The use of measures of modularity in electronic design

  • Ian Armstrong Rogers

    Student thesis: Doctoral Thesis


    This dissertation describes the creation of novel electronic devices using a modular design process, and novel algorithms to evaluate the modularity of the electronic designs.

    Some advantages of modularity are to be able to aim to save time, reduce risk and therefore to save money during the development process.

    Existing algorithms were investigated and five were identified in the literature that had aspects that could have been suitable for electronic designs. A range of different features were extracted from them for use in the research.

    As a start to the research, the modularity of electronic designs created at a collaborating company was evaluated by applying the two most promising algorithms to some existing products. From that work, a new modular design methodology was introduced at the collaborating company and five new products were designed. The modularity of these new products was evaluated using three algorithms selected from the five identified in the literature.

    The design process was revised to improve the measures of modularity and another six new products were designed and their modularity evaluated using all five algorithms identified in the literature.

    A single new algorithm did not appear to be suitable to evaluate the modularity of electronic designs in the different situations identified. Instead, five new and novel algorithms were created. Two of the new algorithms were for two different types of electronic design created at the collaborating company: bespoke design and design for mass manufacture. The other three new algorithms were more general algorithms for: new bespoke design, new products closely based on an existing product (product progression), and design for mass manufacture.

    The five new algorithms were applied to the two products created prior to the start of the research and to the other eleven new products created during the research. The results are presented, discussed and compared to results from the five algorithms selected from the literature.
    Date of AwardMay 2012
    Original languageEnglish
    Awarding Institution
    • University of Portsmouth
    SponsorsDepartment of Trade and Industry & BlueMT Ltd
    SupervisorDavid Sanders (Supervisor), Giles Tewkesbury (Supervisor) & David Robinson (Supervisor)

    Cite this