Question

I’m having some trouble with implementing a breadth-first search in my program that tries to find a Kevin Bacon number. I have created a temporary hashset that stores visited nodes, but the program keeps revisiting nodes despite my removing the nodes from the list that I feed back into the queue.

The information for actors and movies is stored in a HashMap(String, HashSet(String)) – so if you put in an actor as a key, you get the movies they acted in, and if you put in a movie as the key, you get the actors that have acted in it. Attached below is my code and what the System.out.prints print. Thanks for the help!

public List<String> findBaconPath (String actor) throws IllegalArgumentException {
    ArrayList<String> list = new ArrayList<String>();
    HashSet<String> visited = new HashSet<String>();
    visited.add(actor);
    LinkedList<String> bfs = new LinkedList<String>();
    bfs.add(actor);
    while (!bfs.isEmpty()){
        String curr = bfs.remove(); //The current actor we are looking at.
        visited.add(curr);
        HashSet<String> mov = myMovies.get(curr);
        Iterator<String> it = mov.iterator();
        while (it.hasNext()){
            String next = it.next(); //The next movie in the iterator.
            visited.add(next);
            HashSet<String> coStars = myMovies.get(next);
            coStars.removeAll(visited);
            if (coStars.contains("Bacon, Kevin")){
                list.add(curr);
                list.add(next);
                list.add("Bacon, Kevin");
                System.out.println(list.toString());
                return list;
            }
            else {
                list.add(curr);
                list.add(next);
                System.out.println("This is what is in visited: "+visited.toString());
                System.out.println("This is what is in coStars pre removal. "+ coStars.toString());
                coStars.removeAll(visited);
                System.out.println("This is what is in coStars post removal. "+ coStars.toString());
                Iterator<String> cos = coStars.iterator();
                while (cos.hasNext()){
                    bfs.add(cos.next());
                }
            }
        }

    }
    return list;
}

• This is what is in visited: [D, Movie 7]
• This is what is in coStars pre removal. [D, A, B, C]
• This is what is in coStars post removal. [A, B, C]
• This is what is in visited: [D, Movie 7, Movie 2]
• This is what is in coStars pre removal. [D, E, A, B]
• This is what is in coStars post removal. [E, A, B]
• This is what is in visited: [D, A, Movie 7, Movie 2]
• This is what is in coStars pre removal. [A, B, C]
• This is what is in coStars post removal. [B, C]
• This is what is in visited: [D, A, Movie 7, Movie 2]
• This is what is in coStars pre removal. [E, A, B]
• This is what is in coStars post removal. [E, B]
• This is what is in visited: [D, A, Movie 7, Movie 2, Movie 1]
• This is what is in coStars pre removal. [A, B, C]
• This is what is in coStars post removal. [B, C]
• [D, Movie 7, D, Movie 2, A, Movie 7, A, Movie 2, A, Movie 1, A, Movie 0, Bacon, Kevin]

Answer 1

It might help if you better distinguish movies and actors in your visited list instead of mixing them. Something like:

[(Movie1, Actor1, Actor2), (Movie2, Actor1, Actor2, Actor3)]

Then when you do removeAll you can always identify the entry in “visited” by movie title.

Also, removeAll is getting called twice per loop. Once just before ‘if (coStars.contains(“Bacon, Kevin”)){‘ and once again just before you print ‘post removal.’

Btw, is this the answer? Movie 7, Movie 2, Movie 1, Movie 0