About

The Vision

If all of Vermont's tasting rooms could work together, we could attract more visitors to Vermont and provide a much more seamless, fulfilling experience. Vermont is small, but one thing Vermonters do exceptionally well is collaborate with one another and work together. This network of strategic partners will leverage one of Vermont's key strengths and invite newcomers to participate in Vermont's welcoming culture more easily.

Drink the East allows visitors to:

1. Easily view all of the tasting rooms that are open when you're planning to take your tour.
2. Accommodate the tastes and interests of everyone in your group by narrowing the list of open tasting rooms to just the categories that your group is interested in.
3. Accommodate the beer nerd in your group who insists upon visiting one particular tasting room by selecting a mandatory site in your search.
4. Become aware of hotels, restaurants, and other popular destinations along your route.
5. Email yourself an itinerary and link to a google map detailing your route.

How We Do It: The Tech

Genetic algorithms are a powerful tool for finding solutions to complicated problems. Instead of trying to find answers by brute force, using trial and error repeatedly until a good solution reveals itself, genetic algorithms deliberately navigate a search space, continually moving towards better and better solutions until an optimal solution is achieved.
Genetic algorithms are a type of evolutionary computation and use the same principles that geneticists do when trying to understand why different constellations of genes were passed down. They were initially conceived by Holland as a way to study adaptive behavior (Holland, 1992) but have been used extensively to address optimization problems.

Our team created a genetic algorithm in order to find the optimal route to take on a tasting room tour in Vermont. This question is a variation on the Traveling Salesperson Problem where a salesperson has to visit n number of cities without re-visiting any city. By reviewing the literature on the Traveling Salesperson problem, we were able to choose mutation and recombination operators that prioritized adjacencies, allowing increasingly better solutions to emerge during our search.

Genetic algorithms work similarly to natural selection, in that some fitness value is assigned to each combination of alleles and the individuals with a higher fitness are more likely to survive and make it into the next generation. We created a customized fitness function for this problem based off of data we gleaned from google's mapping api. This fitness function allows us to evolve an optimal solution to this mapping problem that can be run quickly and repeatedly.

wWe