Difference between revisions of "Aufgaben:Exercise 4.09Z: Laplace Distributed Noise"

From LNTwww
Line 1: Line 1:
  
{{quiz-Header|Buchseite=Digitalsignalübertragung/Approximation der Fehlerwahrscheinlichkeit}}
+
{{quiz-Header|Buchseite=Digital_Signal_Transmission/Approximation_of_the_Error_Probability}}
  
[[File:P_ID2042__Dig_Z_4_9.png|right|frame|2D–Laplace–WDF]]
+
[[File:P_ID2042__Dig_Z_4_9.png|right|frame|2D Laplacian PDF]]
Wir betrachten zweidimensionales Rauschen  $\boldsymbol{n} = (n_1, n_2)$.  
+
We consider two-dimensional noise  $\boldsymbol{n} = (n_1, n_2)$.  
  
Die beiden Rauschvariablen sind "<i>independent and identically distributed</i>&nbsp;", abgekürzt i.i.d., und besitzen beide jeweils eine Laplace&ndash;Wahrscheinlichkeitsdichte:
+
The two noise variables are "<i>independent and identically distributed</i>&nbsp;", abbreviated i.i.d., and each has a Laplace probability density:
 
:$$p_{n_1}(x)  \hspace{-0.1cm} \ = \ \hspace{-0.1cm}  K \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm} |x|}  \hspace{0.05cm},$$
 
:$$p_{n_1}(x)  \hspace{-0.1cm} \ = \ \hspace{-0.1cm}  K \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm} |x|}  \hspace{0.05cm},$$
 
:$$ p_{n_2}(y)  \hspace{-0.1cm} \ = \ \hspace{-0.1cm}  K \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm} |y|}  \hspace{0.05cm}. $$
 
:$$ p_{n_2}(y)  \hspace{-0.1cm} \ = \ \hspace{-0.1cm}  K \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm} |y|}  \hspace{0.05cm}. $$
  
*Die 2D&ndash;Wahrscheinlichkeitsdichtefunktion&nbsp; $p_{\it \boldsymbol{n}}(x, y)$&nbsp; ist in der Grafik dargestellt.  
+
*The 2D probability density function&nbsp; $p_{\it \boldsymbol{n}}(x, y)$&nbsp; is shown in the graph.
*Zur Vereinfachung der Schreibweise werden hier die Realisierungen von&nbsp; $n_1$&nbsp; und&nbsp; $n_2$&nbsp; mit&nbsp; $x$&nbsp; und&nbsp; $y$&nbsp; bezeichnet.
+
*To simplify notation, the realizations of&nbsp; $n_1$&nbsp; and&nbsp; $n_2$&nbsp; are denoted here by&nbsp; $x$&nbsp; and&nbsp; $y$,&nbsp; respectively.
  
  
  
''Hinweise:''
+
''Notes:''
* Die Aufgabe gehört zum  Kapitel&nbsp; [[Digitalsignal%C3%BCbertragung/Approximation_der_Fehlerwahrscheinlichkeit| Approximation der Fehlerwahrscheinlichkeit]].  
+
* The exercise belongs to the chapter&nbsp; [[Digital_Signal_Transmission/Approximation_of_the_Error_Probability|"Approximation of the Error Probability"]].  
* Wir weisen Sie auf das interaktive Applet &nbsp;[[Applets:2D_Laplace|Zweidimensionale Laplaceverteilung]]&nbsp; hin.
+
* We would like to refer to the interactive applet &nbsp;[[Applets:2D_Laplace|"2D Laplace"]].&nbsp;  
* Das sich in Teilaufgabe '''(6)''' ergebende Integral muss aufgrund der Betragsbildung in mehrere Teilintegrale aufgespalten werden.
+
* The integral resulting in subtask '''(6)''' must be split into several partial integrals due to the magnitude formation.
*Weiterhin gilt:
+
*Furthermore, it holds:
 
:$$\int_{0}^{\infty} x^2 \cdot {\rm e}^{-a \hspace{0.03cm}\cdot \hspace{0.03cm} x} \,{\rm d} x  = {2}/{a^3} \hspace{0.05cm}.$$
 
