University of the Aegean
Numerical analysis of stochastic differential equations with applications in financial mathematics and molecular dynamics
Abstract
dc:descriptionIn this thesis we are interested in the numerical solution of stochastic differential equations (SDE) with solutions in a certain domain. Our goal is to construct explicit numerical schemes that preserve that domain, mainly for cases where the coefficients of the SDEs are non-linear. It is well known that the forward Euler scheme diverges on super-linear problems and the tamed Euler method does not necessarily preserve the structure of the original problem. We propose a new numerical scheme, using the semi-discrete method, for various classes of stochastic differential equations. For some super-linear problems (like the Heston 3/2-model) as well as sub-linear (like the CEV model), which appear in the field of financial mathematics, we are able to construct a positivity preserving scheme. Moreover, we apply our method to problems arising in the field of molecular dynamics, where our structure preserving scheme is able to approximate effectively some SDEs which appear after a coarse graining procedure. We also consider the case of Stochastic Delay Differential Equations (SDDEs) with non-negative solutions. Again we aim for explicit numerical schemes that preserve positivity. We expand the semi-discrete method from the Stochastic Ordinary Differential Equations (SODE) setting and apply it to the constant delay case, for which we prove strong convergence (DGBM model). Numerical experiments support our theoretical results.
Degree
thesis:*- Grantor dc:publisher
- University of the Aegean
- Year dc:date
- 2016
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Stamatiou, Ioannis
- Σταματίου, Ιωάννης
Subjects
dc:subject × 28- Ημι-διακριτή μέθοδος
- Άμεσο αριθμητικό σχήμα
- Υπερ-γραμμική τάση και διάχυση
- Διατήρηση θετικότητας
- Στοχαστικές διαφορικές εξισώσεις
- Αριθμητικές μέθοδοι
- Προσομοίωση Monte Carlo
- Ισχυρό σφάλμα εκτίμησης
- 3/2-μοντέλο
- Τάξη σύγκλισης
- Διαδικασία CEV με ιδιότητα επαναφοράς στο μέσο
- Στοχαστικό μοντέλο μεταβλητότητας
- Semi-discrete method
- Super-linear drift and diffusion
- Explicit numerical scheme
- Positivity preserving
- Stochastic differential equations
- Numerical methods
- Monte Carlo simulation
- Strong approximation error
- 3/2-model
- Order of convergence
- Mean-reverting CEV process
- Stochastic volatility model
- Φυσικές Επιστήμες
- Μαθηματικά
- Natural Sciences
- Mathematics
Rights
dc:rights- Statement dc:rights
-
- BY_ND
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/37499