leff@smu.CSNET.UUCP (08/31/87)
--------------------------------------------------------------------------
ARTIFICIAL INTELLIGENCE IN COMMUNICATION NETWORKS
Andrew Jennings & Phil Nicholson
(Computing Systems and Information Technology Conference,
Brisbane, June 1987)
The AI technologies offer many new and exciting
possibilities for the next generation of communication networks. This paper
introduces the new technologies and discusses their relevance to communication
networks and services. We foresee a greatly expanded role for AI technology,
even to the extent of finding its place in demanding real-time communication
applications such as network configuration and management. Here we suggest that
AI technology offers new means of dealing with the challenges of constructing
the next generation of communication networks. It offers new approaches to
design, and it offers some promise of creating communication facilities that
can be used in creative ways through better service delivery techniques.
---------------------------------------------------------------------------
USING SIMULATORS TO DEVELOP NEW HEURISTICS
(Modelling & Simulation Conference, Melbourne October 1987)
Andrew Jennings
Head, AI Technology, Telecom Australia Research
The use of simulators to develop new heuristics has long been a goal of
artificial intelligence (AI) research. If we can develop new
heuristics by coupling a learning system to repeated simulation of some
complex real world problem then we can greatly assist the process of system
development. This paper reviews some of the research on the use of simulators
and suggests a new avenue for exploration.
-----------------------------------------------------------------------------
A Model Based Approach to Automated VLSI Synthesis.
C.D.Rowles
G. Foster
C. Leckie
( IREECON '87, September 1987 )
A new approach to the automated design of hardware is presented.
The system takes a hardware behavioural specification in the "C"
programming language and produces a detailed architecture design.
The main features are the use of Artificial Intelligence techniques
to emulate the design process, and models for different design
methodologies.
-------------------------------------------------------------------------------
Beyond Expert Systems: Opportunities.
A.J.Jennings
C.D.Rowles
( Australian Computer Conference, September 1987 )
Expert systems are now widely accepted as a software technology capable of
impressive performance in application areas that were formerly considered the
sole
domain of human experts. This success is the result of a careful selection
of applications closely matched to
the technology. But expert systems are inadequate for
solving many important {\it real-world} problems. Instead, other Artificial
Intelligence techniques must be adopted. These are less well developed
but the pay-offs are potentially large.
This paper
presents examples of opportunities that go beyond expert systems and some of theissues involved.
-------------------------------------------------------------------------------
Simulation of VLSI System Architectures Using Prolog.
C.D.Rowles
( Int. Conf. on Modelling and Simulation, October 1987 )
As VLSI circuit complexities increase, the gap between
system-architecture and logic-circuit design widens,
rendering logic design and verification CAD tools
unsuitable for the architectural design problem. The
use of Prolog as a Hardware Description Language
provides an effective and efficient means to overcome
this problem. In addition, Prolog facilitates the
exploration of system behaviour in a way that is not
possible with conventional simulation techniques.
-------------------------------------------------------------------------------
Automated Hardware Synthesis Using Models.
C.D.Rowles
( Int. Conf. on Modelling and Simulation, October 1987 )
A new approach to automating the design of special purpose hardware is
presented. The system described takes a hardware behavioural specification,
in a subset of the ``C" programming language, as its input and produces a
detailed architecture design. The main features of the system are
its use of Artificial
Intelligence techniques to emulate the design process
and the use of heuristics and models to
classify design goals and emulate different design
strategies.
UUCP: ...!{seismo, mcvax, ucb-vision, ukc}!munnari!trlamct.trl!andrew
ARPA: andrew%trlamct.trl.oz@seismo.css.gov
Andrew Jennings Telecom Australia Research eet)
Postal address :
Andrew Jennings,
Telecom Australia Research,
770 Blackburn Rd.,
Clayton 3168,
AUSTRALIA
--
UUCP: ...!{seismo, mcvax, ucb-vision, ukc}!munnari!trlamct.trl!andrew
ARPA: andrew%trlamct.trl.oz@seismo.css.gov
Andrew Jennings Telecom Australia Research Labs
(Postmaster:- This mail has been acknowledged.)