QUANTITATIVE METHODS for ECONOMIC APPLICATIONS TASK 20/4/2020
I M 1) After having determined the complex number which is the solution of the equationD
D D
" 3 " 3 œ #3, calculate its cubic roots $ D.
D D " " " 3 " 3
" 3 " 3 œ D " 3 " 3 œ D " 3 " 3 œ D # œ #3 Ê D œ #3
# .
Since #3 œ # cos 3sin we getÀ
# #
1 1
$ D œ$ #cos1 1 sin1 1
' 5 #$ 3 ' 5 #$ ß ! Ÿ 5 Ÿ # .
For 5 œ ! D œ # # $ 3" à
# #
we get ! $ cos sin $
1 1
' 3 ' œ
for 5 œ " D œ # # $ 3" à
# #
we get " $ cos sin $
& & '1 3 '1 œ
for 5 œ # we get D œ# $ #cos$ sin$ $ # 3 Þ
#1 3 #1 œ
I M 2) Given the matrix œ , since œ " is an eigenvalue of , find
" " "
! 5 5 "
! ! 5
# -
the value of the real parameter , verify that is diagonalizable and then find a matrix that5 diagonalizes .
From we get 1
- ˆ œ ! œ
" " "
! 5 5 "
! ! 5
5 5
œ !
-
-
-
- - -
#
#
For - œ " we get 1 "5 "5# " œ ! Ê 5 " œ !Ê 5 œ ". So the matrix is œ
" " "
! " #
! ! "
. We find all the eigenvalues:
T$ - œ1- "-"-œ !Ê-" œ-# œ "ß-$ œ ". For 5 œ " we study the matrix " œ
! " "
! #
! ! !
and - œ " # and since
ˆ
Rank "ˆœ " Ê 7 œ $ " œ # œ 71" #" and so the matrix is diagonalizable.
Now we go to construct the modal matrix. Firstly we search for two linearly independent eigenvectors corresponding to the eigenvalue - œ ". We must solve the system:
"ˆ †—œ Ê † œ Ê Ê
! " " B !
! # C !
! ! ! D !
C D œ ! a B
a B D œ C
# .
So eigenvectors corresponding to the eigenvalue - œ " are •œ Bß Cß C .
Two linearly independent eigenvectors may be •" œ "ß !ß ! and •# œ !ß "ß " .
Now we search for an eigenvector corresponding to the eigenvalue - œ " solving the sy- stem:
"ˆ †—œÊ † œ Ê Ê
# " " B !
! # C !
! ! # D !
#B C D œ !
#D œ !
!
Ê C œ #B #B
D œ !
. Eigenvectors corresponding to - œ " are •œ Bß ß ! and one of them may be •$ œ "ß #ß ! . Finally we construct the modal matrix:
œ œ ƒ
" ! "
! " #
! " !
. And we can see that † † Ê
Ê
" " " " ! " " ! " " ! !
! " # ! " # ! " # ! " !
! ! " ! " ! ! " ! ! ! "
† œ † .
I M 3) Since the linear system has the solution , calculate
B B B œ !
7B B B œ ! B B 5B œ !
"ß !ß "
" # $
" # $
" # $
the values of the real parameters and and then find a basis for the space of the solutions7 5 of the system.
Substituting we get:
" ! " œ ! 7 ! " œ !
" ! 5 œ !
Ê 7 œ "
5 œ "
and we get the system:
B B B œ ! B B B œ ! B B B œ !
B B B œ ! B œ B B
" # $
" # $
" # $
" # $ $ " #
Ê Ê .
All the solutions of the system are in the form B ß B ß B B" # " #.
A basis for the space of the solutions of the system may be "ß !ß " à !ß "ß " .
I M 4) Determine the vectors parallel to the vector — œ "ß "ß # and with modulus equal to '.
A vector parallel to the vector — œ "ß "ß # is a vector of the type:
˜ œ 5 † "ß "ß # œ 5ß 5ß #5 . And so we get:
˜ œ 5 5 %5 œ# # # '5 œ# ' Ê 5 œ " Ê 5 œ „"# . So we get two vectors cor- responding to the request: ˜" œ "ß "ß # and ˜# œ "ß "ß # .
