|
Not long, in a few seconds I reached the optimal solution, but OptaPlanner can't see that it's optimal (unless you use Brute Force or Branch and Bound which takes too long). William Cook (the TSP expert) confirmed that that tour is optimal. I did enable incrementalJavaCalculation instead of drools, so with drools it might take a minute. In any case, a 50 location TSP is not a challenge. It's when we're dealing with a few thousand or more locations and additional VRP constraints, that it becomes challenging. Here's one with 2750 locations: http://www.optaplanner.org/blog/2015/01/27/ScalingVehicleRoutingAndTSPWithNearbySelection.html On 26-03-15 14:27, Peter wrote:
Hey Geoffrey,I generated the data for this post: http://www.optaplanner.org/blog/2015/03/20/3BugsInTheUltimateAmericanRoadTrip.htmlBTW: How long did OptaPlanner calculate for this?Regards, Peter _______________________________________________ GraphHopper mailing list [email protected] https://lists.openstreetmap.org/listinfo/graphhopper |
_______________________________________________ GraphHopper mailing list [email protected] https://lists.openstreetmap.org/listinfo/graphhopper
