1

Lectures 26-27: Functions of Several Variables (Continuity, Dierentiability, Increment Theorem and Chain Rule)
The rest of the course is devoted to calculus of several variables in which we study continuity, dierentiability and integration of functions from Rn to R, and their applications. In calculus of single variable, we had seen that the concept of convergence of sequence played an important role, especially, in dening limit and continuity of a function, and deriving some properties of R and properties of continuous functions. This motivates us to start with the notion of convergence of a sequence in Rn . For simplicity, we consider only R2 or R3 . General case is entirely analogous. Convergence of a sequence : Let Xn = (x1,n , x2,n , x3,n ) R3 . We say that the sequence (Xn ) is convergent if there exists X0 R3 such that Xn X0 0 as n . In this case we say that Xn converges to X0 and we write Xn X0 . Note that corresponding to a sequence (Xn ), Xn = (x1,n , x2,n , x3,n ), there are three sequences (x1,n )(x2,n ) and (x3,n ) in R, and vice-versa. Thus the properties of (Xn ) can be completely understood in terms of the properties of the corresponding sequences (x1,n )(x2,n ) and (x3,n ) in R. For example, (i) Xn X0 in R3 the coordinates xi,n xi,0 for every i = 1, 2, 3 in R. (ii) (Xn ) is bounded (i.e., M such that Xn M n) each sequence (xi,n ), i = 1, 2, 3, is bounded. Using the previous idea, we can prove the following results. Problem 1: Every convergent sequence R3 is bounded. Problem 2 (Bolzano-Weierstrass Theorem): Every bounded sequence in R3 has a convergent subsequence. In case of a sequence in R, to dene the notion of convergence or boundedness, we use | | in place of , hence it is clear how we generalized the concept of convergence or boundedness of a sequence in R1 to R3 . Moreover, it is also now clear how to dene the concepts of limit and continuity of a function f : R3 R at some point X0 R3 . Limit and Continuity : (i) We say that L is the limit of a function f : R3 R at X0 R3 (and we write limXX0 f (X) = L) if f (Xn ) L whenever a sequence (Xn ) in R3 , Xn = X0 , converges to X0 . (ii) A function f : R3 R is continuous at X0 R3 if limXX0 f (X) = f (X0 ). Examples 1: (i) Consider the function f : R2 R, where f (x, y) = sin (xy) when (x, y) = (0, 0) |x|+|y| and f (0, 0) = 0. We will show that this function is continuous at (0,0). Note that | f (x, y) f (0, 0) | | x y |2 | x | + | y | (or | x y | ) |x|+|y|
2

Therefore, whenever a sequence (xn , yn ) (0, 0), i.e, xn 0 and yn 0, we have f (xn , yn ) f (0, 0). Hence f is continuous at (0, 0). In fact, this function is continuous on the entire R2 . (ii) Consider the function f : R2 R, where f (x, y) = xy 2 This function is continuous at (0, 0), because, | xy 2
x +y 2 x +y 2 |x2 2 +y |2 x2 +y

when (x, y) = (0, 0) and f (0, 0) = 0. = x2 + y 2 0, as (x, y) 0.

2
2xy (iii) Let f (x, y) = x2 +y2 , (x, y) = (0, 0). We will show that this function does not have a limit at 2m (0, 0). Note that f (x, mx) 1+m2 as x 0 for any m. This shows that the function does not have a limit at (0, 0). x y (iv) Let f (x, y) = x4 +y2 when (x, y) = (0, 0) and f (0, 0) = 0. Note that f (x, mx) 0 as x 0. But the function is not continuous at (0, 0) because f (x, x2 ) 1 as x 0. Similarly we can show 2 4 2 that the function f (x, y) dened by f (x, y) = x4 y2 when (x, y) = (0, 0) and f (0, 0) = 0 is not x +y continuous at (0, 0) by taking y = mx2 and allowing x 0.
2

Partial derivatives : The partial derivative of f with respect to the rst variable at X0 = (x0 , y0 , z0 ) is dened by f f (x0 + h, y0 , z0 ) f (x0 , y0 , z0 ) |X0 = lim h0 x h provided the limit exists. Similarly we dene
f y |X0

and

f z |X0

2xy Example 2: The function f dened by f (x, y) = x2 +y2 at (x, y) = (0, 0) and f (0, 0) = 0 is not continuous at (0, 0), however, the partial derivatives exist at (0, 0).

Problem 3: Let f (x, y) be dened in S = {(x, y) R2 : a < x < b, c < y < d}. Suppose that the partial derivatives of f exist and are bounded in S. Then show that f is continuous in S. Solution : Let |fx (x, y)| M and |fy (x, y)| M for all (x, y) S. Then f (x + h, y + k) f (x, y) = f (x + h, y + k) f (x + h, y) + f (x + h, y) f (x, y) = kfy (x + h, y + 1 k) + hfx (x + 2 h, y), (for some 1 , 2 R, by the MVT). Hence, |f (x + h, y + k) f (x, y) | M (|h| + |k|) 2M h2 + k 2 . Hence, for > 0, choose = continuous.
2M

or use the sequential argument to show that the function is

It is clear from the previous example that the concept of dierentiability of a function of several variables should be stronger than mere existence of partial derivatives of the function. Dierentiability : When f : R R, x R we dene f (x) = lim f (x + h) f (x) h0 h ()

provided the limit exists. In case f : R3 R and X = (x1 , x2 , x3 ) R3 the above denition of the dierentiability of functions of one variable (*) cannot be generalized as we cannot divide by an element of R3 . So, in order to dene the concept of dierentiability, what we do is that we rearrange the above denition (*) to a form which can be generalized. Let f : R R. Then f is dierentiable at x if and only if there exists R such that | f (x + h) f (x) h | 0 as h 0. |h| When f is dierentiable at x, has to be f (x). We generalize this denition to the functions of several variables.

