Hi,
I have a potential client who has a handful of trucks that will each handle about 40 stops per day (peaking at around 50 or so). The client wants to optimize the route, however, the API limits the number of addresses to 25 (and indeed, the traveling salesman problem is a computationally difficult problem). Nonetheless, my client would still like an optimal route for the day. Does anyone have any suggestions for how to best make this work? For example, is there a good strategy for pre-processing the addresses so I can do a reasonable job optimizing the route? (Or MapQuest admins, is there a way to up the limit? This particular client would be optimizing about 6 or 7 routes per day - this is not a public-facing application, so it wouldn't be a huge burden.)
Any thoughts would be appreciated. Thanks!
-Michael
Apart from the clustering approach, you can obtain the full distance matrix by successive API calls using different subsets of stops. For instance, a 40x40 distance matrix can be computed using four calls.Then you can apply any heuristic procedure.