Planar and Dual Graphs

Prof. Rupal Kapdi, Assistant Professor

Computer Science and Engineering Department,
Institute of Technology, Nirma University.
Planarity of a graph
Applications:

Design of a circuit without any extra layer of insulation.

Supply of the utilities without crossing the lines.
Combinatorial(abstract) and
Geometric Graph
An abstract graph can be defined as:
G1 = (V, E, Ψ)
where the set V consists of the five objects named a, b, c, d, and e,
that is, V = {a, b, c, d, e}, and
the set E consists of seven objects (none of which is in set V) named
1, 2, 3, 4, 5, 6, and 7, that is, E = {1, 2, 3, 4, 5, 6, 7}, and
the relationship between the two sets is defined by the mapping Ψ,
which consists of
Here, the symbol 1 → (a, c) says that object 1 from set E is mapped
onto the (unordered) pair (a, c) of objects from set V.
Now it so happens that this combinatorial abstract object G1 can
also be represented by means of a geometric figure.
In fact, the sketch in Fig. 2-13 is one such geometric representation
of this graph.
Moreover, it is also true that any graph can be represented by
means of such a configuration in three dimensional Euclidean space.
It will often be necessary to make a distinction between the abstract
(or combinatorial) graph and a geometric representation of a graph.
Planar Graph
A graph G is said to be planar if there exists some geometric representation of G
which can be drawn on a plane such that no two of its edges intersect.
A graph that cannot be drawn on a plane without a crossover between its
edges is called nonplanar.
A drawing of a geometric representation of a graph on any surface such that no
edges intersect is called embedding.
Thus, to declare that a graph G is nonplanar, we have to show that of all
possible geometric representations of G none can be embedded in a plane.
Equivalently, a geometric graph G is planar if there exists a graph isomorphic to
G that is embedded in a plane. Otherwise, G is nonplanar.
An embedding of a planar graph G on a plane is called a plane representation of
G.
Planar graphs non-planar graphs
How to check whether the given graph G
(abstract or geometric) is planar?
Kuratowski’s Two Graphs
THEOREM 5.1 - The complete graph of five vertices is nonplanar.

THEOREM 5.2 - Kuratowski’s second graph is also nonplanar.

Several properties common to the two graphs of Kuratowski are
1. Both are regular graphs.
2. Both are nonplanar.
3. Removal of one edge or a vertex makes each a planar graph.
4. Kuratowski’s first graph is the nonplanar graph with the
smallest number of vertices, and Kuratowski’s second graph is the
nonplanar graph with the smallest number of edges.

Thus both are the simplest nonplanar graphs.

Different representations of a planar
graphs.
THEOREM 5.3 - Any simple planar graph can be embedded in a
plane such that every edge is drawn as a straight line segment.

