Bellman–Ford Algorithm | DP-23

Given a graph and a source vertex src in graph, find shortest paths from src to all vertices in the given graph. The graph may contain negative weight edges.
We have discussed Dijkstra’s algorithm for this problem. Dijkstra’s algorithm is a Greedy algorithm and time complexity is O(VLogV) (with the use of Fibonacci heap). Dijkstra doesn’t work for Graphs with negative weight edges, Bellman-Ford works for such graphs. Bellman-Ford is also simpler than Dijkstra and suites well for distributed systems. But time complexity of Bellman-Ford is O(VE), which is more than Dijkstra.

Algorithm
Following are the detailed steps.

Input: Graph and a source vertex src
Output: Shortest distance to all vertices from src. If there is a negative weight cycle, then shortest distances are not calculated, negative weight cycle is reported.

1) This step initializes distances from source to all vertices as infinite and distance to source itself as 0. Create an array dist[] of size |V| with all values as infinite except dist[src] where src is source vertex.

2) This step calculates shortest distances. Do following |V|-1 times where |V| is the number of vertices in given graph.
…..a) Do following for each edge u-v
………………If dist[v] > dist[u] + weight of edge uv, then update dist[v]
………………….dist[v] = dist[u] + weight of edge uv

3) This step reports if there is a negative weight cycle in graph. Do following for each edge u-v
……If dist[v] > dist[u] + weight of edge uv, then “Graph contains negative weight cycle”
The idea of step 3 is, step 2 guarantees shortest distances if graph doesn’t contain negative weight cycle. If we iterate through all edges one more time and get a shorter path for any vertex, then there is a negative weight cycle

How does this work? Like other Dynamic Programming Problems, the algorithm calculate shortest paths in bottom-up manner. It first calculates the shortest distances which have at-most one edge in the path. Then, it calculates shortest paths with at-most 2 edges, and so on. After the i-th iteration of outer loop, the shortest paths with at most i edges are calculated. There can be maximum |V| – 1 edges in any simple path, that is why the outer loop runs |v| – 1 times. The idea is, assuming that there is no negative weight cycle, if we have calculated shortest paths with at most i edges, then an iteration over all edges guarantees to give shortest path with at-most (i+1) edges (Proof is simple, you can refer this or MIT Video Lecture)

Example
Let us understand the algorithm with following example graph. The images are taken from this source.

Let the given source vertex be 0. Initialize all distances as infinite, except the distance to source itself. Total number of vertices in the graph is 5, so all edges must be processed 4 times.

Let all edges are processed in following order: (B,E), (D,B), (B,D), (A,B), (A,C), (D,C), (B,C), (E,D). We get following distances when all edges are processed first time. The first row in shows initial distances. The second row shows distances when edges (B,E), (D,B), (B,D) and (A,B) are processed. The third row shows distances when (A,C) is processed. The fourth row shows when (D,C), (B,C) and (E,D) are processed.

The first iteration guarantees to give all shortest paths which are at most 1 edge long. We get following distances when all edges are processed second time (The last row shows final values).

The second iteration guarantees to give all shortest paths which are at most 2 edges long. The algorithm processes all edges 2 more times. The distances are minimized after the second iteration, so third and fourth iterations don’t update the distances.

Implementation:

