We consider a general model with structural changes including both non-aggregate and
aggregate models. We change parameters only while keeping a transition matrix.
We do the same thing to make indM1, indM2,. . ., and indMj from j=1 to ns in the same
way as the day before yesterday’s. We set transition matrix tm without any additional
index number.
Here, paras is (ns x dim) x (# of parameters changing in the model).
On the other hand, transition matrix P(or tm in the program) is dim x dim alone.
As mentioned long before, we have two versions with or without offsetting the series
of states between the adjacent two groups of parameters.