Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.458633
Title: Some investigations into the numerical solution of initial value problems in ordinary differential equations
Author: Hayden, G. N.
Awarding Body: Newcastle University
Current Institution: University of Newcastle upon Tyne
Date of Award: 1976
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
In this thesis several topics in the numerical solution of the initial value problem in first-order ordinary differential equations are investigated. Consideration is given initially to stiff differential equations and their solution by stiffly-stable linear multistep methods which incorporate second derivative terms. Attempts are made to increase the size of the stability regions for these methods both by particular choices for the third characteristic polynomial and by the use of optimization techniques while investigations are carried out regarding the capabilities of a high order method. Subsequent work is concerned with the development of Runge-Kutta methods which include second-derivative terms and are implicit with respect to y rather than k. Methods of order three and four are proposed which are L-stable. The major part of the thesis is devoted to the establishment of recurrence relations for operators associated with linear multistep methods which are based on a non-polynomial representation of the theoretical solution. A complete set of recurrence relations is developed for both implicit and explicit multistep methods which are based on a representation involving a polynomial part and any number of arbitrary functions. The amount of work involved in obtaining multistep methods by this technique is considered and criteria are proposed to decide when this particular method of derivation should be employed. The thesis is concluded by using Prony's method to develop one-step methods and multistep methods which are exponentially adaptive and as such can be useful in obtaining solutions to problems which are exponential in nature.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.458633  DOI: Not available
Share: