A network representation of the iterative design process has been presented. The representation is closely based on that previously developed for non-iterative design checking. The representation uses two sets of data values, denoted as Level 1 and Level 2, in a manner similar to first-order iterative computational schemes. Certain data values, designated internally modifiable parameters, serve as a reference between successive cycles. Provisions are made for retaining certain data between cycles, so as to suppress calculations in some iterations, when it is felt that the results are insensitive to changes in certain data values. The representation is illustrated through the application to a typical structural design problem. Detailed data structures and transformation process are presented. In the case of data, it is shown that the subscripted nature of the data must be explicitly taken into account, whereas the non-iterative process can be represented in terms of generic, or unsubscripted, data. The recursive procedures for evaluating and modifying data have been extended to handle the two-level, subscripted data representation. Four additional processes are defined: those to ACCEPT or REJECT the current cycle of results, and those to RETAIN and RELEASE data between cycles.

  • Corporate Authors:

    Carnegie Mellon University

    Department of Civil and Environmental Engineering
    5000 Forbes Avenue
    Pittsburgh, PA  United States  15213
  • Authors:
    • Tamm, M A
    • Fenves, S J
  • Publication Date: 1977-4

Media Info

  • Features: References;
  • Pagination: 44 p.

Subject/Index Terms

Filing Info

  • Accession Number: 00159591
  • Record Type: Publication
  • Source Agency: Office of Naval Research London (England)
  • Report/Paper Numbers: R-77-5 Intrm Rpt.
  • Files: TRIS
  • Created Date: Sep 20 1977 12:00AM