TABLE OF CONTENTS Michael Oberguggenberger
Institut f r Technische Mathematik, Geometrie und Bauinformatik, Universit t Innsbruck
Summary. Queueing problems arise in civil engineering, e. g., in earth work at large construction sites when an excavator serves a number of transport vehicles. Due to a large number of fuzzy side conditions, it is not plausible that a precise estimate for the input parameters can be given, as required in standard probabilistic queueing models. In this article, two alternatives are described that allow to incorporate data uncertainty: a probabilistic queueing model with fuzzy input and fuzzy probabilities, and a purely fuzzy queueing model formulated in terms of network theory.
There is increasing awareness in the engineering community [1, 20, 24, 28] that probability theory alone does not suffice for modelling the uncertainties arising in engineering problems. Indeed, the data commonly available, say in soil mechanics or construction management, are often scarce, vague, ambiguous or in any case in need of interpretation. This necessitates the development of more flexible tools for assessing and processing subjective knowledge and expert estimates.
Using risk analysis, it is usually possible for the planning engineer to provide ranges for the fluctuations of the parameters involved at various risk levels. This opens the door for employing fuzzy sets, possibility theory or random set theory. When these types of methods are employed for describing the input data, it is essential that arithmetical processing is possible in the engineering models and results in output data of the same type. In probability theory, functions...
