Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Minimum Spanning Trees 2704 BOS 867 849 PVD ORD 740 621 1846 LAX 1391 1464 1235 144 JFK 1258 184 802 SFO 337 187 BWI 1090 DFW 946 1121 MIA 2342 © 2010 Goodrich, Tamassia Minimum Spanning Trees 1 Minimum Spanning Trees Spanning subgraph ORD Subgraph of a graph G containing all the vertices of G 1 Spanning tree Spanning subgraph that is itself a (free) tree DEN Minimum spanning tree (MST) Spanning tree of a weighted graph with minimum total edge weight 10 PIT 9 6 STL 4 8 7 3 DCA 5 2 Applications Communications networks Transportation networks © 2010 Goodrich, Tamassia DFW Minimum Spanning Trees ATL 2 Cycle Property Cycle Property: Let T be a minimum spanning tree of a weighted graph G Let e be an edge of G that is not in T and C let be the cycle formed by e with T For every edge f of C, weight(f) weight(e) Proof: By contradiction If weight(f) > weight(e) we can get a spanning tree of smaller weight by replacing f with e 2 4 C © 2010 Goodrich, Tamassia 8 f 6 9 3 e 8 7 7 Replacing f with e yields a better spanning tree f 2 Minimum Spanning Trees 6 8 4 C 9 3 8 e 7 7 3 Partition Property U f Partition Property: Consider a partition of the vertices of G into subsets U and V Let e be an edge of minimum weight across the partition There is a minimum spanning tree of G containing edge e Proof: Let T be an MST of G If T does not contain e, consider the cycle C formed by e with T and let f be an edge of C across the partition By the cycle property, weight(f) weight(e) Thus, weight(f) = weight(e) We obtain another MST by replacing f with e © 2010 Goodrich, Tamassia V 7 4 9 5 2 8 8 3 e 7 Replacing f with e yields another MST U f 2 Minimum Spanning Trees V 7 4 9 5 8 8 3 e 7 4 Kruskal’s Algorithm Maintain a partition of the vertices into clusters Algorithm KruskalMST(G) for each vertex v in G do Create a cluster consisting of v Initially, single-vertex let Q be a priority queue. clusters Insert all edges into Q Keep an MST for each T cluster {T is the union of the MSTs of the clusters} Merge “closest” clusters while T has fewer than n - 1 edges do and their MSTs e Q.removeMin().getValue() [u, v] G.endVertices(e) A priority queue stores the A getCluster(u) edges outside clusters B getCluster(v) Key: weight if A B then Element: edge Add edge e to T mergeClusters(A, B) At the end of the algorithm return T One cluster and one MST © 2010 Goodrich, Tamassia Minimum Spanning Trees 5 Example 8 B 5 1 6 3 H D 9 5 1 2 C 11 7 A 9 C 11 7 10 © 2010 Goodrich, Tamassia E 6 F 3 D H 1 2 A Campus Tour E 6 5 H C 11 10 2 G 9 7 F 3 8 B 4 4 D 10 G 8 5 F G 8 B 4 E C 11 10 B A 9 7 A 1 G 4 E 6 F 3 D H 2 6 Example (contd.) B 5 1 G 8 9 6 H D 10 F 3 C 11 7 A E 8 B 4 9 5 1 2 G C 11 7 A 4 E 6 F 3 H D 10 2 four steps 8 B 1 A G 5 9 C 11 7 10 © 2010 Goodrich, Tamassia E 6 F 3 D H 8 B 4 1 2 A Campus Tour G 5 9 C 11 7 10 4 E 6 F 3 D H 2 7 Data Structures for Kruskal’s Algorithm The graph will be implemented using adjacency lists. The algorithm maintains a forest of trees. A priority queue extracts the edges by increasing weight. The priority queue is implemented as a min heap. An edge is accepted it if connects distinct trees. We need a data structure that maintains a partition, i.e., a collection of disjoint sets, with operations: makeSet(u): create a set consisting of u find(u): return the set storing u union(A, B): replace sets A and B with their union © 2010 Goodrich, Tamassia Minimum Spanning Trees 8 Recall of Data Structures for Disjoint Sets Each set is stored as a linked list represented by its first