Page 295 - Sensors and Control Systems in Manufacturing

The Role of Sensors and Contr ol Technology in CIM
and labor. To install CIM, these databases must be consolidated. 253
Unfortunately, bringing multiple databases into one CIM DB that
remains available to everyone and consistent in all levels presents a
significant obstacle because of the large investment needed in time
and computer hardware and software.
5.5.2 Structure of Multiobjective Support Decision Systems
The success of CIM also depends largely on the ability to incorporate
sensor technology with a database. The database is utilized in mak-
ing decisions on all levels—decisions that are used to update to the
database. Decision support systems can provide a framework for effi-
cient database utilization by allowing storage and retrieval of infor-
mation and problem solving through easy communications.
Decision-making problems in manufacturing can be grouped into
two general classes:

• Structured decisions are those that are constrained by physical
or practical limitations and can be made almost automatically
with the correct input. An example is generating a group
technology part code given the geometry of the part.
• Unstructured decisions typically are those that contain a large
degree of uncertainty. Decisions considered by strategic
planners are almost always unstructured. Deciding whether
or not to expand a certain product line, for example, may
be based on demand forecasting and on the expected
growth of competitors. Due to the predictive nature of these
decisions, they inherently contain more uncertainty than
structured decisions. Long-range planning consists primarily
of unstructured information.

Decision support systems mainly consist of three separate parts:
• Language systems. The function of a language system (LS) is
to provide a means for the user to communicate with the DSS.
Some considerations for the choice of a language are that the
formulation should be easily understood, implementable,
and modifiable. Moreover, processing the language should
be possible on a separate level or on the problem processing
system (PPS) level. An obvious choice for a language would
be the spoken language of the user. This would require little
or no training for the user to interact with a computer, but the
complexity of sentences and the use of words that have mul-
tiple meanings present difficult problems that, when solved,
would introduce unwanted inefficiency into the language
system. An alternative would be to use a more formal lan-
guage based on logic (e.g., PROLOG). The advantage here is