Simulation of clothing manufacture.
There is considerable pressure on the U.K. clothing industry to remain
competitive in the face of foreign competition. Market forces and the
trend of decreasing contract sizes have produced perceived problems
with current methods of production which, coupled with the inertia to
radical change, justify research. Computer simulation is an
established production management tool but its potential in clothing
manufacture could not be inferred. Concentrating on progressive bundle
systems as the dominant method of production, this research considers
the capacity of simulation in this context.
Factory-based studies identified factors affecting system performance
which allowed a conceptual model with high face validity to be defined.
The requirement to handle complex supervisory control strategies led to
the identification of visual interactive simulation as an experimental
route. A computerised model, with an appropriate user interface and
reporting facilities, was developed in the ~Siman si@ulat~£~~Dguage __
This was supported by animated graphics which played a substantial role
in the attainment of face validity.
Replication was considered to be essential for sound estimates of
system performance to be obtained from this stochastic model but, as
interactive control works against replication, steps were taken to
reduce compromise. Software development facilitated an experimental
technique that employed interaction to develop a control strategy,
which then became embedded in the model for replication. By providing
control consistency between replications, a more reliable assessment of
system sensitivity to stochastic variability was possible. Pilot runs
and single factor analysis enabled the effect of controllable factors
on system performance to be quantified.
Supervisory control was found to have a major effect on system
performance so that the need for consistency in interaction was
amplified. Considering alternative experimental methods and the
practical use of the model, application areas for simulation in the
absence of real time data capture were identified and demonstrated.
Each application offered significant advantage over currently
available planning methods and the use of simulation was supported.
Information from the model can be gained about the design and control
of progressive bundle lines at the pre-production phase, and the output
of performance indicators can be useful in assessing real production
lines. The evidence presented by this research illustrates that
animated simulation can provide insight that is otherwise unobtainable.