Next: Reducing IP to SGP
Up: Alternative: Reducing IP to
Previous: Alternative: Reducing IP to
Subsections
An instance of the sequential graph partitioning (SGP) problem
consists of a graph G=(V,A) and a non-negative
integer X that is used as the limit for each partition
fragment. There is a total ordering defined on the vertex set
such that for i<j. A weight w(v)
is assigned to each vertex and a length
l(vi,vj) to each edge . The goal is to partition V into subsets with each Vk holding only consecutive vertices -
thus the name sequential graph partitioning in contrast to
traditional graph partitioning which is NP-complete [Hyafil and Rivest, 1973].
The optimality constraints are that partitioning
- minimises the sum of the lengths of the edges that
start and end in different partition fragments and
- leaves the `weight' of each fragment Vk less than or equal to X.
The `weight' of a fragment Vk is usually the sum of the weights
w(v) for the . For our purpose, however, we have to add
the lengths of incoming edges to this weight. This is a minor
change to the problem tackled in [Kernighan, 1971] and does not
change the complexity of the problem. Figure 5.7
summarises the definition of the SGP problem.
In section 5.6.3, we will give an example of SGP,
actually the instance of SGP that results from reducing the IP example
of figures 5.2, 5.4
and 5.6 to an SGP problem.
- Instance:
- An undirected graph G=(V,A) with a total ordering defined on the vertex set
such that for all i<j.
is the set of edges.
For convenience we assume an additional dummy vertex v0 and
that i<j for all . - A function assigning a weight
w(v) to each vertex ,
- a function assigning a length
l(vi,vj) to each edge and
- a positive number X.
- Question:
-
Is there a partition of V into subsets of
consecutive vertices, i.e.
imposed by a set P of partition vertices
and which minimises
such that
for
?
The sets are defined as
for
. A' is the union of these
Definition of SGP
Next: Reducing IP to SGP
Up: Alternative: Reducing IP to
Previous: Alternative: Reducing IP to
Thomas Zurek