Region: A plane representation of a graph divides the plane into

regions. A region is characterized by the set of edges (or the set of
vertices) forming its boundary.
Note that a region is not defined in a nonplanar graph or even in a
planar graph not embedded in a plane.
Thus a region is a property of the specific plane representation of a
graph and not of an abstract graph.
Infinite Region: The portion of the plane lying outside a graph
embedded in a plane, such as region 4 in Fig. 5-4, is infinite in its
extent. Such a region is called the infinite, unbounded, outer, or
exterior region for that particular plane representation.
Different representations of a Planar Graph
Changing the embedding of a given planar graph, changes the infinite region.
Embedding on a sphere
Embedding on a Sphere: To eliminate the distinction between
finite and infinite regions, a planar graph is often embedded in the
surface of a sphere.
It is accomplished by stereographic projection of a sphere on a
plane. Put the sphere on the plane and call the point of contact SP
(south pole). At point SP, draw a straight line perpendicular to the
plane, and let the point where this line intersects the surface of the
sphere be called NP (north pole).
• Corresponding to any point p on the plane, there exists a unique point
p’ on the sphere and vice versa, where p’ is the point at which the
straight line from point p to point NP intersects the surface of the
sphere.
• Thus there is a one-to-one correspondence between the points of the
sphere and the finite points on the plane, and points at infinity in the
plane correspond to the points NP on the sphere.
• A planar graph embedded in the surface of the sphere divides the
surface into different regions. Each region on the sphere is finite, and
the infinite region on the plane is mapped onto the region containing
the point NP.
• By suitably rotating the sphere we can make any specified region map
onto the infinite region on the plane.
• THEOREM 5.4 - A graph can be embedded in the surface of a sphere if
and only if it can be embedded in a plane.
• Which leads to the theorem:
• Theorem 5.5 - A planar graph may be embedded in a plane such
that any specified region (i.e. specified by the edge forming it) can
be made the infinite region.
• There is no real difference between the infinite region and the
finite region on the plane, the infinite region is also included in the
planar representation of a graph.
• There is no difference between an embedding of a planar graph on
a plane or on a sphere, the term “plane representation” of a graph
is often used to include spherical as well as planar embedding.
• Since a planar graph may have different plane representations, the
number of regions resulting from each embedding is always same.
• Euler’s formula:
• Theorem 5.6 - A connected planar graph with n vertices and e
edges has e-n+2 regions.
• Prove with mathematical induction.
 Proof with direct method:
• any planar graph can be drawn such that each region is a
polygon (a polygonal net). Let the polygonal net representing the
given graph consist of f regions or faces, and let 𝑘𝑝 be the
number of p-sided regions. Since each edge is on the boundary of
exactly two regions,
3 𝑘3 + 4𝑘4 + 5𝑘5 + … + 𝑟𝑘𝑟 = 2𝑒
• where 𝑘𝑟 is the number of polygons, with maximum r edges, and
𝑘 3 + 𝑘4 + 𝑘5 + … + 𝑘𝑟 = 𝑓
• The sum of all angles subtended at each vertex in the polygonal
net is = 2𝜋𝑛.
• Recal that the sum of all interior angles of a p-sided polygon is 𝝅(𝒑 − 𝟐), and the sum of the
exterior angles is 𝝅 𝒑 + 𝟐 .
• Let us compute the summation of all the angles as the grand sum of all interior angles of f −
1 finite regions plus the sum of the exterior angles of the polygon defining the infinite
region as,
• 𝜋 3 − 2 𝑘3 + 𝜋 4 − 2 𝑘4 + 𝜋 5 − 2 𝑘5 + … + 𝜋 𝑟 − 2 𝑘𝑟 + 4𝜋 = 2𝜋𝑛
• 𝜋 3𝑘3 − 2𝑘3 + 4𝑘4 − 2𝑘4 + 5𝑘5 − 2𝑘5 + + … + 5𝑘𝑟 − 2𝑘𝑟 + 4𝜋 = 2𝜋𝑛
• 𝜋 3𝑘3 + 4𝑘4 + 5𝑘5 + … + 5𝑘𝑟 − 2𝑘3 − 2𝑘4 − 2𝑘5 − ⋯ − 2𝑘𝑟 + 4𝜋 = 2𝜋𝑛
• 𝜋 2𝑒 − 2𝑓 + 4𝜋 = 2𝜋𝑛
• 2𝜋 𝑒 − 𝑓 + 4𝜋 = 2𝜋𝑛
• 𝑒−𝑓+2 =𝑛
• 𝑒−𝑛+2 =𝑓
• Corollary: In any simple, connected planar graph with f regions,
n vertices, and e edges (e > 2), the following inequalities must
hold :
For the graphs with the regions made up of at least 3 edges:
3
𝑒 ≥ 𝑓 ⇒ 𝑒 ≤ 3𝑛 − 6
2
For the graphs with the regions made up of at least 4 edges:
2𝑒 ≥ 4𝑓 ⇒ 𝑒 ≤ 2𝑛 − 4
Plane Representation and Connectivity
• In a disconnected graph the embedding of each component can be
considered independently. Therefore, a disconnected graph is planar if
and only if each of its components is planar.
• In a separable graph the embedding of each block can be considered
independently. Therefore a separable graph is planar if and only if each
of its blocks is planar.
• Unique embedding: Two embeddings of a planar graph on sphere are
not distinct if the embeddings can be made to coincide by suitably
rotating one sphere with respect to the other and possibly distorting
regions (without letting a vertex cross an edge). If of all possible
embeddings on a sphere no two are distinct, the graph is said to have a
unique embedding on a sphere.
• Does a nonseparable planar graph G have a unique embedding on a
sphere?
 Two distinct representations of
the same graph do not have
unique embedding on sphere as
(a) does not have any region
with five edges whereas (b) have.

