As a kid, Iain Smith would attend board game parties with his mother and members of her computer science faculty.

"Ticket to Ride has always been a game related to computer science for me," says the fourth-year Bachelor of Science student.

Described as a "cross-country train adventure," award-winning Ticket to Ride allows players to collect cards that enable them to claim railway routes and connect to cities across countries and around the world.

When it came time to begin his capstone project for CMPT 496, Iain found himself thinking back to those games parties and considering projects that could use knowledge of mathematics graph theory.

His initial research and discussions with Dr. Calin Anton, associate professor, led to a previously published paper in which Ticket to Ride is discussed in relation to graph theory and probability and also a paper related to building maps for the game.

As part of the programming stage in his project, Iain used a set of optimizers to choose a set of features, like the value of destinations in the game (e.g., a route from Vancouver to Calgary might be length 5, but a route from Seattle to Vancouver would be a length 1). The optimizers test the game board based on those values.

"The optimizer will make a few of these types of completions, then continue testing but with one of those routes set," Iain explains. Using the example above, it would try new lengths for everything after Seattle to Vancouver but keep Vancouver to Calgary as length 5. "A 'quality' game is one that evens out the win rates between the four agents and that makes for an easy-to-understand quality measure which makes the reasoning behind why such an optimizer could yield boards players find fun."

Looking back on the progress of "Artificial Intelligence Approaches to Build Ticket to Ride Maps," Iain has learned a lot about researching and working in his field, from conceptualizing right through to submitting his paper to the 2022 Educational Advances in Artificial Intelligence conference. (He will be presenting his work at the conference in February.)

"During the programming, I had so many second thoughts about the project and how the work was progressing but had to keep working on the original idea and try to clean up small problems where I could to make it work," says Iain. "But that is how things operate in the working world, and no matter the profession, a project cannot be rolled back after months of work."