• 沒有找到結果。

# f (x) for all x in D

N/A
N/A
Protected

Share "f (x) for all x in D"

Copied!
6
0
0

(1)

Definitions Let c be a number in the domain D of a function f. Then f (c) is the (a) absolute maximum value of f on D, i.e. f (c) = max

x∈D f (x), if f (c) ≥ f (x) for all x in D.

(b) absolute minimum value of f on D, i.e. f (c) = min

x∈Df (x), if f (c) ≤ f (x) for all x in D.

(c) absolute extremum value of f, if f (c) is either an absolute maximum of minimum value of f on D.

(d) local maximum value of f if f (c) ≥ f (x) when x ∈ D is near c, i.e. there exists δ > 0 such that f (c) ≥ f (x) for all x ∈ D ∩ (c − δ, c + δ).

(e) local minimum value of f if f (c) ≤ f (x) when x ∈ D is near c, i.e. there exists δ > 0 such that f (c) ≤ f (x) for all x ∈ D ∩ (c − δ, c + δ).

(f) local extremumvalue of f if f (c) is either a local maximum or minimum value of f.

(g) A graph is said to be concave up at a point if the tangent line to the graph at that point lies below the graph in the vicinity of the point andconcave down at a point if the tangent line lies above the graph in the vicinity of the point.

(h) A point where the concavity changes (from up to down or down to up) is called a point of inflection.

Definitions Let c be a number in the domain D of a function f. Then c is called a critical number of f if either f0(c) = 0 or f0(c) does not exist.

Extreme Value Theorem Let f be continuous on [a, b]. Then there exist x1, x2 ∈ [a, b] such that

min

[a,b] f = f (x1) ≤ f (x) ≤ f (x2) = max

[a,b] f ∀ x ∈ [a, b].

Fermat’s Theorem Let f be continuous on [a, b]. If f has a local extremum at c ∈ (a, b), and if f0(c) exists, then f0(c) = 0,

i.e. If f has a local extremum at an interior point c ∈ (a, b), then either f is not differentiable at c, or f is differentiable at c, and f0(c) = 0.

Proof Suppose that

lim

h→0

f (c + h) − f (c)

h = f0(c) > 0,

since  = f0(c)

2 > 0, there exists δ > 0 such that if 0 < |h| < δ then

(2)

f (c + h) − f (c)

h − f0(c)

< f0(c)

2 ⇔ f0(c)

2 > f (c + h) − f (c)

h − f0(c) > −f0(c) 2

=⇒ f (c + h) − f (c)

h > f0(c) 2 > 0

=⇒

(f (c + h) − f (c) > 0 for all δ > h > 0 f (c + h) − f (c) < 0 for all 0 > h > −δ

(f (c + h) − f (c) > 0 for all δ > h > 0 f (c − h) − f (c) < 0 for all δ > h > 0

=⇒ f (c − h) < f (c) < f (c + h) for all δ > h > 0

=⇒ f increases through c.

Thus f (c) is not a local extremum of f.

Similarly, if f0(c) < 0 then f decreases through c and f (c) is not a local extremum of f.

Hence, if f has a local extremum at an interior point c ∈ (a, b) thenf0(c) = 0.

Remark The local extremum of a continuous function f occurs only at a critical number of f.

Remark To find an absolute extremum of a continuous function f on a closed interval I = [a, b], we note that either it is local or it occurs at an endpoint of the interval I.

Rolle’s Theorem Let f be continuous on I = [a, b] and let f be differentiable on (a, b). Suppose that f (a) = f (b). Then there is a number c ∈ (a, b) such that f0(c) = 0.

Proof

Case 1 Suppose that f (x) = f (a) = f (b) for all x ∈ [a, b] =⇒ f0(x) = 0 for all x ∈ (a, b).

Case 2 Suppose there exits x ∈ (a, b) such that f (x) 6= f (a) = f (b) =⇒ there exists c ∈ (a, b) such that f (c) is an extremum value of f. Therefore, f0(c) = 0 by the Fermat’s Theorem.

The Mean Value Theorem Let f be continuous on I = [a, b] and let f be differentiable on (a, b). Then there is a number c ∈ (a, b) such that

f0(c) = f (b) − f (a)

b − a ⇐⇒ f (b) − f (a) = f0(c)(b − a).

