微積分筆記

lim區

lim目錄
lim(x->a)[f(x) + g(x)] = lim(x->a)f(x) + lim(x->a)g(x)

lim(x->a)[f(x) - g(x)] = lim(x->a)f(x) - lim(x->a)g(x)

lim(x->a)[f(x) * g(x)] = lim(x->a)f(x) * lim(x->a)g(x)

lim(x->a)[f(x) / g(x)] = lim(x->a)f(x) / lim(x->a)g(x)(分母不為零)

lim(x->a)[ c * f(x)] = c * lim(x->a)f(x)(c是常數)

如果f(x)在x=a連續，且lim(x->a)g(x)=L則
lim(x->a)f(g(x)) = f(lim(x->a)g(x)) = f(L)

d/dx f(x) = limh->0 (f(a+h)-f(a))/h = limx->h (f(x)-f(h))/(x-h) 

lim(h->0) sinh/h= 1

im(h->0)(cos h-1))/h= 0

d/dx sin x= cos(x)

d/dx cosx = -sinx

d/dx a^x = lim(h->0)(a^x(a^h-1)/h)

d/dx a^x =f`(0)a^x

lim(x->0) ln(1+x)/x = 1

im(x->0)(ln(1+x)^(1/x)) = 1

d/dx ln(x) = 1/x

S(b,a)f(x)dx= lim(n->infinite) Sum(n,i=1)f(xi) * delta_x

e =  lim(x->0)(1+x)^(1/x)

e = lim(n->infinite)(1+1/n)^(n)

lim(x->0) (e^x - 1)/x = 1

d/dx e^x = e^x

lim(x->infinite)tan^-1 x = pi/2

lim(x->-infinite)tan^-1 x = -pi/2

已知lim(x->a)f(x) =0 and lim(x->a)g(x) =0 
lim(x->a) f(x)/g(x) = lim(x->a) (d/dx f(x)/ d/dx g(x))

已知lim(x->a)f(x) = infinite and lim(x->a)g(x) = infinite
已知lim(x->a)f(x) =-infinite and lim(x->a)g(x) =-infinite
lim(x->a) f(x)/g(x) = lim(x->a) (d/dx f(x)/ d/dx g(x))

====end of 目錄

lim(x->a)[f(x) + g(x)] = lim(x->a)f(x) + lim(x->a)g(x)
proof
    跟εδ有關，自己去查，關鍵字:微積分的嚴格證明，看到εδ這兩個符號就找對了，以下同理。

lim(x->a)[f(x) - g(x)] = lim(x->a)f(x) - lim(x->a)g(x)
proof

lim(x->a)[f(x) * g(x)] = lim(x->a)f(x) * lim(x->a)g(x)
proof

lim(x->a)[f(x) / g(x)] = lim(x->a)f(x) / lim(x->a)g(x)(分母不為零)
proof

lim(x->a)[ c * f(x)] = c * lim(x->a)f(x)(c是常數)
proof

如果f(x)在x=a連續，且lim(x->a)g(x)=L則
lim(x->a)f(g(x)) = f(lim(x->a)g(x)) = f(L)
ex:(用於e =  lim(x->0)(1+x)^(1/x)證明)
    由於e^x在R是連續的，lim(x->a)e^f(x) = e^lim(x->a)f(x)

lim的通用式子
d/dx f(x) = limh->0 (f(a+h)-f(a))/h = limx->h (f(x)-f(h))/(x-h) 
這是一個通用式，很多函數都能通過他找到一個微分函式。

lim(h->0) sinh/h= 1證明，用於對sin微分= cos(x)證明
proof
    關於lim(h->0) sinh/h= 1使用圖型證(自己去找圖，反正就比來比去夾一夾，1就出來了)
    theta > BC = theta > sin theta, so sin theta/theta < 1
    theta < tan theta = theta < sin theta/cos theta ,so cos theta <sin theta/theta . 結合兩個方程，cos theta < sin theta/theta < 1 ,所以lim(h->0) sinh/h = 1, by squeeze theorem

lim(h->0)(cos h-1))/h= 0證明，用於對sin微分= cos(x)證明
proof
    關於lim(h->0)(cos h-1))/h= 0使用算法證
    lim(theta->0)((cos theta)-1)/theta = lim(theta->0)((cos theta)-1)/theta * (cos(theta) +1)/(cos (theta) +1) = lim(theta->0) cos^2theta -1/(theta * cos(theta) +1) = lim(theta->0) -sin^2theta/(theta * cos(theta) +1) =  -lim(theta->0) sin theta/theta * (sin theta/cos theta +1)
=-1 * 1 * (0/1+1) = 0
    所以， sinx * lim(h->0)(cos h-1))/h + cosx * lim(h->0) sinh/h = sinx * 0 + cosx * 1 = cosx，得證。

