Controlling the Heating System of an Intelligent Home with an Artificial Immune System - Artificial Immune Systems - page 346

Figure 13 shows the situation after an additional regulation signal. The temperature
in the 2 nd hour is regulated down by 5° C. This influences the antibodies that are
associated with the 2 nd hour in this room and at this day. The regulation in the 7 th hour
is kept unchanged since the antibodies associated with this hour are not influenced.
The ventilation signals shown in figure 12 are removed from the system because they
had only a temporally suppressing effect and the system can adapt to the normal
behavior very quickly.
Fig. 13. A new regulation signal and the disappearance of the ventilation signals
Finally, figure 14 shows the unchanged effect of the previously done regulations
and how the system has optimally adapted to the desires of the users. This can be seen
by comparing figures 13 and 14. The heating temperature is regulated down in the 1 st
hour and up in the 12 th hour because the system has learned that the users expect a
lower temperature at the 2 nd hour and at the 13 th hour and so it starts in time with
regulating up or down.
The purpose of this scenario was to demonstrate how the system can adapt to
regulation and ventilation signals which are treated as co-stimulative signals. As we
have seen in our experiments, it is rather easy for the system to learn the “normal”
behavior of the users, i.e. the usual course without interrupt signals. But the aim that
we had with this system was to make it able to adapt quickly to special demands from
the users (by interrupt signals) without forgetting the normal course of events and
being able to get back to it as soon as possible. Interrupt signals have only local
effects around the hour where they are sent. In particular the ventilation signals
influence the system only temporally so that it quickly returns to the normal course
when the signals are no longer delivered.