Assisting Design Given Multiple Performance Criteria
Abstract
A design system that accepts multiple performance criteria faces the problem of trading off one criterion for another. Understanding the user's preference about such tradeoffs is essential to producing the most desirable design. An argument is made against specifying preferences over all possibilities before the design process begins. The proposed solution reduces its dependence on the specification's accuracy by encouraging interaction between the user and the system. The system produces quick, high quality information about realizable designs and performance tradeoffs between them. The user interactively evaluates, directs, and terminates the design exploration. Techniques called decoupled design, alteration strategies, and sample search are key elements in the implementation. Decoupled design generates a large number of designs efficiently (an alternative technique that uses linear programming is also presented). Alteration strategies are abstract descriptions of design modifications that assist in directing the design process and creating novel structures. Sample search is a framework to control search using domain specific heuristics. The design system operates in the domain of non-regular, application specific, signal processing architectures. The design process occurs in two phases: First the architectural structure with uninstantiated operators is created then each operator's implementation is selected.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1988
- Accession Number
- ADA200982
Entities
People
- Dennis C. Fogg
Organizations
- Massachusetts Institute of Technology