:$$\int_{0}^{\infty} x^2 \cdot {\rm e}^{-a \hspace{0.03cm}\cdot \hspace{0.03cm} x} \,{\rm d} x  = {2}/{a^3} \hspace{0.05cm}.$$
  
  
  
===Fragebogen===
+
===Questions===
 
<quiz display=simple>
 
<quiz display=simple>
{Wie groß ist die Konstante&nbsp; $K$&nbsp; der 1D&ndash;WDF?
+
{What is the size of the constant&nbsp; $K$&nbsp; of the 1D PDF?
 
|type="[]"}
 
|type="[]"}
 
- $K = 1$.
 
- $K = 1$.
Line 31: Line 31:
 
- $K = 1/a$.
 
- $K = 1/a$.
  
{Es sei&nbsp; $a = 1$. Wie groß sind der Mittelwert&nbsp; ${\rm E}\big[n_i \big]$&nbsp; und die Varianz&nbsp; $\sigma^2 = {\rm E}\big[n_i^2\big]$&nbsp; der beiden 1D&ndash;Zufallsgrößen? &nbsp;$(i = 1, 2)$
+
{Let&nbsp; $a = 1$. What are the mean&nbsp; ${\rm E}\big[n_i \big]$&nbsp; and variance&nbsp; $\sigma^2 = {\rm E}\big[n_i^2\big]$&nbsp; of the two 1D random variables? &nbsp;$(i = 1, 2)$
 
|type="{}"}
 
|type="{}"}
 
${\rm E}\big[n_i\big] \ = \ $ { 0. }
 
${\rm E}\big[n_i\big] \ = \ $ { 0. }
 
${\rm E}\big[n_i^2\big] \ = \ ${ 2 3% }  
 
${\rm E}\big[n_i^2\big] \ = \ ${ 2 3% }  
  
{Welche Form haben die Höhenlinien der 2D&ndash;WDF im ersten Quadranten?
+
{What is the shape of the contour lines of the 2D PDF in the first quadrant?
 
|type="[]"}
 
|type="[]"}
+ Es sind Geraden.
+
+ They are straight lines.
- Es sind Hyperbeln.
+
- They are hyperbolas.
- Es sind Kreise.
+
- They are circles.
  
{Es sei weiterhin&nbsp; $a = 1$. Wie groß ist die Wahrscheinlichkeit, dass sowohl&nbsp; $n_1$&nbsp; als auch&nbsp; $n_2$&nbsp; negativ sind?
+
{Let further&nbsp; $a = 1$. What is the probability that both&nbsp; $n_1$&nbsp; and&nbsp; $n_2$&nbsp; are negative?
 
|type="{}"}
 
|type="{}"}
 
${\rm Pr}\big[(n_1 < 0) ∩ (n_2 < 0)\big]\ = \ $ { 25 3% } $\ \%$
 
${\rm Pr}\big[(n_1 < 0) ∩ (n_2 < 0)\big]\ = \ $ { 25 3% } $\ \%$
  
{Wie groß ist die Wahrscheinlichkeit, dass&nbsp; $n_1$&nbsp; und&nbsp; $n_2$&nbsp; gemeinsam größer als&nbsp; $1$&nbsp; sind?
+
{What is the probability that&nbsp; $n_1$&nbsp; and&nbsp; $n_2$&nbsp; are jointly greater than&nbsp; $1$?&nbsp;  
 
|type="{}"}
 
|type="{}"}
 
${\rm Pr}\big[(n_1 > 1) ∩ (n_2 > 1)\big]\ = \ $ { 3.4 3% } $\ \%$
 
${\rm Pr}\big[(n_1 > 1) ∩ (n_2 > 1)\big]\ = \ $ { 3.4 3% } $\ \%$
  
{Wie groß ist die Wahrscheinlichkeit, dass die Summe&nbsp; $n_1 + n_2 > 2$&nbsp; ist?
+
{What is the probability that the sum is&nbsp; $n_1 + n_2 > 2$?&nbsp;  
 
