Challenge Follies

[From Rick Marken (950918.1300)]

Hans Blom (950918) --

There is a tendency among control engineers -- and not them only -- to reify
control, as in sayings like "control is a fact". No, control isn't a THING.
It is a way of looking at things.

Bill Powers (950918.1130 MDT) --

You're confusing control as an observable process with a model intended
to explain the observations. Control is a phenomenon.

This point is also made in the first paper in "Mind Readings" entitled "The
nature of behavior: Control as fact and theory".

Me:

The feedback function (k) is also limited to values <=1. Any gain
from environmental feedback seems to lead to loss of control.

Hans Blom (950918c) --

I can think of no reason why this could be so except a programming
error. So it cannot be true ;-).

Check out Bill Powers (950918.1130 MDT). It's true, whether you can
think of a reason why it should be true or not. Nature's like that.

Me:

your control system must have a precise measure of its own output

Hans:

Yes. Doesn't yours?

No. Of course not, Hans. You've been making this up. A control model (PCT)
doesn't use u in the computation of the error signal that drives u. If you
don't believe this (which I'm sure you don't) just add a random disturbance
to u in both models. Such a disturbance has little effect on the quality of
controlling done by a control system; it reduces the quality of the
controlling done by your system considerably.

Me:

The difference in the way the PCT and Hans' model dealt with
square wave disturbances suggested a way to test the two models;
just compare the performance of the two models with that of a
human working against a square wave disturbance

Hans Blom (950918d) --

Are you serious?

Yeah. Afraid so.

These models are far too simple to say anything about humans.

How do you know? If the simple PCT model behaves exactly like a human then
how do you know it's too simple? If the simple PCT model behaves exactly like
a human and your somewhat less simple predictive control model doesn't, does
your model win because ours is just too simple?

Best

Rick