Proof

Consider the function g defined by

g(x) = f (x) − f (a) − f (b) − f (a)

b − a (x − a) if x ∈ [a, b].

(3)

Since g is continuous on I = [a, b], differentiable on (a, b) and satisfies that g(a) = g(b) = 0, there exists c ∈ (a, b), by the Rolle’s Theorem, such that

0 = g0(c) = f0(c) − f (b) − f (a)

b − a =⇒ f0(c) = f (b) − f (a) b − a .

Remarks Let f be differentiable on (a, b). Suppose that f0(x) 6= 0 for all x ∈ (a, b). Then (a) f is 1 − 1 on (a, b).

Proof For any x1 6= x2 ∈ (a, b), there exits c lying between x1 and x2, by the Mean Value Theorem, such that

f (x1) − f (x2) = f0(c)(x1− x2) 6= 0 =⇒f (x1) 6= f (x2) =⇒ f is 1 − 1 on (a, b).

(b) If f0(x) > 0 for all x ∈ (a, b), then f is increasing on (a, b), i.e. f (x1) < f (x2) for any x1 < x2 ∈ (a, b).

Proof For any x1 < x2 ∈ (a, b), there exits c ∈ (x1, x2), by the Mean Value Theorem, such that

f (x1) − f (x2) = f0(c)(x1− x2) < 0 =⇒f (x1) < f (x2) =⇒ f is increasing on (a, b).

(c) If f0(x) < 0 for all x ∈ (a, b), then f is decreasing on (a, b), i.e. f (x1) > f (x2) for any x1 < x2 ∈ (a, b).

Proof For any x1 < x2 ∈ (a, b), there exits c ∈ (x1, x2), by the Mean Value Theorem, such that

f (x1) − f (x2) = f0(c)(x1− x2) > 0 =⇒f (x1) > f (x2) =⇒ f is decreasing on (a, b).

(d) Let f be twice differentiable on I = (a, b). If f00(x) > 0 for all x ∈ I = (a, b), then f is concave upward on I, i.e. For each c ∈ I, the graph of y = f (x), for x near c, lies above the tangent line to y = f (x) at (c, f (c)).

Proof For each c ∈ I, since

f00(c) = lim

x→c

f0(x) − f0(c) x − c > 0, there exists δ > 0 such that if 0 < |x − c| < δ then

f0(x) − f0(c) x − c > 0

=⇒

(f0(x) − f0(c) > 0 ∀ x ∈ (c, c + δ) f0(x) − f0(c) < 0 ∀ x ∈ (c − δ, c).

This imples that if 0 < |x − c| < δ, by the Mean Value Theorem, then f (x) − f (c) − f0(c)(x − c)

= [f0(z) − f0(c)] (x − c) for some z lying between x and c

> 0

This proves that the point (x, f (x)) in the graph of y = f (x) lies above the point (x, f (c) + f0(c)(x − c)) in the tangent line to the graph of y = f (x) at (c, f (c)).

(4)

(e) Let f be twice differentiable on I = (a, b). If f00(x) < 0 for all x ∈ I = (a, b), then f is concave downward on I, i.e. For each c ∈ I, the graph of y = f (x), for x near c, lies below the tangent line to y = f (x) at (c, f (c)).

Cauchy Mean Value Theorem Let f, g be continuous on I = [a, b] and let f, g be differen- tiable on (a, b). Then there is a number c ∈ (a, b) such that

f0(c) [g(b) − g(a)] = g0(c)[f (b) − f (a)].

Proof

Case 1 If g(b) = g(a) then there exists c ∈ (a, b), by the Rolle’s Theorem, such that g0(c) = 0.

Hence, the equality holds.

Case 2 If g(b) 6= g(a) the function h defined on I by h(x) = f (x) − f (a) −f (b) − f (a)

g(b) − g(a)[g(x) − g(a)] if x ∈ [a, b].

Since h is continuous on I = [a, b], differentiable on (a, b) and satisfies that h(a) = h(b) = 0, there exists c ∈ (a, b), by the Rolle’s Theorem, such that

0 = h0(c) = f0(c) − f (b) − f (a)

g(b) − g(a) g0(c) =⇒ f0(c) [g(b) − g(a)] = g0[f (b) − f (a)].