d/dx sin x= cos(x)證明
proof
    d/dx sin x = lim(h->0) (sin(x+h)-sinx)/h = lim(h->0) (sinx* cosh + cosx* sinh -sinx)/h =lim(h->0) ((sinx* (cos h-1))/h + lim(h->0)(cosx* sinh)/h = sinx * lim(h->0)(cos h-1))/h + cosx * lim(h->0) sinh/h = sinx * 0 + cosx * 1

d/dx cosx = -sinx證明
proof
     lim(h->0) (cos(x+h)-cosx)/h = lim(h->0) (cosx* cosh - sinx* sinh -cosx)/h = lim(h->0) (cosx(cosh -1)/h - sinx*sinh) = cosx * lim(h->0) (cosh -1)/h - sinx * lim(h->0) sin/h = cosx * 0 - sinx * 1 = -sinx

d/dx a^x = lim(h->0)(a^x(a^h-1)/h)證明，用於d/dx a^x =f`(0)a^x證明
proof
    d/dx a^x =lim(h->0)((f(x+h)-f(x))/h) =lim(h->0)(a^x(a^h-1)/h)

d/dx a^x =f`(0)a^x證明
proof
    f`(0) = lim(h->0)(1 * (a^h-1)/h)
     d/dx a^x lim(h->0)(a^x(a^h-1)/h) =  lim(h->0)((a^h-1)/h) * lim(h->0)(a^x) = f`(0) * a^x

lim(x->0) ln(1+x)/x = 1 證明，用於對d/dx ln(x) = 1/x證明
im(x->0)(ln(1+x)^(1/x)) = 1 證明，用於e =  lim(x->0)(1+x)^(1/x)證明
proof1
    By LHopsital Rule, d/dx ln(1+x) = 1/(1+x), d/dx x = 1, so lim(x->0) ln(1+x)/x = lim(x->0) ( 1/(1+x))/1 = 1/1 = 1

proof2
    d/dx ln(x) = 1/x, f`(1) = 1
    f(x) =ln(x) f`(1)= 1 = lim(x->0)((ln(1+x)-ln(1))/x))=lim(x->0)(ln(1+x)/x)

d/dx ln(x) = 1/x證明
proof
    ln(1+u)/u = 1
    d/dx ln(x) = lim(h->0)((ln(x+h) - ln(x))/h) = lim(h->0)(ln(x+h/x)/h)=lim(h->0) (1/h)*(ln(1+h/x))
    let u=h/x ,lim(h->0) (1/h)*(ln(1+h/x)) = lim(u->0)(1/ux)*(ln(1+u))
    lim(u->0)(1/ux)*(ln(1+u)) = lim(u->0)(1/x)*(ln(1+u)/u) = lim(u->0)(1/x)*1 = 1/x

積分的面積累加定義
S(b,a)f(x)dx= lim(n->infinite) Sum(n,i=1)f(xi) * delta_x
proof
    分成很多塊塊然後底*高加起來。

e =  lim(x->0)(1+x)^(1/x)證明
proof
    lim(x->0)(ln((1+x)^(1/x))) =1
    由於e^x在R是連續的，e^lim(x->a)f(x) = lim(x->a)e^f(x)
    e^1=e^(lim(x->0)(ln((1+x)^(1/x)))) = lim(x->0) e^ln((1+x)^(1/x)) = lim(x->0)(1+x)^(1/x)

e = lim(n->infinite)(1+1/n)^(n)證明
proof
    Let n=1/x,  e = lim(x->0)(1+x)^(1/x) =  lim(n->infinite)(1+1/n)^(n)

lim(x->0) (e^x - 1)/x = 1證明，用於d/dx e^x = e^x證明
proof
    e^x的泰勒級數展開是(x=0) = 1 + x + x^2/2! + x^3/3! + ...，左右減一同除x後 = (e^x - 1 )/x = 1 + x/2! + x^2/3! +  x^3/4! + ...
    lim(x->0) (e^x - 1)/x = 1 + 0 + 0 + ... = 1

d/dx e^x = e^x證明
proof
    d/dx e^x = lim(h->0)( e^(x+h) - e^(x) )/ h = lim(h->0)( e^x*e^h - e^(x) )/ h = lim(h->0) e^x * (e^h-1)/h = lim(h->0) e^x = e^x

lim(x->infinite)tan^-1 x = pi/2證明
lim(x->-infinite)tan^-1 x = -pi/2證明
proof
    把tan圖轉90度，就是tan^-1的圖，可發現x->infinite時y是pi/2，可發現x->-infinite時y是-pi/2。

