Search This Blog

Thursday, April 9, 2015

Genetic Algorithms Have Calculated What the Ultimate U.S. Road Trip Looks Like

As reported by The Washington Post: Who needs an atlas when you have an algorithm? Data tinkerer Randy Olson, who is now known across the internet for developing the optimum search path for Where’s Waldo books, has used this same algorithm to compute the optimal American road trip.
video
At the urging of Tracy Staedter from Discovery News, Olson set out to find the quickest driving route that would stop at a national natural landmark, national historic site, national park or national monument in all of the lower 48 states. He also included Washington, D.C. and added another stop in California to get to a total of 50 stops.

Calculating the fastest way to drive between all 50 landmarks – 2,500 individual routes – could theoretically take forever by hand, but Olson used the same genetic algorithm he used to find a search pattern for Where’s Waldo. This algorithm starts with a handful of solutions, takes the best one, and then compares that to other solutions until it can’t find a better one. The result is the route you see above, which includes stops at the Grand Canyon, Pikes Peak, the Alamo, Mount Vernon, Graceland, the White House, the Statue of Liberty, and much more. You can start in any state and follow the path in either direction.

If you didn’t sleep, stop or hit traffic, Olson calculates that this would take roughly 9.33 days of driving. In reality, you probably need a good two to three months to do justice to this epic American road trip.

Olson also created a bonus map with a route through popular U.S. cities, which you can see here