Fix inSchool alignment boundary solution issues

[LiangShi369]

Previously this branch is called "education alignment", however, it is dedicated to "in school alignment" as defined by
private boolean alignInSchool = true; //Set to true to align share of students among 16-29 age group
in the SimPathsModel class.

I made following fixes:

1. The alignment target should be the share of students among 16-29 age group. Therefore, I modified the Stata do-file line 61-62 to calculate the correct share, and the InSchoolAlignment class line 77-86 to compute the correct simulated share.
2. InSchoolAlignment line 49: In evaluate(double[] args), each person is reset to Les_c4_lag1 before calling person.inSchool(args[0]). This prevents the previous trial state (in root finding) to contaminate the current trial.
3. Person line 1585. I added guard if (demAge > MAX_AGE_TO_ENTER_EDUCATION) return false;, hence person with age >= 29 would not return to education. Perhaps we can revise the relevant flowchart.
4. Warm-start in-school alignment from previous solved year. This strategy reduces the number of iterations in root finding. Specifically, in SimPathsModel line 2113, initial adjustment for inSchool alignment is:
   • previous year’s InSchoolAdjustment when year > startYear
   • otherwise curret-year value.
     I think this warm-start (suggested by Codex) is a good strategy in general, so I added it to the employment alignment pull request (Addressed boundary solution issues in UK employment alignment #333 ).

I also added some diagnotics to show the root finding process during the alignment.

[LiangShi369]

I tried running the simulation without "fix 3", and the alignment looks very good. In fact, the parameter MAX_AGE_TO_ENTER_EDUCATION is no longer used in the simulation.

In our do files (Box\CeMPA shared area_SimPaths_SimPathsUK\input data preparation_Darias backup folder_keep for now\regression_estimates\do), the probability of returning education if condition does not rule out people whose age are greater than MAX_AGE_TO_ENTER_EDUCATION.

[matteorichiardi]

@dariaple and @ANBurdett can confirm, but I recall that the removal of max age from the filtering conditions of returning to education was intentional.
@dav-sonn @ @Mariia-Var can you please double check that the code is consistent with (this and all) filtering conditions?
Also, we need to establish a practice so that when a filtering condition is changed, the related flow diagrams are also changed.
Liang is currently in charge of the flow diagrams, so please let him know.