Habit-based regulation of essential variables

Abstract

A variety of models have been developed to investigate "homeostatic adaptation," a mechanism inspired by Ashby's homeostat, where a plastic control medium is reorganized until one or more essential variables are maintained within predefined limits. In these models, "habits" emerge, defined as behavior-generating mechanisms that rely upon their own influence to maintain the conditions necessary for their own persistence. In this paper, we present a recently developed sensorimotor-habit-based controller that is coupled to a simulated two-wheeled robot with a simulated metabolism. The simulation is used to demonstrate how habits can have the same essential variable(s) as the metabolic or "biological" organism that is performing the behavior, and that in certain conditions when this is the case, the emergent habits will tend to stabilize essential variables within viability limits. The model also demonstrates that an explicit pre-specification of (A) which variables should induce plasticity and (B) which values of those variables should induce plasticity is not always necessary for homeostatic adaptation of behavior.