[Hans Blom, 930620]
(Rick Marken (930619.1400))
could you just tell us -- does the sensory input to a control
system contain information regarding how "plant" outputs should
vary in order to control the sensed variable?
Sorry, but I decline to be arbiter. As a teacher (my other role) I
have to say that having this discussion is by far more fruitful than
knowing who is right. If the two of you stop fighting about who is
right and who is wrong and start to appreciate that the other may have
something useful to say as well, the two of you COMBINED might someday
invent optimal control theory 
It's difficult, sometimes, to tell whether a variable is controlled
or not, especially when 1) you have not built the control system
yourself
YES, YES, YES. This is exactly what I have tried to tell you all
along. Savor this thought and let it reverberate for a while!
I see you and raise you.
I thought, when joining this list, that a scientific discussion was
going on, a common search for deeper understanding. That interests me.
I am not interested in establishing winners and loosers. That is not
the game that I play.
(Bill Powers (930619.1800 MDT))
In order for "robust control" as you define it to be achieved by
an autonomous system (i.e., without an engineer standing by with
test instruments and a screwdriver), something in the system
itself must be able to perceive the quality of control.
YES, YES, YES. That was a discussion I let loose on you some time ago.
Rick Marken shrugged it off as 'the control of Q', but that is what it
really is: the control of the *Quality* of the response.
Now this has wide repercussions when you combine it with evolution:
just a slightly better quality in how an individual controls its basic
evolutionary goal (survival of its descendants) means that slowly by
slowly he will start to dominate the population until no others are
left. Thus maybe the entire evolutionary process can be seen as emer-
gent control.
It also has wide repercussions for the individual. The slogan "it's
all perception" is much too static. It has a connotation of having to
act in an a priori circumprescribed way given a set of perceptions and
a set of top-level (very slowly changing; innate?) reference levels.
Blaming what you see for what you do is usually considered a defense
mechanism or, if more forceful, for criminality. Optimal control
theory applied to humans says that you can fine-tune your actions as
well, that you have control over how you control, i.e. that you have
SELF-control.
The organization of the stabilizing system is thus just like the
basic diagram of a control system in PCT, except for one thing:
it does not act by adjusting the reference signals for lower
systems, but by adjusting the parameters of their components. It
does not perceive a variable that is a report on a higher-level
aspect of the environment, but a variable that represents the way
the control system is controlling.
Yes, see how far this reaches when you consider humans. We do not only
operate on the outside world but on the inside world as well. We can
tune our responses finer and finer, and reach ever higher qualities of
response and perception. Control over control is self-control, per-
ceiving your own perceptions is self-perception, consciousness. See
how fascinating this approach is?
I recognized many years ago that this sort of hierarchical
control relationship had to exist in the human system. However, I
also recognized that modeling it analytically was a very much
more complex undertaking than I would be able to carry out.
Chaos theory might show it impossible. A great many processes have
been discovered where no prediction (and hence no control) seems
possible. In those processes, the only way to know the outcome is to
actually run the process and observe how it develops. Yet, even chaos
theory develops is (you might say 'emergent') laws, that may provide a
little understanding of what goes, and maybe only over short periods
of time. That is why ever better simulations are important: even if it
is impossible to come to analytical grips with the matter, you might
find general tendencies. Maybe only as vague as "flockness", which
doesn't really exist (because it isn't controlled for, you might say),
but still this kind of vagueness may be preferable to not knowing or
misconceptions.
In other words, I could ask the question, "How is competent adult
human behavior organized when its organization is not being
changed?" This is what HPCT is about.
Whereas my focus is more on how human behavior can become even more
competent, i.e. more on learning (and evolution as a kind of learn-
ing).
When I speak of the excellence and tightness of human control
systems, I'm not trying to account for how they got that way.
Neither am I trying to explain how they change to maintain that
excellence when long-term changes in the environment (or the
organism itself) obsolete a formerly competent design.
But those are exactly the things that fascinate me!
Now to your supposed counterexample concerning power gain
Let me summarize our discussion thus far and maybe make it a little
bit more precise. One can, at our level of discourse, see a feedback
amplifier or some such device in two very different ways. The first is
as a device that transforms an input (voltage, current or power) into
an output (voltage, current or power). The feedback is not really
relevant here. The (voltage, current or power) gain may be any value;
both gains and losses can be similarly realized. The second way is to
see the device as a power modulator, where the input modulates the
transfer of power from its power supply to its output. The feedback is
not relevant here either. Here we will always see a power loss, since
the device's efficiency cannot be more than 100%. I often find that
people confuse these two quite independent perspectives, especially
mechanical engineers :-).
Then, internal in the device, we see the LOOP GAIN, which has nothing
to do with power at all, but is the gain in the loop that the SIGNAL
travels. A decent control system has a loop gain much greater than
one, although a loop gain of less than one is not unthinkable. In the
latter case we might not want to consider the system a CONTROL system.
Loop gain is terminology from the analogue past. In a hifi amp's
innards you can still trace the loop with your finger. In a control
system that is described in difference rather than differential
equations, the 'loop' is functional only at a sequence of infinitely
short time intervals. That is why they are also called 'sampled data
systems'. Now, to complicate matters even further, in adaptive control
systems there are (at least) two loops: one to provide the immediate
actions, and another to do the parameter adjustments (learning). These
loops interact and hence become much more difficult to trace with your
finger. All of this causes the term "loop gain" to slowly fade away.
All that really counts
for control is LOOP power gain, not gains or losses in any one
part of the loop. As a control engineer, you surely understand
this.
I understand loop gain, not loop POWER gain.
The basic model for behavior has to be, I think, the
servomechanism, control-of-input, model. The reason is simple.
Organisms are not like commercial control systems; they are not
organized to be used by someone else. The only evolutionary
reason for behavior to occur at all is to produce some effect on
the organism itself, or to prevent something from happening to
it.
No. Quality is the thing that counts in evolution, period. Only those
with the highest quality leave descendants behind, in the long run,
whatever the mechanism. I realize that this usage of the word quality
has a circular definition, but that is because this quality emerges,
and is not the result of a control process. I realize that this must
sound obscure to you, but explaining what I mean would take more time
than I can spare at the moment. I have a huge pile of exams to assess.
Maybe later.
Finally, I think we have agreed that flockness is not under
closed-loop control, but is only under control in the emergent
sense. I still have great difficulty in allowing for the latter
sense of "control:" it is hard for me to conceive of a kind of
control in which the controller never knows about the output and
doesn't care what happens to it.
Maybe the non-controller DOES care but also knows that he alone could
never achieve the emergent thing. Maybe the US is the emergent result
of a great number of people who control for living in peace together
at the level where they CAN control. I think that the notion is VERY
important. Please reconsider your reluctance...
Greetings,
Hans