International Encyclopedia of Systems and Cybernetics

Any device that introduces constraints in the workings of a system.

The first step towards a control mechanism is given when two or more elements are becoming interactive in some definite way.

K. BOULDING gives the following detailed description of the control mechanism:

"A control mechanism always involves a number of parts. There must be: first a receptor of information (1) which picks up information regarding the variable to be controlled and, in particular, picks up information regarding divergence between the actual value of the variable and the limits of tolerance. In the simplest mechanism, all that is necessary is qualitative information about the sign of this divergence: the minimum information is that the variable is above the upper limit or below the lower limit of tolerance.

"This information must be transmitted by a communication system (2)… to an executive or decision maker (3). The function of the executive is to transform the information received into instructions; it (or he) can be thought of as a kind of production function, receiving information by one communications system and sending out instructions over another (4).

"The instructions go to effectors (5) which carry out the instructions, and can produce an effect… "which is carried by another communication system (6) to the controlled variable" (1952, p.37).

At least two elements should be added to this description:

1 – In order to meet, or at least to register any divergence the receptor must be at the same time a comparator, or be connected with one. The comparator must compare the registered input with a previously incorporated standard and, possibly, with a register of former states. Only thus is any information created.

2 – Any effector needs the disposal of some spare resources to be used in order to be able to produce any effect. It should be for example a store of available reserves or a source of energy.

K. BOULDING gives the following example of control mechanism:

"Thus in the thermostatically controlled furnace, we have the thermostat (1) which registers divergence of the temperature from the tolerated limits (the upper and lower limits may in this case be contiguous). This information is transmitted by a wire (2) to the furnace control (3), which transforms the information into instructions sent by another wire(4) to the furnace (5), which is capable of transmitting the effect (heat) by system of pipes and radiators (6) to the environment of the thermostat".

Thereafter BOULDING comments:

"The theory of the control mechanism has been developed in some detail by WIENER, and we need only mention some conclusions here. One is that a control mechanism always involves a cyclical movement of the controlled variable, the amplitude of the cycle being a function of the sensitivity of the mechanism. If the mechanism has pronounced lags, the corrective effect may only be brought into play after an extensive divergence of the controlled variable from the norm (or even after a spontaneous reversal of the divergence), and the correction may go too far before the counter correction comes into play" (Ibid).