Two same graphs as shown in (c)

and (d) can be made to coincide
by rotating one sphere with
respect to other and possibly
distorting regions (without
letting a vertex cross an edge).

• Theorem 5.7 - The spherical embedding of every planar 3-connected graph is

unique.
Detection of Planarity
• Elementary Reduction
• Step 1: Since a disconnected graph is planar if and only if each of its
components is planar, we need consider only one component at a time.
Also, a separable graph is planar if and only if each of its blocks is planar.
Therefore, for the given arbitrary graph G, determine the set
𝐺 = {𝐺1 , 𝐺2 , … , 𝐺𝑘 }
where each 𝐺𝑖 is a nonseparable block of G. Then we have to test each 𝐺𝑖
for planarity.
• Step 2: Since addition or removal of self-loops does not affect planarity,
remove all self-loops.
• Step 3: Since parallel edges also do not affect planarity, eliminate edges in
parallel by removing all but one edge between every pair of vertices.
• Step 4: Elimination of a vertex of degree two by merging two edges in series
does not affect planarity. Therefore, eliminate all edges in series.
• Let the nonseparable connected graph
𝐺𝑖 be reduced to a new graph 𝐻𝑖 after
the repeated application of steps 3
and 4. What will graph 𝐻𝑖 look like?
• Theorem 5.8 -

Graph 𝐻𝑖 is
1. A single edge, or
2. A complete graph of four vertices, or
3. A nonseparable, simple graph with n ≥
5 and e ≥ 7.
• In Theorem 5-8, all 𝐻𝑖 falling in categories 1 or 2 are planar and need
not be checked further.
• For category 3, investigate only simple, connected, nonseparable
graphs of at least five vertices and with every vertex of degree three
or more.
• Check to see if e ≤ 3n − 6. If this inequality is not satisfied, the graph 𝐻𝑖
is nonplanar. If the inequality is satisfied, test the graph further for
the graphs homomorphic to Kuratowski’s 𝐾5 or 𝐾3,3 graphs.
• Homeomorphic Graphs: Two graphs are said to be homeomorphic if
one graph can be obtained from the other by the creation of edges in
series (i.e., by insertion of vertices of degree two) or by the merger of
edges in series.
Below given three graphs are homeomorphic to each other, for
instance. A graph G is planar if and only if every graph that is
homeomorphic to G is planar. (This is a restatement of series
reduction, step 4 in this section.)
• Theorem 5.9 - A necessary and sufficient condition for a graph G
to be planar is that G does not contain either of Kuratowski’s two
graphs or any graph homeomorphic to either of them.
Petersen Graph

