Expected linear time MST algorithm: Difference between revisions

Revision as of 09:51, 16 April 2012

Revision as of 05:24, 22 April 2012

Line 1:

A [[randomized algorithm]] for computing the [[minimum spanning forest]] of a [[weighted graph]] with no [[isolated vertex|isolated vertices]]. It was developed by [[David Karger]], Philip Klein, and [[Robert Tarjan]]

.<ref>{{citation

| last1 = Karger | first1 = David R. | author1-link = David Karger

| last2 = Klein | first2 = Philip N.

Line 11:

| title = A randomized linear-time algorithm to find minimum spanning trees

| volume = 42

| year = 1995}}</ref>. The algorithm relies on techniques from [[Borůvka's algorithm]] along with an algorithm for [[MST verification algorithm|verifying a minimum spanning tree in linear time]]

<ref name=MST-V1>{{cite journal

| last1 = Dixon | first1 = Brandon

Line 24:

| url = http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.49.25

}}</ref>

.<ref name=MST-V2>

{{cite conference

| last1 = King | first1 = Valerie

Line 35:

| ___location = London, UK, UK

}}

</ref>. It combines the design paradigms of [[divide and conquer algorithms]], [[greedy algorithms]], and [[randomized algorithms]] to achieve [[Expected value|expected]] [[linear time|linear performance]].

Deterministic algorithms that find the minimum spanning tree include [[Prim's algorithm]], [[Kruskal's algorithm]], [[Reverse-Delete algorithm]], and [[Borůvka's algorithm]].

Line 74:

===F-heavy and F-light Edges===

In each iteration the algorithm removes edges with particular properties that exclude them from the [[minimum spanning tree]]. These are called '''F-heavy edges''' and are defined as follows. Let ''F'' be a forest on the [[Graph (mathematics)|graph]] ''H''. An F-heavy edge is an edge ''e'' connecting vertices ''u'',''v'' whose weight is strictly greater than the weight of the heaviest edge on the path from ''u'' to ''v'' in ''F''. (If a path does not exist in ''F'' it is considered to have infinite weight). Any edge that is not F-heavy is '''F-light'''. If ''H'' is a [[Glossary_of_graph_theory#Subgraphs|subgraph]] of ''G'' then any F-heavy edge in ''G'' cannot be in the minimum spanning tree of ''G'' by the [[Minimum_spanning_tree#Cycle_property | cycle property]]. Given a forest, F-heavy edges can be computed in [[linear time]] using a [[MST verification algorithm|minimum spanning tree verification algorithm]].<ref name=MST-V1/><ref name=MST-V2/>.

==Algorithm==

Line 81:

2 Create a contracted graph ''G''' by running two successive Borůvka steps on ''G''

3 Create a subgraph ''H'' by selecting each edge in ''G''' with probability 1/2. Recursively apply the algorithm to ''H'' to get its minimum spanning forest ''F''.

4 Remove all F-heavy edges from ''G''' (where ''F'' is the forest from step 3) using a [[MST verification algorithm|linear time minimum spanning tree verification algorithm]].<ref name=MST-V1/><ref name=MST-V2/>.

5 Recursively apply the algorithm to ''G''' to get its minimum spanning forest.

Output: The minimum spanning forest of ''G''' and the contracted edges from the Borůvka steps

==Correctness==

Correctness is proved by induction on the number of vertices in the graph. The base case is trivially true. Let ''T*'' be the minimum spanning tree of ''G''. Every edge selected in a Borůvka step is in ''T*'' by the [[Minimum_spanning_tree#Cut_property | cut property]] and none of the edges removed to form the contracted graph are in ''T*'' by the [[Minimum_spanning_tree#Cut_property | cut property]] (for redundant edges) and the [[Minimum_spanning_tree#Cycle_property | cycle property]] (for self loops). The remaining edges of ''T*'' not selected in step 2 form the [[minimum spanning tree]] of the contracted graph by the [[Minimum_spanning_tree#Cut_property | cut property]] (let each cut be a supernode). Every [[#F-heavy and F-light Edges|'''F-heavy edge''']] deleted is not in the minimum spanning tree by the [[Minimum_spanning_tree#Cycle_property | cycle property]]. Finally ''F<nowiki>'</nowiki>'' is the minimum spanning tree of the contracted graph by the inductive hypothesis. Thus ''F<nowiki>'</nowiki>'' and the edges contracted edges from the Borůvka steps form the minimum spanning tree.

==Performance==

The expected performance is a result of the random sampling step. The effectiveness of the random sampling step is described by the following lemma which places a bound on the number of [[#F-heavy and F-light Edges| '''F-light''']] edges in ''G'' thereby restricting the size of the second subproblem.

