Intelligent Manufacturing € January € 1995 € Vol. 1 € No. 1
Intelligent Manufacturing € January € 1995 € Vol. 1
€ No. 1
The concept of intelligent manufacturing is not a new concept. It has
been discussed for more than 10 years. However, what we mean today by
intelligent manufacturing is very different from what was being
discussed a decade ago. In the mid-1980s, the concept of intelligent
manufacturing conjured up images of plants populated by robotics,
automated guided vehicles, and run 24 hours a day - the so-called
"lights out" factories.
In these settings, software played an important role. Software was
designed to control and manage activities. It was designed to replace
the worker. We did not need workers to do detailed planning and
scheduling - that would be done by sophisticated MRP II and
dispatching packages. We would capture the intelligence of key
employees involved in critical activities and make that knowledge
available to anyone through artificial intelligence and expert
systems.
However, today's version of intelligent manufacturing demands
different software capabilities. These demands reflect the new
challenges now facing manufacturing. To understand the new face of
intelligent manufacturing software, we must understand the
challenges.
The first is the challenge of the knowledge factory. In the past, we
saw factories as places where only goods and services were produced.
That view is changing. We now see factories as places where we
produce goods, services and, most importantly, knowledge. But what is
knowledge? It is the ability of the employee to apply information and
data to various situations. It involves the ability of the employee
not only to execute plans but to identify problems, analyze them,
formulate solutions and implement them.
The emergence of the knowledge factory, a place where we purposely
try to create factors encouraging the creation and growth of
knowledge, is a result of factors such as downsizing, Total Quality
Management and Just-in-Time manufacturing. These developments provide
a stronger and more active role for the employee. The knowledge
factory demands that we, as managers, improve the ability of our
employees to gain knowledge and act as decision-makers.
The second challenge is the reduction of overhead and the breakup of
central staff. Traditionally, many firms have had a large
headquarters staff populated by experts or in-house consultants
occupying staff positions. Everyone in the firm contributed to the
maintenance of these experts through overhead contributions. In
return, they were provided with access to a source of knowledge which
could be used for addressing difficult problems. However, these
headquarters are now being broken up with members of the central
staff being moved into line positions. Why? Because of the need to
reduce overhead and increase overall corporate flexibility. However,
with this breakup has come a new problem - that of making the
knowledge of these experts quickly available to anyone within the
firm who needs it. Often the source of the demand for the expertise
and the source of the expertise are geographically separated with the
urgency of the need making impractical the old strategy of putting
the expert on the latest airplane.
The third challenge is that of complexity, a development
characterized by short product life cycles, increasing variety and
greater uncertainty. It describes an environment where learning takes
place. People look at what they have done in the past and use this
information to shape the choices that they make in the future. There
is an increasing amount of information being generated. Every action
creates more information than can be easily and quickly analyzed.
Furthermore, every action that we take eliminates from future
consideration certain options currently available. These actions also
create new opportunities not previously available. Finally, the lead
time available for analysis and reaction is constantly shrinking. In
such an environment, we emphasize strategies such as simplifying the
basic processes, standardizing processes and components whenever
possible, encouraging activities in parallel, and integrating
activities across functional and corporate boundaries.
The final challenge is that of maintaining focus. Increasingly, firms
see themselves as embodying certain critical core competencies. Core
competencies are those unique skills and knowledge sets which enable
the firm to create value for its customers.
For example, Honda produces cars, generators, motorcycles and lawn
mowers because of its core competencies in engine design. Chrysler is
able to bombard the automotive marketplace with a large variety of
new product offerings because of its core competencies in design and
assembly. When we recognize core competencies, we are forced to
rethink how we organize manufacturing. The all-inclusive factory
which can make everything we need for our products becomes obsolete.
Rather, we focus only on those areas linked to our core competencies
- those we build internally. Everything else not related to our core
competencies, we subcontract out.
If we are a Chrysler, we now subcontract out the construction of
radios and car seats. These are not our core competencies. There are
suppliers who can design and build these items faster, better and
cheaper than we can. By subcontracting, we can eliminate the overhead
associated with these activities. However, once we start down this
path, we become faced with a manufacturing process which is a mixture
of our own manufacturing elements integrated with those of our
suppliers. The challenge is to manage this mixture so that it behaves
as one seamless integrated system - a virtual factory.
These are the challenges. They demand an environment that is more
organic in nature and in which people work together in teams. They
also demand streamlined and simplified processes. They demand systems
in which the voice of the customer (whoever that is) can be clearly
heard and listened to at every stage of the design and execution
process.
Several new categories of software are now emerging in response to
these challenges. The first is that of groupware. A groupware
product, such as Lotus Notes, allows different people located in
geographically separated areas to work together. Groupware allows
different people to add their comments and observations and make
those insights available to anyone else working on the problem.
Groupware is consistent with the teamwork environments now emerging
in response to these challenges. Groupware allows for learning
because the users can benefit from the insights and experiences of
others.
The second category is visually oriented simulation systems (e.g.,
Arena, Extend and Witness). In the past, using a simulation language
required that a user be extensively trained in a language such as
SLAM or in a general purpose computer language (e.g., FORTRAN or
Pascal). Model construction and debugging formed significant and
time-consuming obstacles to the use of these packages by a
manufacturing manager. Output took the form of long lists of
numbers.
However, this is now changing as a result of the emergence of
visually oriented simulation packages. With these packages, the user
simply builds a visual representation of the factory (either real or
proposed) on the computer screen. Icons are chosen from a pallet and
brought down and positioned at the appropriate places on the screen.
Options are now filled in using a fill-in-the-blank approach. There
is now a real correspondence between what is on the screen and the
system to be studied. These computer models can also be quickly
built.
Once built, the operation of these systems can be studied using
animation. We can see how the systems would run by observing their
operation on the screen. We can track the movement of jobs from one
operation to another. Potential bottlenecks and problems can be
quickly identified. Alternatives ("what ifs") can be tested quickly
before actually being implemented. As a result, overall response time
drops.
The third category is that of business process reengineering
packages. With these packages, we can now document and analyze
various business processes (be they production or information
related). We can identify the activities involved and pinpoint the
location of potential bottlenecks or inefficiencies. We can identify
redundancies. The result is that we now have the tools for
simplifying processes and for ensuring that they are always
consistent with the needs of our customers.
There are still other categories of software tools. Some of these
tools involve rapid prototyping whereby we can develop and evaluate
quickly proposed product designs by creating smaller scale, physical
representations of them. Still others involve procedures such as
Quality Function Development (QFD). With QFD, we can ensure that the
products and processes we are designing are consistent with the voice
of the customer (i.e., the requirements and needs of the customer).
Instead of focusing on technical features and feasibility, we first
focus on customer-driven concerns and then evaluate the fit with the
technical features. Still another set of tools involves finite
scheduling systems whereby we can continuously ensure that the plans
we are generating are feasible and being met in the system.
The software tools of intelligent manufacturing in the mid-1990s are
indeed different from those of 10 years ago. They reflect a
perspective in which the intelligence lies in the user and the
software supports, expands and empowers the user. As a result,
intelligent manufacturing today is far more exciting than it was 10
years ago since it represents a merger and expansion of capabilities.
It is truly intelligent manufacturing.
Steven Melnyk (517-432-3506) is a professor in the Department of
Management at Michigan State University, East Lansing, Mich. Mike
Martin is the intellectual properties officer at MSU.