COMMENT ⊗   VALID 00004 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.REQUIRE "ABSTRA.CMD[BIB,CSR]" SOURCE FILE
C00003 00003	.once center <<general instructions>>
C00018 00004	%4----------------------------------------------------------------------%1
C00022 ENDMK
C⊗;
.REQUIRE "ABSTRA.CMD[BIB,CSR]" SOURCE FILE
.every heading (January,1979 Reports,{Page})
.at "∞" ⊂"%4α∞%*"⊃
.next page
.once center <<general instructions>>
%3MOST RECENT CS REPORTS - JANUARY 1979

@Listed below are abstracts of the most recent reports published by the Computer
Science Department of Stanford University.

@TO REQUEST REPORTS:##Check the appropriate places on the enclosed order form,
and return the entire order form page (including mailing label) by
January 19, 1979.
In many cases we can print only a limited number of copies, and requests will
be filled on a first come, first serve basis.  If the code (FREE) is printed
on your mailing label, you will not be charged for hardcopy.  This exemption
from payment is limited primarily to libraries.  (The costs shown include all
applicable sales taxes.  PLEASE SEND NO MONEY NOW, WAIT UNTIL YOU GET AN 
INVOICE.)

@ALTERNATIVELY:  Copies of most Stanford CS Reports may be obtained by writing
(about 2 months after MOST RECENT CS REPORTS listing) to NATIONAL TECHNICAL
INFORMATION SERVICE, 5285 Port Royal Road, Springfield, Virginia 22161.  Stanford
Ph.D. theses are available from UNIVERSITY MICROFILMS, 300 North Zeeb Road,
Ann Arbor, Michigan 48106.
.begin nofill
%4----------------------------------------------------------------------%1
%4----------------------------------------------------------------------%1
%3STAN-CS-79-691  THE CONSTRUCTION OF INITIAL DATA FOR HYPERBOLIC SYSTEMS
			FROM NONSTANDARD DATA
%3Author:%1  Kenneth P. Bube↔(Thesis)
.end
@%3Abstract:%1@We study the first order systems of hyperbolic partial differential
equations with periodic boundary conditions in the space variables for which
complete initial data are not available.  We suppose that we can measure
%2u↑I%1 , the first %2j%1 components of a solution %2u%1 of the system, perhaps
with its time derivatives, but cannot measure %2u∩[II]%1 , the rest of the
components of %2u%1 , completely and accurately at any time level.  Such problems
arise in geophysical application where satelites are used to collect data.
We consider two questions.  How much information do we need to determine the
solution %2u%1 uniquely in a way which depends continuously on the data?  How
do we use these data compuationally to obtain complete initial data at some
time level?

@We investigate several approaches to answering these questions.  We show that
under certain hypotheses %2u∩[II]%1 at the initial time is determined uniquely
by and depends continuously on the data obtained by measuring either %2u↑I%1 over
a whole time interval or %2u↑I%1 and its first time derivative at the initial
time, together ith either %2u∩[II]%1 on a hyperplane in space of one lower
dimension or a finite number of Fourier coefficients of %2u∩[II]%1 at
the initial time.  Our results demonstrate that it is possible to reduce the
data requirements of %2u∩[II]%1 if sufficient information about %2u↑I%1 is
available.

@One appliction we examine is the effect of the Coriolis term in the linearized
shallow water equations on the possibility of recovering the wind fields from
the geopotential height.

@We present algorithms and computational results for these approaches for a model
two-by-two system, and examine the method for intermittent updating currently
being used in numerical weather prediction as a method for the assimilation of
data.  Our results suggest that the use of different frequencies of updating
is important to avoid slow convergence.
.begin nofill
No. of pages:  119
Available in microfiche only.
%4----------------------------------------------------------------------%1
%3HPP-77-39  MODELS OF LEARNING SYSTEMS
%3Authors:%1  Bruce G. Buchanan, Tom M. Mitchell, Reid G. Smith & C. Richard Johnson Jr.
.end
@%3Abstract:%1@The terms adaptation, learning, concept-formation, induction,
self-organization, and self-repair have all been used in the context of learning
system (LS) research.  In this article, three distinct approaches to machine
learning and adaptation are considered:  (%2i%1) the adaptive control approach,
(%2ii%1) the pattern recognition approach, and (%2iii%1) the artificial intelligence
approach.