|type="{}"}
 
|type="{}"}
 
$ {\rm Pr}\big[n_1 + n_2 > 2)\big] \ = \ $ { 13.5 3% } $\ \%$
 
$ {\rm Pr}\big[n_1 + n_2 > 2)\big] \ = \ $ { 13.5 3% } $\ \%$
 
</quiz>
 
</quiz>
  
===Musterlösung===
+
===Solution===
 
{{ML-Kopf}}
 
{{ML-Kopf}}
'''(1)'''&nbsp; Richtig ist der <u>Lösungsvorschlag 2</u>:
+
'''(1)'''&nbsp; <u>Solution 2</u> is correct:
*Die Fläche unter der WDF muss $1$ ergeben:
+
*The area under the PDF must equal $1$:
 
:$$\int_{-\infty}^{+\infty}  p_{n_1}(x) \,{\rm d} x  = 1 \hspace{0.3cm}\Rightarrow \hspace{0.3cm}
 
:$$\int_{-\infty}^{+\infty}  p_{n_1}(x) \,{\rm d} x  = 1 \hspace{0.3cm}\Rightarrow \hspace{0.3cm}
 
  \int_{0}^{+\infty}  p_{n_1}(x) \,{\rm d} x  = 0.5  \hspace{0.3cm}  
 
  \int_{0}^{+\infty}  p_{n_1}(x) \,{\rm d} x  = 0.5  \hspace{0.3cm}  
Line 65: Line 65:
  
  
'''(2)'''&nbsp; Der <u>lineare Mittelwert</u> ist aufgrund der WDF&ndash;Symmetrie <u>gleich 0</u>.  
+
'''(2)'''&nbsp; The <u>linear mean</u> is <u>equal to 0</u> due to PDF symmetry.  
  
Damit ist die Varianz $\sigma^2$ tatsächlich &ndash; wie bereits in der Fragestellung angegeben &ndash; gleich dem quadratischen Mittelwert:
+
Thus, the variance $\sigma^2$ is actually equal to the root mean square, as already stated in the question:
 
:$$\sigma^2 = {\rm E}[n_1^2] = 2 \cdot \frac{a}{2} \cdot \int_{0}^{\infty} x^2 \cdot {\rm e}^{-a \hspace{0.03cm} \cdot \hspace{0.03cm} x} \,{\rm d} x  = a \cdot {2}/{a^3}=
 
:$$\sigma^2 = {\rm E}[n_1^2] = 2 \cdot \frac{a}{2} \cdot \int_{0}^{\infty} x^2 \cdot {\rm e}^{-a \hspace{0.03cm} \cdot \hspace{0.03cm} x} \,{\rm d} x  = a \cdot {2}/{a^3}=
 
  {2}/{a^2} \hspace{0.05cm}. \hspace{0.2cm}{\rm Mit}\hspace{0.15cm}a = 1\text{:} \hspace{0.2cm}\hspace{0.1cm}\underline {\sigma^2 = 2 }\hspace{0.05cm}.$$
 
  {2}/{a^2} \hspace{0.05cm}. \hspace{0.2cm}{\rm Mit}\hspace{0.15cm}a = 1\text{:} \hspace{0.2cm}\hspace{0.1cm}\underline {\sigma^2 = 2 }\hspace{0.05cm}.$$
  
  
[[File:P_ID2045__Dig_Z_4_9c.png|right|frame|Höhenlinien der 2D-Laplaceverteilung]]
+
[[File:P_ID2045__Dig_Z_4_9c.png|right|frame|Contour lines of the 2D Laplace distribution]]
'''(3)'''&nbsp; Richtig ist der <u>Lösungsvorschlag 1</u>:
+
'''(3)'''&nbsp; <u>Solution 1</u> is correct:
*Im ersten Quadranten ($x &#8805; 0, y &#8805; 0$) kann auf die Betragsbildung verzichtet werden. Dann gilt für die 2D&ndash;WDF:
+
*In the first quadrant ($x &#8805; 0, y &#8805; 0$), the magnitude formation can be omitted. Then the 2D PDF is given by:
 
:$$\boldsymbol{ p }_{\boldsymbol{ n }}  (x,\hspace{0.15cm} y) = {a^2}/{4} \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}x} \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}y }= {a^2}/{4} \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}(x+y)}\hspace{0.05cm}.$$
 