洛必達0版本證明
已知lim(x->a)f(x) =0 and lim(x->a)g(x) =0 
lim(x->a) f(x)/g(x) = lim(x->a) (d/dx f(x)/ d/dx g(x))
proof
    f(a)=g(a)=0
    lim(x->a) (d/dx f(x)/ d/dx g(x)) = d/dx f(a)/ d/dx g(a) = lim(x->a)(f(x)-f(a)/x-a )/lim(x->a)(g(x)-g(a)/x-a ) = lim(x->a)(f(x)-f(a)/g(x)-g(a)) =lim(x->a)(f(x)/g(x))

洛必達infinite版本證明
已知lim(x->a)f(x) = infinite and lim(x->a)g(x) = infinite
已知lim(x->a)f(x) =-infinite and lim(x->a)g(x) =-infinite
lim(x->a) f(x)/g(x) = lim(x->a) (d/dx f(x)/ d/dx g(x))
proof
    窩不會，好像跟Cauchy's Mean Value Theorem有關。
    
Cauchy's Mean Value Theorem
f`(c)/g`(c) = (f(b)-f(a))/(g(b)-g(a))

=====
d/dx區
d/dx的加法規則
d/dx ( f(x) + g(x) ) = d/dx f(x) + d/dx g(x)證明 
proof
    d/dx (f(x)+g(x)) = lim(h->0) ([f(x+h)+g(x+h)] - [f(x) + g(x)])/h = lim(h->0) (f(x+h) -f(x))/h + (g(x+h) -g(x))/h 
    By lim的加法規則，lim(h->0) (f(x+h) -f(x))/h + (g(x+h) -g(x))/h  = lim(h->0) (f(x+h) -f(x))/h + lim(h->0) (g(x+h) -g(x))/h = d/dx f(x) + d/dx g(x)

d/dx的減法規則
d/dx ( f(x) - g(x) ) = d/dx f(x) - d/dx g(x)證明
proof
    d/dx (f(x)-g(x)) = lim(h->0) ([f(x+h)-g(x+h)] - [f(x) - g(x)])/h = lim(h->0) (f(x+h) -f(x))/h - (g(x+h) -g(x))/h 
    By lim的加法規則，lim(h->0) (f(x+h) -f(x))/h - (g(x+h) -g(x))/h  = lim(h->0) (f(x+h) -f(x))/h - lim(h->0) (g(x+h) -g(x))/h = d/dx f(x) - d/dx g(x)

d/dx的乘法規則
d/dx ( f(x) * g(x) ) = d/dx f(x) * g(x) + f(x) * d/dx g(x)證明
proof
    d/dx (f(x) * g(x)) = lim(h->0) (f(x+h) * g(x+h) - f(x) * g(x))/h = lim(h->0) (f(x+h) * g(x+h) - f(x) * g(x) + f(x)*g(x+h) -  f(x)*g(x+h))/h = lim(h->0)( g(x+h) * (f(x+h) - f(x))/h  + f(x) * (g(x+h)-g(x))/h) = lim(h->0)( g(x) * (f(x+h) - f(x))/h  + lim(h->0)(f(x) * (g(x+h)-g(x))/h) = g(x) * d/dx f(x) + f(x) * d/dx g(x)

d/dx的除法規則
d/dx ( f(x) / g(x) ) = d/dx f(x) * g(x) - f(x) * d/dx g(x) / g(x) * g(x)證明 (分母不為零)
proof
    d/dx ( f(x) / g(x) ) =  lim(h->0) ( f(x+h) / g(x+h) - f(x)/g(x) )/h =  lim(h->0) (f(x+h)g(x) -f(x)g(x+h)) / g(x)g(x+h)h = lim(h->0) ([f(x+h)g(x) - f(x)g(x)] - [f(x)g(x+h) - f(x)g(x) ]) / g(x)g(x+h)h = lim(h->0) (f(x+h) - f(x))/h / * g(x)/g(x)g(x+h) - lim(h->0)(g(x+h) - g(x)) / h * f(x)/g(x)g(x+h) = (d/dx f(x) * g(x)) - (f(x) * d/dx g(x)) / g(x)g(x)

d/dx的倍數規則
d/dx ( c * f(x) ) = c * d/dx f(x)證明
proof
    d/dx ( c * f(x) ) = lim(h->0)  (cf(x+h) - cf(x)) /h = lim(h->0)  c(f(x+h) - f(x)) /h = c * lim(h->0) (f(x+h) - f(x)) /h = c * d/dx f(x)

