I got an undirected graph with such that all the edges with same weight and all the vertices are connected. I want to find path between any given pair of vertices.
A less efficient solution is:
To perform BFS starting with one of the vertices, keep track of the visited vertices till the destination vertex is reached. This would perform in O(V + E). But this will have to be done for every pair of vertices given. Hence if there are M number of queries to find path, complexity would be O(M *(E+V)).
Can we do it better? Is it possible to leverage the output a BFS and solve the rest?
As you state that the graph is connected, it is not necessary to call a searching algorithm for the graph more than once. It is sufficient to use a single call to BFS (or DFS, this should make no major difference) to generate a spanning tree for the input. As from your problem statement it is apparently not necessary to find a shortest path between the vertices, any pair (a,b) of vertices is connected via the path from a to the root r of the spanning tree and the path from r to b.
The runtime would be O(V+E), namely the runtime of te searching algorithm; an additional computational cost might be necessary for the generation of the actual paths themselves.
Related
I am trying to understand the time complexity of some efficient methods of detecting cycles in a graph.
Two approaches for doing this are explained here. I would assume that time complexity is provided in terms of worst-case.
The first is union-find, which is said to have a time complexity of O(Vlog E).
The second uses a DFS based approach and is said to have a time complexity of O(V+E). If I am correct this is a more efficient complexity asymptotically than O(Vlog E). It is also convenient that the DFS-based approach can be used for directed and undirected graphs.
My issue is that I fail to see how the second approach can be considered to run in O(V+E) time because DFS runs in O(V+E) time and the algorithm checks the nodes adjacent to any discovered nodes for the starting node. Surely this would mean that the algorithm runs in O(V2) time because up to V-1 adjacent nodes might have to be traversed over for each discovered node? It is obviously impossible for more than one node to require the traversal of n-1 adjacent nodes but from my understanding this would still be the upper bound of the runtime.
Hopefully someone understands why I think this and can help me to understand why the complexity is O(V+E).
The algorithm, based on DFS, typically maintains a "visited" boolean variable for each vertex, which contains one bit of information - this vertex was already visited or not. So, none of vertices can be visited more than once.
If the graph is connected, then starting the DFS from any vertex will give you an answer right away. If the graph is a tree, then all the vertices will be visited in a single call to the DFS. If the graph is not a tree, then a single call to the DFS will find a cycle - and in this case not all the vertices might be visited. In both cases the subgraph, induced by all the already visited vertices, will be a tree at each step of the DFS lookup - so the total number of traversed edges will be O(V). Because of that we can reduce the time complexity estimate O(V+E) of the cycle detection algorithm to O(V).
Starting the DFS from all vertices of the graph is necessary in the case when the graph consists of a number of connected components - the "visited" boolean variable guarantees that the DFS won't traverse the same component again and again.
I have weighted undirected graph. I need to find spanning tree with minimal possible cost, so that distance between point A and B will be as low as possible.
For example, I have this graph: graph.
Minimal distance between A and B is 2.
Minimal spanning tree would look like this. But that would make distance between A and B = 3.
Right now I am doing this:
Find distance between AB in graph, using BFS.
Find all paths between AB with length from step 1 using DFS.
Generate spanning tree from every path from step 2.
Compare them and get minimal one.
Everything is OK until I got graph with A-B distance = 12.
Second step then take too much time. Is there any faster way of doing this? Thanks.
The fastest/most efficient way to solve this problem is to use Dijkstra's Shortest Path algorithm. This is a greedy algorithm that has the following basic structure:
1-all nodes on the graph start "infinity" distance apart
2-start with your first node (node A in your example) and keep track of each edge weight to get from this node A to each of its neighbors.
3-Choose the shortest current edge and follow it to your next node, let's call it node C for now
4-Now, for each of C's neighbors, compare the current distance (including infinity if applicable) with the sum of A's edge to C and C's shortest edge to the current neighbor. If it is shorter than the current distance, update it to the new distance.
5-Continue this process until all nodes have been visited and you reach the node you were looking for the shortest path to (i.e. B in your example)
This is a pretty efficient way of finding the shortest path between two nodes, with a running time of O(V^2), not O(nlogn) as mentioned above. As you can see, by making this a greedy algorithm, we are constantly choosing the optimal solution based on the available local information, and therefore we will never have to go back and change our decisions.
