To (hopefully) answer:

I have more (or the same number of) locations than people. We (hopefully) wouldn't book more sales calls than we have sales people for a given day. Therefore, I guess this might equate to your flight crew situation, where M>=X.

I think I want to minimise the mean distance. (I guess the sales team won't be happy with one poor guy having to travel the length of the country, even if the overall distance travelled for the whole group is less.)

Presumably this latter item will be defined in my scoring/fitness routine - whether the score is per individual or for the whole group. Incidentally, it did occur to me that my "score" might not simply be the distance travelled (lower being obviously better). Perhaps I could consider returning the distance travelled squared, so that greater distances get penalised even more? (e.g. two reps travelling 50 miles each is better than one travelling 25 miles and one travelling 75.) I get a nagging feeling that this "off-the-top-of-my-head" solution may be troublesome, however. Perhaps instead I might introduce a large penalty if the distance travelled for one individual exceeds a certain threshold? Any thoughts?

To be honest, I was envisaging trying to get the core routine working, and then "playing" with the scoring routine later. (After, for example, I run it, and find one rep travelling the length of the country. :-D)

I hope this explains things a little better.

SmugX