GREEDY ALGORITHMS
Tree Vertex Splitting Problem(TVSP)
• Consider a Directed and weighted binary tree
• Consider a network of power line transmission
• The transmission of power from one node to the other
results in some loss, such as drop in voltage
• Each edge is labeled with the loss that occurs (edge weight)
• Network may not be able to tolerate losses beyond a certain
level
• You can place boosters in the nodes to account for the losses
• Definition: Given a network and a loss tolerance level, the
TVSP is to determine an optimal placement of boosters.
In graph terminology
• Let T = (V,E,w) be a weighted directed tree
• V is the set of vertices
• E is the set of edges
• w is the weight function for the edges
• w(i,j) is the weight of the edge <i, j> ε E
• We say that w(i,j) is undefined if there is no edge between i and j.
• A vertex with in-degree zero is called a source vertex
• A vertex with out-degree zero is called a sink vertex
• For any path P in tree T, its delay d(P) is defined to be the sum of
the weights (w(i,j) ) of that path.
• Delay of tree, d(T) is the maximum of all the path delays.
Splitting vertex
• Splitting vertices to create forest
• Let T/X be the forest that results when each vertex
u in X is split into two nodes ui and uo such that all
the edges <u,j> є E (<j, u> є E) are replaced by edges
of the form <uo; j> є E (<j, ui > є E).
• Outbound edges from u now leave from uo
• Inbound edges to u now enter at ui
• Split node is the booster station
Example
1 1
2 3 2 3i
4 5 6 4
3o
5 6
The Problem Statement
• The Tree Vertex Splitting Problem is to identify a
set X V of minimum cardinality for which
d (T / X )
• Where δ is the specified tolerance limit.
• Note: TVSP has a solution only if the maximum
edge weight is less than equal to δ .
Feasible Solution
• Feasible Solution: Any subset X of V if d (T / X )
• Trivial Way:
– Find all the subsets of V.
– For each subset compute d(T/X).
– Select X with minimum cardinality.
– How many subsets?
A greedy approach
• For each node u є V, compute that maximum
delay d(u) from u to any other node in its
subtree.
• If u has a parent v such that d(u) + w(u,v) > δ,
then the node u gets split and d(ui) is set to
zero.
• Computation proceeds from the leaves toward
the root.
Example
1
4 2
2 3
1 3
2
4 5 6
2 3
1 4
7 9 10
8
Example
1
4 2
2 3
1 3
2
4 5 6
2 3
1 4
7 9 10
8
Let δ = 5.
For each leaf node 5, 7,8, 9, 10 delay is 0.
Computing Delay
• Let C(u) be the children of u.
• Then d(u) is given by
d (u ) max d (v) w(u , v)
vC ( u )
Example
1
4 2
2 3
1 3
2
4 5 6
2 3
1 4
7 9 10
8
d(4) = 4
Since d(4) + w(2,4) = 6 > δ=5, Node 4 has to be split.
Example (after splitting node 4)
1
4 2
2 3
1 3
2
4o 4i 5 6
2 3
1 4
7 9 10
8
Example (after splitting node 4)
1
4 2
2 3
1 3
2
4o 4i 5 6
2 3
1 4
7 9 10
8
d(4) = 0
d(2) = 2
Since d(2) + w(1,2) = 6 > δ=5, Node 2 has to be split.
Example (after splitting node 2)
1
4 2
2i 2o 3
1 3
2
4o 4i 5 6
2 3
1 4
7 9 10
8
Example (after splitting node 2)
1
4 2
2i 2o 3
1 3
2
4o 4i 5 6
2 3
1 4
7 9 10
8
d(6) = 3
Since d(6) + w(3,6) = 6 > δ=5, Node 6 has to be split.
Example (after splitting node 6)
1
4 2
2i 2o 3
1 3
2
6i
4o 4i 5
1 4 6o
2 3
7 8
9 10
Example (after splitting node 6)
1
4 2
2i 2o 3
1 3
2
6i
4o 4i 5
1 4 6o
2 3
7 8
9 10
d(3) = 3
d(1) = 5
No further split is required.
Algorithm
TVS(T, δ)
//Determine and output the nodes to be split
//w() is the weighting function for the edges
{
If ( T ≠ 0) then
{
d[T] = 0;
For each child v of T do
{
TVS(v, δ);
d[T] = max{d[T], d[v] + w(T,v)};
}
If((T is not the root) and
(d[T] + w(parent(T),T) > δ)) then
{
Write(T); d[T] = 0;
}
}
}
• Theorem: Algorithm TVS outputs a minimum cardinality set U such
that d(T/U) ≤δ on any tree T, provided no edge of T has weight > δ.
• Proof:
• Base case. If the tree has only one node, the theorem is true.
• Induction hypothesis. Assume that the theorem is true for all
trees of size n.
• Induction step. Consider a tree T of size n + 1
• – Let U be the set of nodes split by TVS.
• – Let W be a minimum cardinality set such that d(T/W) ≤δ
• – We need to show that |U| ≤ |W|
• – If |U|= 0, the above is indeed true
• Otherwise
• Let x be the first vertex split by TVS
• Let Tx be the subtree rooted at x.
• Let T’ = T - Tx + x (Delete Tx from T except for x)
• W has to have at least one node, y, from Tx
• Let W’ = W – {y}.
• If there exists W* such that |W*| < |W’| and d(T’/W*) ≤δ, then since
d(T/(W*+{x})) ≤δ, W is not minimum cardinality split set for T.
• Thus W’ has to be minimum cardinality split set such that d(T’/W’) ≤δ.
• If TVS is run on tree T’, the set of split nodes output is U – {x}
• – Since T’ has ≤ n nodes, U – {x} is a minimum cardinality set split for T’.
• – This means that |W’| ≥ |U| -1, or |W| ≥ |U|
Reference
• E Horowitz, S Sahni, S. Rajsekaran,
Fundamental of Computer Algorithms , 2nd Ed,
University Press.
• http://www.cis.upenn.edu/~matuszek/cit594-
2005/Lectures/36-greedy.ppt
• http://www.cs.umsl.edu/~sanjiv/classes/
cs5130/lectures/gm.pdf