:$$\boldsymbol{ p }_{\boldsymbol{ n }}  (x,\hspace{0.15cm} y) = {a^2}/{4} \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}x} \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}y }= {a^2}/{4} \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}(x+y)}\hspace{0.05cm}.$$
*Eine Höhenlinie mit dem Faktor $\beta$ gegenüber dem Maximum hat dann den folgenden Verlauf ($0 < \beta < 1$):
+
*A contour line with factor $\beta$ versus maximum then has the following shape ($0 < \beta < 1$):
 
:$${\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}(x+y)} = \beta  
 
:$${\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}(x+y)} = \beta  
 
  \hspace{0.3cm} \Rightarrow \hspace{0.3cm} x + y = \frac{{\rm ln}\hspace{0.15cm}1/\beta}{a}
 
  \hspace{0.3cm} \Rightarrow \hspace{0.3cm} x + y = \frac{{\rm ln}\hspace{0.15cm}1/\beta}{a}
 
   \hspace{0.05cm}.$$
 
   \hspace{0.05cm}.$$
  
*Die Grafik zeigt die Höhenlinien für $a = 1$ und einige Werte von $\beta$, die jeweils ein um $45^\circ$ gedrehtes Quadrat ergeben &nbsp;&#8658;&nbsp; die Höhenlinien sind also Gerade..
+
*The graph shows the contour lines for $a = 1$ and some values of $\beta$, each giving a square rotated by $45^\circ$ &nbsp;&#8658;&nbsp; the contour lines are thus even.
  
  
  
  
'''(4)'''&nbsp; Das hier betrachtete Wahrscheinlichkeitsereignis entspricht genau dem dritten Quadranten der oben skizzierten Verbund&ndash;WDF. Aufgrund der Symmetrie ist diese Wahrscheinlichkeit:
+
'''(4)'''&nbsp; The probability event considered here corresponds exactly to the third quadrant of the composite PDF sketched above. Because of symmetry, this probability is:
 
:$${\rm Pr}[(n_1 < 0) ∩ (n_2 < 0)]\hspace{0.15cm}\underline {=25\%}.$$
 
:$${\rm Pr}[(n_1 < 0) ∩ (n_2 < 0)]\hspace{0.15cm}\underline {=25\%}.$$
  
  
  
'''(5)'''&nbsp; Dafür kann mit der Verbund&ndash;WDF geschrieben werden:
+
'''(5)'''&nbsp; For this, the composite PDF can be written:
 
:$${\rm Pr} \left [ (n_1 > 1)\cap (n_2 > 1)\right ] = {1}/{4} \cdot \int_{1}^{\infty} \int_{1}^{\infty}{\rm e}^{-  (x+y)} \,{\rm d} x  \,{\rm d} y == {1}/{2} \cdot \int_{1}^{\infty} {\rm e}^{-  x} \,{\rm d} x  \hspace{0.15cm} \cdot \hspace{0.15cm}
 
:$${\rm Pr} \left [ (n_1 > 1)\cap (n_2 > 1)\right ] = {1}/{4} \cdot \int_{1}^{\infty} \int_{1}^{\infty}{\rm e}^{-  (x+y)} \,{\rm d} x  \,{\rm d} y == {1}/{2} \cdot \int_{1}^{\infty} {\rm e}^{-  x} \,{\rm d} x  \hspace{0.15cm} \cdot \hspace{0.15cm}
 
   {1}/{2} \cdot \int_{1}^{\infty} {\rm e}^{-  y} \,{\rm d} y $$   
 
   {1}/{2} \cdot \int_{1}^{\infty} {\rm e}^{-  y} \,{\rm d} y $$   
 
:$$ \Rightarrow \hspace{0.3cm}{\rm Pr} \left [ (n_1 > 1)\cap (n_2 > 1)\right ] =  \left [ {\rm Pr} (n_1 > 1)\right ] \cdot \left [ {\rm Pr} (n_2 > 1)\right ]\hspace{0.05cm}. $$
 