神秘微分三角小口訣(自己創的，<是找的次數)
∨sin sec<> <<tan 是正的
∧cos csc<> <<cot 是負的

d/dx sinx = cosx(已於前面證過)
d/dx cosx = -sinx(已於前面證過)

d/dx secx = secx * tanx證明1/c s/c
proof
    d/dx secx = d/dx 1/cosx = (0 * cosx - 1 * -sin) / cosx * cosx = sin/cos * 1/cos = tanx * secx

d/dx tanx = secx * secx證明
proof
    d/dx tanx = d/dx sinx/cosx = (1/cos^2(x)) * (-sinx *-sinx + cosx cosx) 
    = tan^2x +1 = sec^2

d/dx cscx = -cscx * cotx證明
proof
    d/dx cscx = 1/sinx = (0 * sinx - 1 * cosx) / sin^2(x) = -cosx/sinx * 1/sinx = -cotx * cscx

d/dx cotx = -cscx * cscx證明
proof
    d/dx cotx = d/dx cosx/sinx = -sinx*sinx - cosx*cosx / sin^2(x) = -1 - cot^2(x) = -csc^2(x)

d/dx 有chain rule
d/dx 的chain rule常與cos(u)、e^u、ln(u)
dy/dx = dy/du du/dw dw/dz dz/da da/dn dn/dx, 把每個都算出來, 然後瘋狂帶入。
d/dx 能對等式兩邊d/dx
ln 能對等式兩邊ln
d/dx f的-+交接有最大，+-交接有最小
(d/dx)(d/dx) f +是凹向上，-是凹向下

xn-f(xn)/f`(xn)不斷執行，得到f(x)的解

反函數規則
y=f(x)
x=(f^-1)(y)
x=(f^-1)(f(x))

d/dy (f^-1)(y) = 1 / d/dx f((f^-1)(y))證明(可以取得反函數的某個微分值)
proof
    y=f(x)
    x=(f^-1)(y)
    dy/dy = 1 =dy/dx * dx/dy = d/dx y * d/dy x =  d/dx f(x) * d/dy (f^-1)(y) = d/dx f((f^-1)(y)) * d/dy (f^-1)(y)
    d/dy (f^-1)(y) = 1 / d/dx f((f^-1)(y))

d/dx a^x = lim(h->0)(a^x(a^h-1)/h)證明(已於前面證過)

d/dx a^x =ln(a) * a^x 證明
proof
    d/dx a^x=d/dx e^ln(a^x) = e^ln(a^x) * d/dx ln(a^x)=  e^ln(a^x) * d/dx x * ln(a)  = a^x * ln(a)
    請注意: d/dx ln(a^x)  = 1/ ln(a^x) * d/dx a^x是錯誤的，以及 ln(a) 是常數

d/dx a^x =f`(0)a^x證明(已於前面證過)

d/dx e^x = e^x 證明(已於前面證過)
    
e =  lim(x->0)(1+x)^(1/x)證明(已於前面證過)
e = lim(n->infinite)(1+1/n)^(n)證明(已於前面證過)

e=2.71828...
proof
    e^x=2^cx f`(x)=f`(0)a^x = a^x * k
    f`(0) = 0.693147 =k
    e^x = d/dx e^x = 2^cx * k * d/dx cx =ck2^cx
    x=0 1=ck c=1/k
    e=2^1/k = 2.271828

d/dx ln(x) = 1/x證明(已於前面證過)

d/dx loga(x) = 1/x*ln(a)
proof
    d/dx loga(x)=d/dx (ln(x)/ln(a)) = 1/ln(a) * d/dx ln(x) =1/ln(a) * 1/x = 1/x*ln(a)

====
S區
加法規則
S f(x) + g(x) dx = S f(x) dx + S g(x) dx
proof
    [a,b]切n個小區間, delta(x) = (b-a)/n, 區間[x(i-1), x(i)]中選擇一個代表點c(i)
    S f(x) + g(x) dx = sum(n, i=1)(f(c(i)) + g(c(i))) * delta(x) = sum(n, i=1)f(c(i)) * delta(x) - sum(n, i=1) g(c(i)) * delta(x) = S f(x) dx + S g(x) dx