Petersen graph satisfies both

the conditions 𝑒 ≤ 3𝑛 − 6 and
Petersen Graph 𝑒 ≤ 2𝑛 − 4 ) but still the Petersen subgraph homeomorphic to 𝐾3,3
graph is non planar as it has
the subgraph homeomorphic
to 𝐾3,3 ).
• An elegant and simple-looking criterion (Theorem 5.9) for
planarity of a graph, the theorem is difficult to apply in the actual
testing of a large graph (say, a simple, nonseparable graph of 25
vertices, each of degree three or more).
• There have been several alternative characterizations of a planar
graph.
• One of these characterizations, the existence of a dual graph.
 Geometric Dual
• Consider the plane
representation of a graph as
shown in Fig (a), with six regions
or faces 𝐹1 , 𝐹2 , 𝐹3 , 𝐹4 , 𝐹5 , and 𝐹6 .
• Let us place six points 𝑝1 , 𝑝2 , 𝑝3 ,
𝑝4 , 𝑝5 , and 𝑝6 one in each of the
regions, as shown in Fig (b).
• Next let us join these six points
according to the procedure:
If two regions 𝐹𝑖 and 𝐹𝑗 are adjacent (i.e.,
have a common edge), draw a line
joining points 𝑝𝑖 and 𝑝𝑗 that intersects
the common edge between 𝐹𝑖 and 𝐹𝑗
exactly once. If there is more than one
edge common between 𝐹𝑖 and 𝐹𝑗 , draw
one line between points 𝑝𝑖 and 𝑝𝑗 for
each of the common edges.
For an edge e lying entirely in one
region, say 𝐹𝑘 , draw a self-loop at point
𝑝𝑘 intersecting e exactly once.
By this procedure we obtain a
new graph 𝐺 ∗ [in broken lines in
Fig. (c)] consisting of six vertices,
𝑝1 , 𝑝2 , 𝑝3 , 𝑝4 , 𝑝5 , 𝑝6 and edges
joining these vertices. Such a
graph 𝐺 ∗ is called a geometric
dual of 𝐺.
• Clearly, there is a one-to-one correspondence between the edges
of graph 𝐺 and its dual 𝐺 ∗ −one edge of 𝐺 ∗ intersecting one edge
of 𝐺. Some simple observations that can be made about the
relationship between a planar graph 𝐺 and its dual 𝐺 ∗ are
1. An edge forming a self-loop in 𝐺 yields a pendant edge in 𝐺 .
∗

2. A pendant edge in 𝐺 yields a self-loop in 𝐺 .

∗

3. Edges that are in series in 𝐺 produce parallel edges in 𝐺 .

∗

4. Parallel edges in 𝐺 produce edges in series in 𝐺 .

∗
5. Remarks 1-4 are the result of the general observation that the number
of edges constituting the boundary of a region 𝐹𝑖 in 𝐺 is equal to the
degree of the corresponding vertex 𝑝𝑖 in 𝐺 ∗ , and vice versa.
6. Graph 𝐺 ∗ is also embedded in the plane and is therefore planar.
7. Considering the process of drawing a dual 𝐺 ∗ from 𝐺, it is evident that
𝐺 is a dual of 𝐺 ∗ . Therefore, instead of calling 𝐺 ∗ a dual of 𝐺, we usually
say that 𝐺 and 𝐺 ∗ are dual graphs (symmetric property).
8. If 𝑛, 𝑒, 𝑓, 𝑟, and 𝜇 denote as usual the numbers of vertices, edges,
regions, rank, and nullity of a connected planar graph 𝐺, and if 𝑛∗ , 𝑒 ∗ ,
𝑓 ∗ , 𝑟 ∗ , and 𝜇∗ are the corresponding numbers in dual graph 𝐺 ∗ , then
𝑛∗ = 𝑓, 𝑒 ∗ = 𝑒, 𝑓 ∗ = 𝑛, 𝑟 ∗ = 𝜇 , 𝜇∗ = 𝑟
Is a (geometric) dual of a graph unique? OR Are
all duals of a given graph isomorphic?

• Find the duals of the following graphs.

• Duals of the graphs are:

Dual of (a)

The same graph (isomorphic) which has two

distinct embeddings, (a) and (b) leads to the
duals of these isomorphic graphs which are
nonisomorphic. However these duals are 2- Dual of (b)

isomorphic.
Uniqueness of dual graphs
• Is a (geometric) dual of a graph unique? OR Are all duals of a given
graph isomorphic? ANSWER: A planar graph G will have a unique
dual if and only if it has a unique plane representation or unique
embedding on a sphere.
• THEOREM 5.10 - All duals of a planar graph G are 2-isomorphic; and
every graph 2-isomorphic to a dual of G is also a dual of G.
• Since a 3-connected planar graph has a unique embedding on a
sphere, its dual must also be unique. In other words, all duals of a 3-
connected graph are isomorphic.
• It is quite appropriate to refer to a dual as the dual of a planar graph.
The unique planar embedding of a cycle graph
divides the plane into only two regions, the inside
and outside of the cycle (Jordan curve theorem).
However, in an n-cycle, these two regions are
separated from each other by n different edges. Cycle graph

