Implementing Randomized Kruskal’s Algorithm

Kruskal's algorithm in mazes

To implement Kruskal’s algorithm in our mazes, we’re going to create a new class, State, to represent the state of the algorithm—the collection of information that it will use internally. Once we’ve got this State class written, the rest of Kruskal’s algorithm becomes almost trivial to implement.

Kruskal’s algorithm initially requires that each cell be associated with a unique set. This suggests that, at a minimum, our State class is going to need to track that association. We’ll need a way to query the set that a particular cell belongs to, as well as the list of cells that belong to a given set. Also, the algorithm wants to select random pairs of neighboring cells on each iteration, so we could probably help things along by keeping a list of those cell pairs. Let's get things started by creating a new class called Kruskals.

The Kruskals class