Optimization of A* in Python

Question

I am working on an IA (in Python) for small board game, and I need to do a lot of pathfinding per round in order to find the best move I can do (minmax implementation).

My problem is that even if the board is small (9*9), I am able to run it (only) approx. 10000 per second, and I don't know how to speed it. Do you have some advise?

def shortest_path(self, start, goal):
    elements = []
    heapq.heappush(elements, (0, start))

    came_from = {}
    cost_so_far = {}
    came_from[start] = None
    cost_so_far[start] = 0
    current = None
    while len(elements) > 0:
        current = heapq.heappop(elements)[1]
        if goal.is_on_goal(current):
            break
        for next in self.edges[current].neighbors:
            new_cost = cost_so_far[current] + 1
            if next not in cost_so_far or new_cost < cost_so_far[next]:
                cost_so_far[next] = new_cost
                lowest_priority = new_cost + goal.heuristic(next)
                heapq.heappush(elements, (lowest_priority, next))
                came_from[next] = current
    way = []
    if goal.is_on_goal(current):
        reverse_edge = current
        while reverse_edge in came_from:
            way.insert(0, reverse_edge)
            reverse_edge = came_from.get(reverse_edge, None)
    return way

I would like (not sure it is possible) to run it 200 000 times per seconds — trnsnt, Feb 17 '15 at 10:33
Are you running some kind of simulation? You wouldn't need to run it that fast in a game. You'd only have to re-run A* if something in the field changed. — House, Feb 17 '15 at 15:13
200 000 times per second is quite unusual. Can you describe more what you are trying to do here? — amitp, Feb 17 '15 at 15:38

Optimization of A* in Python

0 Answers0