Implementation of sub-structuring within an object-oriented framework

R. I. Mackie

    Research output: Contribution to journalArticlepeer-review

    11 Citations (Scopus)


    This paper describes how the UtDU decomposition method and sub-structuring algorithms can be implemented using object-oriented techniques. It is shown that this enables the algorithms to be implemented very concisely. Moreover, there is no increase in code complexity when the algorithms are extended to take account of sparsity. The sub-structuring, or domain decomposition algorithms are expressed in block matrix terms, and classes are used to represent each of these matrices. The solution processes are incorporated in a finite element program. The finite element program uses a distributed data structure, and this facilitates a straightforward interface between the finite element program and the mathematics. Moreover, the program possesses a clear control structure for responding to user changes to the finite element model.
    Original languageEnglish
    Pages (from-to)749-758
    Number of pages10
    JournalAdvances in Engineering Software
    Issue number10-11
    Publication statusPublished - 2001


    • Finite elements
    • Object-oriented programming
    • Sub-structuring
    • Domain decomposition
    • Equation solution


    Dive into the research topics of 'Implementation of sub-structuring within an object-oriented framework'. Together they form a unique fingerprint.

    Cite this