Given the graph:
With nodes:
A, B, C, D.
and Edges:
A->B, A->C, B->C, B->D, C->D
A path is (under my understanding) a sequence of edges that connect vertices together.
An simple path would be A->B, B->C, C->D.
Is it possible to have a path that has a branch/fork/split?
Example:
A->B, B->C, B->D
No, that's not permitted. The standard definition of a path is a series of one or more nodes v1, v2, ..., vn such that for any two consecutive nodes in the path, there's an edge from the first node to the second. That would rule out the sort of object that you're talking about from being a path.
That's not to say that what you're looking at isn't interesting or couldn't be studied, though. There may be some other definition that captures what you're looking for.
Related
I'm trying to solve this work problem, but I'm having some difficuly, since I suck at recursion.
My question is: is there a way where I can pick a node in a graph and traverse the graph all the way through back to the same node where I started, but the nodes can only be visited once? And of course to save the resulting edges traversed.
The graph is unweighted, but it is coordinates in a 2D x and y coordinate system, so each coordinate has an x and y value, meaning the edges can weighted by calculating the distance between the coordinates. If that helps...
I'm not completely sure I understand, but here is a suggestion: pick a node n0, then an edge e=(n0,n1). Then remove that edge from the graph, and use a breadth-first search to find the shortest path from n1 back to n0 if it exists.
Another suggestion, which might help you to control the length of the resulting path better: Pick a starting node n0 and find a spanning tree T emanating from n0. Remove n0, and T will (hopefully) break into components. Find an edge e=(n1,n2) from one component to another. Then that edge, plus the edges in T connecting the n1 to n0, plus the edges in T connecting n2 to n0, is a cycle with the properties you desire.
Heyo!
So I've got this directed and/or undirected graph with a bunch of vertices and edges. In this graph there is a start vertex and an end vertex. There's also a subset of vertices which are coloured red (this subset can include the start and end vertices). Also, no pair of vertices can have more than one edge between them.
What I have to do is to find:
A) The shortest path that passes no red vertices
B) If there is a path that passes at least one red vertex
C) The path with the greatest amount of red vertices
D) The path with the fewest amount of red vertices
For A I use a breadth first search ignoring red branches. For B I simply brute force it with a depth first search of the graph. And for C and D I use dynamic programming, memoizing the number of red vertices I find in all paths, using the same DFS as in B.
I am moderately happy with all the solutions and I would very much appreciate any suggestions! Thanks!!
For A I use a breadth first search ignoring red branches
A) is a Typical pathfinding problem happening in the sub-graph that contains no red edges. So your solution is good (could be improved with heuristics if you can come up with one, then use A*)
For B I simply brute force it with a depth first search of the graph
Well here's the thing. Every optimal path A->C can be split at an arbitrary intermediate point B. A Nice property of optimal paths, is that every sub-path is optimal. So A->B and B->C are optimal.
This means if you know you must travel from some start to some end through an intermediary red vertex, you can do the following:
Perform a BFS from the start vertex
Perform a BFS from the endvertex backwards (If your edges are directed - as I think - you'll have to take them in reverse, here)
Alternate expanding both BFS so that both their 'edge' (or open lists, as they are called) have the same distance to their respective start.
Stop when:
One BFS hits a red vertex encountered by (or in the 'closed' list of) the other one. In this case, Each BFS can construct the optimal path to that commen vertex. Stitch both semi-paths, and you have your optimal path with at least a red vertex.
One BFS is stuck ('open' list is empty). In this case, there is no solution.
C) The path with the greatest amount of red vertices
This is a combinatorial problem. the first thing I would do is make a matrix of reachability of [start node + red nodes + end nodes] where:
reachability[i, j] = 1 iff there is a path from node i to node j
To compute this matrix, simply perform one BFS search starting at the start node and at every red node. If the BFS reaches a red node, put a 1 in the corresponding line/column.
This will abstract away the underlying complexity of the graph, and make an order of magnitude speedup on the combinatorial search.
The problem is now a longest path problem through that connectivity matrix. dynamic programming would be the way to go indeed.