減法規則
S f(x) - g(x) dx = S f(x) dx - S g(x) dx
proof
    [a,b]切n個小區間, delta(x) = (b-a)/n, 區間[x(i-1), x(i)]中選擇一個代表點c(i)
    S f(x) + g(x) dx = sum(n, i=1)(f(c(i)) - g(c(i))) * delta(x) = sum(n, i=1)f(c(i)) * delta(x) - sum(n, i=1) g(c(i)) * delta(x) = S f(x) dx - S g(x) dx

倍數規則
S c * f(x) dx = c * S f(x) dx
proof
    [a,b]切n個小區間, delta(x) = (b-a)/n, 區間[x(i-1), x(i)]中選擇一個代表點c(i)
    S c * f(x)  dx = sum(n, i=1) c * f(c(i)) * delta(x) = c * sum(n, i=1) f(c(i)) * delta(x) = c * S f(x) dx

S有substitution rule，使用此法須找到u和u的微分，u的微分在函數最外層\
S的substitution rule常和1/u使用
S有1/x = ln|x| +C  (log中x不能為負的)
S+一個基於x的函數+一個x範圍=三圍面積

由於微分和積分是逆運算，以下公式的證明通過微分右側來處理
S cos(x) dx = sin(x) +C(已於前面證過)
S sin(x) dx = -cos(x) +C(已於前面證過)
S sec(x)tan(x) dx = sec(x) +C(已於前面證過)
S csc(x)cot(x) dx = -csc(x) +C(已於前面證過)
S sec^2(x) dx = tan(x) +C(已於前面證過)
S csc^2(x) dx = -cot(x) +C(已於前面證過)

S sec(x) dx = ln|sec(x)+tan(x)| +C 證明
proof
    d/dx ln|sec(x)+tan(x)| = 1/(sec(x)+tan(x)) * d/dx (sec(x)+tan(x)) = 1/(sec(x)+tan(x)) *(sec(x)*tan(x) + sec(x)*sec(x)) = sec(x) * (sec(x)+tan(x)) * 1/(sec(x)+tan(x)) = sec(x)

S tan(x) dx = -ln|cos(x)| +C 證明
proof
    d/dx -ln|cos(x)| = -1/cos(x) * d/dx cos(x) = -1/cos(x) * -sin(x) = tan(x)

S csc(x) dx = -ln|csc(x) + cot(x)| +C 證明
proof
    d/dx -ln|csc(x) + cot(x)| = -1/(csc(x) +cot(x)) * d/dx (csc(x) +cot(x)) = -1/(csc(x) +cot(x)) * ( -csc(x)cot(x) + -csc(x)csc(x) ) = csc(x) * 1/(csc(x) +cot(x)) * (cot(x) + csc(x) = csc(x)

S cot(x) dx = ln|sin(x)| +C 證明
proof
    d/dx ln|sin(x)| = 1/sin(x) *d/dx sin(x) =  1/sin(x) * cos(x) = cot(x)

當x的次方!= -1
S x^n dx = x^(n+1)/(n+1) +C 證明
proof
    d/dx x^(n+1)/(n+1)  = (n+1) * x^n /(n+1)= x^n

當x的次方=-1
S x^-1 dx = ln|x| + C
proof

S a^x dx = a^x/ln(a) +C 證明
proof
    d/dx a^x/ln(a) = 1/ln(a) *d/dx a^x = 1/ln(a) * ln(a) * a^x =a^x

S ln(x) dx = xln(x) - x + C
proof
    d/dx xln(x) - d/dx x = x*(1/x) + ln(x) * 1  -1 = ln(x)

積分奇技淫巧區1號區

Inetgration by Parts
S f(x)g`(x)dx + S f`(x)g(x)dx = f(x)g(x) 證明
S f(x)g`(x)dx = f(x)g(x) -  S f`(x)g(x)dx(常用)
proof
    d/dx f(x)g(x) = f(x)g`(x) + f`(x)g(x)

f(x)g(x) = S f(x)dg + S g(x) df 證明
proof 
    S f(x)dg * dx/dx + S g(x) df *dx/dx = S f(x)d/dx g * dx + S g(x) d/dx f * dx = S f(x)g`(x) dx + S g(x)f`(x) dx = f(x)g(x)

