The game of “Geography” is played by 2 (or more, but let’s stick with 2) players. The idea is that one player lists out the name of a place (say, “California”). The next person needs to list out a name whose first letter is the same as the last letter of the previous name (So, since “California” ends with an “A”, the next name needs to star with an A.  Perhaps the “Alps”). The game continues until one player can’t find a name that hasn’t been used already, and they lose.

Your job is to think about how you’d design an agent to play this game. DO NOT create a program! This is just a design.

You can assume that the agent has access to a “place list” of all legal names.  It may be a directed graph, with places as vertices and an edge (x, y) existing if the last letter of the place name x corresponds to the first letter in the place name y. You can think of it as represented in other ways if you want. 

  • If the agent has multiple choices for the next location, what should it pick?
  • What strategy will you use to pick a move?
  • How will you evaluate whether a move is good or not?



Leave a Reply