II M 1) Given the equation 0 Bß C œ / B C# # /BC œ ! satisfied at the point !ß ! , verify that an implicit function B Ä C B can be defined with it, and then calculate the first order derivative of this function.
The function 0 Bß C is a differentiable function a Bß C − ‘#. It also turns out:
f0 B ß Cœ #B /B C# # /BCà #C /B C# # /BC for which f0 !ß ! œ "à ". Since 0Cw !ß ! œ " Á ! it is possible to define an implicit function B Ä C B . For its derivative we have: C ! œ " œ " Þ
"
w
II M 2) Solve the problem Max min . s v
Î 0 Bß C œ B #C $D Þ Þ B C D œ "%# # #
The objective function of the problem is a continuous function, the constraint defines a feasi- ble region which is a compact set since it consists of only boundary points (the surface of aX sphere) and therefore surely maximum and minimum values exist.
For a problem with equality constraints we construct the Lagrangian function and apply to it the first and second order conditions. We have:
ABß Cß Dß-œ B #C $D-B C D "%# # # . Applying the first order conditions we have:
f Ê Ê Ê
œ " # B œ ! œ # # C œ ! œ $ # œ !
œ œ œ
A -
A -
A -
A -
Bß Cß Dß œ
D B C D œ "%
B C D
œ "%
wB wC wD
# # #
"
# "
$
" # " *
% %
- - -
-# -# -#
Ê Ê Ê ∪
œ œ œ
œ œ œ
œ " œ "
œ # œ #
œ $ œ $
B C D
œ "% œ
B C D
B B
C C
D D
œ œ
Þ
"
# "
$
"% # "
%
"
# "
$
# " "
- - - -
- - -
# - - -
# % # #
We have only two stationary points: T œ "ß #ß $" and T œ "ß #ß $# .
By Weierstrass' theorem, since 0 "ß #ß $ œ "% and 0 "ß #ß $ œ "% it obviously turns out that is the maximum point while is the minimum point.T" T#
If we want to apply the second order conditions we have to built the bordered Hessian matrix:
‡
Bß Cß Dß œ
! #B #C #D
#B # ! !
#C ! # !
#D ! ! #
- -
-
-
to calculate then, in the points and , theT" T#
minors ‡$ and ‡%. Since ‡
T" œ À
! # % '
# " ! !
% ! " !
' ! ! "
it is
‡$ T" œ #! ! and ‡% T" œ &' ! and so is the maximum point;T"
Since ‡
T# œ À
! # % '
# " ! !
% ! " !
' ! ! "
it is
‡$ T# œ #! ! and ‡% T# œ &' ! and so is the minimum point.T#
II M 3) Given 0 Bß C œ B C and 1 Bß C œ B C # # determine for which values of parameter α it is W@0 "ß " œ W@1 "ß " , with @œcosαßsinα.
The functions 0 Bß C œ B C and 1 Bß C œ B C # # are polynomials and therefore are diffe- rentiable in any order a Bß C − ‘#.
So H 0@ "ß " œ f0 "ß " † @ and H 1@ "ß " œ f1 "ß " † @. Since:
f0 B ß C œ Cà BÊf0 "ß " œ "ß " and f1 B ß C œ #Bà #CÊf1 "ß " œ #à # , to get W@0 "ß " œ W@1 "ß " it will be:
"ß " cosαßsinαœ # à # cosαßsinαÊcosαsinαœ ! Êcosαœ sin and so:α αœ $1 αœ (1
% or % .
II M 4) Given the function 0 Bß Cß D œ B CD # $ logC B / CD, determine the gradient vector of the function at T œ "ß #ß #! .
It is f BCD " à B D " / à $B CD /
C B C B
0 Bß Cß D œ # $ # $ CD # # CD and so:
f "ß #ß # " à ) " / à #% /
# " # "
0 œ $# ## ## œ $$à )à #$.