Beyond curse of dimensionality with transition matrix

 

We now set the aggregate model with A and Q changing according to transition matrix.

 

We set

 

fc(K,K1,para)=para[.,1]*K^alpha-(K1-(1-delta)*K)/para[.,2]

 

rather than

 

fc(K,K1,A,Q)=A*K^alpha-(K1-(1-delta)*K)/Q

 

or

 

fc(k,k1,para)=para*A*k^alpha+(1-delta)*k-k1

 

rather than

 

fc(k,k1,theta)=theta*A*k^alpha+(1-delta)*k-k1

 

in general.

 

We also change the value of Kstar(or kstar) by xstar.

 

All we have to modify is to change the calling part of fc in terms of para[j,.] rather than

 

A[j] and Q[j] for aggregate model or theta[j] for non-aggregate model.