3 Denition : Let f : R3 R and X = (x1 , x2 , x3 ). We say that f is dierentiable at X if there exists = (1 , 2 , 3 ) R3 such that the error function (H) = tends to 0 as H 0. In the above denition H is the scalar product. Note that the derivative f (X) = (1 , 2 , 3 ). Theorem 26.1: Let f : R3 R, X R3 . If f is dierentiable at X then f is continuous at X. Proof : Suppose f is dierentiable at X. Then there exists = (1 , 2 , 3 ) R3 such that | f (X + H) f (X) H | = Hence | f (X + H) f (X) | H (
i=1

f (X + H) f (X) H H

(H) and (H) 0 as H 0.

| i |)+

(H)

and (H) 0 as H 0. Therefore f (X + H) f (X) as H 0. This proves that f is continuous at X. How do we verify that a given function is dierentiable at a point in R3 ? The following result helps us to answer this question. Theorem 26.2: Suppose f is dierentiable at X. Then the partial derivatives
f z |X f f x |X , y |X

and

exist and the derivative f (X) = (1 , 2 , 3 ) = ( f f f | , | , | ). x X y X z X

Proof : Suppose f is dierentiable at X and f (X) = (1 , 2 , 3 ). Then by taking H = (t, 0, 0), we have (H) = f (X + H) f (X) 1 t f (X + H) f (X) 1 t 0 as t 0, i.e., 0 |t| t
f x |X .

This implies that 1 = Example 3 : Let

Similarly we can show that 2 =

f y |X

and 3 =

f x |X .

f (x, y) = xy

x2 y 2 at (x, y) = (0, 0) x2 + y 2 and verify that (H) 0

= 0 at (0, 0) To verify that f is dierentiable at (0, 0), let us choose = ( f , f ) | x y as H = (h, k) 0. In this case = (0, 0) and | (H) | = | hk f (0 + H) f (0) (0, 0) H || | 2 + k2 H h h2 + k 2 0 as H 0.
(0,0)

Hence f is dierentiable at (0, 0). Example 2 illustrates that the partial derivatives of a function at a point may exist but the function need not be dierentiable at that point. The previous theorem says that if the function is

4 dierentiable at X then the derivative f (X) can be expressed in terms of the partial derivatives of f at X. Since nding partial derivatives is easy because they are based on one variable and it is related to the derivative, one naturally asks the following question: Under what additional assumptions on the partial derivatives the function becomes dierentiable. The following criterion answer this question. Theorem 26.3: If f : R3 R is such that all its partial derivatives exist in a neighborhood of X0 and continuous at X0 then f is dierentiable at X0 . We omit the proof of this result. We will see in a tutorial class that the converse of the previous result is not true. Chain Rule: We have seen that the chain rule which deals with derivative of a function of a function is very useful in one variable calculus. In order to derive a similar rule for functions of several variables we need the following theorem called Increment Theorem. For simplicity we will state this theorem only for two variables. We will employ the notation fx =
f x

and fy =

f x .

Theorem 26.4: Let f (x, y) be dierentiable at (x0 , y0 ). Then we have f (x0 + x, y0 + y) f (x0 , y0 ) = fx (x0 , y0 )x + fy (x0 , y0 )y + 1 x + 2 y where 1 (x, y), 2 (x, y) 0 as x 0 and y 0. Proof (*): Let H = (x, y). Since the function is dierentiable at (x0 , y0 ), we have f (x0 + x, y0 + y) f (x0 , y0 ) = fx (x0 , y0 )x + fy (x0 , y0 )y+ We have to show that (H) H H = H (H), (H) 0 as H 0. (H) = 1 x + 2 y for some functions 1 and 2 . Note that (H) (H) (H) (x2 + y 2 ) = (x )x + (y )y. H H H

Dene 1 (H) = x (H) and 2 (H) = y (H) . Note that H H | 1 (H) | = | x (H) | | (H) | 0 as H 0. H

Similarly we can show that 2 (H) 0 as H 0. This proves the result. In the next result we present the chain rule. Theorem 26.5: Let f (x, y) be dierentiable (or f has continuous partial derivatives) and if x = x(t), y = y(t) are dierentiable functions on t, then the function w = f (x(t), y(t)) is dierentiable at t and df df f dx f dy = fx (x(t), y(t))x (t) + fy (x(t), y(t))y (t), i.e., = + . dt dt x dt y dt Proof : By increment theorem we have f = fx (x0 , y0 )x + fy (x0 , y0 )y + 1 x + 2 y, This implies that 1 , 2 0 as x, y 0

x y x y f = fx + fy + 1 + 2 . t t t t t Allow t 0, which implies that 1 , 2 0 because x, y 0. Therefore, we get f dx f dy x dt + y dt .

df dt