Scheduling policies can change without rewriting the safety-related state transitions in the simulation. The controller should still close doors before moving and open them only at a served stop.

Pending floors form a set, so repeated button presses do not duplicate stops. Each tick first closes open doors; otherwise it selects the nearest pending floor, moves one level, and opens on arrival. Ties favor the lower floor. The example serves finite batches; continuous nearby requests could starve a distant request. Passenger direction, capacity, emergencies, multiple cars, and physical interlocks are explicitly not modeled.

What the function promises: Floors are integers 0..top. request adds a stop; tick returns (floor, doors_open). This is a single-car educational simulator, never suitable for real elevator control.

class Elevator:
    def __init__(self, top, start=0):
        if type(top) is not int or top < 0 or not 0 <= start <= top:
            raise ValueError("Invalid floor range")
        self.top, self.floor = top, start
        self.pending = set()
        self.doors_open = False

    def request(self, floor):
        if type(floor) is not int or not 0 <= floor <= self.top:
            raise ValueError("Floor out of range")
        self.pending.add(floor)

    def tick(self):
        if self.doors_open:
            self.doors_open = False
        elif self.pending:
            target = min(self.pending, key=lambda f: (abs(f-self.floor), f))
            if target != self.floor:
                self.floor += 1 if target > self.floor else -1
            if self.floor == target:
                self.pending.remove(target)
                self.doors_open = True
        return self.floor, self.doors_open

car = Elevator(5)
car.request(2)
car.request(5)
car.request(2)
assert car.tick() == (1, False)
assert car.tick() == (2, True)
assert car.tick() == (2, False)
assert car.tick() == (3, False)
assert car.tick() == (4, False)
assert car.tick() == (5, True)
assert not car.pending

Work and storage: Each scheduling decision scans P pending floors: O(P) time, O(P) storage. More advanced indexes and policies can reduce work or improve fairness, but must follow stated requirements.

When to recognize it: Separate selecting the next stop from moving the car and opening its doors.

Check before applying it: This is not a safety-certified controller. Real elevators require hardware interlocks, fault handling, passenger direction, and multi-car dispatch.

No. Separate hall calls from car destinations, define dispatch ownership and fairness, and coordinate assignments. A directional sweep may reduce reversal, while multi-car dispatch introduces another policy layer.