D) The path with the fewest amount of red vertices
Simply perform a Dijkstra search, but use the following metric when sorting the nodes in the 'open' list:
dist(start, a) < dist(start, b) if:
numRedNodesInPath(start -> a) < numRedNodesInPath(start -> b)
OR (
numRedNodesInPath(start -> a) == numRedNodesInPath(start -> b)
AND
numNodesInPath(start -> a) < numNodesInPath(start -> a)
)
For this, when discovering new vertices, you'll have to store the path leading up to them (well, just the nb of nodes in the path, as well as the nb of red nodes, separately) in a dedicated map to be fetched. I mention this because usually, the length of the path is stored implicitly as the position of the verrtex in the array. You'll have to enforce it explicitely in your case.
Note on length optimality:
Even though you stated you did not care about length optimality outside of problem A), the algorithm I provided will produce shortest-length solutions. In many cases (like in D) it helps Dijkstra converge better I believe.
I'm doing a breadth-first search on a digraph. I'm lost at nodes c and f, and I'm not sure if and how they should be in the BF-tree or if you only go as far as reachable from the source node and don't start at another node in order to get all the vertices.
Here's what I'm getting so far. As you can see, letters mark the nodes. Distance and predecessor are marked by d and pi:
This was helpful BFS traversal of directed graph from a given node but I'm not familiar with graphs enough to understand how that applies to this situation. From what I'm getting from that question, it seems like in this case I would not include c and f at all.
In fact, it seems like I have the maximal number of nodes included already, just because I started at i. I think that the d=4 at the g node (also at k but that doesn't even connect to any other nodes), this is the greatest distance and max-depth possible in BFS in this graph.
I want to find the cheapest path between 2 vertices and I can choose one path which I can go for free., e.g:
Cheapest path between vertices 1 and 6 is 1-3-4-5-6 - I go on edge 1-3 (cost 30) for free and it gives me total cost 21.
Is there any other way than checking all paths one by one?
One way to do it would be to do the following:
Duplicate your graph, so that you have your original graph G with nodes 1,2,etc. and a copy G' with nodes 1',2',etc. and corresponding edges.
For each edge (a,b) of G, make an arc (directed!) from a to b' and another from b to a', both with cost 0.
Finally, use any shortest path algorithm from any vertex of subgraph G to any vertex of G' (1 to 6' in your example).
Basically what happens is that you switch from subgraph G to G' when you use your joker.
You can generalise from there to any number of joker edges by adding extra copies and linking each new one to the last. In that case however you may have to add targets using less jokers in order to account for the cases where the shortest path requires less edges than you have jokers.
Let's say we have a fully connected directed graph G. The vertices are [a,b,c]. There are edges in both directions between each vertex.
Given a starting vertex a, I would like to traverse the graph in all directions and save the path only when I hit a vertex which is already in the path.
So, the function full_paths(a,G) should return:
- [{a,b}, {b,c}, {c,d}]
- [{a,b}, {b,d}, {d,c}]
- [{a,c}, {c,b}, {b,d}]
- [{a,c}, {c,d}, {d,b}]
- [{a,d}, {d,c}, {c,b}]
- [{a,d}, {d,b}, {b,c}]
I do not need 'incomplete' results like [{a,b}] or [{a,b}, {b,c}], because it is contained in the first result already.
Is there any other way to do it except of generating a powerset of G and filtering out results of certain size?
How can I calculate this?
Edit: As Ethan pointed out, this could be solved with depth-first search method, but unfortunately I do not understand how to modify it, making it store a path before it backtracks (I use Ruby Gratr to implement my algorithm)
Have you looked into depth first search or some variation? A depth first search traverses as far as possible and then backtracks. You can record the path each time you need to backtrack.
If you know your graph G is fully connected there is N! paths of length N when N is number of vertices in graph G. You can easily compute it in this way. You have N possibilities of choice starting point, then for each starting point you can choose N-1 vertices as second vertex on a path and so on when you can chose only last not visited vertex on each path. So you have N*(N-1)*...*2*1 = N! possible paths. When you can't chose starting point i.e. it is given it is same as finding paths in graph G' with N-1 vertices. All possible paths are permutation of set of all vertices i.e. in your case all vertices except starting point. When you have permutation you can generate path by:
perm_to_path([A|[B|_]=T]) -> [{A,B}|perm_to_path(T)];
perm_to_path(_) -> [].
simplest way how to generate permutations is
permutations([]) -> [];
permutations(L) ->
[[H|T] || H <- L, T <- permutations(L--[H])].
So in your case:
paths(A, GV) -> [perm_to_path([A|P]) || P <- permutations(GV--[A])].
where GV is list of vertices of graph G.
If you would like more efficient version it would need little bit more trickery.