Multivariate calculus

Definition
• A relation f that relates one output variable to
several input variables : is called a
multivariate function if for each ( )  exactly
one output f( ).
 Examples
• f()= 3+4-5

• =4

• g(x)=
 Definition of limit
• Let f(x, y): R2 → R be a function. The number
L is said to be the limit of the function f at
(a,b) , that is :

• If ∧),(x, y)≠(a,b) (x≠a, y≠b) (whenever 0<<δ

,Іf(x)-LІ<є.
 Definition of continuity
• A function f(x,y): R2 →R is continuous at (a,b) if

• Polynomials are continuous everywhere .

• Rational functions are continuous everywhere
they are defined
 Partial derivatives
• Let f:R2R. The partial derivative of f
evaluated at (a, b) are:
• Partial derivatives with respect to x:
• (a, b)=
• Partial derivative with respect to y:
• (a, b)=
 Clairaut’s Theorem
• Let f: R2 R be a function. the partial
derivatives , , and exists at (a,b) and are
continuous at (a,b) then :
• (a,b)= (a,b)

• If mixed partial derivatives are continuous

then they are equal.
 Total derivative
• The total derivative of a function f at a point is the best
linear approximation near this point of the function with
respect to its arguments. Unlike partial derivatives, the total
derivative approximates the function with respect to all of its
arguments, not just a single one. In many situations, this is the
same as considering all partial derivatives simultaneously.
 Chain rule
• Suppose z=f(x, y) , f is differentiable , x=g(t),
and y=h(t). Assuming all relevant derivatives
exist :
 Total derivative

• dz= dy
 Implicit function theorem
• f(x, y)= 2y+3x+7
• f(x,y)=
• f(x,y)=
 Implicit function theorem
• Given an implicit equation G(x,y)=c and a
specific solution ( such that G()=c we want to
know the answer to the following questions:
• (1) Does there exist a continuous function
y=f(x) x in an interval I around so that
F(x,y)=cx∊I and =y()?
• (2)If f(x) exists and is differentiable , what is
f’(x) at x=?
 Implicit function theorem
• Let G(x,y) : R2 →R be a function such that
• G(x0 ,y0 )-c=0 and consider the expression
G(x,y)-c=0 . If δG(x0 ,y0 )/δy≠0, then ꓱa
function y=f(x):R→R defined in an open
interval I about x0 such that :
• (1) G(x,f(x))-c=0ⱯxєI
• (2) f(x0 )= y0
• (3)= -
 Implicit function extended to several
variables
• Let F(y)=c be an equation corresponding to the function F: →R.
Let (,……..,, y) satisfy :
• F(y)=c &
• ≠0
• Then there is a function y=f() defined on an open ball B around
(,……..,) such that :
• (1) F( f() )=c () ∊B
• (2) =f(,……..,)
• (3) = -i=1,2,3,…….,k
Interpretation of implicit function theorem

• When the set of points defined by the

function G(x,y)= c can be considered as the
graph of a function y=f(x)
 Theorem
• Let ) be a solution to the equation G(x,y)=c G:
R. If ) then G(x,y)=c defines a smooth curve
around ) which can be considered the graph of
a function y=f(x) , with slope :
• .
 Home work
• Write down the implicit function theorem for
5 variables (, y)
Definition of an optimum for a multivariate
function
• Let f:→R f is said to have an absolute
maximum on if ∊such that
)≤f(.
• A real valued function f defined on is said to
have a relative maximum at a point
• If there is some open ball
• Such that f≤) ∊
 Home work
• Define the relative minimum and absolute
minimum for f:→R
First order condition for local extremum
• at ()
 Quadratic form
• A quadratic form on is a real valued function
of the form:
• Q(
 Hessian matrix
• Hessian matrix or Hessian is a square matrix
of second-order partial derivatives of a scalar-
valued function.
 Necessary & sufficient condition for local
maximum
• Maximize :Z=f(x,y)
• First order condition :
• at (
• 2nd order condition : 0⇒

• Alternatively:
• >0
 Necessary & sufficient condition for local
minimum
• Minimize :Z=f(x,y)
• First order condition :
• at (
• 2nd order condition : ⇒

