WHAT ASPECT OF THE PROBLEM WAS MOST DIFFICULT?
to control the mass during the optimization

WHAT INSIGHTS (IF ANY) DID YOU GAIN FROM WORKING ON THE PROBLEM?
It was an interesting problem.

ADDITIONAL COMMENTS:
My solution is not optimal. I believe the calculation
of the relative illumination is incorrect in the program, I used.
It is only the relation of the areas. The factor
     cos(Sigma)*cos(Sigma)*cos(Sigma')*cos(Sigma')
was probably forgotten by the programmer. I did not check the
value of the relative illumination.