EnglishCancella i cookie per ripristinare le impostazioni di lingua associate al browser in uso
Titolo/Abstract/Parole chiave

STOCHASTIC AND DETERMINISTIC SIMULATION TECHNIQUES FOR TRAFFIC AND ECONOMICS

Foscari W. R., Piero (2009) STOCHASTIC AND DETERMINISTIC SIMULATION TECHNIQUES FOR TRAFFIC AND ECONOMICS. Tesi di Dottorato , Università degli studi di Ferrara.

[img]
Anteprima
File PDF
Foscari_2006_PhD_Thesis.pdf

Download (6MB) | Anteprima

    Abstract

    In this work I present the result of different investigations conducted in the last years in the context of stochastic modeling for decision making in the areas of traffic simulation and economics. Traffic simulation has seen us from the Center for Modeling, Computing and Statistics involved in a project for the evaluation and planning of two highway stretches in the area around Ferrara. In particular we conducted the modeling and numerical simulation of the highway network, in collaboration with Michael Herty. Later the study of kinetic analysis and simulation techniques proved useful in another related setting, that is agent based models in economics, a discipline of growing importance in understanding the workings of markets, be they financial or centered on tangible goods. Due to my job in the asset management industry some of the research activity has been tilted towards practical methods for financial simulations, and in particular that of parallel random number generation is a topic that has been gaining importance during these last years. While at Eurizon Capital I developed a novel fast algorithm for moving over certain widely used random number streams, and at NEC Labs Europe this was further reimplemented as a core block of a professional C++ library for parallel Monte Carlo simulation in finance. Finally I present a small note on a common numerical artifact arising in Monte Carlo simulations when only a limited number of kinetic particles are used. Already with simple kernels the resulting probability distributions differ significantly from those predicted by theory and obtained with large particle sets.

    Tipologia del documento:Tesi di Dottorato (Tesi di Dottorato)
    Data:3 Marzo 2009
    Relatore:Pareschi, Lorenzo
    Coordinatore ciclo:Trillo, Stefano
    Istituzione:Università degli studi di Ferrara
    Dottorato:XXI Anno 2006 > SCIENZE DELL'INGEGNERIA
    Struttura:Dipartimento > Ingegneria
    Soggetti:Area 01 - Scienze matematiche e informatiche > MAT/08 Analisi numerica
    Parole chiave:traffico stradale, distribuzioni di ricchezza, numeri pseudocasuali
    Depositato il:04 Lug 2009 16:02

    Staff:

    Accesso riservatoAccesso riservato