文档介绍：In telligen t Sim ulation T o ols for Mining Large Scien ti c Data Sets 1
In telligen t Sim ulation T o ols for Mining
Large Scien ti c Data Sets
F eng ZHA O
Xer ox Palo A lto R ese ar ch Center
3333 Coyote Hil l R o ad, Palo A lto, CA 94304
******@ m
Chris BAILEY-KELLOGG
Dartmouth Col le ge
6211 Sudiko L ab or atory, Hanover, NH 03755
******@ du
~
Xingang HUANG and Iv
an ORD O NEZ
The Ohio State University
2015 Neil A venue, Columbus, OH 43210
f huang,iordonez g ***@c io-s tate. edu
Receiv ed 24 F ebruary 1999
A bstr act
This pap er describ es problems, c hallenges, and opp ortunities for in-
tel ligent simulation of ph ysical systems. Protot yp e in telligen t sim ulation
to ols ha v e b een constructed for in terpreting massiv e data sets from ph ys-
ical elds and for designing engineering systems. W e iden tify the c harac-
teristics of in telligen t sim ulation and describ e sev eral concrete application
examples. These applications, whic h include w eather data in terpretation,
distributed con trol optimization, and spatio-temp oral di usion-reaction
pattern analysis, demonstrate that in telligen t sim ulation to ols are indis-
p ensable for the rapid protot yping of application programs in man yc hal-
lenging scien ti c and engineering domains.
Keyw ords In telligen t sim ulation, Scien ti c data mining, Qualitativ e
reasoning, Reasoning ab out ph ysical systems, Programmi ng en vironmen ts.
~
2 F eng ZHA O, Chris BAILEY-KELLOGG, Xingang HUANG, and Iv
an ORD O NEZ
x 1 In tro duction
Information tec hnology has fundamen tally c hanged the w a yw e conduct
scien ti c exp erimen ts and syn thesize engineering artifacts. F or instance, p o w er-
puters ha v e routinely b een used to extract in teresting features in satel-
lite images, diagnose abnormalities in n uclear reactors, allo cate resources in air
trac con trol, and discern subtle trends in sto c k mark ets, t