Speedup algorithm: error bounds

Recall the following m states model:

.

.

.

where

and

.

.

.

and

.

.

.

as model 3 in 0816.

We firstly compute m constant labors.

We then compute m policy functions corresponding to each theta and then pick a

policy function according to the probability matrix.

We implement the same algorithm of error bounds as yesterday’s when we compute

m policy functions respectively.

Model 1 and 2 can be computed if we keep one of the parameters unchanged.