APPLICATIONS OF THE SPACE DIFFERENT IAL GEOMETRY

AT THE STUDY OF PRODUCTION FUNCTIONS

Catalin Angelo IOAN, Lecturer, Phd.

“DANUBIUS”University of Galati

Abstract:
This paper is a new onset about production functions. Because all papers on this subject use the projections
of production functions on a plan, the analysis becomes heavy and less general in conclusions, and for this reason
we made a treatment from the point of view of differential geometry in space.
On the other hand, we generalise the Cobb -Douglas, CES and Sato production functions to a unique form
and we made the analysis on this.
The conclusions of the paper allude to the principal directions of the surface (represented by the graph of
the production function) i.e. the directions in whic h the function varies the best. Also the concept of the total
curvature of a surface is applied here and we obtain that it is null in every point, that is all points are parabolic.
We compute also the surface element which is useful to finding all product ion (by means the integral) when
both labour and capital are variable

Key words: production function, differential geometry, curvature, principal direction,

Cobb-Douglas
JEL Classification: C65

1. INTRODUCERE 1. INTRODUCTION
Fie funcţia de producţie Q=Q(K,L) unde: Let a production function Q=Q(K,L)
 Q=producţia; where:
 K=capitalul;  Q=product;
 L=munca  K=capital;
FuncţiaQ:R+R+R+ trebuie să satisfacă  L=labour
condiţiile: The function Q:R+R+R+ must satisfiy
1. Q(0,0)=0; the conditions:
2. Q este diferenţiabilă de ordinul 2 în 6. Q(0,0)=0;
orice punct interior al domeniului de 7. Q is differentiable of order 2 in an y
definiţie; interior point of the production set;
Q Q Q Q
3.  0,  0; 8.  0,  0;
K L K L
 2Q  2Q  2Q  2Q
4.  0, 2  0 9.  0, 2  0
K 2
L K 2
L
5. Q este o funcţie omogenă d e grad 1, 10. Q is a homogenous function of
adică Q(tK,tL)=tQ(K,L) tR degree 1, that is Q(tK,tL)=tQ(K,L)
Semnificaţia primei condiţii este aceea că tR
la anularea unui factor, producţia devine nulă. The meaning of the first condition is that
A doua condiţie este utilă numai din at a vanishing of one factor the product is null.
considerente pur matematice. The second condition is useful just for
Condiţia a treia semnifică faptul că o mathematical calculus.
creştere a unuia dintr e factori (munca sau The third means that at an increase of one
capitalul) producţia creşte, de asemenea. factor (labour or capital) the product also grow.

30
 Cea de-a patra condiţie, deoarece derivata The fourth, because the second derivative
a doua reprezintă viteza de variaţie a primeia, is the speed of variation of the first, means that
semnifică faptul că producia are o viteză mai the product has a slower speed when one factor
mică atunci când unul din factori rămâne becomes constant and the other varies.
constant, iarcelălalt variază. The graph representation of a production
Reprezentarea grafică a unei funcţii de function is a surface.
producţie este o suprafaţă. Let:
Fie deci: Q Q  2Q  2Q  2Q
p= , q= , r= 2 , s= , t= .
Q Q  2Q  2Q  2Q L K L LK K 2
p= , q= , r= 2 , s= , t= .
L K L LK K 2 For a constant value of one parameter we
Pentru o valoare constantă a unuia dintre obtain a curve on the surface. For exemple:
parametri, obţinem o curbă pe o suprafaţă. De Q=Q(K,L0) or Q=Q(K 0,L) are both curves on the
exemplu: Q=Q(K,L 0) sau Q=Q(K 0,L) sunt production surface. They are obtained from the
amândouă curbe ale suprafeţei de producţie. intersection of the plane L=L 0 or K=K0 with the
Ele sunt obţinute din intersecţia planuli surface Q=Q(K,L).
L=L0 sau K=K 0 cu suprafaţa Q=Q(K,L). The curvature of a curve is from an
Curbura unei curbe este, dintr -un elementary point of view the degree of deviation
punctelementar de vedere, gradul de deviaţie al of the curve relative to a straight line.
acesteia de la o linie dreaptă. In the study of the surfaces, two quadratic
În studiul suprafeţelor, două forme forms are very useful.
pătratice sunt de un deosebit folos: The first fundamental quadratic form of
Prima formă fundamentală a suprafeţei the surface is:
este: g=EdL2+2FdLdK+GdK 2
g=EdL2+2FdLdK+GdK 2 where:
unde:  E=1+p2;
 E=1+p2;  F=pq;
 F=pq;  G=1+q2.
 G=1+q2. The area element is d= EG  F 2 dKdL
Elementul de arie este:
and the surface area A when (K,L) R (a region in
d= EG  F 2 dKdL, iar cel de suprafaţă A unde the plane K-O-L) is A=  ddKdL .
(K,L)R (o regiune din planul K-O-L) este R