@Progress in each of these areas is summarized in the first part of the article.
In the next part a general model for learning systems is presented that allows
characterization and comparison of individual algorithms and programs in all these
areas.  The model details the functional components felt to be essential for any
learning system, independent of the techniques used for its construction, and the
specific environment in which it operates.  Specific examples of learning systems
are described in terms of the model.
.begin nofill
No. of pages:  38
Cost:  $ 2.80
%4----------------------------------------------------------------------%1
%3STAN-CS-79-693  A CLASS OF SOLUTIONS TO THE GOSSIP PROBLEM
%3Author:%1  Douglas B. West
.end
@%3Abstract:%1@We characterize and count optimal solutions to the gossip
problem in which no one hears his own information.  That is, we consider graphs
with %2n%1 vertices where the edges have a linear ordering such that an
increasing path exists from each vertex to every other, but there is no
increasing path from any vertex to itself.  Such graphs exist only when %2n%1
is even, in which case the fewest number of edges is %22n-4%1, as in the original
gossip problem.  We characterize optimal solutions of this sort (NOHO-graphs)
using a correspondence with a set of permutations and binary sequences.  This
correspondence enables us to count these solutions and several subclasses of
solutions.  The numbers of solutions in each class are simple powers of %22%1
and %23%1, with exponents determined by %2n%1.  We also show constructively that
NOHO-graphs are planar and Hamiltonian, and we mention applications to related
problems.
.begin nofill
No. of pages:  61
Cost:  $ 3.45
%4----------------------------------------------------------------------%1
%3STAN-CS-79-694  COMPUTER SCIENCE AT STANFORD, 1977-1978
%3Editor:%1  Jonathan King
.end
@%3Abstract:%1@This report is a review of Stanford Computer Science Department activities
for the academic year 1977-78.  The report includes:  (1) The Chairman's 
overview of the important events of the year, (2) Recent activities and 
interests of faculty members and research associates arranged by research 
specialty, (3) Advanced doctoral students in each specialty and their research 
topics, (4) Recipients of M.S. and Ph.D. degrees, (5) Speakers at seminar and 
colloquia series, and (6) Bibliography of Computer Science Department technical 
reports.
.begin nofill
No. of pages:  27
Cost:  $ 2.50
%4----------------------------------------------------------------------%1
%3AIM-321  RECENT RESEARCH IN ARTIFICIAL INTELLIGENCE AND FOUNDATIONS 
			OF PROGRAMMING
%3Authors:%1  Lester Earnest et al.
.end
@%3Abstract:%1@Summarizes recent research in the following areas:  artificial
intelligence and formal reasoning, mathematical theory of computation and
program synthesis, program verification, image understanding, and knowledge
based programming.
.begin nofill
No. of pages:  94
Cost:  $ 4.35
%4----------------------------------------------------------------------%1
%3HPP-78-22  CONSIDERATIONS FOR MICROPROCESSOR-BASED TERMINAL DESIGN
%3Authors:%1  Reid G. Smith & Tom M. Mitchell
.end
@%3Abstract:%1@We discuss the design of hardware and software for inexpensive
microprocessorr-based terminal/microcomputers.  Such devices are fundamentally
microcomputers that have been adapted, with specialized software, to operate as
remote terminals for a host compuer.

@The discussion centers on a specific video terminal designed and constructed
by the authors.  This terminal is based on the INTEL 8080 microprocessor and is
equipped with software sufficient to emulate the characteristics of standard
video terminals required by several available %2screen-oriented%1 text editors
in common use at sites throughout the ARPAnet.

@We have found that the microprocessor adequately serves as the controller for
such terminals, and that a %2software-based%1 approach to the design of such
terminals offers substantial advantages in capabilities, flexibility, and cost
over the %2hardware-based%1 approach.  We suggest guidelines for future designs
of microprocessor-based terminals on the basis of our experience designing and
using the terminal described here.

In order to take full advantage of the flexibility afforded by microprocessor-based
designs, we have implemented the capability to %2download%1 and execute 8080
programs written and assembled on a host computer.  This allows the user to
customize and extend the microcomputer with the software development tools and
mass storage provided by the host computer.  The terminal is thus a complete,
stan-alone microcomputer system specially configured for its role as a terminal.
.begin nofill
No. of pages:  14
Cost:  $ 2.10
%4----------------------------------------------------------------------%1
%3
%3Author:%1  Dick Sweet↔(thesis)
.end
@%3Abstract:%1@
.begin nofill
No. of pages:  
Cost:  $ 
%4----------------------------------------------------------------------%1
%4----------------------------------------------------------------------%1
.END
.begin center
Available from
Stanford Systems Optimization Laboratory
.end
.begin nofill

%3SOL 78-19  A BIDIAGONALIZATION ALGORITHM FOR SPARSE LINEAR EQUATIONS AND
			LEAST-SQUARES PROBLEMS
%3Authors:%1  Christopher C. Paige & Michael A. Saunders
.end
@%3Abstract:%1@A method is given for solving A%2x = b%1 and 
%2min||%1A%2x-b||%1↓2 where the matrix A is large and sparse.  The method
is based on the bidiagonalization procedure of Golub and Kahan.  It is
analytically equivalent to the method of conjugate gradients (CG) but possesses
more favorable numerical properties.  The Fortran implementation of the method
(subroutine LSQR) incorporates reliable stopping criteria and provides
estimates of various quantities including standard errors for %2x%1 and the
condition number of A.  Numerical tests are described comparing LSQR with
several other CG algorithms.  Further results for a large practical problem
illustrate the effect of pre-conditioning least-squares problems using a sparse
LU factorization of A.

Available by writing directly to:  Gail L. Stein, Systems Optimization Laboratory,
Department of Operations Research, Stanford University, Stanford, California
94305 U.S.A.
.BEGIN NOFILL
%4----------------------------------------------------------------------%1

%4********************************************************************%1
%4********************************************************************%1
.end
.once center
%3ATTENTION!

%3@Due to the rising costs of printing, postage and handling, we find it is now
necessary to prune our mailing list.  We will NOT drop anyone who specifically
asks not to be, but your cooperation is essential.  
Here is how to determine
if you are in danger of being dropped from our mailing list:  

@@(1) you have not ordered any reports from us in the last year, and 

@@(2) you do not check the appropriate box
on the attached order form and return it to us by the deadline.%1

@The management wishes to extend its thanks for your cooperation in helping 
us to curb our expenses.
.break
.begin nofill
%4********************************************************************%1
%4********************************************************************%1
.end