ABSTRACT
An archer can learn to hit a target by following a very simple rule. If he misses to the left, he can aim a little more to the right. If he misses to the right, he can aim a little more to the left.
This sort of directional learning applies to a wide range of situations. Only three requirements must be met:
1 As in any learning situation, there must be time to learn, and here we assume discrete time periods t=1,…, T.