INTELLIGENT CONTROL OF MATERIAL HANDLING SYSTEMS - Environmentally Conscious Materials Handling - page 98

The curve is dynamical, due to the various and changing load on the sys-
tem, and the maximum cannot be calculated exactly. Thus, the strategy is to
quickly respond to minor observations, which indicate that the maximum has
been reached, and then block new inputs to the area. This approach is called
saturation management , where the toploaders will be blocked if the routes from
the toploader are overloaded.
Queues close to the toploader are most critical, as the toploaders have great
impact on filling up those queues, whereas the parts of the route far from the
toploader could easily have been resolved before the new totes arrive. Instead
of blocking the toploader completely, it can just slow down the release of new
totes using the following fraction of full speed:

i

i w i q i α

d i q i

i d i
where v t is the full speed of the toploader, w i is weight of the queue statuses, q i ,
along the routes. The weight is given by a fitted coefficient, α , and the distance
from the toploader d i . Queue statuses, q i , are always a number between 0 and
1, where 1 indicates no queue.
The effect of the saturation management strategy is clearly documented by the
graph in Figure 3.23. Thus, the decision taken by the toploader agent is highly
dependent on the current configuration of the environment around the toploader.

v t = i w i =

12 CASE STUDY 2: MATERIAL HANDLING IN AN

ANODIZATION SYSTEM
The second case is a material handling system that moves bars of items between
different chemical baths. Each bar has its own recipe to process the system, and
system scheduling is modeled by simple reactive agents that influence each other
through their actions on the environment. Therefore, there is no direct negotiation
between the agents; it is more a matter of coordinating their activities.
Timetabling of classes is a classical constraint satisfaction problem, which is
known to be hard or even NP-complete for large schools or universities. But

300

250

200

150

100

50 no saturation management

saturation management

0

0 100 200 300 400 500 600 700 800

Time

Figure 3.23 Result of a test scenario with and without the saturation management strategy