:$$ \Rightarrow \hspace{0.3cm}{\rm Pr} \left [ (n_1 > 1)\cap (n_2 > 1)\right ] =  \left [ {\rm Pr} (n_1 > 1)\right ] \cdot \left [ {\rm Pr} (n_2 > 1)\right ]\hspace{0.05cm}. $$
  
Berücksichtigt ist die statistische Unabhängigkeit zwischen $n_1$ und $n_2$ sowie die Gleichheit $p_{\it n1}(x) = p_{\it n2}(y)$.  Für $a = 1$ gilt:
+
Considered is the statistical independence between $n_1$ and $n_2$ and the equality $p_{\it n1}(x) = p_{\it n2}(y)$.  For $a = 1$ holds:
 
:$${\rm Pr} (n_1 > 1) = {1}/{2} \cdot \int_{1}^{\infty} {\rm e}^{-  x} \,{\rm d} x = {1}/({2{\rm e}})\approx 0.184\hspace{0.3cm}
 
:$${\rm Pr} (n_1 > 1) = {1}/{2} \cdot \int_{1}^{\infty} {\rm e}^{-  x} \,{\rm d} x = {1}/({2{\rm e}})\approx 0.184\hspace{0.3cm}
 
\Rightarrow \hspace{0.3cm} {\rm Pr} \left [ (n_1 > 1)\cap (n_2 > 1)\right ] = {1}/({4{\rm e}^2)}\hspace{0.1cm}\hspace{0.15cm}\underline {\approx 3.4\%}\hspace{0.05cm}.$$
 
\Rightarrow \hspace{0.3cm} {\rm Pr} \left [ (n_1 > 1)\cap (n_2 > 1)\right ] = {1}/({4{\rm e}^2)}\hspace{0.1cm}\hspace{0.15cm}\underline {\approx 3.4\%}\hspace{0.05cm}.$$
  
  
[[File:P_ID2046__Dig_Z_4_9f.png|right|frame|Aufteilung des Integrationsbereichs]]
+
[[File:P_ID2046__Dig_Z_4_9f.png|right|frame|Division of the integration domain]]
'''(6)'''&nbsp; Die hier betrachtete Region ist in der folgenden Grafik farbig markiert.  
+
'''(6)'''&nbsp; The region considered here is highlighted in color in the following graph.
*Die Regionen erstrecken sich aber nach rechts und oben bis ins Unendliche.  
+
*However, the regions extend to the right and above to infinity.
*Die gesuchte Wahrscheinlichkeit ergibt sich zu
+
*The searched probability results in
  
 
:$${\rm Pr}  [ n_1 \hspace{-0.2cm} \ + \ \hspace{-0.2cm} n_2 > 2 ]  =\frac{1}{4} \cdot \int\limits_{-\infty}^{+\infty} {\rm e}^{-|x|} \int\limits_{2-x}^{\infty}{\rm e}^{-|y|} \,{\rm d} y  \,{\rm d} x = I_1 + I_2 + I_3 + I_4 \hspace{0.05cm}.$$
 
:$${\rm Pr}  [ n_1 \hspace{-0.2cm} \ + \ \hspace{-0.2cm} n_2 > 2 ]  =\frac{1}{4} \cdot \int\limits_{-\infty}^{+\infty} {\rm e}^{-|x|} \int\limits_{2-x}^{\infty}{\rm e}^{-|y|} \,{\rm d} y  \,{\rm d} x = I_1 + I_2 + I_3 + I_4 \hspace{0.05cm}.$$
  
Aufgrund der Betragsbildung ist eine Aufspaltung in Teilintegrale vorzunehmen. Nach oben und rechts erstrecken sich alle Gebiete bis ins Unendliche. Aufgrund der Symmetrie gilt $I_4 = I_3$.
+
Because of the magnitude, a splitting into partial integrals has to be done. Upwards and to the right all areas extend to infinity. Because of the symmetry $I_4 = I_3$ is valid.
 
