I don't believe the minimum exposure algorithm works in all cases:
Specific example:
FanDuel NBA (has two roster slots for PG)
Setup three PGs with minimum exposures
Point Guard A - 0.8
Point Guard B - 0.6
Point Guard C - 0.2
Attempt to build 10 lineups. The optimizer will build 8 lineups but will find 'problem is infeasible' on lineup number 9.
The minimum exposure constraints for these three players together should be feasible. I think the algorithm needs to somehow identify that these players need to be forced into lineups sooner than it is currently doing. I noticed that after lineup number 8, all three PGs still needed to be in two more lineups to meet their minimum exposure requirement--but at that point there's not enough remaining slots in the last two lineups for all of them.
I don't believe the minimum exposure algorithm works in all cases:
Specific example: FanDuel NBA (has two roster slots for PG) Setup three PGs with minimum exposures Point Guard A - 0.8 Point Guard B - 0.6 Point Guard C - 0.2
Attempt to build 10 lineups. The optimizer will build 8 lineups but will find 'problem is infeasible' on lineup number 9.
The minimum exposure constraints for these three players together should be feasible. I think the algorithm needs to somehow identify that these players need to be forced into lineups sooner than it is currently doing. I noticed that after lineup number 8, all three PGs still needed to be in two more lineups to meet their minimum exposure requirement--but at that point there's not enough remaining slots in the last two lineups for all of them.