這是有關「可導性」(differentiablity)的討論,寫給那天沒有上復活假期補課班的中六同學。注:討論純粹以中學數學的觀點出發。
先請同學回答下面三道是非題:
1. Put into the expression of and it is undefined, then is not differentiable at . True?
2. If , then is differentiable at . True?
3. If is finite, then the value of is also finite. True?
=======================================
上面三題的答案皆「不一定是」。
=======================================
首先,要描述一個函數是否「可導」(differentiable),我們不是看 這個表達式是什麼,最安全的,都是返回「可導」的定義。
再看看上述的三題
Question 1:
代 入表達式 ,而「計不到數」(undefined),並不一定代表” is not differentiable at ”.
以下是一個經典的例子:
for and .
對於 ,恆有:
– – – – – – (*)
若把 代入上式 (*),但因為當 , 是「計不到數」(沒有定義 undefined),同學往往以為 在 處也是 undefined,故他們聲稱” is not differentiable”。
錯!
判別「可導」與否,一定要回歸「可導」的定義:
” is differentiable at ”的意思是「極限 存在」。
一旦極限存在,我們才可寫 ,而並不是先用中四五的手法,找出 的表達式,再代入 ,就得出 。
返回那經典例子:
for and .
當我們考慮極限
根據定義,
, ,故
因 有界(bounded),且 ,故 ,即
亦即此極限存在(等於 0,即 )是故
is differentiable at
Question 2 和 Question 3 可以一併考慮,隨便舉例
for and
for
對於 ,我們易知
for and
for
易知
於是
但,明顯地, 在 之處是不可導(not differentiable at )。
最簡單的原因是, 在 之處根本不連續(not continuous at ),見下
可見
是故 在 之處不連續,從而不可導。
現在,具體地計算一下 在 處的左導數(left-hand derivative),即
左導數不存在,導數也自然地不存在,順便回答了 Question 3:
(finite)但
(infinite)
重申一次,縱使存在以下關係:
我們也不能知道 存在與否,
亦即不能保證 在 之處是可導。
習題
1. Let
for and
for
Suppose is differentiable at , evaluate .
2. Let
for and
for
Will be differentiable at for some real number ?
3. Let
for and
for
Given that is differentiable at .
Is is true to write ””?
sir, here is my ans:
1. a=+1
2. differentiable when b=+1
3. T
but i actually don’t know what exactly what f'(x) is
what is the different in meaning between
lim f'(x)
x->0-
and
lim (f(h)-f(0))/h
h->o-
and also
lim f(x)
x->0-
i don’t know which one to consider when doing calculation
迴響 由 harrison — 2009/04/17 @ 8:16 下午 |
Harrison,
The answer to Q.2 is FALSE.
is NOT differentiable for ANY real number .
Just think about the graph of , it is ALWAYS discontinuous at (even for ). (Or prove it by simple algebraic computation)
What is ?
It is a limit.
By definition,
If the limit EXISTS, we define it as
.
What is ?
It is NOT in general.
By the definition above,
.
hence
You see, it is NOT in general, because, by the definition, is simply:
SOMETIMES, is really equal to , as for example, when the function is continuous at .
Also, can you prove that the answer to Q.3 is TRUE?
迴響 由 johnmayhk — 2009/04/18 @ 6:12 下午 |
For Question 1, I used to teach my students in the same way as yours before at the beginning, but later I found that the answer of Q1 is true, and I also got a very simple proof.
Let me first point out the problem of your counter-example.
If f(x) = x^2 sin 1/x if x is not 0 and f(0) = 0,
then what is f'(x)?
Answer:
f'(x) = (the expression you have) if x is not 0 and f'(0) = 0
Therefore, when we put x = 0 , f'(0) = 0 which is defined.
[ What you showed in your counter-example is in fact,
lim {x -> 0} f'(x) . ]
Back to the proof of the statement.
Assume f(x) is differentiable at x = 0.
Then f'(0) exists (defined), which leads to a contradiction.
Therefore f'(x) is not differentiable at x = 0.
迴響 由 Cheng Wing Kuen — 2009/04/21 @ 8:09 上午 |
Thank you Cheng Wing Kuen.
是我的表達出了問題。
在 Question 1 中,
Put into the expression of
當中我希望說的,不是 。
當中的 expression,其實指,把 進行求導後得出的 expression。
如上例
利用求導法,得出的 expression 就是
Question 1 想說的,就是 put into the expression above.
它是 undefined 的。
它不是 。
我只是想指出,同學往往誤以為
「把 D 完,再代個零入去,計唔到,就即係函數係 x = 0 個位 D 唔到」。
是我的表達出問題,”expression of ”確實可以指
when and when
那麼,當”Put into the expression of ” 就被理解為 時,
誠如 Cheng Wing Kuen 所言,
Question 1 is trivially true.
We can prove it simply by contrapositivity.
“ is differentiable at exists"
is equivalent to
“ does not exist is not differentiable at "
P.S.
Oh, Mr. cheng, are you the writer of “永權網頁"? I love the website, it is very useful.
迴響 由 johnmayhk — 2009/04/21 @ 2:46 下午 |