The problem of defining a switching boundary for a learning control system using the concept of reinforcement of a subgoal is discussed. If the state space is quantized into sample set regions, it is shown that the boundary sets are unevenly reinforced because they do not lie symmetrically on the subgoal contours.
No takes yet. Share an insight, caveat, or question.
Lambert et al. (1968) studied this question.
Synapse has enriched one closely related paper. Consider it for comparative context: