Value function approximation and collocation method

Today we implement collocation method for the model

by switching model with infinite many states in order:

.

.

.

as in 1104 and 1106 for non-aggregate and aggregate models. As in there, we assume

that each policy function is realized for each parameter set instantaneously.

We now value function approximation for each policy function in order to speed up

the convergence.

Of course, it may take longer time to converge because we use the same model and

its parameters for all the approximations.