Chinese Postman Problem is a variation of Eulerian circuit problem for undirected graphs. An Euler Circuit is a closed walk that covers every edge once starting and ending position is same. Chiniese Postman problem is defined for connected and undirected graph. The problem is to find shortest path or circuity that visits every edge of the graph at least once.
An undirected and connected graph has Eulerian cycle if “all vertices have even degree“.
It doesn’t matter whether graph is weighted or unweighted, the Chinese Postman Route is always same as Eulerian Circuit if it exists. In weighted graph the minimum possible weight of Postman tour is sum of all edge weights which we get through Eulerian Circuit. We can’t get a shorter route as we must visit all edges at-least once.
In this case, the task reduces to following.
1) In unweighted graph, minimum number of edges to duplicate so that the given graph converts to a graph with Eulerian Cycle.
2) In weighted graph, minimum total weight of edges to duplicate so that given graph converts to a graph with Eulerian Cycle.
Algorithm to find shortest closed path or optimal Chinese postman route in a weighted graph that may not be Eulerian. step 1 : If graph is Eulerian, return sum of all edge weights.Else do following steps. step 2 : We find all the vertices with odd degree step 3 : List all possible pairings of odd vertices For n odd vertices total number of pairings possible are, (n-1) * (n-3) * (n -5)... * 1 step 4 : For each set of pairings, find the shortest path connecting them. step 5 : Find the pairing with minimum shortest path connecting pairs. step 6 : Modify the graph by adding all the edges that have been found in step 5. step 7 : Weight of Chinese Postman Tour is sum of all edges in the modified graph. step 8 : Print Euler Circuit of the modified graph. This Euler Circuit is Chinese Postman Tour.
3 (a)-----------------(b) 1 / | | \1 / | | \ (c) | 5 6| (d) \ | | / 2 \ | 4 | /1 (e)------------------(f) As we see above graph does not contain Eulerian circuit because is has odd degree vertices [a, b, e, f] they all are odd degree vertices . First we make all possible pairs of odd degree vertices [ae, bf], [ab, ef], [af, eb] so pairs with min sum of weight are [ae, bc] : ae = (ac + ce = 3 ), bf = ( ad + df = 2 ) Total : 5 We add edges ac, ce, ad and df to the original graph and create a modified graph. Optimal chinese postman route is of length : 5 + 23 = 28 [ 23 = sum of all edges of modified graph ] Chinese Postman Route : a - b - d - f - d - b - f - e - c - a - c - e - a This route is Euler Circuit of the modified graph.
This article is contributed by Nishant Singh . If you like GeeksforGeeks and would like to contribute, you can also write an article using contribute.geeksforgeeks.org or mail your article to email@example.com. See your article appearing on the GeeksforGeeks main page and help other Geeks.
Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.
- Max Flow Problem Introduction
- Graph Coloring | Set 1 (Introduction and Applications)
- Push Relabel Algorithm | Set 1 (Introduction and Illustration)
- Karger's algorithm for Minimum Cut | Set 1 (Introduction and Implementation)
- Hungarian Algorithm for Assignment Problem | Set 1 (Introduction)
- Hopcroft–Karp Algorithm for Maximum Matching | Set 1 (Introduction)
- Vertex Cover Problem | Set 1 (Introduction and Approximate Algorithm)
- Introduction to Data Structures | 10 most commonly used Data Structures
- Why Prim’s and Kruskal's MST algorithm fails for Directed Graph?
- D'Esopo-Pape Algorithm : Single Source Shortest Path
- Count ways to change direction of edges such that graph becomes acyclic
- Check if a Tree can be split into K equal connected components
- Minimum value of distance of farthest node in a Graph
- Minimum number of colors required to color a graph