Professional Documents
Culture Documents
Overview of Some Solved NP-complete Problems in Graph Theory
Overview of Some Solved NP-complete Problems in Graph Theory
net/publication/266861264
CITATIONS READS
0 1,549
1 author:
Milena Bogdanović
Univerzitet Metropolitan
25 PUBLICATIONS 28 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Milena Bogdanović on 10 May 2015.
All problems in NP have a deterministic algorithm in polynomial time, which returns true only
when he is in and submit evidence to the entrance within the language. These algorithms check
potential solutions to the problem. Therefore, the challenge of NP problems is to find the answer
in an efficient manner, it is an effective way to check the solution already found. Since many
important problems in this class, there have been intensive efforts to find a polynomial time
algorithms for problems in NP. However, a large number of NP problems that defy such efforts,
and they seem to require super-polynomial time. One of the biggest open questions in computer
science is: are these problems really are not solvable in polynomial time?
An important notion in this context is a set of NP- complete decision problems, which are a
subset of NP, and can be informally described as the hardest problems in NP. If there were a
polynomial-time algorithm for at least one of them, there would exist a polynomial algorithm for
all problems in NP. Because of this, and because up to now (in spite of all efforts) has not found
a polynomial algorithm for any NP-complete problem, when the show for a problem to be NP-
complete, usually considered to be likely that there is a polynomial algorithm for this problem.
At first, it seemed completely unexpected that NP-problems do exist, but in the famous Cook-
Levin theorem (independently proved by Levin Leonid), Cook proved that SAT problem is NP-
complete. Richard Karp 1972. was proved that several other problems also NP-complete, and
therefore there is a class of NP-complete problems (and the SAT problem is not alone in it).
From Cook’s SAT problems, for thousands of other problems has been shown to be NP-
complete, and many of these problems are collected Gary and Johnson 1979. in the book
Computers and Intractability: A Guide to NP-Completeness.
The problem that satisfies the condition that it be reduced to a NP-complete problem, but not
necessarily that of the NP is called NP-hard problem. Informally, the NP-hard problem is at least
as hard as any NP-complete problem, if not harder. For example, choosing the perfect moves in
certain games of arbitrarily large board is NP-hard or even harder than strictly NP-complete
problems.
The NP-complete problems are classified into the following groups:
The MST (Minimum Spanning Tree) problem is the problem of finding the Minimum Spanning
tree. This is a problem much studied and solved using combinatorial optimization and computer
techniques. Gengui Zhou and Mitsuo Gen, in the paper (Gengui Z., Mitsuo G., 1997.) represent a
new approach to solve this problem using genetic algorithm.
The problem of level-restricted Spanning tree has great practical importance. For the
construction of genetic algorithm to solve MST problem, the authors gave the following model.
For given a finite connected graph to find the minimum spanning subgraph of all nodes.
Observes a connected, undirected graph G = (V, E), where the set of vertices
representing terminals or telecommunications stations, etc.., and a set of edges
that represent connections between the terminals and stations. Each edge is
associated with a positive real number in the set , which represents the
distance, price and more.
Then the spanning tree of the graph G can be represented using the vector x. Let T denote the set
of all such vectors corresponding to the spanning tree in the graph G. Then the MST problem can
be formulated as
min . (2.2.1)
If we assume that there are some limitations on the degree of each node, so the value of the
degree of the each node , some (number of edges incident to each node) assigned to a
highest value , then the problem (2.2.1) and can be formulated as follows:
min . (2.2.2)
The genetic algorithm for the MST problem is formulated thus made which is based on the
experimental results and can be said to provide the optimum solution or near optimal solution
for solving the other „hardly solvable“ the MST problems.
Matsuda and Matsumura in the paper (H. Matsuda, H. Matsumura, 2008.) give sufficient
conditions for the graph was limited degree of tree. All the graphs which they consider are
simple and finite.
Let G is a connected graph and A is a subset of nodes of the graph G. With is denoted the
minimum value of the degree sum in G of arbitrary k independent vertices in A, and
denote the number of components in the induced subgraph G-A. Main results of this paper are:
O. Amines and others in (Amidi O., Peleg D., P rennes S., San J., Saurabh S., 2008.) describe
the two NP-hard problem, as follows: The Minimum Subgraph of Minimum Degree problem
(MSMD) and the Dual Degree-Dense k-Subgraph problem (DDDkS).
