VS: Behavior is (part of) Control (was RE: Mad idea?)

[From Rick Marken (2017.06.13.0950)]


[Eetu Pikkarainen 2017-06-13 2]


EP: Thank you Martin for instructive account. But my exclamation was only because Rick seemed to explicitly state that the iron filings have a reference value to be near
the magnet.

RM: I’m afraid you didn’t understand my point at all. What I said was that the position of the iron fillings *appears *to be a variable that is brought to a reference state (near the magnet) because the filings move to the magnet and stay there. But one can easily show that the position of the filings near the magnet is not a reference state by applying a disturbance to the filings, such as a gentle push away from the magnet, and observing that the position of the filings is not protected from this disturbance.Â

RM: If you (or anyone) really want to learn PCT you have to understand the difference between a reference state (which is an objectively observable state of a variable; see my recent reply to Martin Taylor in the “Controlling observables” thread) and a reference signal, which is the theoretical explanation of the existence of reference states.Â




Of course the resistance against the disturbances is crucial. Rick seems to infer so that if there is a reference state AND the resistance for disturbances then there is control, but there can be references without resistance (like in the case
of iron filings and magnet) and then there is no control. I, instead, think that references are always set by some living subject and they are as such unobservable. So if there is observable resistance against disturbances THEN there can be a case of control
AND if it is control THEN there must be a reference (a preselected goal). Conceptually all depends on reference and follows from it.




Lähettäjä: Martin Taylor [mailto:mmt-csg@mmtaylor.net]
Lähetetty: 9. kesäkuutata 2017 19:55
Vastaanottaja: csgnet@lists.illinois.edu
Aihe: Re: VS: Behavior is Control (was RE: Mad idea?)


[Martin Taylor 2017.]

[Eetu Pikkarainen 2017-06-09]

EP: In strict and technical use the definition of control – if I have understood it right  “ is something like: “causing the value (measure) of some variable towards
some reference value and keeping it near it�.


RM: That’s good. Though I think it’s good to add “protected from disturbance” in order to be explicit about why “keeping it there” is an important part of the definition of control. It distinguishes “control”
from “cause”. For example, a magnet causes the position of iron filings to move towards some reference value (the position of the magnet) and keeps it near it, but the magnet doesn’t control this variable, as is easily demonstrated by disturbing the iron filings
by pushing them away from the magnet. If the position of the filings were controlled the pushing would have been resisted and, thus, the position of the filings would be protected from that disturbance.

EP: Are you serious?

I think we have another kind of language issue here, not saying what the writer thinks to be self-evident, but is not necessarily self-evident to the reader. I should never speak for Rick, but on this occasion I will.

What Rick omitted saying is that if there is control, the controlled variable will tend to return toward its “neutral” pre-disturbance value, which we can identify with a reference value, while the disturbance continues to act with a constant force. If there
is no control, nothing much will happen while the disturbance continues unchanged. The variable will remain where it went under the influence of the disturbance. No reference value could be observed, no value toward which the variable tends while the disturbance
continues. The stone sits where it is moved, the cloud goes where the wind takes it, with no countervailing force acting to but them back where they were.

from an energy viewpoint, without control, different forces have been acting on the variable and (in the example case) have brought it to some kind of equilibrium state, like a ball at the bottom of a hemispherical bowl. When the disturbance influences the
value of the variable, it adds potential energy, like pushing the ball up the rim of the bowl. When the disturbance goes away in the absence of control, the ball falls down, dissipating the energy supplied by the disturbance. If there is control and the disturbance
goes away suddenly, the countervailing force will add energy to the ball pushing it up the opposite side of the bowl. That energy does not come from the disturbance, but must be dissipated (or reabsorbed by the control system) before the ball can return to
the bottom of the bowl.

The key difference is whether the variable tends to have changing forces applied that oppose changing disturbance forces, or in other words, whether the energy involved in the changes of the variable are supplied by the disturbance,.


Richard S. MarkenÂ

"Perfection is achieved not when you have nothing more to add, but when you
have nothing left to take away.�
                --Antoine de Saint-Exupery