A=  ddKdL . The second fundamental form of the

R surface is:
Forma a doua fundamentală a suprafeţei h=dL2+2dLdK+dK2
este: where:
h=dL2+2dLdK+dK2 r
unde:  = ;
1  p2  q 2
r
 = ; s
1  p2  q 2  = ;
1  p2  q 2
s
 = ; t
1  p2  q 2  = .
1  p2  q 2
t
 = . Considering the quantity =-2 we
1  p2  q 2 have that:
 If >0 in each point of the surface, we
Considerând cantitatea =-2 avem: will say that it is eliptical. Such surfaces
 Dacă >0 în fiecare punct al are the hyperboloid with two sheets, the
suprafeţei, vom spune că aceasta este eliptical paraboloid and the elypsoid.
eliptică. Astfel de suprafeţe sunt:
hiperboloidul cu două pânze,
paraboloidul eliptic şi elipsoidul;

31
  Dacă 0 în fiecare punct al
suprafeţei, vom spune că aceasta este
hiperbolică. Astfel de suprafeţe sunt:
hiperboloidul cu o pânză şi
hiperboloidul parabolic;
 Dacă =0 în fiecare punct al
suprafeţei, vom spune că aceasta este
parabolică. Astfel de suprafeţ e sunt
cele conice şi cele cilindrice.
Considerând o suprafaţă S şi o curbă
arbitrară printr-un punct P al suprafeţei ce are
vectorul tangent v în P, fie planul  determnat de
vectorul v şi nrmala N în P la S.Intersecţia lui 
cu S este o curbă C n numită secţiunea normală a
lui S. Curbura acesteia se num eşte curbură
normală.
Figure-1: The normal section of a curve
dL
Dacă avem o direcţie m= în planul
dK
tangent la suprafaţă în P, rezultă că formula
curburii normale este dată prin:
m 2  2m  
k(m)=
Em 2  2Fm  G
Valorile extreme k 1 şi k2 ale funcţiei k(m)
se numesc curburi principale ale suprafeţei în acel
punct. Ele satisfac, de asemenea, ecuaţia:
(EG-F2)k2-(E-2F+G)k+(-2)=0
Valorile lui m ce furnizează aceste
extreme se nmesc direcţii principale în acel punct.
Ele satisfac, de asemenea, ecuaţia:
(E-F)m2+(E-G)m+(F-G)=0
32
 If 0 in each point of the surface, we
will say that it is hyperbolic. Such
surfaces are the the hyperbolid with one
sheet and the hyperbolic paraboloid.
 If =0 in each point of the surface, we
will say that it is parabolic. Such surfaces
are the cone surfaces and the cylinder
surfaces.
Considering a surface S and an arbitrary
curve through a point P of the surface who has
the tangent vector v in P, let the pl ane 
determined by the vector v and the normal N in P
at S. The intersection of  with S is a curve C n
named normal section of S. Its curvature is called
normal curvature.
dL
If we have a direction m= in the
dK
tangent plane of the surface in an arbitrary point P
we have that the normal curvature is given by:
m 2  2m  
k(m)=
Em 2  2Fm  G
The extreme values k 1 and k2 of the
function k(m) call the principal curvatures of the
surface in that point. They satisfy also the equation:
