The Expectations and Practicality

wildlifeplaincityΔιαχείριση

6 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

42 εμφανίσεις

The
Expectations
and
Practicality

of Knowledge Management

in Industry


Colin Piddington

colin@tanet.org.uk

TANET Association

Who am I


BAE systems 1962 to 1994


CSC 1994 to 2004


MD for
Cimmedia

2004 to 2010


Contract with Salford University 2006
-

2010


Director of
TANet

2004 to date


Associate of Control 2K manufacturing SME

KM the pathway


Look at the beginning through to today from an
Industrial perspective


Not an OWL computer expert


We will start with expert systems


Look at projects using knowledge data bases


Discuss the experiences of each and the limitations
with aspects of


Data Collection


Human Interaction


Collaborative Working


We will end with a summary of work to be done and
give some insight to the next steps that are being taken
by the industrial working group of the I
-
VLab

Expert Systems


1989 Investigations into expert systems for glue
and
airmotor

selection


These were based on data sheets transferred to
data bases to serve the application


Tests worked successfully with good results


BUT


the range covered was limited


AND the cost of populating the data bases exceeded
the efficiency gains.


Left a legacy of distrust in industry as
expectations could not be met


Fashion business

ebusiness

garment selection


EU research project


This initiative was prompted by the
emergence of the business internet.


Needed to pass design requirements quickly
across the potential supply chains to optimise
the selection


Worked to a degree but need to develop a
common set of words
-

Glossary
-

difficult

Commercial package deployment

of I2 software for component

selection in electronics


A different business model


Software was free


Commonisation of the data from various
suppliers was sold as a service

Enterprise bus deployment

for PDM to ERP


Enterprise bus technology


single end application adaptors


cuts change costs


Uses PDM Schema approach for information translation


Allows for multiple glossaries with a rule based mapping
capability



Still requires a lot of consultation to establish mapping rules


Maintenance in line with business change is still a problem


Solutions still operate in a limited domain

KM and shop floor management


Flexquar

-

Enterprise organisation


Principle


to give the shop floor manager the
ability to impact his knowledge to give tactical
decision making and schedules


Has access to computer data at his workplace
controller


Allows him to input rules to capture his
knowledge


it is his decision


By changing the rule set the organisation can
change easily to represent the enhanced
responsibilities of the individual


Organisational Management Position

demand

report

delegation

Feedback collection

Manager

Window to
Digital world

Automation Layer

Execution of rules

Execution of rules

Input receiving and display to Manager

Distribution
of commands

Interop



NoE

KMAP


The need was to develop a competence map of
the Enterprise Interoperable domain to see who
was doing what, collaborating with whom and
where the gaps were in research across Europe


Multiple views


People


Organisation


Papers


digital library


Needed a common ontology


No existing standards

KMap


Provided a
unique

opportunity as there was
no defined Interoperability domain


3 Phases were determined


To create a list of terms


To create a Glossary


To determine a taxonomical relationship between
Glossary terms


To create a set of analysis tools

KMap

Process


Search the current papers in the depository for
common used terms


Web based Expert voting to reach agreement


Using the list of terms search the internet for
meanings of terms


Web based Expert voting to reach agreement


Expert inputs to form groups



Web based Expert voting to reach agreement


Reclassification of papers in the depository



KMap

competencies


A competence MAP of the
Interop

NoE

partners was created using Protégé as an
initial starting point


this was to be replaced
at a later date to better reflect the Taxonomy


All partners were requested in input both the
personal and corporate involvement on
projects and papers written

KMap


Analysis tools were then used to identify the
gaps and the concentrations of research and
people and institutional relationships.


Pictorial views were used to clearly identify
these relationships

Collaborative working & KM


A multi dimensional problem


Need to understand the relationship between
organisations, processes, skill silos, tacit and
explicit knowledge.


Have to understand both formal and informal
structures


Identification of Collaboration Spaces

Product Life cycle


Birth to death

Concept to recycling

Project management process
-

Workflow

Decisional
gate

Decisional
gate

Decisional
gate

Decisional
gate

Functional organisation

Project organisation

Discipline

Application

Information
store

API

Discipline

Project lead

Discipline

Application

Information
store

API

Discipline

Project lead

Discipline

Application

Information
store

API

Discipline

Project lead

A Collaborative Work Space

Decisional Gates

Design optimisation

Risk analysis


next stage start

Problem resolution

Represents the design,

engineering, procurement,

contracts etc

Agrees,.

approach to problem solving,

Information Sharing

Lead is supported

by Discipline experts

Discipline related

Automated tools

Captive data base

When data is required

to support problem resolution


it is via the vendors API


Information

based on need to know

transferred


mapped across discipline

Same for many disciplines

Each contribute to a different level

depending on maturity

Collaborative Domain


Formal

Experience


Difficult to motivate people to input data


Data inputs had differing granularity


Domain was in constant evolution


Maintaining the taxonomy was hard to justify


People asked why we could not take inputs from
institutional sites for their data (duplication of
effort)


The creation of the
KMap

was a major step
forward from previous experiences and gave
renewed hope for KM futures

Other EU projects


STASIS
-

Completed 2009



ADVENTURE
-

http://www.fp7
-
adventure.eu



Both these projects used the basis of agreed ontology's and
enterprise bus approaches to provide common order
exchanges to support dynamic supply chains(STASIS) and
provide feedback information (ADVENTURE)


These are federated approaches where the users ‘buy into’
a common ontology


Problems as always is the maintenance of information in
line with the changing global business and IT evolution

Overall Observations


Knowledge is a moving feast


Some knowledge will be inaccurate


Maintenance of data is a major problem


50% of effort to create the projects is data
collection


Justification is still difficult to encourage
implementations in Industry


Specific implementations are possible e.g.
Security, personnel with specific objectives


Next Steps
-
FLEXINET


Intelligent Systems Configuration Services for
FLEXI
ble

Dynamic Global Production
NET
works



Basic premise
-

reference ontologies


Any multi
-
system design or configuration method
needs a common base of concepts from which to
build and to share knowledge


Text based concepts are not good enough


see
next slide


Formal languages like OWL are NOT sufficiently
expressive to represent the complexities of
manufacturing


Common Logic based formalisms provide both a
semantic capability and also an inference
capability for compliance checking

Conclusions


The problem domain is increasing faster than
the science capability


Knowledge is not static


Data can be proved wrong in the light of
experience


The costs of data collection must be an
integral part of the user processes to avoid
unjustifiable costs

Thank You for listening

Can we Meet the Challenge?


With multi
-
skills in collaborative working the
human application of experience is critical
assisted by technology


Reconciliation of multiple skill taxonomies is a
steep hill to climb


The world is increasingly GLOBAL


language
and its use and customs play a part