===Random Sampling Lemma===

'''Lemma'''- Let ''H'' be a subgraph of ''G'' formed by including each edge of ''G'' independently with probability ''p'' and let ''F'' be the minimum spanning forest of ''H''. The [[Expected value|expected number]] of [[#F-heavy and F-light Edges| '''F-light''']] edges in ''G'' is at most ''n/p'' where ''n'' is the number of vertices in ''G''

To prove the lemma examine the edges of ''G'' as they are being added to ''H''. The number of [[#F-heavy and F-light Edges| F-light]] edges in ''G'' is independent of the order in which the edges of ''H'' are selected since the minimum spanning forest of ''H'' is the same for all selection orders. For the sake of the proof consider selecting edges for ''H'' by taking the edges of ''G'' one at a time in order of edge weight from lightest to heaviest. Let ''e'' be the current edge being considered. If the endpoints of ''e'' are in two disconnected components of ''H'' then ''e'' is the lightest edge connecting those components and if it is added to ''H'' it will be in ''F'' by the [[Minimum_spanning_tree#Cut_property | cut property]]. This also means ''e'' is [[#F-heavy and F-light Edges| F-light]] regardless of whether or not it is added to ''H'' since only heavier edges are subsequently considered. If both endpoints of ''e'' are in the same component of ''H'' then it is (and always will be) F-heavy by the [[Minimum_spanning_tree#Cycle_property | cycle property]]. Edge ''e'' is then added to ''H'' with probability ''p''.

The maximum number of [[#F-heavy and F-light Edges| F-light]] edges added to ''H'' is ''n''-1 since any minimum spanning tree of ''H'' has ''n''-1 edges. Once ''n''-1 F-light edges have been added to ''H'' none of the subsequent edges considered are F-light by the [[Minimum_spanning_tree#Cycle_property | cycle property]]. Thus, the number of F-light edges in ''G'' is bounded by the number of F-light edges considered for ''H'' before ''n''-1 F-light edges are actually added to ''H''. Since any F-light edge is added with probability ''p'' this is equivalent to flipping a coin with probability ''p'' of coming up heads until ''n''-1 heads have appeared. The total number of coin flips is equal to the number of F-light edges in ''G''. The distribution of the number of coin flips is given by the [[negative binomial distribution|inverse binomial distribution]] with parameters ''n''-1 and ''p''. For these parameters the expected value of this distribution is (''n''-1)/''p''.

===Expected Analysis===

Ignoring work done in recursive subproblems the total amount of work done in a single invocation of the algorithm is [[linear time|linear]] in the number of edges in the input graph. Step 1 takes constant time. Borůvka steps can be executed in time linear in the number of edges as mentioned in the [[#Borůvka steps|Borůvka step]] section. Step 3 iterates through the edges and flips a single coin for each one so it is linear in the number of edges. Step 4 can be executed in linear time using a modified [[MST verification algorithm|linear time minimum spanning tree verification algorithm]].<ref name=MST-V1/><ref name=MST-V2/>. Since the work done in one iteration of the algorithm is linear in the number of edges the work done in one complete run of the algorithm (including all recursive calls) is bounded by a constant factor times the total number of edges in the original problem and all recursive subproblems.

Each invocation of the algorithm produces at most two subproblems so the set of subproblems forms a [[binary tree]]. Each [[#Borůvka steps|Borůvka step]] reduces the number of vertices by at least a factor of two so after two Borůvka steps the number of vertices has been reduced by a factor of four. Thus, if the original graph has ''n'' vertices and ''m'' edges then at depth ''d'' of the the tree each subproblem is on a graph of at most ''n''/4^''d'' vertices. Also the tree has at most log₄''n'' levels.

Line 108 ⟶ 109:

Each edge in a left child problem is selected from the edges of its parent problem (less the edges contracted in the [[#Borůvka steps|Borůvka steps]]) with probability 1/2. If a parent problem has ''x'' edges then the [[Expected value|expected number]] of edges in the left child problem is at most ''x''/2. If ''x'' is replaced by a random variable ''X'' then by the [[Linearity_of_expectation#Linearity|linearity of expectation]] the expected number of edges in the left child problem ''Y'' is given by <math>E[Y] \leq E[X]/2</math>. Thus if the expected number of edges in a problem at the top of a left path is ''k'' then the sum of the expected number of edges in each subproblem in the left path is at most <math>\sum_{d=0}^{\infty} \frac{k}{2^d}=2k</math> (see [[Geometric series]]). The root has ''m'' edges so the [[Expected value|expected number]] of edges is equal to 2''m'' plus twice the expected number of edges in each right subproblem.

The expected number of edges in each right subproblem is equal to the number of [[#F-heavy and F-light Edges|F-light]] edges in the parent problem where ''F'' is the minimum spanning tree of the left subproblem. The number of F-light edges is less than or equal to twice the number of vertices in the subproblem by the [[#Random Sampling Lemma| sampling lemma]]. The number of vertices in a subproblem at depth ''d'' is ''n''/4^''d'' so the total number of vertices in all right subproblems is given by <math>\sum_{d=1}^{\infty}\frac{2^{d-1}n}{4^d}=n/2</math>. Thus, the expected number of edges in the original problem and all subproblems is at most 2''m''+''n''. Since ''n'' at most 2''m'' for a graph with no isolated vertices the algorithm runs in expected time ''O''(''m'').

===Worst Case Analysis===

The worst case runtime is equivalent to the runtime of [[Borůvka's algorithm]]. This occurs if all edges are added to either the left or right subproblem on each invocation. In this case the algorithm is identical to [[Borůvka's algorithm]] which runs in ''O''(min{''n''², ''m''log''n''}) on a graph with ''n'' vertices and ''m'' edges.

==See Alsoalso==

*[[Kruskal's Algorithm]]

*[[Borůvka's algorithm]]

Line 119 ⟶ 120:

*[[Reverse-Delete algorithm]]

==Further Readingreading==

[http://www.cs.technion.ac.il/~idddo/mstverif.pdf| Minimum Spanning Tree Verification in Linear Time]

==References==