(EG-F2)k2-(E-2F+G)k+(-2)=0
The values of m who give the extremes
call principal directions in that point.
They also satisfy the equation:
(E-F)m2+(E-G)m+(F-G)=0
sau or
(Es-Fr)m2+(Et-Gr)m+(Ft-Gs)=0 (Es-Fr)m2+(Et-Gr)m+(Ft-Gs)=0
dL dL
Curba =m (unde m este una din The curve =m (where m is one of the
dK dK
direcţiile principale) se numeşte linie de curbură a principal directions) is called line of curvature on
suprafeţei. Pe o astfel de curbă avem maximul sau the surface. On such a curve we have the
minimul variaţiei lui Q într -o vecinătate a lui P. maximum or minimum variation of the value of
Cantitatea K=k 1k2 se numeşte curbură Q in a neighbourhood of P.
k  k2 The quantity K=k 1k2 is named the total
totală în punctul considerat, iar H= 1 se
2 k  k2
curvature in the considered point and H= 1
numeşte curbura medie a suprafeţei în acel punct. 2
Avem, de asemenea; is named the mean curvature of the surface in that
   2 E  2F  G point.
K= şi H= We have therefore:
EG  F 2
EG  F2
O suprafaţă cu K=constant se numeşte    2 E  2F  G
K= and H=
suprafaţă de curbură totală constan tă, iar dacă EG  F 2
EG  F2
H=0 vom spune că avem o surafaţă minimală. A surface with K=constant call surface
Considerând acum planul tangent  la with constant total curvature and if H=0 call
suprafaţă într-un punct P şi o direcţie m, dacă minimal surface.
m2+2m+=0 vom spune că m este direcţie Considering now in the tangent plane  at
 dL 
2
dL the surface in a point P a direction m, if
asimptotică, iar ecuaţia:    2 0 m2+2m+=0 we will say that m is an
 dK  dK
asymptotic direction, and the equation:
va furniza curbele asimp totice ale suprafeţei în 2
 dL  dL
punctul P.    2    0 gives the asymptotic
 dK  dK
2. FUNCŢIA GENERALĂ DE PRODUCŢIE curves of the surface in the point P.
Fie funcţia de producţie:
K  L 2. THE GENERAL PRODUCTION
Q=A , ,,[0,1], R, +0 FUNCTION
( K   L ) 
Let the production function:
 Pentru =0, ,,=arbitrari,
K  L
,[0,1] avem funcţia Cobb - Q=A , ,,[0,1], R, +0
Douglas: Q=AK L; ( K   L ) 
1  For =0, ,,=arbitrary, ,[0,1]
 Pentru =0, =0, =- avem we have the Cobb-Douglas function:

1
Q=AKL;

funcţia CES: Q=A ( K   L ) ;  For =0, =0, =-
1
we have the
 Pentru =2, =2, =3 şi =1 avem 
1
K 2 L2 
funcţia SATO: Q=A . CES function: Q=A ( K   L ) ;
K 3  L3
 For =2, =2, =3 and =1 we have
 Pentru a avea o funcţie omogenă de
grad 1, avem că: Q(tK,tL)=tQ(K,L) K 2 L2
the SATO function: Q=A .
tR K 3  L3
Avem, de asemenea:  In order to have a homogenous
K  L function of degree 1, we have that:
Q(tK,tL)=A t  = t  Q(K,L) Q(tK,tL)=tQ(K,L) tR
( K   L ) 
We have therefore:
 +-=1.
K  L
Q(tK,tL)= A t  = t  Q(K,L)
( K   L )
  

 +-=1.

