Cases 36 - 39 are required, since each could have a different number of associated cycles, so they could each have a different "cyclic reduction factor". Therefore you just can't "look at" the temperatures and decide which case is the worst.

Cases 40 and 41 are "summary" cases, basically reporting the max or min (as specified) from the listed cases. This just saves you from having to individually review cases 36 - 39.

Case 42 doesn't make sense. This will compute a stress range based on the max/min displacements of cases 40 and 41, which don't represent a real situation. Get rid of case 42 - everything else makes sense.