C++

 `// A C++ program for Bellman-Ford's single source  ` `// shortest path algorithm. ` `#include ` ` `  `// a structure to represent a weighted edge in graph ` `struct` `Edge ` `{ ` `    ``int` `src, dest, weight; ` `}; ` ` `  `// a structure to represent a connected, directed and  ` `// weighted graph ` `struct` `Graph ` `{ ` `    ``// V-> Number of vertices, E-> Number of edges ` `    ``int` `V, E; ` ` `  `    ``// graph is represented as an array of edges. ` `    ``struct` `Edge* edge; ` `}; ` ` `  `// Creates a graph with V vertices and E edges ` `struct` `Graph* createGraph(``int` `V, ``int` `E) ` `{ ` `    ``struct` `Graph* graph = ``new` `Graph; ` `    ``graph->V = V; ` `    ``graph->E = E; ` `    ``graph->edge = ``new` `Edge[E]; ` `    ``return` `graph; ` `} ` ` `  `// A utility function used to print the solution ` `void` `printArr(``int` `dist[], ``int` `n) ` `{ ` `    ``printf``(````"Vertex   Distance from Source "````); ` `    ``for` `(``int` `i = 0; i < n; ++i) ` `        ``printf``(````"%d %d "````, i, dist[i]); ` `} ` ` `  `// The main function that finds shortest distances from src to ` `// all other vertices using Bellman-Ford algorithm.  The function ` `// also detects negative weight cycle ` `void` `BellmanFord(``struct` `Graph* graph, ``int` `src) ` `{ ` `    ``int` `V = graph->V; ` `    ``int` `E = graph->E; ` `    ``int` `dist[V]; ` ` `  `    ``// Step 1: Initialize distances from src to all other vertices ` `    ``// as INFINITE ` `    ``for` `(``int` `i = 0; i < V; i++) ` `        ``dist[i]   = INT_MAX; ` `    ``dist[src] = 0; ` ` `  `    ``// Step 2: Relax all edges |V| - 1 times. A simple shortest  ` `    ``// path from src to any other vertex can have at-most |V| - 1  ` `    ``// edges ` `    ``for` `(``int` `i = 1; i <= V-1; i++) ` `    ``{ ` `        ``for` `(``int` `j = 0; j < E; j++) ` `        ``{ ` `            ``int` `u = graph->edge[j].src; ` `            ``int` `v = graph->edge[j].dest; ` `            ``int` `weight = graph->edge[j].weight; ` `            ``if` `(dist[u] != INT_MAX && dist[u] + weight < dist[v]) ` `                ``dist[v] = dist[u] + weight; ` `        ``} ` `    ``} ` ` `  `    ``// Step 3: check for negative-weight cycles.  The above step  ` `    ``// guarantees shortest distances if graph doesn't contain  ` `    ``// negative weight cycle.  If we get a shorter path, then there ` `    ``// is a cycle. ` `    ``for` `(``int` `i = 0; i < E; i++) ` `    ``{ ` `        ``int` `u = graph->edge[i].src; ` `        ``int` `v = graph->edge[i].dest; ` `        ``int` `weight = graph->edge[i].weight; ` `        ``if` `(dist[u] != INT_MAX && dist[u] + weight < dist[v]) ` `            ``printf``(``"Graph contains negative weight cycle"``); ` `    ``} ` ` `  `    ``printArr(dist, V); ` ` `  `    ``return``; ` `} ` ` `  `// Driver program to test above functions ` `int` `main() ` `{ ` `    ``/* Let us create the graph given in above example */` `    ``int` `V = 5;  ``// Number of vertices in graph ` `    ``int` `E = 8;  ``// Number of edges in graph ` `    ``struct` `Graph* graph = createGraph(V, E); ` ` `  `    ``// add edge 0-1 (or A-B in above figure) ` `    ``graph->edge[0].src = 0; ` `    ``graph->edge[0].dest = 1; ` `    ``graph->edge[0].weight = -1; ` ` `  `    ``// add edge 0-2 (or A-C in above figure) ` `    ``graph->edge[1].src = 0; ` `    ``graph->edge[1].dest = 2; ` `    ``graph->edge[1].weight = 4; ` ` `  `    ``// add edge 1-2 (or B-C in above figure) ` `    ``graph->edge[2].src = 1; ` `    ``graph->edge[2].dest = 2; ` `    ``graph->edge[2].weight = 3; ` ` `  `    ``// add edge 1-3 (or B-D in above figure) ` `    ``graph->edge[3].src = 1; ` `    ``graph->edge[3].dest = 3; ` `    ``graph->edge[3].weight = 2; ` ` `  `    ``// add edge 1-4 (or A-E in above figure) ` `    ``graph->edge[4].src = 1; ` `    ``graph->edge[4].dest = 4; ` `    ``graph->edge[4].weight = 2; ` ` `  `    ``// add edge 3-2 (or D-C in above figure) ` `    ``graph->edge[5].src = 3; ` `    ``graph->edge[5].dest = 2; ` `    ``graph->edge[5].weight = 5; ` ` `  `    ``// add edge 3-1 (or D-B in above figure) ` `    ``graph->edge[6].src = 3; ` `    ``graph->edge[6].dest = 1; ` `    ``graph->edge[6].weight = 1; ` ` `  `    ``// add edge 4-3 (or E-D in above figure) ` `    ``graph->edge[7].src = 4; ` `    ``graph->edge[7].dest = 3; ` `    ``graph->edge[7].weight = -3; ` ` `  `    ``BellmanFord(graph, 0); ` ` `  `    ``return` `0; ` `} `

Java

 `// A Java program for Bellman-Ford's single source shortest path ` `// algorithm. ` `import` `java.util.*; ` `import` `java.lang.*; ` `import` `java.io.*; ` ` `  `// A class to represent a connected, directed and weighted graph ` `class` `Graph ` `{ ` `    ``// A class to represent a weighted edge in graph ` `    ``class` `Edge { ` `        ``int` `src, dest, weight; ` `        ``Edge() { ` `            ``src = dest = weight = ``0``; ` `        ``} ` `    ``}; ` ` `  `    ``int` `V, E; ` `    ``Edge edge[]; ` ` `  `    ``// Creates a graph with V vertices and E edges ` `    ``Graph(``int` `v, ``int` `e) ` `    ``{ ` `        ``V = v; ` `        ``E = e; ` `        ``edge = ``new` `Edge[e]; ` `        ``for` `(``int` `i=``0``; i

