Discuss introducing new block for setting starting point for evaluate_planner_objective

Currently, there are two cases for setting the initial condition for the states at time 0:

1. For perfect foresight, it will be the content of initval unless histval is present.
2. For stochastic simulations, we employ histval. There are two issues:
3. Both cases do not allow setting the value of lagged Lagrange multipliers that are state variables. I am not sure this is much of an issue as there are only two natural values (0 and steady state) that we already compute by default.
4. With histval in the stochastic case, we do not allow for setting the value of the shock at time 0. But these shocks are part of the information set so that one should be able to set them.

For 4.7, the easiest way is to probably stick to histval and to not allow for shock values.