This also eliminates the need for a BFS and a DFS like in your example. Hope this helps!
While your step two is correct, I think that the problem is that you are doing too many operations.
You are doing both a BFS and a DFS which is going to be very costly. Instead, what I would recommend doing is using one of several different traversal techniques that will minimize in the computational costs.
This is a common problem to find the shortest path, and one of the popular solutions is Dijkstra's algorithm. Here is an article that expounds on this topic. https://www.geeksforgeeks.org/dijkstras-shortest-path-algorithm-greedy-algo-7/
In short, what this algorithm does, is it takes the starting point A, and then it generates the minimal spanning tree until point B is hit, then there is a single path in which A can get to B, and that is the shortest path.
Both this and your algorithm both run in O(nlogn), but in practice this solution can be thought of as running a single BFS instead of both a BFS and a DFS.
I´ve been trying to find an algorithm to search if a graph is connected. The graph is undirected and I only want to find a solution (there can be multiple) or if there is none. I was looking for a alg. that performs near linear time, maybe O(logN) or O(NlogN).
Can DFS be up to the task or is there another alternative for this specific problem?
It's going to depend on how you define N, if N is number of vertices, the input itself can be of size O(N^2), and you will be needing to read all of it (unless you have some specific ordering of the input, and than that might change).
DFS runs in O(|V|+|E|) (number of nodes + number of edges), and can find if the graph is connected by simply counting the number of new vertices you discover, and when done, checking if this number is |V|.
Given a directed graph, I need to find all vertices v, such that, if u is reachable from v, then v is also reachable from u. I know that, the vertex can be find using BFS or DFS, but it seems to be inefficient. I was wondering whether there is a better solution for this problem. Any help would be appreciated.
Fundamentally, you're not going to do any better than some kind of search (as you alluded to). I wouldn't worry too much about efficiency: these algorithms are generally linear in the number of nodes + edges.
The problem is a bit underspecified, so I'll make some assumptions about your data structure:
You know vertex u (because you didn't ask to find it)
You can iterate both the inbound and outbound edges of a node (efficiently)
You can iterate all nodes in the graph
You can (efficiently) associate a couple bits of data along with each node
In this case, use a convenient search starting from vertex u (depth/breadth, doesn't matter) twice: once following the outbound edges (marking nodes as "reachable from u") and once following the inbound edges (marking nodes as "reaching u"). Finally, iterate through all nodes and compare the two bits according to your purpose.
Note: as worded, your result set includes all nodes that do not reach vertex u. If you intended the conjunction instead of the implication, then you can save a little time by incorporating the test in the second search, rather than scanning all nodes in the graph. This also relieves assumption 3.
I have a directed acyclic weighted graph which I want to traverse.
The constraints for a valid solution route are:
The sum of the weights of all edges traversed in the route must be the highest possible in the graph, taking in mind the second constraint.
Exactly N vertices must have been visited in the chosen route (including the start and end vertex).
Typically the graph will have a high amount of vertices and edges, so trying all possibilities is not an option, and requires quite an efficient algorithm.
Looking for some pointers or a suitable algorithm for this problem. I know the first condition is easily fulfilled using Dijkstra's algorithm, but I am not sure how to incorporate the second condition, or even where to begin to look.
Please let me know if any additional information is needed.
I'm not sure if you are interested in any path of length N in the graph or just path between two specific vertices; I suspect the latter, but you did not mention that constraint in your question.
If the former, the solution should be a trivial Dijkstra-like algorithm where you sort all edges by their potential path value that starts at the edge weight and gets adjusted by already built adjecent paths. In each iteration, take the node with the best potential path value and add it to an adjecent path. Stop when you get a path of length N (or longer that you cut off at the sides). There are some other technical details esp. wrt. creating long paths, but I won't go into details as I suspect this is not what you are interested in. :-)
If you have fixed source and sink, I think there is no deep magic involved - just run a basic Dijkstra where a path will be associated with each vertex added to the queue, but do not insert vertices with path length >= N into the queue and do not insert sink into the queue unless its path length is N.