S(b,a) f(x)dg = f(x)g(x)](b,a) - S(b,a) g(x) df

用於S sin(x) + cos^(2k+1)(x) dx
S sin^m(x) cos^(2k+1)(x) dx = S u^m * (1-u^2)^k du
proof
    d(sin(x)) = cos(x) dx
    S sin^m(x) cos^(2k+1)(x) dx = S sin^m(x)(1-sin^2(x))^k*cos(x)dx 
    = S sin^m(x) * (1-sin(x)^2)^k d(sin(x))

用於S cos(x) + sin^(2k+1)(x) dx
S cos^m(x) sin^(2k+1)(x) dx = S u^m * (1-u^2)^k du
    proof
    d(cos(x)) = -sin(x) dx
    S cos^m(x) sin^(2k+1)(x) dx = S cos^m(x) *(1-cos^2(x))^k *sin(x) dx = -S cos(x)^m * (1-cos^2(x))^k d(cos(x))

S 2sin(x)cos(x) dx = S sin(2x)dx
sin^2(x) = 1/2*(1-cos(2x))和cos^2(x) = 1/2*(1+cos(2x))
proof
    cos(x+x) = cos^2(x) - sin^2(x) = 1-2sin^2(x)
    so,2*sin^2(x) = (1-cos(2x))