:$$I_1 = {1}/{4} \cdot \int_{2}^{+\infty} \hspace{-0.15cm}{\rm e}^{-x} \int_{0}^{\infty}\hspace{-0.15cm}{\rm e}^{-y} \,{\rm d} y  \,{\rm d} x =  
 
:$$I_1 = {1}/{4} \cdot \int_{2}^{+\infty} \hspace{-0.15cm}{\rm e}^{-x} \int_{0}^{\infty}\hspace{-0.15cm}{\rm e}^{-y} \,{\rm d} y  \,{\rm d} x =  
 
  {1}/{4} \cdot \int_{2}^{+\infty} {\rm e}^{-x}  \,{\rm d} x ={1}/({4{\rm e}^2})\hspace{0.05cm},$$
 
  {1}/{4} \cdot \int_{2}^{+\infty} {\rm e}^{-x}  \,{\rm d} x ={1}/({4{\rm e}^2})\hspace{0.05cm},$$
Line 122: Line 122:
 
   {1}/({8{\rm e}^2}) = I_3\hspace{0.05cm}.$$
 
   {1}/({8{\rm e}^2}) = I_3\hspace{0.05cm}.$$
  
Insgesamt ergibt sich somit:
+
Thus, the overall result is:
 
:$${\rm Pr} \left [ n_1 + n_2 > 2 \right ] = {\rm e}^{-2} \cdot ({1}/{4} +{1}/{2} +{1}/{8} +{1}/{8})= {\rm e}^{-2} \hspace{0.1cm}\hspace{0.15cm}\underline {\approx 13.5\%}\hspace{0.05cm}.$$
 
:$${\rm Pr} \left [ n_1 + n_2 > 2 \right ] = {\rm e}^{-2} \cdot ({1}/{4} +{1}/{2} +{1}/{8} +{1}/{8})= {\rm e}^{-2} \hspace{0.1cm}\hspace{0.15cm}\underline {\approx 13.5\%}\hspace{0.05cm}.$$
 
{{ML-Fuß}}
 
{{ML-Fuß}}

Revision as of 09:57, 18 July 2022

2D Laplacian PDF

We consider two-dimensional noise  $\boldsymbol{n} = (n_1, n_2)$.

The two noise variables are "independent and identically distributed ", abbreviated i.i.d., and each has a Laplace probability density:

$$p_{n_1}(x) \hspace{-0.1cm} \ = \ \hspace{-0.1cm} K \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm} |x|} \hspace{0.05cm},$$
$$ p_{n_2}(y) \hspace{-0.1cm} \ = \ \hspace{-0.1cm} K \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm} |y|} \hspace{0.05cm}. $$
  • The 2D probability density function  $p_{\it \boldsymbol{n}}(x, y)$  is shown in the graph.
  • To simplify notation, the realizations of  $n_1$  and  $n_2$  are denoted here by  $x$  and  $y$,  respectively.


Notes:

  • The exercise belongs to the chapter  "Approximation of the Error Probability".
  • We would like to refer to the interactive applet  "2D Laplace"
  • The integral resulting in subtask (6) must be split into several partial integrals due to the magnitude formation.
  • Furthermore, it holds:
$$\int_{0}^{\infty} x^2 \cdot {\rm e}^{-a \hspace{0.03cm}\cdot \hspace{0.03cm} x} \,{\rm d} x = {2}/{a^3} \hspace{0.05cm}.$$


Questions

1

What is the size of the constant  $K$  of the 1D PDF?

$K = 1$.
$K = a/2$
$K = 1/a$.

2

Let  $a = 1$. What are the mean  ${\rm E}\big[n_i \big]$  and variance  $\sigma^2 = {\rm E}\big[n_i^2\big]$  of the two 1D random variables?  $(i = 1, 2)$

${\rm E}\big[n_i\big] \ = \ $

${\rm E}\big[n_i^2\big] \ = \ $

3

What is the shape of the contour lines of the 2D PDF in the first quadrant?

They are straight lines.
They are hyperbolas.
They are circles.

4

Let further  $a = 1$. What is the probability that both  $n_1$  and  $n_2$  are negative?

${\rm Pr}\big[(n_1 < 0) ∩ (n_2 < 0)\big]\ = \ $

$\ \%$

5

What is the probability that  $n_1$  and  $n_2$  are jointly greater than  $1$? 

${\rm Pr}\big[(n_1 > 1) ∩ (n_2 > 1)\big]\ = \ $

$\ \%$

6

What is the probability that the sum is  $n_1 + n_2 > 2$? 

