Policy function calculation from value function iterations

We approximate the following random model without labor:

for

and

where and are error terms with mean zero.

A path of A

A path of Q

The corresponding behavior of K