member. Each list element (set member) has a reference to the set representative. Operation find(u) takes O(1) time, and returns the set of which u is a member. In operation union(A,B), we move the elements of the smaller set to the sequence of the larger set and update their references The time for operation union(A,B) is min(|A|, |B|) Whenever an element is processed, it goes into a set of size at least double, hence each element is processed at most log n times © 2010 Goodrich, Tamassia Minimum Spanning Trees 9 Partition-Based Implementation Partition-based version of Kruskal’s Algorithm Cluster merges as unions Cluster locations as finds Running time O((n + m) log n) PQ operations O(m log m) =O(m log n) UF operations O(n log n) © 2010 Goodrich, Tamassia Algorithm KruskalMST(G) Initialize a partition P for each vertex v in G do P.makeSet(v) let Q be a priority queue. Insert all edges into Q T {T is the union of the MSTs of the clusters} while T has fewer than n - 1 edges do e Q.removeMin().getValue() [u, v] G.endVertices(e) A P.find(u) B P.find(v) if A B then Add edge e to T P.union(A, B) return T Minimum Spanning Trees 10 Prim-Jarnik’s Algorithm Similar to Dijkstra’s algorithm We pick an arbitrary vertex s and we grow the MST as a cloud of vertices, starting from s We store with each vertex v label D(v) representing the smallest weight of an edge connecting v to a vertex in the cloud At each step: We add to the cloud the vertex u outside the cloud with the smallest distance label We update the labels of the vertices adjacent to u © 2010 Goodrich, Tamassia Minimum Spanning Trees 11 Prim-Jarnik’s Algorithm (cont.) We will use adjacency lists for the representation of the input graph. We will use a priority queue to store, for each vertex v, the pair (v,e) with key D(v) where e is the edge with the smallest weight connecting v to the cloud and D(v) is that weight. The priority queue will be implemented as a min heap. © 2010 Goodrich, Tamassia Minimum Spanning Trees 12 Prim-Jarnik’s Algorithm (cont.) Algorithm PrimJarnikMST(G) Pick any vertex v of G D[v] 0 for each vertex u v do D[u] + Initialize T . Initialize a priority queue Q with an entry ((u,null),D[u]) for each vertex u, where (u,null) is the element and D[u] is the key. while Q is not empty do (u,e) Q.removeMin() Add vertex u and edge e to T. for each vertex z adjacent to u such that z is in Q do if weight((u,z)) < D[z] then D[z] weight((u,z)) Change to (z,(u,z)) the element of vertex z in Q Change to D[z] the key of vertex z in Q Return the tree T © 2010 Goodrich, Tamassia Minimum Spanning Trees 13 Example 2 7 B 0 2 B 5 C 0 © 2010 Goodrich, Tamassia 2 0 A 4 9 5 C 5 F 8 8 7 E 7 7 2 4 F 8 7 B 7 D 7 3 9 8 A D 7 5 F E 7 2 2 4 8 8 A 9 8 C 5 2 D E 2 3 7 7 B 0 Minimum Spanning Trees 3 7 7 4 9 5 C 5 F 8 8 A D 7 E 3 7 14 4 Example (contd.) 2 2 B 0 4 9 5 C 5 F 8 8 A D 7 7 7 E 4 3 3 2 B 2 0 © 2010 Goodrich, Tamassia Minimum Spanning Trees 4 9 5 C 5 F 8 8 A D 7 7 7 E 3 3 15 4 Analysis Let n and m denote the number of vertices and edges of the input graph respectively. Since the priority queue is implemented as a heap, we can extract the vertex u in O(log n) time. We can update each D[z] value in O(log n) time as well (why?). This update is done at most once for each edge (u,z). Hence, Prim-Jarnik’s algorithm runs in O((n + m) log n) time. © 2010 Goodrich, Tamassia Minimum Spanning Trees 16 Readings M. T. Goodrich, R. Tamassia and D. Mount. Data Structures and Algorithms in C++. 2nd edition. John Wiley. 2011. Chapter 13 © 2010 Goodrich, Tamassia Minimum Spanning Trees 17