Good regulator theorem

<p>The good regulator theorem is a <a href="/facts/Theorem/f2Pe1FMD">theorem</a> conceived by Roger C. Conant and <a href="/facts/W._Ross_Ashby/3soK059G">W. Ross Ashby</a> that is central to <a href="/facts/Cybernetics/4qb58zbt">cybernetics</a>. It was originally stated as "every good regulator of a system must be a model of that system". That is, any <a href="/facts/Regulator_(automatic_control)/8hhSALdF">regulator</a> that is maximally simple among optimal regulators must behave as an image of that system under a <a href="/facts/Homomorphism/0gyqdEse">homomorphism</a>.
</p><p>More accurately, every good regulator must contain or have access to a model of the system it regulates. And while the authors sometimes say the regulator and regulated are 'isomorphic', the mapping they construct is only a homomorphism, meaning the model can lose information about the entity that is modeled. So, while the system that is regulated is a pattern of behavior in the world, it is not necessarily the only pattern of behavior observable in a regulated entity.
</p>

Good regulator theorem open-in-new

Good regulator theorem