Solving the MSMD problem is to find the smallest subgraph of the graph G (in terms of number
of vertices) with minimum degree at least d. O. Amines and oth. have proven to that the MSMD
problem is not in the class APX (class APX consists of all NP-hard optimization problems so that
they can be approximated by a constant factor), for an arbitrary , and whether they were -
approximate algorithm for the class of graphs excluding a fixed subgraph as a minor,
using a technique dynamic programming and known results on the minor of graph. This
algorithm can be applied to planar graphs and to graphs of limited gender.
In the paper (Kumlander D., 2006.) Kumlander provides an exact algorithm to the problem of
determining the maximum weight clique of an arbitrary undirected graph. The algorithm is based
on the fact that nodes of the same independent set can be included in the same maximum clique.
These independent sets are obtained based on heuristics coloring nodes, where each node is the
color class. The class colors and backtrack-search are used to “cut off” edges broad maximum
clique tree search. These strategies used together provide a very efficient algorithm for finding
maximum weight clique. Experimental calculations of random graphs have shown that this
algorithm is better than previously published algorithms, especially for dense graphs.
Let G = (V, E, W) represented an undirected graph, where V is the set of vertices, E - a set of
edges and W-set of weight at each node. Clique is a complete subgraph of G, ie. the subgraph
consists of those nodes that are adjacent in pairs MC problem (Maximum Clique Problem), is the
problem of finding the maximum complete subgraph of the graph G, ie. to finding a set of
vertices of G that are adjacent in pairs. Independent set is a set of vertices which are adjacent in
pairs with all other nodes. Graph coloring problem is defined so that no pair of vertices a graph
can not be painted the same color. All these problems are computationally equivalent,
respectively, each of which can be transformed into one another.
Maximum weight clique problem requires a maximum weight clique. Weighted clique number is
the total weight of the weighted maximum clique. Typically, the weights assigned to the positive
integers, because it does not diminish the complexity of the problem. This problem is a NP-hard
problem.
MC problem has important economic implications in different applications, such as the theory of
codes, pattern recognition, molecular biology, diagnostic set, and so on.
The complete algorithm can be found at http://www.simpleconcepts.ee/dk/index.html.
Another advantage of this algorithm is that it shows how one can use heuristics to obtain
accurate results: it is common for research results set lower or upper limit, which very quickly
lose their actuality, since replaced by new better result. Moreover, the algorithm allows, without
changing the basic steps in it, use better and more efficient heuristics for node coloring.
N. Derhy and C. Picouleau into the paper (Derhy N., Picouleau C., 2009.) solve the problem of
finding a subgraph in a graph is a tree. Detecting induced subgraph which tree plays an important
role in the theoretical combinatorial optimization, but in many applications. In a large series of
problems, such subgraph occurring at finding specific graph structure, complete graph, graph or
cycles and the like.
All graphs in the paper (Derhy N., Picouleau C., 2009.) are finite, simple and connected. For a
graph G, then with V (G) it is mark the set of vertices, and with E (G) the set of edges a graph G.
If x is a node in G, then the set of its neighbors in G is denoted by . If denote a
subgraph with the set of vertices X and set of edges that contains all the edges of G with both
ends in X, called the induced subgraph of X. The graph G is H-free if it does not exist, so that the
induced subgraph on X is isomorphic to H. For an arbitrary integer , the denote the
cycle of k nodes. Accordingly, the graph G is -free, if it do not contain as an induced
subgraph. In the case of a weighted graph, the graph G is defined an WEIGHT function
. If it hold , then we say that X is covered by a set of mandatory
nodes W.
In the paper (Beinstock D., 1992.) Beinstock proves that the next problem is NP-complete:
• INDUSED-TREE:
Input: Graph G and .
• WEIHTED induced-TREE:
2.7. Finding the maximum common subgraph partial k-tree and graph from polynomial
limit the number of spanning tree
In the paper (Yamaguachi A., Aoki KF, Mamitsuka H., 2004.), Yamaguchi and others solve NP-
complete problem, so called. Maximum Common Subgraph problem (MCS) and the problem of
finding a common subgraph, which is quite applied in various fields. For example, in molecular
biology it is possible to reduce the problem space analysis of the structure of chemical bonds.
The authors found that the width of tree chemical components limited by a constant, and that it is
possible that the spanning tree polynomial of each component is limited. The paper (Yamaguachi
A., Aoki KF, Mamitsuka H., 2004.) presented an algorithm that runs in polynomial time, and
finds common induced subgraph of the limited degree of partial k-tree and connected graph
whose number of possible spanning tree is polynomial.
MCS problem is applied in many areas of research, such as video indexing and pattern
recognition, for example. Yamaguachi and oth. have reported this problem to the field of
molecular biology, and using the chemical bond and other appropriate terminology, resolve this
problem.