S tan^m(x)*sec^2k(x) dx

~~~~let x=a*sin(theta) ,dx = a*cos(theta) * d(theta)
(a^2 - x^2)^(1/2) = a*cos(theta)  for -pi/2<=theta<=pi/2
note:
此處的dx=...是為了把積分的dx替換掉
最後反導數結果會出現theta,可以使用反三角函數表達theta，表達方式是通過x=a*???(theta)變為???(theta) = x/a,然後畫三角
proof for dx = a*cos(theta) * d(theta)
    d/d(theta) x = d/d(theta) a*sin(theta)
    d * x = a*cos(theta) * d(theta)
    so, dx = a*cos(theta) * d(theta)

proof for (a^2 - x^2)^(1/2) = a*cos(theta)
    (a^2 - x^2)^(1/2) = (a^2 - (a*sin theta)^2)^(1/2) = (a^2(1 - sin^2theta))^(1/2) = a * cos(theta) by 1=cos^2theta+sin^2theta
問題:dx/d(theta) 嚴謹上是不可拆分的?(chatgpt)
在嚴謹的數學微積分領域中，導數表示為 dx/d(theta)，這裡 dx 表示函數 x 隨變數 theta 的變化量，而 d(theta) 表示變數 theta 的一個無窮小變化。導數 dx/d(theta) 描述了當 theta 變化一個微小量時，函數 x 隨之變化的率。

在這個表示中，d 來自拉丁語 differencia 或 差異，意味著變化的量。導數符號中的斜線 / 雖然看起來像是除法操作，但實際上，整個 dx/d(theta) 符號指的是一個極限過程，即當 theta 的變化趨近於零時，x 的變化量 dx 與 theta 的變化量 d(theta) 的比率。

因此，儘管符號中包含除法的形式，dx/d(theta) 應被視為一個不可分割的單元，它代表了變化率，而不是實際的除法操作。在某些數學和物理的應用中，為了解決特定問題，可能會在計算過程中操作這些微分，比如通過分離變數法來解微分方程時將它們分開，但這種操作基於微積分的嚴格原理，並遵循特定的規則。儘管如此，在概念上，dx/d(theta) 依然是一個表示變化率的整體。

If
    let x=a*tan(theta)
    -pi/2<theta< pi/2
then
    dx = a*sec^2(theta) * d(theta)
    (a^2 + x^2)^(1/2) = a * sec(theta) for 
proof
    (a^2 + (a*tan theta)^2)^(1/2) = (a^2(1 + tan^2theta))^(1/2) = a * sec(theta) by (1 +tan^2theta = sec^2theta)

If
    let x=a*sec(theta)
    0<=theta<pi/2 or pi<=theta<3pi/2
then
    dx = a*sec(theta)*tan(theta) * d(theta)
    S (-a^2 + x^2)^(1/2) dx = a * tan(theta) 
proof 
    (-a^2 + x^2)^(1/2) = (a^2(-1 + sec theta^2))^(1/2) = a * tan(theta by( sec^2theta -1 = tan^2theta)

積分奇技淫巧區2號區
1.多項式長除法，發生在分子子數很高時
S (2x^3 + 3x^2 + 4x + 5 )/(x+3) dx = 2x^2 - 3x + 13 - 34/(x+3)
(2x^3 + 3x^2 + 4x + 5 )/(x+3) = 2x^2 - 3x + 13 ... -34

2.多項式分解，發生在分母很多(x-1)*(x-2)*(x-3)^2*(x-4)^2時
S k/((x-1)*(x-2)*(x-3)^2*(x-4)^2) dx = S A/(x-1) + B/(x-2) +  C/(x-3) + D(x-3)^2 + E/(x-4) + F/(x-5)^2

3.自增替換，發生在有更號時
S 1/(x+2)^(1/2) dx = S 1/(u^2 -2 +2)^(1/2) * 2u du = S 2u/u du
u=(x+2)^(1/2), u^2 = x+2, x=u^2-2, dx/du = 2u, dx = 2u * du 

4.強湊du，發生在某個東西的微分剛好滿足能夠湊出du時
S x/(x^2-x+2) dx = S (2x -1 + 1)/2/(x^2-x+2) dx = 1/2 * S 1/u du + 1/2 * S (1/ (x^2-x+2)dx

5.在分母的二項式強行轉為1/(x^2+1)後變為arctan，發生在1/(x+a)時

6. S f(x)g`(x) dx = f(x)g(x) - Sf`(x)g(x)dx，讓g`(x)=1, g(x)=x,再計算f`(x)...，發生在ln(fuck)時

====三角反函數
sinx=y and -pi/2 <= x <=pi/2
sin^(-1)y=x
ps:這些範圍限定是為了使函數one-to-one
ps:sin^-1的改動在於y變成角度 x變成數值

theta = arcsin 1/2 => sin (theta) = 1/2

sin^-1(sinx) =x for -pi/2 <= x <=pi/2
sin(sin^-1 x) =x for -1 <= x <=1

以下兩公式說明了cosy=(1-x^2)^(1/2)
d/dx (sin^-1 x = y for -pi/2 <= x <=pi/2 ) =>d/dx sin^-1  = 1/cosy
proof
    d/dx siny = d/dx x => 1 = d/dy dy/dx siny => 1= cosy * dy/dx =>dy/dx = 1/cosy

let y=sin^-1 x, d/dx sin^-1 x = 1/(1-x^2)^(1/2) for -1<x<1
proof
    y = sin^-1 => x = sin y 
    cos y = (1-sin^2 y)^(1/2) = (1-x^2)^(1/2)
    so,d/dx sin^-1 x = 1/(1-x^2)^1/2

cos^-1 x =y => (cos y = x for 0<=y<=pi)

cos^-1(cos(x)) = x for 0<=x<=pi,domain [–1, 1],range [0, pi]
cos(cos^-1(x)) = x for -1<=x<=1

d/dx cos^-1(x) = -1/(1-x^2)^(1/2) for -1<x<1
proof
    cos^-1(x) =y => cos(y) = x
    
    d/dx cos(y) = d/dx x => dy/dx d/dy cos(y) = 1
    d/dx * cos^-1(x)  = -1/(sin y)
    siny = (1-cosy)^(1/2) = (1-x)^(1/2)
    so, d/dx * cos^-1(x)  = -1/(1-x)^(1/2)

lim(x->infinite)tan^-1 x = pi/2