33
    1    1
În consecinţă: = şi expresia In consequence: = and the
 
generală a lui Q va fi: general expression of Q will be:
K  L K  L
Q=A  1
, ,,[0,1], +0 Q=A  1
, ,,[0,1], +0
( K   L )
  
( K   L )
  

Avem acum: We have now:

Q  (  ) K   L Q  (  ) K   L
-1  -1 
=AK L ( K   L ) 1 . =AK L ( K   L ) 1 .
K K
Deoarece AK  L
Because ( K   L )  = we obtain:
AK  L Q
( K   L )  = obţinem:
Q Q (1  ) K   L
q= =Q
Q (1  ) K   L K K ( K   L )
q= =Q
K K ( K   L ) Through analogy:
Prin analogie: Q (1   )L   K 
p= =Q
Q (1  )L  K  L L( K   L )
p= =Q
L L( K   L ) With the upper relations we have now:
Din relaţiile de mai sus avem acum:  2Q Q
t= =- 2 [(1-)2L2+[(-
 2Q Q K 2
K ( K   L ) 2
t= =- 2 [(1-)2L2+[(-
K 2
K ( K   L ) 2 1)(+-1)+2]KL+(1-)2K2]
1)(+-1)+2]KL+(1-)2K2]  2Q Q
r= 2 =- 2 [(1-)2L2+[(-
 2Q Q L L ( K   L )
  2
r= 2 =- 2 [(1-)2L2+[(-
L L ( K  L )
  2
1)(+-1)+2]KL+(1-)2K2]
1)(+-1)+2]KL+(1-)2K2]  2Q Q
s= = [(1-)2L2+[(-
 2Q Q KL KL( K   L ) 2
s= = [(1-)2L2+[(-
KL KL( K   L ) 2 1)(+-1)+2]KL+(1-)2K2]
1)(+-1)+2]KL+(1-)2K2] Let note now:
Să notăm acum: P=(1-)2L2+[(-1)(+-1)+2]KL+(1-
P=(1-)2L2+[(-1)(+-1)+2]KL+(1- )2K2
)2K2 U=(1-)L+K
U=(1-)L+K V=L+(1-)K
V=L+(1-)K from where:
de unde: U+V=L+K.
U+V=L+K. If +-10 we have:
Dacă +-10 avem: (1   ) V  U (1  ) U   V
K = and L= .
 (1   ) V  U (1  ) U   V (1    )  (1    )
K= şi L= .
(1    )  (1    )

Avem acum: We have now:

Q QU Q QU
p= = ; p= = ;
L L( U  V ) L L( U  V )
Q QV Q QV
q= = ; q= = ;
K K ( U  V ) K K ( U  V )
U2 U2
2 2 2 2
E=1+p =1+Q L2 ( U  V ) 2 ; E=1+p =1+Q L2 ( U  V ) 2 ;
UV UV
F=pq=Q2 KL( U  V ) 2 ; F=pq=Q2 KL( U  V ) 2 ;

34
 V2 V2
2 2 . 2 2 .
G=1+q =1+Q K ( U  V ) 2 2 G=1+q =1+Q K ( U  V ) 2 2

K 2 U 2  L2 V 2 K 2 U 2  L2 V 2
Cu =1+p2+q2=1+Q2 With =1+p2+q2=1+Q2
K 2 L2 ( U  V ) 2 K 2 L2 ( U  V ) 2
avem: we have:
r s t r s t
= , = , = . = , = , = .
     
QP QP QP QP
t=- 2 , r=- 2 , t=- 2 , r=- 2 ,
K ( U  V) 2
L (U  V) 2 K ( U  V) 2
L ( U  V) 2
QP QP
s= . s= .
KL( U  V ) 2 KL( U  V ) 2
După calcule simple, avem EG - After an easy computing we have EG -
2 Q 2 ( K 2 U 2  L2 V 2 ) 2 Q 2 ( K 2 U 2  L2 V 2 )
F =1+ F =1+
K 2 L2 ( U  V ) 2 K 2 L2 ( U  V ) 2
de unde: from where:
K 2 L2 ( U  V ) 2  Q 2 ( K 2 U 2  L2 V 2 ) K 2 L2 ( U  V ) 2  Q 2 ( K 2 U 2  L2 V 2 )
d= dKdL d= dKdL
KL( U  V ) KL( U  V )
şi suprafaţa va fi calculată prin: and the surface area will be compute by:
K 2 L2 ( U  V ) 2  Q 2 ( K 2 U 2  L2 V 2 ) K 2 L2 ( U  V ) 2  Q 2 ( K 2 U 2  L2 V 2 )
A=  dKdL A=  dKdL
R KL( U  V ) R KL( U  V )
Direcţiile principale vor fi date de: The principal directions will be given by:
K2[L2(U+V)2+Q2U2+Q2UV]m2+KL[-L2(U+V)2- K2[L2(U+V)2+Q2U2+Q2UV]m2+KL[-L2(U+V)2-
Q2U2+K2(U+V)2+Q2V2]m- Q2U2+K2(U+V)2+Q2V2]m-
L2[K2(U+V)2+Q2V2+Q2UV]=0 L2[K2(U+V)2+Q2V2+Q2UV]=0
de unde: from where:
L L ( U  V ) K 2  VQ 2 L L ( U  V ) K 2  VQ 2
m1= , m2=- . m1= , m2=- .
K K ( U  V ) L2  UQ 2 K K ( U  V ) L2  UQ 2
Pentru o direcţie m avem: For a direction m we have:
m 2  2m   m 2  2m  
k(m)= = k(m)= =
Em 2  2Fm  G Em 2  2Fm  G
QPLK ( U  V ) QPLK ( U  V )
 
