Section 21.2

Supplementary Problems

[3/0%]
LaTeX
Problem 21.2.1

Compare the two adjacency matrices

A=(0110100110010110),B=(1001011001101001). A = \begin{pmatrix} 0 & 1 & 1 & 0 \\ 1 & 0 & 0 & 1 \\ 1 & 0 & 0 & 1 \\ 0 & 1 & 1 & 0 \end{pmatrix}, \quad B = \begin{pmatrix} 1 & 0 & 0 & 1 \\ 0 & 1 & 1 & 0 \\ 0 & 1 & 1 & 0 \\ 1 & 0 & 0 & 1 \end{pmatrix} .

Do they admit an Euler path?

?
Problem 21.2.2

Consider the adjacency matrix

A=(0111010001100011000101110). A = \begin{pmatrix} 0 & 1 & 1 & 1 & 0 \\ 1 & 0 & 0 & 0 & 1 \\ 1 & 0 & 0 & 0 & 1 \\ 1 & 0 & 0 & 0 & 1 \\ 0 & 1 & 1 & 1 & 0 \end{pmatrix} .

Is there a Hamilton cycle?

?
Problem 21.2.3

Let GG be a graph with vertices VG={1,…,nG}V_{G} = \left\{ 1,\ldots ,n_{G}\right\} and edges EG⊆VG×VGE_{G} \subseteq V_{G} \times V_{G}. The nG×nGn_{G} \times n_{G} adjacency matrix AGA_{G} is given by (AG)ij=χEG(i,j)∈{0,1}(A_{G})_{ij} = \chi_{E_{G}}(i,j) \in \left\{ 0,1\right\} where i,j∈Vi,j \in V. In the following products of graphs G1G_{1} and G2G_{2} with nG1nG2×nG2nG2n_{G_{1}} n_{G_{2}} \times n_{G_{2}} n_{G_{2}} adjacency matrices. The rows and columns are indexed by VG1×VG2V_{G_{1}} \times V_{G_{2}}, where we use the ordering (i,j)≤(k,l)(i,j) \leq (k,l) if i<ki<k or, i=ki=k and j≤lj \leq l.

The cartesian product G1×G2G_{1} \times G_{2} of two graphs G1G_{1} and G2G_{2} is the graph with vertices

VG1×G2=VG1×VG2 V_{G_{1} \times G_{2}} = V_{G_{1}} \times V_{G_{2}}

and edges

EG1×G2={((a,b),(a,b′)):a∈VG1 and (b,b′)∈EG2} E_{G_{1} \times G_{2}} = \left\{ ((a,b),(a,b')) : a \in V_{G_{1}} \text{ and } (b,b') \in E_{G_{2}} \right\} EG1×G2=  ∪{((a,b),(a′,b)):(a,a′)∈EG1 and b∈VG2}. \phantom{E_{G_{1} \times G_{2}} =} \; \cup \left\{ ((a,b),(a',b)) : (a,a') \in E_{G_{1}} \text{ and } b \in V_{G_{2}} \right\} .

It follows that

(AG1×G2)(i,j),(k,l)=δij(AG2)(k,l)⊞δkl(AG1)(i,j) (A_{G_{1} \times G_{2}})_{(i,j),(k,l)} = \delta _{ij} (A_{G_{2}})_{(k,l)} \boxplus \delta _{kl} (A_{G_{1}})_{(i,j)}

where ⊞\boxplus is the usual addition with the convention that 1⊞1=11 \boxplus 1 = 1. Thus

AG1×G2=(AG1⊗InG2)⊞(InG2⊗AG2). A_{G_{1} \times G_{2}} = (A_{G_{1}} \otimes I_{n_{G_{2}}}) \boxplus (I_{n_{G_{2}}} \otimes A_{G_{2}}) .