$ {\rm Pr}\big[n_1 + n_2 > 2)\big] \ = \ $

$\ \%$


Solution

(1)  Solution 2 is correct:

  • The area under the PDF must equal $1$:
$$\int_{-\infty}^{+\infty} p_{n_1}(x) \,{\rm d} x = 1 \hspace{0.3cm}\Rightarrow \hspace{0.3cm} \int_{0}^{+\infty} p_{n_1}(x) \,{\rm d} x = 0.5 \hspace{0.3cm} \Rightarrow \hspace{0.3cm} K \cdot \int_{0}^{\infty} {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}x} \,{\rm d} x = - {K}/{a} \cdot \left [ {\rm e}^{- a \hspace{0.03cm} \cdot \hspace{0.03cm} x} \right ]_{0}^{\infty}= {K}/{a} = 0.5 \hspace{0.3cm}\Rightarrow \hspace{0.3cm} K = {a}/{2}\hspace{0.05cm}.$$


(2)  The linear mean is equal to 0 due to PDF symmetry.

Thus, the variance $\sigma^2$ is actually equal to the root mean square, as already stated in the question:

$$\sigma^2 = {\rm E}[n_1^2] = 2 \cdot \frac{a}{2} \cdot \int_{0}^{\infty} x^2 \cdot {\rm e}^{-a \hspace{0.03cm} \cdot \hspace{0.03cm} x} \,{\rm d} x = a \cdot {2}/{a^3}= {2}/{a^2} \hspace{0.05cm}. \hspace{0.2cm}{\rm Mit}\hspace{0.15cm}a = 1\text{:} \hspace{0.2cm}\hspace{0.1cm}\underline {\sigma^2 = 2 }\hspace{0.05cm}.$$


Contour lines of the 2D Laplace distribution

(3)  Solution 1 is correct:

  • In the first quadrant ($x ≥ 0, y ≥ 0$), the magnitude formation can be omitted. Then the 2D PDF is given by:
$$\boldsymbol{ p }_{\boldsymbol{ n }} (x,\hspace{0.15cm} y) = {a^2}/{4} \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}x} \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}y }= {a^2}/{4} \cdot {\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}(x+y)}\hspace{0.05cm}.$$
  • A contour line with factor $\beta$ versus maximum then has the following shape ($0 < \beta < 1$):
$${\rm e}^{- a \hspace{0.03cm}\cdot \hspace{0.03cm}(x+y)} = \beta \hspace{0.3cm} \Rightarrow \hspace{0.3cm} x + y = \frac{{\rm ln}\hspace{0.15cm}1/\beta}{a} \hspace{0.05cm}.$$
  • The graph shows the contour lines for $a = 1$ and some values of $\beta$, each giving a square rotated by $45^\circ$  ⇒  the contour lines are thus even.



(4)  The probability event considered here corresponds exactly to the third quadrant of the composite PDF sketched above. Because of symmetry, this probability is:

$${\rm Pr}[(n_1 < 0) ∩ (n_2 < 0)]\hspace{0.15cm}\underline {=25\%}.$$


(5)  For this, the composite PDF can be written:

$${\rm Pr} \left [ (n_1 > 1)\cap (n_2 > 1)\right ] = {1}/{4} \cdot \int_{1}^{\infty} \int_{1}^{\infty}{\rm e}^{- (x+y)} \,{\rm d} x \,{\rm d} y == {1}/{2} \cdot \int_{1}^{\infty} {\rm e}^{- x} \,{\rm d} x \hspace{0.15cm} \cdot \hspace{0.15cm} {1}/{2} \cdot \int_{1}^{\infty} {\rm e}^{- y} \,{\rm d} y $$
$$ \Rightarrow \hspace{0.3cm}{\rm Pr} \left [ (n_1 > 1)\cap (n_2 > 1)\right ] = \left [ {\rm Pr} (n_1 > 1)\right ] \cdot \left [ {\rm Pr} (n_2 > 1)\right ]\hspace{0.05cm}. $$

