Return to search

Efficient Simulation, Accurate Sensitivity Analysis and Reliable Parameter Estimation for Delay Differential Equations

Delay differential equations (DDEs) are a class of differential equations that have received considerable recent attention and
been shown to model many real life problems, traditionally formulated as systems of ordinary differential equations (ODEs),
more naturally and more accurately. Ideally a DDE modeling package should provide facilities for approximating the solution,
performing a sensitivity analysis and estimating unknown parameters. In this thesis we propose new techniques for efficient simulation, accurate sensitivity analysis and reliable parameter estimation of DDEs.

We propose a new framework for designing a delay differential equation (DDE) solver which works with any supplied initial value
problem (IVP) solver that is based on a general linear method (GLM) and can provide dense output. This is done by treating a
general DDE as a special example of a discontinuous IVP. We identify a precise process for the numerical techniques used when solving the implicit equations that arise on a time step, such as when the underlying IVP solver is implicit or the delay vanishes.

We introduce an equation governing the dynamics of sensitivities for the most general system of parametric DDEs. Then, having a similar view as the simulation (DDEs as discontinuous ODEs), we introduce a formula for finding the size of jumps that appear at discontinuity points when the sensitivity equations are integrated. This leads to an algorithm which can compute
sensitivities for various kind of parameters very accurately.

We also develop an algorithm for reliable parameter identification of DDEs. We propose a method for adding extra constraints to the
optimization problem, changing a possibly non-smooth optimization to a smooth problem. These constraints are effectively handled
using information from the simulator and the sensitivity analyzer.

Finally, we discuss the structure of our evolving modeling package DDEM. We present a process that has been used for incorporating
existing codes to reduce the implementation time. We discuss the object-oriented paradigm as a way of having a manageable design with reusable and customizable components. The package is programmed in C++ and provides a user-friendly calling sequences. The numerical results are very encouraging and show the effectiveness of the techniques.

Identiferoai:union.ndltd.org:TORONTO/oai:tspace.library.utoronto.ca:1807/19253
Date03 March 2010
CreatorsZivariPiran, Hossein
ContributorsEnright, Wayne
Source SetsUniversity of Toronto
Languageen_ca
Detected LanguageEnglish
TypeThesis

Page generated in 0.0022 seconds