GervaisCouple
Mouseketeer
- Joined
- Dec 11, 2012
I just subscribed to TouringPlans...any one have tips on how to use it to plan our WDW trip?
They like to think we're just taking the day as it comes, when really I have it mostly mapped out
LOVE IT! THANKS ALL!
So, the app will allow me to update my plan in the park and recalculate?
I need to test out this at home...
The touringplans algorithm is similar to TSP in that it calculates a route to a list of points, but it is far more complex in every way than TSP. Whereas TSP only requires a list of vertices to visit and a single edge value for each connection, touringplans is evaluating each point of interest's location, wait time as a function of day/time/crowds, FP+ availability, distance to other POIs, and more. It's a combinatorial model that employs the assignment problem and queueing theory. To use a taxonomical comparison, I'd argue that the TP algorithm is more in the same family as TSP than the same species.
Then again, nerds being nerds...
So one of the things the TP optimizer does is takes your touring plan, marries it to another touring plan, they have a touring plan baby, and then it compares the baby to the parent to see which is more efficient. Then it repeats millions of times. Yes really.
Nice analysis. The first approach used queuing theory and combinatorial optimization. The current version thinks of it as a Time-Dependent Traveling Salesman Problem.
Lots of regular TSP heuristics don't work on the time-dependent version. Things like the triangle inequality and convex hull property apply to the TSP but not the TDTSP. So yeah, same family, not same species.
This is true. Even in the code we refer to "parent" and "child" touring plans. There may even be a comment that starts "When a mommy touring plan and daddy touring plan love each other very much..."
...t alas Scholar and JSTOR came up empty.