Browse Prior Art Database

Loop Deduction and Analysis of Directed Network Problems

IP.com Disclosure Number: IPCOM000111523D
Original Publication Date: 1994-Feb-01
Included in the Prior Art Database: 2005-Mar-26
Document File: 4 page(s) / 210K

Publishing Venue

IBM

Related People

Larsson, C: AUTHOR

Abstract

Disclosed is a network loop deduction and analysis algorithm (DLA) which provides OPC/ESA users with relevant data on loops with good performance resolutions. DLA reports only those nodes on the loop path(s), analyzes the loop for irregularities in the network definition, according to a set of predefined criteria, and resolves the loop condition(s) by removing the most probable arc(s) to be invalidly defined. The report from DLA will consist of the nodes on the loop path(s), but also contain a set of arcs that was depicted as irregular by DLA analysis.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 27% of the total text.

Loop Deduction and Analysis of Directed Network Problems

      Disclosed is a network loop deduction and analysis algorithm
(DLA) which provides OPC/ESA users with relevant data on loops with
good performance resolutions.   DLA reports only those nodes on the
loop path(s), analyzes the loop for irregularities in the network
definition, according to a set of predefined criteria, and resolves
the loop condition(s) by removing the most probable arc(s) to be
invalidly defined.   The report from DLA will consist of the nodes on
the loop path(s), but also contain a set of arcs that was depicted as
irregular by DLA analysis.   Also, the user may in critical
situations resolve the loop condition by modifying the network in the
same manner as the DLA report suggest, reducing time spent by the
OPC/ESA user in resolving the loop, by automatically providing
information about nodes contained in the loop, and irregularities in
the network definitions.

      The DLA may be viewed as a matrix-oriented implementation,
where the network is represented as an arc-matrix or adjacency
matrix.   The DLA reduced the network by removing all arcs and nodes
not part of the loop path during the reduction phase.   Since any
matrix-oriented algorithm for this type of problem would produce its
output in 0(n2), the DLA does not use a matrix as the actual
representation of the network.  As the algorithm is most easily
understood by this view, this description will continue to refer to
the DLA data representation as a matrix.  The analyze phase of DLA
scans a loop network for node-ordering failures, loop path entry node
closeness, and finally minimum through-put node.  DLA removes the arc
matching the highest weighted analyze criterion above an exit from
the analyze phase, and iterate the process of reduction, analyzing
until all loops within the network are removed.  In the Operations
Planning Control (OPC) product series, the Daily Planning programs
are a set of programs that use network algorithms to produce
operating plans.   These are - at program execution - a set of
directed network graphs, that represent dependent work within the
plan.   In any such directed network, a loop will be fatal to the
program, since the directional property is then no longer kept, and a
loop in the network is created by unintentional user input.   When a
loop is detected in the OPC Daily Planning programs, the program will
fail, and no operating plan is produced.   Since installations rely
on OPC managing their daily production load, this situation is
critical to the user.   Previous releases of OPC fail to provide
accurate and sufficient information to aid the user when attempting
to resolve the loop problem.  Particularly for installation with
complex networks, this might result in hours of tedious manual work,
and the possibility of increased costs since the production work
cannot start until the problem is resolved.   In worst cases, a full
network of...