The optimal road trip across the U.S. according to machine learning

Careful -- this is only locally optimal! You can do better :). [1] is the proven globally optimal solution of a very similar problem from one of the foremost experts in the TSP, Bill Cook at U of Waterloo. It is solvable in less than a second on your iPhone [2] -- yes your iPhone can solve a TSP to &quot;true&quot; optimality.<p>[1] <a href="http:&#x2F;&#x2F;www.math.uwaterloo.ca&#x2F;tsp&#x2F;usa50&#x2F;road.html" rel="nofollow">http:&#x2F;&#x2F;www.math.uwaterloo.ca&#x2F;tsp&#x2F;usa50&#x2F;road.html</a><p>[2] <a href="https:&#x2F;&#x2F;twitter.com&#x2F;wjcook&#x2F;status&#x2F;575762813345480705" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;wjcook&#x2F;status&#x2F;575762813345480705</a>

Very curious about the decision metric for &quot;major landmark&quot; here. The San Benito County mark, for example, leads you to a dirt road in the middle of nowhere, and the major point for Delaware is just Delaware. Neither of those seem to fit a reasonable metric for a major landmark.

If you&#x27;re wondering why the route has some &#x27;strange&#x27; segments (why get off I-10 to head down Orlando way if you&#x27;re going to drive to Jacksonville later?):<p>The link goes to the &quot;Major U.S. landmarks&quot; road trip.<p>Other trips (including across Canada, South America, and Europe) are listed here: <a href="http:&#x2F;&#x2F;rhiever.github.io&#x2F;optimal-roadtrip-usa&#x2F;" rel="nofollow">http:&#x2F;&#x2F;rhiever.github.io&#x2F;optimal-roadtrip-usa&#x2F;</a>

What does &quot;machine learning&quot; mean here? How does the algorithm guarantee the trip is optimal?

This is clearly not the optimal road trip, since it passes through Nebraska.

WTF! Skips Chicago? Made a huge detour for the Wright Bros.

Ok. Following this &quot;algorithm&quot;, California coast it is not interesting.<p>I pass it.

The cable car museum? Ok, that makes sense for SF. <a href="http:&#x2F;&#x2F;www.cablecarmuseum.org&#x2F;info.html" rel="nofollow">http:&#x2F;&#x2F;www.cablecarmuseum.org&#x2F;info.html</a>

I&#x27;m being pedantic, but the problem being solved here is not a learning problem, but a combinatorial optimization problem.

Not clear to me is whether the machine learning is for the destinations, the sequencing or am I routed on scenic roads?

Skips LA - clever girl.