Now, I think I'm understanding your problem better. The actual problem isn't so much the assignment of sales calls as the scheduling of sales calls. Let me rephrase it, if you don't mind:

• You have a set of resources, each with a starting location.
• You have a set of locations those resources have to cover, with both an expected arrival time and a duration at that location.
• You have a set of estimated travel costs between locations. (This is the hardest part.)

The question is how to optimize the assignment of the arrival times at each location. For example, let's say you have two sales calls in the same building, each expected to take an hour. You can assign them both to the same salesrep and schedule one for 9am and the other for 10.30 am.

Of course, I haven't seen a good definition for the cost of travel. For example, it takes longer to go from Albany to New York vs. Philadelphia to New York, even though Albany is closer. The reason is that there is excellent train service to/from Philly, but only highways to/from Albany. And, if you want to be useful, you need to take into account plane schedules (though you could assume that everyone takes the red-eye, but you won't be popular after a few months).

You may also want to estimate the expected ROI for a sales call. For example, if you expect $1000 of business from a call, but the salesrep's time is worth $320 ($40/hr) and it costs $600 to get there, is that salescall worth it? But, I think we're getting too complicated.

Essentially, this is really starting to approach the travelling salesman problem in complexity.