Considered is the statistical independence between $n_1$ and $n_2$ and the equality $p_{\it n1}(x) = p_{\it n2}(y)$. For $a = 1$ holds:

$${\rm Pr} (n_1 > 1) = {1}/{2} \cdot \int_{1}^{\infty} {\rm e}^{- x} \,{\rm d} x = {1}/({2{\rm e}})\approx 0.184\hspace{0.3cm} \Rightarrow \hspace{0.3cm} {\rm Pr} \left [ (n_1 > 1)\cap (n_2 > 1)\right ] = {1}/({4{\rm e}^2)}\hspace{0.1cm}\hspace{0.15cm}\underline {\approx 3.4\%}\hspace{0.05cm}.$$


Division of the integration domain

(6)  The region considered here is highlighted in color in the following graph.

  • However, the regions extend to the right and above to infinity.
  • The searched probability results in
$${\rm Pr} [ n_1 \hspace{-0.2cm} \ + \ \hspace{-0.2cm} n_2 > 2 ] =\frac{1}{4} \cdot \int\limits_{-\infty}^{+\infty} {\rm e}^{-|x|} \int\limits_{2-x}^{\infty}{\rm e}^{-|y|} \,{\rm d} y \,{\rm d} x = I_1 + I_2 + I_3 + I_4 \hspace{0.05cm}.$$

Because of the magnitude, a splitting into partial integrals has to be done. Upwards and to the right all areas extend to infinity. Because of the symmetry $I_4 = I_3$ is valid.

$$I_1 = {1}/{4} \cdot \int_{2}^{+\infty} \hspace{-0.15cm}{\rm e}^{-x} \int_{0}^{\infty}\hspace{-0.15cm}{\rm e}^{-y} \,{\rm d} y \,{\rm d} x = {1}/{4} \cdot \int_{2}^{+\infty} {\rm e}^{-x} \,{\rm d} x ={1}/({4{\rm e}^2})\hspace{0.05cm},$$
$$I_2 = {1}/{4} \cdot \hspace{-0.1cm} \int_{0}^{2} \hspace{-0.15cm}{\rm e}^{-x} \int_{2-x}^{\infty}\hspace{-0.15cm}{\rm e}^{-y} \,{\rm d} y \,{\rm d} x = {1}/{4} \cdot \hspace{-0.1cm} \int_{0}^{2} {\rm e}^{-x}\hspace{-0.1cm} \cdot {\rm e}^{x-2} \,{\rm d} x$$
$$\Rightarrow \hspace{0.3cm}I_2 = {1}/{4} \cdot \hspace{-0.1cm}\int_{0}^{2} {\rm e}^{-2} \,{\rm d} x = {1}/({2{\rm e}^2})\hspace{0.05cm},$$
$$I_3 \hspace{-0.1cm} \ = \ \hspace{-0.1cm} {1}/{4} \cdot \int_{-\infty}^{0} {\rm e}^{x} \int_{2-x}^{\infty}{\rm e}^{-y} \,{\rm d} y \,{\rm d} x = {1}/{4} \cdot \int_{-\infty}^{0} {\rm e}^{x} \cdot {\rm e}^{x-2} \,{\rm d} x = {1}/{4} \cdot \int_{-\infty}^{0} {\rm e}^{2x-2} \,{\rm d} x = \frac{{\rm e}^{-2}}{4} \cdot \int_{0}^{\infty} {\rm e}^{-2x} \,{\rm d} x = {1}/({8{\rm e}^2})\hspace{0.05cm},$$
$$I_4 ={1}/{4} \cdot \int_{-\infty}^{0} {\rm e}^{y} \int_{2-y}^{\infty}{\rm e}^{-x} \,{\rm d} x \,{\rm d} y = ... = {1}/({8{\rm e}^2}) = I_3\hspace{0.05cm}.$$

Thus, the overall result is:

$${\rm Pr} \left [ n_1 + n_2 > 2 \right ] = {\rm e}^{-2} \cdot ({1}/{4} +{1}/{2} +{1}/{8} +{1}/{8})= {\rm e}^{-2} \hspace{0.1cm}\hspace{0.15cm}\underline {\approx 13.5\%}\hspace{0.05cm}.$$