K L ( U  V ) 2  Q 2 ( K 2 U 2  L2 V 2 )
2 2
K L ( U  V ) 2  Q 2 ( K 2 U 2  L2 V 2 )
2 2

 K 2 m 2  2KLm  L2  K 2 m 2  2KLm  L2
K [ L ( U  V )  Q U ]m 2  2KLQ 2 UVm  L2 [ K 2 ( U  V ) 2  Q 2 V 2 ]
2 2 2 2 2
K [L ( U  V)  Q U ]m 2  2KLQ2 UVm  L2 [K 2 ( U  V) 2  Q 2 V 2 ]
2 2 2 2 2

L
For m1= we have that k 1=k(m1)=0 and for
L K
Pentru m1= avem că k 1=k(m1)=0 şi pentru
K L ( U  V ) K 2  VQ 2
m2=- we have k 2=k(m2)=
L ( U  V ) K 2  VQ 2 K ( U  V ) L2  UQ 2
m2=- avem k2=k(m2)=
K ( U  V ) L2  UQ 2 QPLK ( U  V )
- 
QPLK( U  V) K L ( U  V ) 2  Q 2 ( K 2 U 2  L2 V 2 )
2 2
- 
K L ( U  V) 2  Q 2 (K 2 U 2  L2 V 2 )
2 2

35
 [( U  V ) K 2  VQ 2 ]2  [( U  V ) K 2  VQ 2 ][( U  V ) L2  UQ 2 ]  [( U  V ) L2  UQ 2 ]2
[ L ( U  V ) 2  Q 2 U 2 ][( U  V ) K 2  VQ 2 ]2  2Q 2 UV[( U  V ) K 2  VQ 2 ][( U  V ) L2  UQ 2 ] 
2

[ K 2 ( U  V ) 2  Q 2 V 2 ][( U  V ) L2  UQ 2 ]2

Curbura totală a suprafeţei este: K=k 1k2=0. The total curvature of the surface is K=k 1k2=0.
Curbura medie este, de asemene a: The mean curvature is also:
E  2F  G E  2F  G
H= =- H= =-
2( EG  F 2 ) 2( EG  F 2 )
KLQP( U  V )L2  K 2  Q 2  KLQP( U  V )L2  K 2  Q 2 
3 3
2 K 2 L2 ( U  V ) 2  Q 2 [ K 2 U 2  L2 V 2 ] 2 K 2 L2 ( U  V ) 2  Q 2 [ K 2 U 2  L2 V 2 ]
Obţinem că suprafaţa de producţie este cu We obtain that the production surface is
curbură totală nulă, dar u este minimală în niciun with null total curvature but it is not minimal in
punct. any point.
Linia e curbură este: The line of curvature equation is:
2 2
 dL   dL   dL   dL 
(E-F)   +(E-G)   +(F-G)=0 (E-F)   +(E-G)   +(F-G)=0
 dK   dK   dK   dK 
Ca mai sus, obţinem uşor că: Like at upper, we obtain easy that:
dL L dL dK dL L dL dK
= că este:   L=CK cu = that is:   L=CK with
dK K L K dK K L K
C(0,) C(0,)
respectiv: respectively:
dL dL
=- =-
dK dK
L L
K K
2 (  1) 2 (  1)