lim(x->-infinite)tan^-1 x = -pi/2

d/dx tan^-1x= 1/(1+x^2)
    y=tan^-1 x => x=tan y
    d/dx x = d/dx tan y => 1 = dy/dx d/dy tany =>1 = dy/dx sec(y^2)
    d/dx tan^-1 x = 1/sec(y^2) = 1/(1+tan^2y) = 1/(1+x^2)
    
csc^-1x sec^-1x cot^-1x。這三個使用上沒這麼常用
d/dx csc^-1x = -1/(x(x^2-1)^1/2)
d/dx sec^-1x = 1/(x(x^2-1)^1/2)
d/dx cot^-1x = 1/(1+x^2)
對反三角函數微分總結，反三角函數是用來表示角度如theta
d/dx sin^-1x = 1/(1-x^2)^(1/2)
d/dx cos^-1x = -1/(1-x^2)^(1/2)
d/dx sec^-1x = 1/(|x| * (x^2-1)^1/2)
d/dx csc^-1x = -1/(|x| * (x^2-1)^1/2)
d/dx tan^-1x = 1/(1+x^2)
d/dx cot^-1x = -1/(1+x^2)
對反三角積分
這兩個是正確的
S1/(1-x^2)^(1/2) dx = sin^-1x +C
S1/(1+x^2) dx = tan^-1x +C
S1/(1-x^2)^(1/2) dx = cos^-1x +C
S1/(|x| * (x^2-1)^1/2) dx = sec^-1x +C 
S-1/(|x| * (x^2-1)^1/2) dx = csc^-1x +C
S-1/(1+x^2) dx = cot^-1x +C

S1/(x^2+a^2)dx = 1/atan^-1(x/a)+C , (少見的反函數應用)
proof
    d/dx 1/a*tan^-1(x/a)+C = 1/a * (1/(1+(x/a)^2)) *d/dx x/a = 1/a^2(1+(x/a)^2) = 1/(a^2+x^2)

====其他

Trapezoidal Rule
Midpoint Rule
Simpson's Rule

====Approximate Integration

S(b, a) f(x) dx ~= sum(n,i=1) f(x(i*) * delta x
x(i*) is any point in the [x(i-1), x(i)], i is 1 to n,i ∈ ℕ,range of f(x) is [x(0), x(n)]

L(n)
For left endpoint approximate
x(i*) = x(i-1), i is 0 to n
S(a,b) f(x)dx ~= sum(n,i=1) f(x(i*) * delta x

R(n)
For right endpoint approximate
x(i*) = x(i), i is 0 to n
S(a,b) f(x)dx ~= sum(n,i=1) f(x(i*) * delta x

M(n)
For midpoint rule endpoint approximate
delta x  = (b-a)/n
x(i*) = 1/2(x(i-1) + x(i)), i is 1 to n
S(a,b) f(x)dx ~= sum(n,i=1) f(x(i*) * delta x

K >= | f``(x) |, for x is a to b
|E(M)| <= ( K(b-a)^3 )/( 24n^2 )

T(n)
For Trapezoidal Rule endpoint approximate
delta x  = (b-a)/n
x(i*) = a + i * delta x
S(a,b) f(x)dx ~= T(n) = delta x/2 * [ f(x0) + 2f(x1) + 2f(x2) + ... + 2f(n-1) +f(n) ]
proof
    通過梯形(上底+下底)*高/2，而非相對傳統之長方型。

K >= | f``(x) |, for x is a to b
|E(T)| <= ( K(b-a)^3 )/( 12n^2 )

S(n)
For Simpson's Rule endpoint approximate
delta x  = (b-a)/n, n is even
S(a,b) f(x)dx ~= S(n) = delta x /3 * [f(x0) + 4f(x1) + 2f(x2) + 4f(x3) + ... + 2f(x(n-2)) + 4f(x(n-1)) + f(x(n))]

K >= | f````(x) |, for x is a to b
|E(S)| <= ( K(b-a)^5 )/( 180n^4 )

====improper integral
==infinite intervals
S(infinite, a) f(x) dx = lim(t->infinite) S(t,a) f(x) dx

S(a, -infinite) f(x) dx = lim(t->-infinite) S(t,a) f(x) dx

S(infinite, -infinite) f(x) dx = S(infinite, a) f(x) dx + S(a, -infinite) f(x) dx =  lim(t->infinite) S(t,a) f(x) dx + lim(t->-infinite) S(t,a) f(x) dx

==discontinuous integrands
If 
    f is continuous on [a, b)
    discontinuous at b
then
    S(b, a) f(x) dx = lim(t->b^(-)) S(t, a) f(x) dx

If
    f is continuous on (a, b]
    discontinuous at b
then
    S(b, a) f(x) dx = lim(t->a^(+)) S(b, t) f(x) dx

If
    f is discontinuity at c, where a < c < b
    S(b, c) f(x) dx, S(c, a) f(x) dx are convergent
then
    S(b, a) f(x) dx = S(b, c) f(x) dx + S(c, a) f(x) dx

==Comparison test
If
    f(x) and g(x) are continuous functions
    f(x) >= g(x) >= 0, where x >= a
then
    If
        S(infinite, a) f(x) dx is convergent
    then
        S(infinite, a) g(x) dx is convergent
    else if
        S(infinite, a) g(x) dx is divergent
    then
        S(infinite, a) f(x) dx is divergent

arc length
L = lim(n->infinite) sum(n,i=1) | P(i-1) * P(i) |
L = S(b, a) sqrt( 1 + ( d/dx f(x) )^2 ) dx
L = S(b, a) sqrt( 1 + ( dy/dx )^2 ) dx