Python

 `# Python program for Bellman-Ford's single source  ` `# shortest path algorithm. ` ` `  `from` `collections ``import` `defaultdict ` ` `  `#Class to represent a graph ` `class` `Graph: ` ` `  `    ``def` `__init__(``self``,vertices): ` `        ``self``.V``=` `vertices ``#No. of vertices ` `        ``self``.graph ``=` `[] ``# default dictionary to store graph ` `  `  `    ``# function to add an edge to graph ` `    ``def` `addEdge(``self``,u,v,w): ` `        ``self``.graph.append([u, v, w]) ` `         `  `    ``# utility function used to print the solution ` `    ``def` `printArr(``self``, dist): ` `        ``print``(``"Vertex   Distance from Source"``) ` `        ``for` `i ``in` `range``(``self``.V): ` `            ``print``(``"%d %d"` `%` `(i, dist[i])) ` `     `  `    ``# The main function that finds shortest distances from src to ` `    ``# all other vertices using Bellman-Ford algorithm.  The function ` `    ``# also detects negative weight cycle ` `    ``def` `BellmanFord(``self``, src): ` ` `  `        ``# Step 1: Initialize distances from src to all other vertices ` `        ``# as INFINITE ` `        ``dist ``=` `[``float``(``"Inf"``)] ``*` `self``.V ` `        ``dist[src] ``=` `0`  ` `  ` `  `        ``# Step 2: Relax all edges |V| - 1 times. A simple shortest  ` `        ``# path from src to any other vertex can have at-most |V| - 1  ` `        ``# edges ` `        ``for` `i ``in` `range``(``self``.V ``-` `1``): ` `            ``# Update dist value and parent index of the adjacent vertices of ` `            ``# the picked vertex. Consider only those vertices which are still in ` `            ``# queue ` `            ``for` `u, v, w ``in` `self``.graph: ` `                ``if` `dist[u] !``=` `float``(``"Inf"``) ``and` `dist[u] ``+` `w < dist[v]: ` `                        ``dist[v] ``=` `dist[u] ``+` `w ` ` `  `        ``# Step 3: check for negative-weight cycles.  The above step  ` `        ``# guarantees shortest distances if graph doesn't contain  ` `        ``# negative weight cycle.  If we get a shorter path, then there ` `        ``# is a cycle. ` ` `  `        ``for` `u, v, w ``in` `self``.graph: ` `                ``if` `dist[u] !``=` `float``(``"Inf"``) ``and` `dist[u] ``+` `w < dist[v]: ` `                        ``print` `"Graph contains negative weight cycle"` `                        ``return` `                         `  `        ``# print all distance ` `        ``self``.printArr(dist) ` ` `  `g ``=` `Graph(``5``) ` `g.addEdge(``0``, ``1``, ``-``1``) ` `g.addEdge(``0``, ``2``, ``4``) ` `g.addEdge(``1``, ``2``, ``3``) ` `g.addEdge(``1``, ``3``, ``2``) ` `g.addEdge(``1``, ``4``, ``2``) ` `g.addEdge(``3``, ``2``, ``5``) ` `g.addEdge(``3``, ``1``, ``1``) ` `g.addEdge(``4``, ``3``, ``-``3``) ` ` `  `#Print the solution ` `g.BellmanFord(``0``) ` ` `  `#This code is contributed by Neelam Yadav `

C#

 `// A C# program for Bellman-Ford's single source shortest path  ` `// algorithm.  ` ` `  `using` `System; ` ` `  `// A class to represent a connected, directed and weighted graph  ` `class` `Graph  ` `{  ` `    ``// A class to represent a weighted edge in graph  ` `    ``class` `Edge {  ` `        ``public` `int` `src, dest, weight;  ` `        ``public` `Edge() {  ` `            ``src = dest = weight = 0;  ` `        ``}  ` `    ``};  ` ` `  `    ``int` `V, E;  ` `    ``Edge []edge;  ` ` `  `    ``// Creates a graph with V vertices and E edges  ` `    ``Graph(``int` `v, ``int` `e)  ` `    ``{  ` `        ``V = v;  ` `        ``E = e;  ` `        ``edge = ``new` `Edge[e];  ` `        ``for` `(``int` `i=0; i

Output:

```Vertex   Distance from Source
0                0
1                -1
2                2
3                -2
4                1```

Notes
1) Negative weights are found in various applications of graphs. For example, instead of paying cost for a path, we may get some advantage if we follow the path.

2) Bellman-Ford works better (better than Dijksra’s) for distributed systems. Unlike Dijksra’s where we need to find minimum value of all vertices, in Bellman-Ford, edges are considered one by one.

Exercise
1) The standard Bellman-Ford algorithm reports shortest path only if there is no negative weight cycles. Modify it so that it reports minimum distances even if there is a negative weight cycle.

2) Can we use Dijksra’s algorithm for shortest paths for graphs with negative weights – one idea can be, calculate the minimum weight value, add a positive value (equal to absolute value of minimum weight value) to all weights and run the Dijksra’s algorithm for the modified graph. Will this algorithm work?

Bellman Ford Algorithm (Simple Implementation)