Therefore, the dual graph of the n-cycle is a

multigraph with two vertices (dual to the regions),
connected to each other by n parallel edges. Such a
graph is called a dipole graph.
Conversely, the dual to an n-edge dipole graph is
an n-cycle. Dipole graph

https://en.wikipedia.org/wiki/Dual_graph
Self-dual graphs
• A plane graph is said to be self-dual if it is isomorphic to its dual
graph. The wheel graphs provide an infinite family of self-dual
graphs. Some of the wheel graphs are:

• The four-vertex complete graph is a self-dual graph.

• Which wheel graph is same as four vertex complete graph?
https://en.wikipedia.org/wiki/Dual_graph
Combinatorial Dual
• Independent definition of duality independent of the geometric
representation is :
• THEOREM 5.11 - A necessary and sufficient condition for two
planar graphs 𝐺1 and 𝐺2 to be duals of each other is as follows:
There is a one-to-one correspondence between the edges in 𝐺1
and the edges in 𝐺2 such that a set of edges in 𝐺1 forms a circuit if
and only if the corresponding set in 𝐺2 forms a cut-set.
 Dual of (b)
• Proof: Let us consider a plane representation of a planar graph G. Let
us also draw (geometrically) a dual G* of G. Then consider an
arbitrary circuit Γ in G.
• Clearly, Γ will form some closed simple curve in the plane
representation of G dividing the plane into two areas. (Jordan Curve
Theorem). Thus the vertices of G* are partitioned into two nonempty,
mutually exclusive subsets−one inside Γ and the other outside.
• In other words, the set of edges Γ* in G* corresponding to the set Γ in
G is a cut-set in G*. Likewise it is apparent that corresponding to a cut-
set S* in G* there is a unique circuit consisting of the corresponding
edge-set S in G such that S is a circuit. This proves the necessity
portion of Theorem 5.11.
• To prove the sufficiency, let G be a planar graph and let G’ be a
graph for which there is a one-to-one correspondence between
the cut-sets of G and circuits of G′, and vice versa.
• Let G* be a dual graph of G.
• There is a one-to-one correspondence between the circuits of G′
and cut-sets of G, and also between the cut-sets of G and circuits
of G*. Therefore there is a one-to-one correspondence between
the circuits of G′ and G*, implying that G′ and G* are 2-isomorphic
(Theorem 4-15).
• According to Theorem 5-10, G′ must be a dual of G.
How to obtain Dual of a Subgraph?

• Let G be a planar graph and G* be its dual. Let a be an edge in G, and

the corresponding edge in G* be a*. Suppose that we delete edge a
from G and then try to find the dual of G − a.
• If edge a was on the boundary of two regions, removal of a would
merge these two regions into one. Thus the dual (G − a)* can be
obtained from G* by deleting the corresponding edge a* and then
fusing the two end vertices of a* in G* − a*. On the other hand, if edge
a is not on the boundary, a* forms a self-loop. In that case G* − a* is
the same as (G − a)*.
• Thus if a graph G has a dual G*, the dual of any subgraph of G can be
obtained by successive application of this procedure.
How to obtain Dual of a Homeomorphic
Graph?
• Let G be a planar graph and G* be its dual.
• Let a be an edge in G, and the corresponding edge in G* be a*.
• Suppose that we create an additional vertex in G by introducing a
vertex of degree two in edge a (i.e., a now becomes two edges in
series). How will this addition affect the dual?
• It will simply add an edge parallel to a* in G*. Likewise, the reverse
process of merging two edges in series (step 4 in Section 5-5) will
simply eliminate one of the corresponding parallel edges in G*.
• Thus if a graph G has a dual G*, the dual of any graph homeomorphic
to G can be obtained from G* by the specified procedure.
Find the dual of the following graphs
Find the dual of the following graph.
Find the dual of the following graph. Check if
the graph is self dual or not.
Check whether the below given graph is planar
or not based on the ‘elementary reduction’
graph operations.
Calculate number of edges of a connected
planar graph G with 10 vertices and number of
edges on each face is three.
r = e-n+2 …. (1)
e = 3/2 r …… (2)

