I am against the idea of picking locations based on the locations of the player's mechs. While it could prevent the situation described it also allows the player to control the initial location, which is exactly what I don't want.
For essentially all escort missions, I now play in such a way that the mechs that spawn when the convoy reaches the destination always appear with their backs to my short ranged mechs (as an example I offer up the last video in Edmond's "There are four lights" campaign). Forcing them to be further from my mechs would let me do that and place them in a location of my preference, which would warp play even more. After all, I can place one mech (with LRMs so it's not out of the fight) in the spot I don't want them to spawn, then place the rest slightly farther away from the other location.
Making the enemy mechs appear farther away also doesn't prevent you from having the rear of your units facing them when they appear.
Also, you'll need a very good definition of what is farther away. I can think of three offhand and none of them can completely prevent what you want.
1) Find the closest mech to each location and pick the location which has the greatest distance. What happens if I split my forces to attack from both sides and end up with a mech in both reinforcement locations?
2) Use the greatest distance from the shape made by the mechs (the centroid). Now one mech very far away from both locations can distort the shape so they still appear close to the majority of your units.
3) Maximize the sum of the squares of the distances to all the mechs. Now the one mech separated from the others is the one that gets dropped on, at which point it doesn't even matter if I go second because it's now 4v1 against me.
In other words, choosing the "farther" location can still be gamed, has some difficult questions about implementation, and cannot offer a complete solution without having a lot more than two reinforcement locations (I think 5 would be the minimum in the unlikely event that the player occupied the other 4).