• Alternatively:
• >0
Saddle point
 Saddle point
• A differentiable function f(x, y) has a saddle point at a
critical point (a, b) if in every open disk centered at
(a, b) there are domain points (x, y) where f(x, y) >
f(a, b) and domain points where f(x, y) < f(a, b). The
corresponding point (a, b, f(a, b)) on the surface z =
f(x, y) is called a saddle point of the surface
• A saddle point is identified by: fx =0 fy =0 ,
 Classification of critical points
• Critical points are identified by : fx =0; fy =0
• 1. a local maximum
• 2. A local minimum
• 3. saddle point
• 4. A point of inflexion
 Three variable case
• z=f(

• dz=

• z=

q=
 Example
• Find the relative extremum of the function :
 Multi-variable case Concavity
Definition 1
• Definition : Concavity:
• A function f:DnR DnRn is concave iff
• (i)Dn is a convex set
• (ii) for every , Dn, and for any , 0 <<1 ,
• f()+(1-)f()f[+(1-)]
 Multi-variable case Convexity
Definition 1
• Definition : Convexity:
• A function f:DnR DnRn is convex iff
• (i)Dn is a convex set
• (ii) for every , Dn, and for any , 0 <<1 ,
• f()+(1-)f()≥f[+(1-)]
 Convex set
• A set is a convex set if the line segment
between any two points in C lies in C.

• If x,yC ,and any [0,1] , x+(1-)yC

then C is a convex set.
Convex and non convex sets
 Definition 2
• A function is convex if its epigraph is a convex
set.

• A function is concave if its sub graph is a

convex set.
 Sub graph & Epigraph of a convex set
• A concave/convex function can also be
defined through convex sets.
• Sub graph of a function f :sub(f)
• sub(f) = {(x,y) ϵ S XR І f(x)≥ y}
• epigraph of f:
• epi(f):{(x,y) ϵ S XR І f(x)≤ y}
Sub graph ;
f(x)≥y
Epigraph f(x)≤y
 Definition 3
• If the set ={x⃓ f(x)≤ k} is convex for any
constant k then f(x) is a convex function.

• If the set = {x⃓ g(x)≥k} is convex for any

constant k then g(x) is a concave function.
Differentiable function-Single variable
• A differentiable function f(x) is convex iff for
any given points (u,v) in the domain of the
function u≠v
• f(v)≥f(u)+f’(u)(v-u)
• A differentiable function f(x) is concave iff for
any given points (u, v) in the domain of the
function u≠v
• f(v)≤f(u)+f’(u)(v-u)
 Differentiable function- multi variable
• A differentiable function is concave iff for any
two distinct given points ( in the domain of
the function :
• f(v)≤f(u)+
• A differentiable function is convex iff for any
two distinct given points ( in the domain of
the function :
• f(v)≥f(u)+
Twice differentiable functions: Convexity

• Let ⊂ n be an open convex set and let f : → R

be twice continuously differentiable in D. Then f
is convex if and only if d2f0( positive semi
definite)  Dn .
• f is strictly convex if (but not only if) d2f>0
• The condition for positive definiteness is that
the Hessian matrix of f is positive semi-definite
throughout :
• H1>0, H2>0, ……..Hn>0
 Twice differentiable functions: Concavity

• Let Dn Rn be an open convex set and let f : Dn →

R be twice continuously differentiable in D. Then f
is concave if and only if d2f0(negative semi
definite ) Dn.
• f is strictly concave if (but not only if) d2f<0.
• The condition for negative definiteness is that the
Hessian matrix of f alternates in sign through out :
• H1<0, H2>0, ……..(-1)n Hn>0 if n is even,
……..(-1)n Hn<0 if n is odd.
 Home work
• Create the same diagram given in Chiang for
minimization.
 Local vs global optimum
• 1: If f() is concave everywhere in its domain
and * is a local maximum then * is also a global
maximum.
– If f() is strictly concave the global maximum is also
unique.
• 2:If f() is convex everywhere in its domain and
* is a local minimum then * is also a global
minimum.
– If f() is strictly convex the global minimum is unique
 Unconstrained optimization
• Example : Profit maximization
• Joint profit maximization
 Constrained optimization
• 1. Utility maximization subject to budget
constraint.

• 2. Output maximization subject to cost

constraint .

• 3. Cost minimization subject to output

constraint,