(L  K  )K 2 ( K   L ) 
 (L  (1  ) K  )A 2 K 2  L2 (L  K  ) K 2 ( K   L ) 
 (L  (1  ) K  ) A 2 K 2  L2
2 (  1) 2 (  1)

(L  K  )L2 ( K   L )
 
 ((1  )L  K  )A 2 K 2  L2 (L  K  ) L2 ( K   L )
 
 ((1   )L   K  ) A 2 K 2  L2

Direcţiile asimptotice satisfac: The asymptotic directions satisfy:

m2+2m+=0 m2+2m+=0
adică: that is:
rm2+2sm+t=0 rm2+2sm+t=0
de unde: from where:
L L
-K2m2+2KLm-L2=0 deci m 1=m2= . -K2m2+2KLm-L2=0 therefore m 1=m2= .
K K
Curbele asimptotice au ecuaţia: The asymptotic curves have the equation:
dL dL
=m (cu m – direcţie asimptotică) deci ele =m (with m asymptotic direction) therefore
dK dK
sunt: L=CK cu C(0,). they are: L=CK with C(0,).

3. APLICAŢII PENTRU FUNCŢIA COBB - 3. APPLICATIONS FOR THE COBB -

DOUGLAS DOUGLAS FUNCTION
Pentru funcţia de pr oducţie Cobb- For the Cobb-Douglas production
Douglas, adică pentru +=1, =1, =0, =1 function, that is for +=1, =1, =0, =1 we
avem: have:
U=K U=K
V=K V=K
U+V=K U+V=K
P=K2 P=K2
36
 L L K 3  KA 2 K 2  L2 L L K 3  KA 2 K 2  L2
m1 = , m2=- = m1 = , m2=- =-
K K KL2   KA 2 K 2  L2 K K KL2   KA 2 K 2  L2
K  A 2 K 2  1L2 K  A 2 K 2  1L2
-
L   A 2 K 2  L21 L   A 2 K 2  L21
K K
şi notând cu g= obţinem: and denoting with g= the endowment with
L L
1 1   A 2 g 2  capital we obtain:
m1= , m2=- g .
g 1  A 2g 2 1 1  A 2 g 2
m1= , m2=- g .
k1=0 g 1  A 2g 2
Q KL k1=0
k2=- 
K L  Q 2 ( 2 K 2   2 L2 ) Q KL

2 2
k2=-
[K 2  Q2 ]2  [K 2  Q2 ][L2  Q2 ]  [L2  Q2 ]2 K L  Q 2 ( 2 K 2   2 L2 )
2 2

[L2  Q22 ][K 2  Q2 ]2  2Q2[K 2  Q2 ][L2  Q2 ]  [K 2  Q2 2 ][L2  Q2 ]2 [K 2  Q 2 ]2  [K 2  Q 2 ][L2  Q 2 ]  [L2  Q 2 ]2
[L2  Q 22 ][K 2  Q 2 ]2  2Q 2[K 2  Q 2 ][L2  Q 2 ]  [K 2  Q 2 2 ][L2  Q 2 ]2
Curbura totală a suprafeţei este:
K=k1k2=0 şi curbura medie: The total curvature of the surface is
 LQL2  K 2  Q 2  K=k1k2=0 and the mean curvature is:
 LQL2  K 2  Q 2 
H=- 3
.
2 K 2 L2  Q 2 [ 2 K 2   2 L2 ] H=- 3
.
2 K 2 L2  Q 2 [ 2 K 2   2 L2 ]

Selective bibliography:

1. Ioan Catalin Angelo (2004), Applied mathematics in economy, EDP, Bucharest

2. Nerlove M. (1965) Estimation and Identification of Cobb -Douglas Production Functions. Skokie, Ill:
Rand-McNally.
3. Nerlove M. (1967) "Recent Empirical Studies of the CES and Related Production Functions", in M.
Brown, editor, Theory and Empirical Analysis of Production. New York: Columbia University Press.

37
38