20-07-2012, 03:32 PM
Application of Fuzzy Logic in Computer-Aided VLSI Design
36Application of Fuzzy Logic.pdf (Size: 170.45 KB / Downloads: 38)
Abstract
Application of fuzzy logic structures in computeraided
design (CAD) of digital electronics substantially improves
quality of design solutions by providing designers with flexibility
in formulating goals and selecting tradeoffs. In addition, the
following aspects of a design process are positively impacted
by application of fuzzy logic: utilization of domain knowledge,
interpretation of uncertainties in design data, and adaptation of
design algorithms. We successfully applied fuzzy logic structures
in conjunction with constructive and iterative algorithms for
selecting of design solutions for different stages of the design
process.We also introduced fuzzy logic software development tool
to be used in CAD applications.
INTRODUCTION
FUZZY set theory has been recently applied in many
areas of science and engineering [24], [30], [35], [37],
[38]. Works raising possibility of application of fuzzy logic
in computer-aided design of digital electronics started to
appear in late 1980’s and early 1990’s. In [7]–[9], the authors
proposed fuzzy logic structures for decision making in architectural
synthesis of VLSI and transistor network optimization.
In these papers, the authors considered fuzzy logic as a method
of enhancement in the knowledge-based reasoning process, but
do not present implementation of the theoretical considerations
into working CAD tools.
CLASSIFICATIONS OF FUZZY
LOGIC APPLICATIONS IN CAD TOOLS
Human factors involved in design selection and evaluation
make design problems natural targets for application of fuzzy
logic. This does not preclude an application of fuzzy logic for
certain types of analysis when fuzziness in input data should
be taken into consideration.
The ultimate criteria for effectiveness of fuzzy logic in
design is practice, i.e., whether or not quality of solutions
obtained by fuzzy logic tools is higher than solutions generated
by tools, which do not employ fuzzy logic. The gain certainly
should be weighted against efforts required for incorporation
of fuzzy logic into design tools, their flexibility to adaptation
and user friendliness. Based on our experience, such efforts
are justifiable. In the following text, we will generalize our
experience and propose a plan for further developments.
Design of complex digital systems is a multistage process,
which involves application of variety of algorithms. Design
systems differ by how they organize design flow and formulate
subtasks as well as how they solve them. In this paragraph,
we propose a classification useful for describing fuzzy logic
applications in CAD.
APPLICATION OF FUZZY LOGIC IN
ITERATIVE ALGORITHMS FOR PHYSICAL DESIGN
Unlike constructive algorithms for physical design, which
produce a solution only at the end of the design process,
iterative algorithms operate with design solutions defined at
each iteration.
Many popular algorithms used in physical design automation
are iterative in nature (simulated annealing genetic algorithms).
A value of the objective function is used to compare
results of consecutive iterations and to select a solution based
on the maximal (minimal) value of the objective function [4],
[6], [10], [15], [22], [25], [28].
In the situations with the multiple objectives, they are
included in the objective function with the weight coefficients
reflecting hierarchy of goals in the eyes of the designer.
CONCLUSION
This paper presents characterization, justification, and classification
of fuzzy logic-based CAD tools for digital design. It
is demonstrated that application of fuzzy logic is beneficial on
all hierarchical levels of the design process. It allows to merge
traditional numeric techniques with the domain information
available in the linguistic form and, as a result, to improve
quality of design solutions.
Fuzzy logic structures are shown to be used in constructive
and iterative procedures. Short descriptions of several
CAD tools are accompanied by the experimental data and
comparison of design solutions is obtained by different tools
for popular benchmarks.