https://www.randalolson.com/2015/03/08/computing-the-optimal-road-trip-across-the-u-s/ Could this have been the Genesis for IBR 2015? Interesting programming solutions utilized to solve the problem.
It's an interesting exercise, but it's not really how LD Riders do it.
That algorithm is another attempt to solve the traveling salesman problem with an eclectic bunch of landmarks thrown in. The problem is that it is, at its heart, a 48/10 type route that is over 13000 miles long. I know that a few extra miles would be needed to accommodate the landmarks, but my 48/10 route was less than 7300 miles.
So I'm thinking, although I haven't tried to route it, the the solution the software came up with might not be all that efficient.
I'd wager that for a route hitting those landmarks, the solution was pretty good. It's not the same problem as a minimum distance 48/10. I suspect that if you picked a set of points along your route and used them as waypoints, the solution would look very similar to the actual route you chose. Or it might even be better
If you have a gpx file you can send me of your route, I'll plug points from it into the software I've written for the TSP, and see how it compares.
I've seen that in this rally and the last, some riders head out with a very loose plan or none at all, and only a few of them have a very good innate sense of what the optimum route is. Or could be. Which is not at all unexpected. The problem is vastly more difficult than any human can even hope to solve, even on the simple basis of shortest distance. Throw in the additional variables of actual time-of-day factors, such as traffic, pit stops, and rest periods, and the problem becomes even larger. I'm very impressed with what some of the riders come up with without the aid of TSP algorithms and other computational help.
Eventually, software will be developed that includes an actual start time, and all the other factors that will follow as the rider travels both through space and time, with appropriate adjustments for rush hour in large cities, rest stops, and time spent doing other things such as eating, collecting bonuses, etc. Adding that to the problem will result in routes that might be less efficient in the purely linear, full-speed A to B to C to D sense, but that are actually achievable by a person out on the road.
I've done some work on it. I'll keep doing more. It's a fun problem.