An Indeterminate Form 00 and l’Hˆopital’s Rule Let f, g be continuous on I = [a, b] and let f, g be differentiable on (a, b). Suppose that

(a) f (c) = 0 = g(c) for some c ∈ (a, b), i.e. lim

x→cf (x) = 0 = lim

x→cg(x), (b) g0(x) 6= 0 for all x ∈ (a, c) ∪ (c, b),

(c) lim

x→c

f0(x)

g0(x) = L ∈ R exists.

Then

limx→c

f (x)

g(x) = lim

x→c

f0(x) g0(x) = L.

Outline of the Proof

x→clim f (x)

g(x) = lim

x→c

f (x) − f (c)

g(x) − g(c) since f (c) = g(c) = 0

= lim

x→c

f0(t)

g0(t) for some t lying between x and c by the Cauchy Mean Value Theorem

= lim

t→c

f0(t)

g0(t) by observng that t → c whenever x → c

= L.

(5)

An Indeterminate Form ± and l’Hˆopital’s Rule Let c be a point in the interval (a, b) and let f, g be be differentiable on (a, c) ∪ (c, b). Suppose that

(a) lim

x→cf (x) = ±∞ and lim

x→cg(x) = ±∞, (b) g0(x) 6= 0 for all x ∈ (a, c) ∪ (c, b),

(c) lim

x→c

f0(x)

g0(x) = L ∈ R exists.

Then

limx→c

f (x)

g(x) = lim

x→c

f0(x) g0(x) = L.

Outline of the Proof Case 1 Suppose that lim

x→c

f0(x)

g0(x) = L = 0.

Given  > 0, since lim

x→c

f0(x)

g0(x) = L = 0, there exists δ1 > 0 such that if 0 < |x − c| < δ1 then

f0(x) g0(x)

< .

For any points x and x1 satisfying

either c < x < x1 < c + δ1, or c − δ1 < x1 < x < c

there exists x2 ∈ (c − δ1, c) ∪ (c, c + δ1) lying between x and x1, by the Cauchy Mean Value Theorem, such that

f (x) − f (x1) g(x) − g(x1)

=

f0(x2) g0(x2)

< .

Also, since

limx→c

1 −f (x1) f (x) 1 −g(x1)

g(x)

= 1,

there exists δ2 > 0 such that

if 0 < |x − c| < δ2 then

1 − f (x1) f (x) 1 −g(x1)

g(x)

> 1 2.

Let δ = min{δ1, δ2} such that if 0 < |x − c| < δ then

(6)

1 2

f (x) g(x)

<

f (x) g(x)





1 −f (x1) f (x) 1 −g(x1)

g(x)





=

f (x) − f (x1) g(x) − g(x1)

=

f0(x2) g0(x2)

<  which implies that

f (x) g(x)

< 2.

Hence,

limx→c

f (x)

g(x) = 0 = L = lim

x→c

f0(x) g0(x). Case 2 Suppose that lim

x→c

f0(x)

g0(x) = L 6= 0.

Consider the function h defined by h(x) = f (x) − L g(x) for all x ∈ (a, c) ∪ (c, b) and note that limx→c

h0(x) g0(x) = lim

x→c

f0(x)

g0(x) − L = 0.

Applying Case 1, we get limx→c

f (x)

g(x) − L = lim

x→c

h(x)

g(x) = 0 =⇒ lim

x→c

f (x) g(x) = L.

Figure 6 shows the relationship between the increment ∆y and the differential dy: ∆y represents the change in height of the curve y = f(x) and dy represents the change in height

4 上午 9 時起至 100.1.11 下午 5 時止。. ＊請使用 IE 5.0

the larger dataset: 90 samples (libraries) x i , each with 27679 features (counts of SAGE tags) (x i ) d.. labels y i : 59 cancerous samples, and 31

Biases in Pricing Continuously Monitored Options with Monte Carlo (continued).. • If all of the sampled prices are below the barrier, this sample path pays max(S(t n ) −

Biases in Pricing Continuously Monitored Options with Monte Carlo (continued).. • If all of the sampled prices are below the barrier, this sample path pays max(S(t n ) −

The best way to picture a vector field is to draw the arrow representing the vector F(x, y) starting at the point (x, y).. Of course, it’s impossible to do this for all points (x, y),

In addition, geometric engineering also suggests an absence of walls conjecture stating an equivalence between refined DT invariants of large radius limit stable objects of D b (X)

If x or F is a vector, then the condition number is defined in a similar way using norms and it measures the maximum relative change, which is attained for some, but not all