The lexicographic product G1∙G2G_{1} \bullet G_{2} of two graphs G1G_{1} and G2G_{2} is the graph with vertices

VG1∙G2=VG1×VG2 V_{G_{1} \bullet G_{2}} = V_{G_{1}} \times V_{G_{2}}

and edges

EG1∙G2={((a,b),(a,b′)):a∈VG1 and (b,b′)∈EG2} E_{G_{1} \bullet G_{2}} = \left\{ ((a,b),(a,b')) : a \in V_{G_{1}} \text{ and } (b,b') \in E_{G_{2}} \right\} EG1∙G2=  ∪{((a,b),(a′,b′)):(a,a′)∈EG1 and b,b′∈VG2}. \phantom{E_{G_{1} \bullet G_{2}} =} \; \cup \left\{ ((a,b),(a',b')) : (a,a') \in E_{G_{1}} \text{ and } b,b' \in V_{G_{2}} \right\} .

Thus EG1×G2⊆EG1∙G2E_{G_{1} \times G_{2}} \subseteq E_{G_{1} \bullet G_{2}}. It follows that

(AG1×G2)(i,j),(k,l)=δij(AG2)(k,l)⊞(AG1)(i,j). (A_{G_{1} \times G_{2}})_{(i,j),(k,l)} = \delta _{ij} (A_{G_{2}})_{(k,l)} \boxplus (A_{G_{1}})_{(i,j)} .

Hence

AG1×G2=(AG1⊗1→nG2)⊞(InG2⊗AG2). A_{G_{1} \times G_{2}} = (A_{G_{1}} \otimes \overrightarrow {1}_{n_{G_{2}}}) \boxplus (I_{n_{G_{2}}} \otimes A_{G_{2}}) .

Here 1→nG1\overrightarrow {1}_{n_{G_{1}}} is the nG2×nG2n_{G_{2}} \times n_{G_{2}} with every entry equal to 1.

The tensor product G1⊗G2G_{1} \otimes G_{2} of two graphs G1G_{1} and G2G_{2} is the graph with vertices

VG1⊗G2=VG1×VG2 V_{G_{1} \otimes G_{2}} = V_{G_{1}} \times V_{G_{2}}

and edges

EG1⊗G2={((a,b),(a′,b′)):(a,a′)∈EG1 and (b,b′)∈EG2}. E_{G_{1} \otimes G_{2}} = \left\{ ((a,b),(a',b')) : (a,a') \in E_{G_{1}} \text{ and } (b,b') \in E_{G_{2}} \right\} .

It follows that

(AG1⊗G2)(i,j),(k,l)=(AG1)(i,j)(AG2)(k,l)=(AG1⊗AG2)(i−1)nG2+k,(j−1)nG2+l. (A_{G_{1} \otimes G_{2}})_{(i,j),(k,l)} = (A_{G_{1}})_{(i,j)} (A_{G_{2}})_{(k,l)} = (A_{G_{1}} \otimes A_{G_{2}})_{(i-1)n_{G_{2}}+k,(j-1)n_{G_{2}}+l} .

Thus AG1⊗G2=AG1⊗AG2A_{G_{1} \otimes G_{2}} = A_{G_{1}} \otimes A_{G_{2}} where ⊗\otimes is the Kronecker product of matrices.

The normal product G1⋆G2G_{1} \star G_{2} of two graphs G1G_{1} and G2G_{2} is the graph with vertices

VG1⋆G2=VG1×VG2 V_{G_{1} \star G_{2}} = V_{G_{1}} \times V_{G_{2}}

and edges

EG1⋆G2=EG1×G2∪EG1⊗G2. E_{G_{1} \star G_{2}} = E_{G_{1} \times G_{2}} \cup E_{G_{1} \otimes G_{2}} .

Thus

AG1⋆G2=AG1×G2⊞AG1⊗G2=(AG1⊗InG2)⊞(InG2⊗AG2)⊞AG1⊗AG2. A_{G_{1} \star G_{2}} = A_{G_{1} \times G_{2}} \boxplus A_{G_{1} \otimes G_{2}} = (A_{G_{1}} \otimes I_{n_{G_{2}}}) \boxplus (I_{n_{G_{2}}} \otimes A_{G_{2}}) \boxplus A_{G_{1}} \otimes A_{G_{2}} .
?
(i)

Show that the Euler path is not preserved (in general) under these operations.

(ii)

Show that the Hamilton path is not preserved (in general) under these operations.