You are on page 1of 1

CALPOTURA, Vea Nicole N.

BS Mechanical Engineering
2014 - 00660

Module 0 Numerical Differentiation


n
> Limits and continuity ‣ Solving for f (x)
|f(x) - ​L ​| < Ɛ, whenever 0 < |x − xo | < : 2 Point Forward Scheme Formula
f (xj+1 )−f (j)
F ′(xj ) = h
+ O(h)
>Differentiability O(h) = − h2 f 2 [E(x)]
f (x)−f (x )
f ′(xo ) = x−xo o
: 3 Point Midpoint (x2 )
f (xj+2 )−f (xj )
F ′(xj ) = + O(h2 )
>Mean Value Theorem 2h
2
f (b)−f (a)
f ′(xo ) = b−a O(h2 ) =− h 3
6
f [E(x)]
: 3 Point Endpoint
−3f (xj )−4f (xj+1 )−f (xj+2 )
>Extreme Value Theorem F ′(xj ) = 2h
+ O(h2 )
2
f (c1 ) ≤ f (x) ≤ f (c2 ) O(h2 ) = h3 f 3 [E(x)]
: 5 Point Midpoint (x3 )
>Intermediate Value Theorem f (xj )−8f (xj+1 )+8f (xj+2 )−f (xj+3 )
F ′(xj ) = 12h
+ O(h4 )
If f 𝝐 C [a,b] and K is any number between f(a) and f(b) then there 4

exist a number c 𝝐 (a,b) for which f(c)=k O(h4 ) =− 30 h 5


f [E(x)]

f(a) ≤ f(c) = k ≤ f(b) Numerical Integration


‣ ​Solving for ∫ f (x)dx
>Taylor’s Polynomial (Order n)
f (x) = P n (x) + Rn (x) : Composite Midpoint
f ′ (a) f ′′ (a) 2 f n (a) n At bounds a, b, a < b
P n (x) = f (a) + 1!
(x − a) + 2!
(x − a) + ... n!
(x − a) n
f n+1 (a) b−a xj +xj+1
n+1 h = ; I = ∑ h f( )
Rn (x) = (n+1)!
(x − a) n 2
j=1
: Trapezoidal
Module 1 n−1
‣ f (x) = 0 [Soln to Eqs w/1 Variable] I = ∑ ( h2 )(f i + f i+1 )
i=0
: Fixed Point Iteration Method
: Simpson’s Rule
Rewrite f (x) to x = g(x) , start with a number x
6 [f (a) + 4f ( 2 ) + f (b)]
I = b−a a+b
xk+1 = g (xk ) where ||g | (x)|| < 1 h
Ĩ = 3 (y 1 + 4y 2 + 2y 3 + 4y 4 + ... + y n ) , if given data points
: Bisection Method (Interval Halving)
a0 = a, b0 = b
0 0
x′ = a +b
2
, f (a0 )f (x′) < 0
b0 = x| , else ao = x|
: Newton-Raphson Method ‣ Matrix Solving
f (x )
xk+1 = xk − f ′(xk ) , where k is the iteration : Jacobi Method
k
N
: Secant Method bi
f (x )(x −x ) xi k+1 = −1
Aii
∑ Aij + Aii
xk+1 = xk − f (xk )−fk (x k−1) , where k is the iteration j=1,j≠ 1
k k−1
: Regula Falsi (Method of False Position) : Gauss Seidel Method
Applicable only if f(x) passes IVT Same with Jacobi but use forward subs i.e subs xk=0 = {0, 0, ..., 0}
Assume a and b is near root to Eq. 1, to get x1k=1 , then subs xk=0 = {x1k=1 , 0, ..., 0} to Eq. 2
f (a) [(b−a)]
c = a − f (b)*− f (a) , f (a)f (c) < 0 to get x2k=1 , repeat till last equation. Repeat for k = 1, 2, ..., n
b = c, else a = c : Gauss Seidel Method with Relaxation
Solve as with GS, but rewrite the equations as:
‣ f (x) = polynomial ​(Interpolation & Fxn Approximation) xn k+1 = (1 − w)(xn k ) + w* f n (x1 k, x2 k, x3 k, ..., xn k)
: Vandermonde Polynomial SOR (over relaxation) w > 1 & SUR (under relaxation) w < 1
xno an + xon−1 an−1 + ... + xo a1 + ao = y o = P v
Solve for the a using matrices: X a = y
: Lagrange Polynomials
f (x0 )(x−x1 )(x−x2 )...(x−xn ) f (x )(x−x )(x−x2 )...(x−xn )
(x −x )(x −x )...(x −xn )
+ (x 1−x )(x0 −x )...(x −xn )
+...
0 1 0 2 0 1 0 1 2 1

: Newton’s Interpolating Divided Differences


n+1
P n (x) = f [x1 ] + ∑ F [x1 , x2 , x3 , ..., xk ](x − x1 )(x − x2 )...(x − xk−1 )
k=2
f [xj+1 ,xj+2 ,...,xj+k ]−f [xj ,xj+1 ,...,xj+k ]
F [xj , xj+1 , ..., xj+k ] = xj+k −xj

ME 91 Midterms Formula Sheet

You might also like