Advent of Code 2020 - Day 24Lobby Layout

| Problem statement | Source code | Tags: Cellular automataGrid walking

Part 1

The hexagonal grid looks like this:

1      A   B
2    C   D   E
3  F   G   H   I
   1 2 3 4 5 6 7

We can still label them by rows and columns, and now the neighbors are:

• e: (r, c + 2)
• w: (r, c - 2)
• ne: (r - 1, c + 1)
• nw: (r - 1, c - 1)
• se: (r + 1, c + 1)
• sw: (r + 1, c - 1)

This encoding turns this problem back into a grid problem, where we can use a set of coordinates to track black tiles.

Part 2

Yet another cellular automaton. The only twist is that the grid is hexagonal instead of square, but that just requires tweaks to the neighbor calculation. Again, the template remains unchanged:

def evolve(active: set[Point], get_neighbors: Callable[[Point], list[Point]]) -> set[Point]:
    new_active = set[Point]()
    candidates = active.union(
        neighbor for p in active for neighbor in get_neighbors(p)
    )
    for p in candidates:
        active_neighbors = sum(
            (neighbor in active) for neighbor in get_neighbors(p)
        )
        if p in active and should_stay_active(active_neighbors):
            new_active.add(p)
        elif p not in active and should_become_active(active_neighbors):
            new_active.add(p)
    return new_active

The neighbors are just the six directions defined above.