ENGR 2422 Engineering Mathematics 2
Brief Notes on Chapter 2
Contents:
2.3 Differentials and the Chain Rule
2.4 The Jacobian
2.1 Partial Derivatives
If f is a function of the independent variables x, y and z, then the rate of change of f with respect to one of the independent variables (in a slice through ú3 in which the other two independent variables are constant) is given by the appropriate partial derivative:
or
The definition can be extended to functions of any number of variables.
The usual rules and techniques of differentiation (product rule, quotient rule, implicit, logarithmic, etc.) extend to partial differentiation in an obvious way.
Examples:
1.
2.
2.2 Higher Derivatives
Examples of higher order partial derivatives include
Clairaut’s theorem:
If, on a disc D containing the point (a, b) the function f is defined and the partial derivatives fxy and fyx are both continuous, (which is the case for most functions of interest), then
fxy(a, b) = fyx(a, b)
that is, the order of differentiation doesn’t matter.
One of the most important partial differential equations involving second partial derivatives is Laplace’s equation, which arises naturally in many applications, including electrostatics, fluid flow and heat conduction:
or its equivalent in ú3:
2.3 Differentials and the Chain Rule
If z = f (x, y) where x and y are both functions of t only, then
and
More generally, for z = f (x1(t1, t2, ... , tm), x2(t1, t2, ... , tm), ... , xn(t1, t2, ... , tm) ),
To use the general form of the chain rule, follow every path from z to the dependent variable tj via all of the xi variables.
Example 1:
u = xy + yz + zx, x = st, y = est, z = t2 . Find us in terms of s and t only.
= t ( t(s + t) + (1 + t2)est)
This derivative could also be found directly by replacing x, y and z by the respective functions of s and t before differentiating u.
Note that there is no need to
find because z
is not a function of s.
Differentials may be used to estimate changes in f caused by small changes in the independent variables xi.
Example 2:
If the errors in measuring a 30 cm ´ 24 cm rectangle are at most 0.1 cm in each of length and width, then the maximum error in the area A = LW is approximately
[The exact value is 5.41 cm2.]
The maximum relative error is
Implicit functions:
If z is defined implicitly as a function of x and y by F (x, y, z) = c , then
dF = Fx dx + Fy dy + Fz dz = 0
Example 3:
Curves of Intersection
Example 4:
Find both partial derivatives with respect to z on the curve of intersection of the sphere centre the origin, radius 4, and the circular cylinder, central axis on the y-axis, radius 3.
Sphere: f = x2 + y2 + z2 = 16
Cylinder: g = x2 + z2 = 9
Þ df = 2x dx + 2y dy + 2z dz = 0
and dg = 2x dx + 2z dz = 0
which leads to the linear system
Therefore
[The intersection is the pair of circles x2 + z2 = 9, y = ±4 .
Because y never changes on each circle, x is actually a function of z only.]
2.4 The Jacobian
A transformation from one orthogonal coordinate system (x, y, z) to another, (u, v, w), often leads to a need to know how the [differential] volume element dV = dx dy dz transforms into the new coordinate system. The relationship is
where is the Jacobian.
The two dimensional equivalent is
where .
The above is the explicit method for determining the Jacobian of the transformation.
It requires all of the old coordinates (x, y, z) to be known as explicit functions of the new coordinates (u, v, w).
The implicit method can be used even when only an implicit relationship between (x, y, z) and (u, v, w) is known:
Let (x, y, z) and (u, v, w) be related by
f (x, y, z, u, v, w) = c1
g (x, y, z, u, v, w) = c2
h (x, y, z, u, v, w) = c3
then
df = fx dx + fy dy + fz dz + fu du + fv dv + fw dw = 0
dg = gx dx + gy dy + gz dz + gu du + gv dv + gw dw = 0
dh = hx dx + hy dy + hz dz + hu du + hv dv + hw dw = 0
which leads to the matrix equation
where
The Jacobian is the magnitude of the determinant of the matrix A-1B .
Also det (A-1B) = det B / det A .
For a transformation from Cartesian to plane polar coordinates in ú2,
For cylindrical polar coordinates in ú3,
For spherical polar coordinates in ú3,
Explicit method for plane polar coordinates:
x = r cos q , y = r sin q
= r cos2q + r sin2q = r .
