从点列表中找到最短路径

Question

If I have a list of points returned from a breadth-first-search through a type of 2D array maze in Java, how could I find the shortest path in that list of points?

For example, if my potential target points are [2, 4] and [6, 0] and I have a list of points that goes to each target point, how can I find out which route is shortest?

I'm thinking I might have to use Djikstra's Algorithm or something, but I'm unsure.

Thanks very much

Answer 1

You can use Dijkstra's algorithm here. Given the array maze you mentioned, the first step would be to get all the edges in your maze between 2 adjacent nodes, and the edge costs. For the nodes, you would need to know if each node has been visited, and the current cost of visiting that node.

Now, if you are not concerned with getting the shortest path, you might not need Dijkstra's. Instead, you can use a DP/Recursion approach to get possible paths from a source coordinate to a target coordinate. If you want to see a Dijkstra's implementation, this is something I had written. To get the shortest path, you would obviously need the distance between nodes.

To just find a path between 2 points, I would suggest something like this implementation. It includes both DP and recursion implementations and compares the running times for both (recursion can take very long to run for large datasets).

I feel this much should be enough to get you started. Let me know if you need some other information.

Answer 2

I didn't end up using Dijkstra's, but instead changed my breadth-first-search to add a "level" value or "distance" value from the origin point that would count upward with each visited node. The counts would stay consistent if the path ever branched, and since I already knew the end point, all it would take is to check out what the "count" was in my end points and compare.

Thank you for your help though! You'll get an upvote if the site lets me.

For those facing a similar problem: I made a simple class called PointC which inherits from Point, but with an added "count" value. The count was updated appropriately with each step of the breadth first search, then the end points were compared at the end to obtain the most optimal path.