BISC Seminar Announcement Oct 23rd, 4-5pm, 310 Soda Hall

Frank Hoffmann (fhoffman@cs.berkeley.edu)
Sun, 19 Oct 1997 15:22:25 +0200 (MET DST)

************************************************************
B I S C S e m i n a r A n n o u n c e m e n t
************************************************************

INDUSTRIAL AND FINANCIAL APPLICATIONS OF SOFT COMPUTING

Piero P. Bonissone
Information Technology Laboratory
General Electric Corporate Research and Development
Schenectady, NY 12309

Email: Bonissone@crd.ge.com

Date: Thursday, October 23, 1997
Time: 4-5pm
Location : 310 Soda Hall

Abstract

Soft Computing (SC) represents the combination of emerging problem-solving
technologies such as Fuzzy Logic, Probabilistic Reasoning, Neural
Networks, and Genetic Algorithms. Each of these technologies
provide us with complementary reasoning and searching methods to solve complex,
real-world problems. After a brief description of each of these technologies,
we will analyze two SC applications to real-world problems.

* GA-TUNED FUZZY CONTROLLERS FOR FREIGHT TRAIN CRUISE CONTROL.
We will describe the design and tuning of a controller for enforcing compliance
with a prescribed velocity profile for a rail-based transportation system.
This
requires following a trajectory, rather than fixed setpoints (as in
automobiles). We synthesize a fuzzy controller for tracking the velocity
profile, while providing a smooth ride and staying within the prescribed speed
limits. We use a genetic algorithm to tune the fuzzy controller's performance
by adjusting its parameters (the scaling factors and the membership functions)
in a sequential order of significance. We show that this approach results in a
controller that is superior to the manually designed one, and with only modest
computational effort. This makes it possible to customize automated tuning to a
variety of different configurations of the route, the terrain, the power
configuration, and the cargo.

* RESIDENTIAL PROPERTY EVALUATION USING SOFT COMPUTING TECHNIQUES.
Residential property valuation is the process of determining a dollar estimate
of the property value for given market conditions. A residential property is a
single family residence designed or intended for owner- occupancy. The value of
a property changes with market conditions, so any estimate of its value
(appraisal) must be periodically updated to reflect those market changes. Any
appraisal must also be supported by current evidence of market conditions, e.g.
recent real estate transactions.

We will briefly present some related works and competitive approaches followed
by a description of our proposed fuzzy CBR process. We will then illustrate the
process with an example, providing an estimate value for the subject property
and computing a confidence value for the estimate. Finally, we will present a
statistical analysis of the overall process, followed by a summary of our
findings, conclusions, and future work.

**********************************************************************
Please direct questions with regard to the contents of the talk
and request for papers to the speaker.
**********************************************************************


---------------------------------------------------------------------------
Frank Hoffmann UC Berkeley
Computer Science Division Department of EECS
Email: fhoffman@cs.berkeley.edu phone: 1-510-642-8282
URL: http://http.cs.berkeley.edu/~fhoffman fax: 1-510-642-5775
---------------------------------------------------------------------------