e = 24
• The duality depends on the graph being embedded in a plane.
However, now that Theorem 5-11 provides us with an equivalent
abstract definition of duality (namely, the correspondence
between circuits and cut-sets), which does not depend on a plane
representation of a graph.
• Can the concept of duality be extended to nonplanar graphs also.
In other words, given a nonplanar graph G, can we find another
graph G′ with one-to-one correspondence between their edges
such that every circuit in G corresponds to a unique cut-set in G′,
and vice versa? ANSWER: No
• Theorem 5.12: (Whitney’s theorem) A graph has a dual if and only
if it is planar.
We only need to prove that a nonplanar graph does not have a dual. Let
G be a nonplanar graph. Then according to Kuratowski’s theorem, G
contains 𝐾5 or 𝐾3,3 or a graph homeomorphic to either of these.
We have already seen that a graph G can have a dual only if every
subgraph g of G and every graph homeomorphic to g has a dual. Thus if
we can show that neither 𝐾5 nor 𝐾3,3 has a dual, we have proved the
theorem.
Prove by contradiction that neither 𝐾5 nor 𝐾3,3 has a dual.
 𝐾3,3

• Suppose that 𝐾3,3 has a dual D. Observe that the cut-sets in 𝐾3,3
correspond to circuits in D and vice versa (Theorem 5.10).
• Since 𝐾3,3 has no cut-set consisting of two edges, D has no circuit
consisting of two edges. That is, D contains no pair of parallel edges.
• Since every circuit in 𝐾3,3 is of length four or six, D has no cut-set with
less than four edges. Therefore, the degree of every vertex in D is at
least four.
• As D has no parallel edges and the degree of every vertex is at least
four, D must have at least five vertices each of degree four or more.
• That is, D must have at least (5 x 4)/2 = 10 edges. This is a
contradiction, because 𝐾3,3 has nine edges and so must its dual. Thus
𝐾3,3 cannot have a dual.
• Suppose that the graph 𝐾5 has a dual H. 𝐾5
• Note that 𝐾5 has: (1) 10 edges, (2) no pair of parallel edges, (3) no cut-set
with two edges, and (4) cut-sets with only four or six edges.
• Consequently, graph H must have (1) 10 edges, (2) no vertex with degree
less than three, (3) no pair of parallel edges, and (4) circuits of length four
and six only.
• Now graph H contains a hexagon (a circuit of length six), and no more than
three edges can be added to a hexagon without creating a circuit of length
three or a pair of parallel edges.
• Since both of these are forbidden in H and H has 10 edges, there must be at
least seven vertices in H.
• The degree of each of these vertices is at least three. This leads to H having
at least 11 edges. A contradiction.
• There is yet another equivalent combinatorial definition of
duality, also given by Whitney and proved equivalent to the
earlier two definitions [5-10].
• Two planar graphs G and G* are said to be duals (or
combinatorial duals) of each other if there is a one-to-one
correspondence between the edges of G and G* such that if g is
any subgraph of G and h is the corresponding subgraph of G*,
then
• 𝑟𝑎𝑛𝑘 𝑜𝑓 𝐺 ∗ − ℎ = 𝑟𝑎𝑛𝑘 𝑜𝑓 𝐺 ∗ − 𝑛𝑢𝑙𝑙𝑖𝑡𝑦 𝑜𝑓 𝑔.
Example

Dual of (a)