Implicit method for plane polar coordinates:
f = x - r cos q = 0 Þ df = dx - cos q dr + r sin q dq = 0
g = y - r sin q = 0 Þ dg = dy - sin q dr - r cos q dq = 0
2.5 The Gradient Vector
When F(r) is a scalar function of position (x, y,
z) in ú3 and all
coordinates are, in turn, functions of a single parameter t, then the
chain rule becomes
and
.
ÑF is the gradient
vector of the scalar function F.
Ñ is the gradient
operator. (The symbol Ñ is pronounced “nabla” or “del”.)
Let â
be the unit vector in the direction of a non-zero vector a , (so that a = aâ).
Then the
rate of change of F at point
Po in the
direction of a is the directional derivative
Applications of the Directional Derivative and the Gradient:
(1) The directional derivative D at the point Po = (xo, yo, zo) is maximized by choosing
a
to be parallel to ÑF
at Po, so that .
(2) A normal vector to the surface F(x, y, z) = c at the point Po = (xo, yo, zo) is
.
(3) The equation of the line normal to the surface F(x, y, z) = c at the point
Po = (xo,
yo, zo)
is
(4) The equation of the tangent plane to the surface F(x, y, z) = c at the point
Po = (xo,
yo, zo)
is
(5) If the point Po = (xo, yo, zo) lies on both of the surfaces F(x, y, z) = c and
G(x, y, z) = k, then the angle of intersection q of the surfaces at the point is given by
If f(r) is the potential
function for some force F(r), then
F = Ñf (if like charges
or masses repel; F = -Ñf if like charges or masses attract).
For a
central force law f = k rn
it follows that .
2.6 Maxima and Minima
For a function f (x, y) defined on some domain D in ú2, the point P(xo, yo) is a critical point [and the value f (xo, yo) is a critical value] of f if
1) P is on any boundary of D; or
2) f (xo, yo) is undefined; or
3) fx and/or fy is undefined at P; or
4) fx and fy are both zero at P (Þ Ñf = 0 at P).
To determine the nature of a critical point:
1) Examine the values of f in the neighbourhood of P; or
2) [First derivative test:] Examine the changes in fx and fy at P; or
3) Use the second derivative test:
At all points (a, b) where Ñf = 0 , find all second partial derivatives, then find
and evaluate D at (x, y) = (a, b).
D(a, b) > 0 and fxx(a, b) > 0 Þ a relative minimum of f is at (a, b)
D(a, b) > 0 and fxx(a, b) < 0 Þ a relative maximum of f is at (a, b)
D(a, b) < 0 Þ a saddle point of f is at (a, b)
D(a, b) = 0 Þ test fails (no information).
Example:
Find all extrema of f (x, y) = x2 + y2 + 4x - 6y .
f (x, y) is a polynomial function of x and y and is therefore defined and differentiable in all of ú2. Any critical points will therefore be of type (4) only.
fxx = 2, fxy = 0, fyy = 2
D > 0 and fxx > 0 at (-2, 3) Þ there is a relative minimum at (-2, 3) and
the minimum value is f (-2, 3) = -13.
As there are no other critical points, f (x, y) has an absolute minimum value of -13 at (-2, 3) and has no maxima.
[z = f (x, y) is a circular paraboloid, vertex at (-2, 3, -13) and axis of symmetry parallel to the z-axis.]
Various other examples will appear both in class and on the problem sets.
2.7 Lagrange Multipliers
To find the maximum or minimum value(s) of a function f (x1, x2, ... , xn) subject to a constraint g(x1, x2, ... , xn) = k, solve the system of simultaneous (usually non-linear) equations in (n + 1) unknowns:
Ñf = l Ñg
g = k
where l is the Lagrange multiplier.
Then identify which solution(s) gives a maximum or minimum value for f.
See example 1 from the class notes.
In the presence of two constraints g(x1, x2, ... , xn) = k and h(x1, x2, ... , xn) = c , solve the system in (n + 2) unknowns:
Ñf = l Ñg + m Ñh
g = k
h = c
See example 2 from the class notes.
END OF CHAPTER 2
[Return to your previous page]