields (p 413)
The periodic demands of a single product are forecasted and given by a distribution function for each period. The product can be manufactured in n plants with heterogeneous characters. Each plant has its specific stochastic production capability. The expected capability and the standard deviation of each plant can be increased by allocation of additional budgets. The problem is to determine the total budget needed and its distribution among the n plants in order to ensure a complete fulfillment of the demands according to the due dates and the pre-given confidence levels.
Productivity Assessment and Improvement Measurement of Decision Making Units - An Application for Ranking Cities in Israel (p 421)
In this paper we will demonstrate how productivity and improvement rate of urban organizational units (called also Decision Making Units - DMU's) may be assessed when measured along several time periods. The assessment and subsequent ranking of cities is achieved by means of the Data Envelopment Analysis (DEA) methodology to determine DMU's efficiency for each period, the Cross Efficiency ranking method to rank DMU's and the Malmquist Index approach which measures changes in productivity relative to a base period. The above combined methodology will be applied to a case study of 70 Israeli cities in years 2006, 2007 and 2008.
Aestetics, Usefulness and Performance in User-Search – Engine Interaction (p 436)
Issues of visual appeal have become an integral part of designing interactive systems. Interface aesthetics may form users' attitudes towards computer applications and information technology. Aesthetics can affect user satisfaction, and influence their willingness to buy or adopt a system. This study follows previous studies that found that users associate aesthetics with other system attributes, e.g. usability. In this study, we asked whether the well-known phenomenon that beautiful things are perceived as good applies to the perception of the system’s usefulness. A controlled laboratory experiment tested the relationships between users’ perception of aesthetics, usefulness and user performance in tasks performed by participant using an interactive application that surrogated a search engine. We measured users’ perceptions of the search engine before and after they used the system to solve information-seeking tasks, and measured user task performance. As expected, significant correlations were found between perceived aesthetics and perceptions of usability and usefulness prior to actual use of the system. We did not find a relation between perceived aesthetics and usefulness after use; and we did not find an expected effect for aesthetic perceptions neither on perceived usefulness nor on performance. We conclude that there is need for a deeper understanding of aesthetic perceptions; a finer grain perspective of perceived aesthetics that differentiates between aesthetic dimensions may reveal that some aesthetic aspects have greater influence on the relations between aesthetics and usefulness.
Online Hoeffding Bound Algorithm for Segmenting Time Series Stream Data (p )
In this paper we introduce the ISW (Interval Sliding Window) algorithm, which is applicable to numerical time series data streams and uses as input the combined Hoeffding bound confidence level parameter rather than the maximum error threshold. The proposed algorithm has two advantages: first, it allows performance comparisons across different time series data streams without changing the algorithm settings, and second, it does not require preprocessing the original time series data stream in order to determine heuristically the reasonable error value. The proposed algorithm was implemented in two modes: off line and online. Finally, an empirical evaluation was performed on two types of time series data: stationary (normally distributed data) and non stationary (financial data).
Resource Reallocation Models for Deterministic Network Construction Projects (p )
Hierarchical budget reallocation models for a portfolio of construction network projects with deterministic activity durations are considered. Optimal reallocation models both at the company level and at the project level are developed.
Implementing Beta-Distribution in Project Management (p )
A research is undertaken to justify the use of beta-distribution p.d.f. for man-machine type activities under random disturbances. The case of using one processor, i.e., a single resource unit, is examined. It can be proven theoretically that under certain realistic assumptions the random activity – time distribution satisfies the beta p.d.f.
Changing more or less the implemented assumptions, we may alter to a certain extent the structure of the p.d.f. At the same time, its essential features (e.g. asymmetry, unimodality, etc.) remain unchanged.
The outlined above research can be applied to semi-automated activities, where the presence of man-machine influence under random disturbances is, indeed, very essential. Those activities are likely to be considered in organization systems (e.g. in project management), but not in fully automated plants.
Fuzzy Probabilistic Models for Structural Serviceability (p )
Structural serviceability is of uttermost importance for the overall performance of many common structures. As a rule, both the load effects (serviceability indicators due to loading) and admissible constraints (ensuring required structural performance) are random variables of considerable scatter and significant vagueness. Common experience indicates that a structure does not loose its ability to comply with specified performance requirements abruptly at a distinct point of the serviceability indicator, but gradually within a certain transition interval. Fuzzy-probabilistic methods are therefore employed to analyze the structural serviceability.
As an example, serviceability limit states of water retaining structures with respect to cracking are investigated in detail. Fuzzy probabilistic models are proposed to derive theoretical models for the limiting crack width. It is shown that the fuzzy probabilistic distribution of serviceability requirements may be used similarly as classical distribution function to specify the characteristic value of limiting crack width, to analyze reliability of crack width and to optimize structural design to achieve the minimum total costs.
Perinatal Assistance Network Planning Via Simulation (p )
Consider a geographical region where population is distributed in health districts, and there exists a neonatal care network, which includes birth centres able to supply assistance at three levels, respectively basic assistance, mild pathology care and intensive care. Each mother-to-be is admitted to a facility where the assistance level corresponds to the expected newborn conditions; newborn transfers from a lower to a higher-level facility are affected if conditions worsen. Each district has a known probabilistic demand for each care level previously mentioned and each facility is characterized by its capacity, i.e., the amount of patients simultaneously admittable there. A simulation model describing mothers and newborns movements from districts to birth centres and among centres has been built up, with the aim of revealing inadequacies of the assistance network and of obtaining useful suggestions about network resizing to improve service quality and reduce trouble due to distance. The model has been applied to Veneto region in North-East Italy but its use may be extended to other similar situations.
Semi-Markov Reliability Model of the Cold Standby System (p )
The semi-Markov reliability model of the cold standby system with renewal is presented in the paper. The model is some modification of the model that was considered by Barlow & Proshan (1965), Brodi & Pogosian (1978). To describe the reliability evolution of the system, we construct a semi-Markov process by defining the states and the renewal kernel of that one. In our model the time to failure of the system is represented by a random variable that denotes the first passage time from the given state to the subset of states. Appropriate theorems from the semi-Markov processes theory allow us to calculate the reliability function and mean time to failure. As calculating an exact reliability function of the system by using Laplace transform is often complicated we apply a theorem which deals with perturbed semi-Markov processes to obtain an approximate reliability function of the system.
Analytical and Numerical Studies of Perturbed Renewal Equations with Multivariate Non-Polynomial Perturbations (p )
The object of study is a model of nonlinearly perturbed continuous-time renewal equation with multivariate non-polynomial perturbations. The characteristics of the distribution generating the renewal equation are assumed to have expansions in a perturbation parameter with respect to a non-polynomial asymptotic. Exponential asymptotics for such a model as well as their applications are given. Numerical studies are performed to gain insights into the asymptotical results.
A Non-Parametric Test for a Change-Point in Linear Profile Data (p )
We propose a change-point approach for testing the constancy of regression parameters in a linear profile data set (panel data in econometrics).
Each sample collected over time in the historical data set consists of several multivariate observations for which a linear regression model is appropriate. The question now is whether all of the profiles follow a linear regression model with the same parameter vector or whether a change occurred in one or more model parameters after a special sample.
We use the partial sum operator in several dimensions to test the null hypothesis "H0: no change-point occurred" and propose a non-parametric size α-test.
In Bischoff and Gegg (2010) we compared our proposed method with the likelihood-ratio-test by Mahmoud et al. (2007) in a simulation study. By these simulations we could show that our procedure can, in contrast to the likelihood-ratio-test, even be applied to the non-normal case. In this paper, however, we show how to compute our proposed test statistic step-by-step by considering an artificial data set.