Consider a subgraph {𝑒4 , 𝑒5 , 𝑒6 , 𝑒7 } from G and {𝑒4 ∗ ,𝑒5 ∗ ,𝑒6 ∗ ,𝑒7 ∗ } from 𝐺 ∗ .
𝑟𝑎𝑛𝑘 𝑜𝑓 𝐺 ∗ − 𝑒4 ∗ ,𝑒5 ∗ ,𝑒6 ∗ ,𝑒7 ∗ = 𝑟𝑎𝑛𝑘 𝑜𝑓 𝑒1 ∗ ,𝑒2 ∗ ,𝑒3 ∗ = 2
𝑟𝑎𝑛𝑘 𝑜𝑓𝐺 ∗ = 3
𝑛𝑢𝑙𝑙𝑖𝑡𝑦 𝑜𝑓{𝑒4 , 𝑒5 , 𝑒6 , 𝑒7 } = 1
And 2 = 3 − 1
Clearly, this definition is also independent of the geometric connotation. It is
therefore often preferred for proving results in purely algebraic fashion.
The third classic planarity criterion by
MacLane
• Set of Basic Circuits: A set C of circuits in a graph is said to be a
complete set of basic circuits if
(i) every circuit in the graph can be expressed as a ring sum of some or
all circuits in C, and
(ii) no circuit in C can be expressed as a ring sum of others in C.
It may, however, be mentioned here that whereas a set of fundamental
circuits (as defined in Chapter 3 with respect to a spanning tree) always
constitutes a complete set of basic circuits, the converse does not hold
for all graphs (Problem 5-15).
In a planar graph a complete set of basic circuits has an additional
property.
• In a plane representation of a planar, connected graph G the set of circuits
forming the interior regions constitutes a complete set of basic circuits.
• For any circuit Γ in G can be expressed as the ring sum of the circuits
defining the regions contained in Γ.
• Observe that every edge appears in at most two of these basic circuits. Thus
for every planar graph G we can find a complete set of basic circuits such
that no edge appears in more than two of these basic circuits.
• This result and its converse lead to another well-known characterization of
planar graphs.
• Theorem 5.13 - A graph G is planar if and only if there exists a complete set
of basic circuits (i.e., all μ of them, μ being the nullity of G) such that no edge
appears in more than two of these circuits.
• All three of these classic characterizations suffer from two shortcomings.
First, they are extremely difficult to implement for a large graph. Second, in
case the graph is planar they do not give a plane representation of the
graph.
• These drawbacks have prompted recent discoveries of several
mapconstruction methods, where the testing of planarity itself is based on
an attempt to produce a plane representation of the graph.
• One such method is given by Tutte [5-9]. Several other construction
methods, some of them quite similar, have been implemented on digital
computers [5-2, 5-8].
• In most of these methods, the given graph is first reduced to one or more
simple, nonseparable graphs with every vertex of degree three or more and
with e ≤ 3n − 6. Then the construction algorithm is applied such that either
one succeeds in obtaining a planar realization of the graph or the graph is
nonplanar.
Thickness
• Having found that a given graph G is nonplanar, it is natural to ask, what is
the minimum number of planes necessary for embedding G?
• The least number of planar subgraphs whose union is the given graph G is
called the thickness of G.
• In a printed-circuit board, for instance, the number of insulation layers
necessary is the thickness of the corresponding graph.
• By definition, then, the thickness of a planar graph is one.
• The thickness of each of Kuratowski’s graphs is clearly two.
• The thickness of the complete graph of eight vertices is two, while the
thickness of the complete graph of nine vertices is three.
• Although there are several results available on the thickness of special
types of graphs, the thickness of an arbitrary graph is in general, difficult to
determine.
• Another question one might ask about a nonplanar graph is:
What is the fewest number of crossings (or intersections)
necessary in order to “draw” the graph in a plane?
• The crossing number of a planar graph is, by definition, zero, and
of either of Kuratowski’s graphs, it is one.
• The crossing numbers of only a few graphs have been
determined. No formula exists to give the crossing number of an
arbitrary graph.