Faloutsos, McCurley and Tomkins in the paper (Faloustos C., McCurley KS, Tomkins A., 2004.)
are defined connection subgraph as a small subgraph large graph which best reflects the
relationship between the two node graph. The main motivation for this work was to provide a
paradigm for exploration and knowledge discovery in large sociological network graph. In
addition to the formal definition of the problem, given the “ideal” solution based on the analogy
with electricity. It was shown how to accelerate the calculation, how to get a close but high-
quality connection graph in real time on very large graphs.
The authors are working on this issue failed on the basis of small graphs get a lot of information
about the relationship between pairs of nodes. They came up the following results:
• have found display graph generation algorithm, which using the technique of dynamic
programming to find the best leads to a connected graph with nodes, where b is an integer
given in advance.
• They offered candidate graph generation algorithm, with fast heuristics that can handle large
graphs, in almost real time with not lead to decreased accuracy.
REFERENCES
1. Amidi O., Peleg D., P rennes S., San J., Saurabh S., “Degree-Constrained Subgraph
Problems Hardness and Approximation Results”, http://hal.inria.fr/inria-
00331747/en/,(2008) 1-29.
2. Chudnovsky M., Seymour P., „The three-in-a-tree problem“, (2006), (submitted for
publication).
3. Beinstock D., „On the complexity of testing for even holes and induced odd path“,
Discrete Mathematics 90, pp. 85-92 (1991). Corrigendum in Discrete Math. 102 (1992),
109.
4. Edwards K., Farr G., „An Algorithm for finding large induced planar subgraph“, In P.
Mutzel, M. Jünger, S. Leipert, editors, Graph Drawing: 9th International symposium, GD
2001, Lecture Note in Computer Science 2265, pp. 75-83, Springer-Verlag, Berlin (2002).
5. Edwards K., Farr G., „Fragmentability of graphs“, J.Combin. Theory Ser. B, 82:30-37
(2001).
6. Edwards K., Farr G., „Planarization and fragmentability of some classes of graphs“,
Technical Report 2003/144, School of Computer Science and Software Engineering,
MonachUniversity,(2003).
http://www.csse.monash.edu.au/~gfarr/research/publication.html
7. Faloustos C., McCurley K. S., Tomkins A., „Fast Discovery of Connection Subgraphs“,
KDD’04, August 22-25, 2004, Seattle, Washington, USA.
8. Garey M., Johnson D., „Computers and Intractability“, W. H. Freeman, San Francisco,
1979.
9. Halldórsson M. M., Lau H., „Low-degree graph partitioning via local search with
application to constraint sattisfaction, max sut and coloring“, Journal of graph
Algorithms and Applications, 1:1-13 (1997).
10. Kumlander D., „A new exact algorithm for the maximum-weight clique problem based
on a heuristic vertex-coloring and a backtrack search“, 2006.,
http://www.simpleconcept.ee/dk/index.html
11. Lévêque B., Lin D., Maffray F., Trotingnon N., „Detected induced subgraphs“,
Discrete Appl. Math., in press, (doi:10.1016/j.dam.2009.02.015), (special issue GO VI).
12. Lovász L., „On decomposition of graph“, Studia Sci Math. Hungar, 1:237-238 (1986).
13. Matsuda H., Matsumura H., „Degree conditions and degree bounded trees“, Discrete
Mathematics,Volume 309, Issue 11, pp. 3653-3658 (2009).
14. Morgan K., Farr G., „Approximattion Algorithms for the Maximum Induced Planar and
Outerplanar Subgraph Problems“, Journal of Graph Algorithms and Applications,
Vol.11, No.1, pp. 165-193 (2007).
15. Yamaguachi A., Aoki K. F., Mamitsuka H., „Finding the maximum common subgraph
of a partial k-tree and a graph with a polynomially bounded number of spanning trees“
Information Processing Letters 92, pp. 57-63 (2004).
16. Yannakakis M., „Node-and-edge-deletion NP-complete problems“, In STOC ’78:
Proceedings of the Tenth Annual ACM Symposium on Theory of Computing, pp. 253-
264, ACM Press, New York (1978).
17. Gengui Z., Mitsuo G., „A note on genetic algorithms for degree-constrained spanning
tree problems”, networks, 30(2):91-95, 1997. doi: 10.1002/(SICI)1097-
0037(199709)30:2<91::AID-NET3>3.0.CO;2-F.
18. Derhy N., Picouleau C., „Finding induced trees“, Discrete Applied Mathematics, 157,
pp. 3552-3557, (2009).