 |
 |
In assembly lines and mass manufacturing plants,
base sequences or
schedules are established for horizons like a week or a month. In
practice, very few plants are able to operate with 100% FIFO
conformance - i.e. where the actual final off sequence or schedule
matches the base sequence or schedule.
The deviations from base sequence or schedule can
occur due to several reasons:
- Quality
problems in a particular process require pulling units out of sequence;
for example, in the paint shop.
- Some of the more
detailed constraints were not considered when generating the base
sequence or schedule; for example, no back to back cars with side air
bags.
- Parts supply shortages or quality problems.
- Re-sequencing
or in process buffers may exist.
- For
example: ASRS or mixing banks in auto assembly plants.
The objective of re-sequencing or
rescheduling in such situations is:
- To
maintain the integrity of the base solution so as to minimize changes
elsewhere on the line, while adjusting for any problems or required
changes.
- To generate a new optimal solution so that
costs and inventories are controlled.
In many environments, business
practices or manufacturing technology require frequent real-time
re-sequencing and rescheduling.
Example 1: Mass manufacturing
plant operating in a vendor managed
inventory (VMI) environment or as a "shopping center"
A
reschedule is
run upon every new customer order or change to any customer order.
Example
2: Semiconductor cluster tools / equipment
A
semiconductor cluster tool is a self-contained unit with typical
configurations including Loadlocks, Reactors and Robots. A cluster or
system Controller controls each unit in the cluster. During operation,
the cluster tool or equipment has to be re-scheduled at frequent
intervals based on signals received from the Controller. An optimal
schedule, rapidly generated within hard time limits is required.
A
real time sequencing or scheduling optimizer generates solutions upon
receiving a message from a controlling system. The schedule is based on
a "state" represented by data. This state also includes a copy of the
most recent solution or the base solution so as to enable a bias to
stability in the solutions.
A signal is generated by
the system controlling the process, as in the
following examples:
- ASRS or mixing bank in
an auto assembly plant: ALCS or line control
system.
- Semiconductor cluster tool: system
Controller.
- VMI: stock signals.
- "Shopping
center": order entry system.
The figure
below is a conceptual image of an event driven real time
sequencing / scheduling setup.
|
|
|
|
|
