User:Egm6321.f09.Team1/HW7

From Wikiversity
Jump to: navigation, search

Contents

Problem 1: Odd and Even Solutions of Legendre DE [edit]

From Lecture Slide 37-1.

Given [edit]

From lecture slide 31-3, general expression for Legendre Polynomial is given by,

\displaystyle 
P_n(x) = \sum_{i=0}^{\lfloor n/2 \rfloor}
\frac{1 \cdot 3 \dots \left(2n - 2i - 1\right)}{2^i i! (n-2i)!} (-1)^i x^{n-2i},

\displaystyle (Eq. 1)

and from lecture slide 37-1, the general expression for the second solution of Legendre Differential Equation is given as,

\displaystyle 
Q_n(x) = P_n(x) \tanh^{-1}(x) - 2 \sum_{j=1,3,5...}^{J}
\frac{2n-2j+1}{(2n-j+1)j} P_{n-j}(x),

\displaystyle (Eq. 2)

where \displaystyle J is given by,

\displaystyle 
J := 1 + 2 \left\lfloor \frac{n-1}{2} \right\rfloor .

\displaystyle (Eq. 3)

Find [edit]

Using Equation (2), to show when \displaystyle Q_n is even or odd, depending on whether "\displaystyle n " is even or odd.

Solution [edit]

We know, from Odd and Even property of \displaystyle P_n , that \displaystyle P_n is odd, when \displaystyle n is odd and even, when \displaystyle n is even and the fact that

\displaystyle 
\tanh^{-1}(x),

\displaystyle (Eq. 4)

is odd.

(i). When \displaystyle n is odd.

When \displaystyle n is odd, \displaystyle P_n is odd and so,

\displaystyle 
P_n(x)\tanh^{-1}(x),

is even.

So, when \displaystyle n is odd, first term in Equation (2) is even.

When \displaystyle n is odd, value of \displaystyle n - j is even since \displaystyle j values are odd, and so, \displaystyle P_{n-j} is even.

So, when \displaystyle n is odd, all the terms in the summation of Equation (2) are even.

And so,

\displaystyle \Rightarrow So, when \displaystyle n is odd, \displaystyle Q_n is even.

(ii). When \displaystyle n is even.

When \displaystyle n is even, \displaystyle P_n is even and so,

\displaystyle 
P_n(x)\tanh^{-1}(x),

is odd.

So, when \displaystyle n is even, first term in Equation (2) is odd.

When \displaystyle n is even, value of \displaystyle n - j is odd since \displaystyle j values are odd, and so, \displaystyle P_{n-j} is odd.

So, when \displaystyle n is even, all the terms in the summation of Equation (2) are odd.

And so,

\displaystyle \Rightarrow So, when \displaystyle n is even, \displaystyle Q_n is odd.

Problem 2: Legendre Solutions Plot [edit]

From Lecture Slide 37-1.

Given [edit]

From lecture slide 31-3, general expression for Legendre Polynomial is given by,

\displaystyle 
P_n(x) = \sum_{i=0}^{\lfloor n/2 \rfloor}
\frac{1 \cdot 3 \dots \left(2n - 2i - 1\right)}{2^i i! (n-2i)!} (-1)^i x^{n-2i} ,

\displaystyle (Eq. 1)

and from lecture slide 37-1, the general expression for the second solution of Legendre Differential Equation is given as,

\displaystyle 
Q_n(x) = P_n(x) \tanh^{-1}(x) - 2 \sum_{j=1,3,5...}^{J}
\frac{2n-2j+1}{(2n-j+1)j} P_{n-j}(x),

\displaystyle (Eq. 2)

where \displaystyle J is given by,

\displaystyle 
J := 1 + 2 \left\lfloor \frac{n-1}{2} \right\rfloor .

\displaystyle (Eq. 3)

Find [edit]

Plot {\displaystyle P_0, P_1, P_2, P_3, P_4} and {\displaystyle Q_0, Q_1, Q_2, Q_3, Q_4} using Matlab.

Solution [edit]

Please refer Problem 6 Solution of HW6, for \displaystyle 
P_0(x),\ P_1(x),\ P_2(x),\ P_3(x),\ P_4(x)
.

Please refer [1], p. 33, for expressions of \displaystyle Q_0(x), Q_1(x), Q_2(x), Q_3(x).

The expression for \displaystyle Q_4(x)is given by,

\displaystyle 
Q_4= P_4(x) \tanh^{-1}(x) - 2 \sum_{j=1,3}^{}
\frac{2n-2j+1}{(2n-j+1)j} P_{n-j}(x)
\displaystyle 
\Rightarrow Q_4= \left[\frac{35}{8} x ^ 4 - \frac{15}{4} x ^ 2 + \frac{3}{8}\right] \tanh^{-1}(x) - 2 \left[ \frac{2(4)-2+1}{(2(4)-1+1)1} P_{3}(x) \right] - 2 \left[ \frac{2(4)-2(3)+1}{(2(4)-3+1)3} P_{1}(x) \right]
\displaystyle 
\Rightarrow Q_4= \left[\frac{35}{8} x ^ 4 - \frac{15}{4} x ^ 2 + \frac{3}{8}\right] \tanh^{-1}(x) - \frac{7}{8} \left[ 5 x ^ 3 - 3 x\right] -  \frac{1}{3} x
.

Matlab Code:

clear all;
x = -1:0.01:1;
P_0 = 1*ones(1,length(x));
P_1 = x;
P_2 = (1/2)*(3*x.^2 - 1);
P_3 = (1/2)*(5*x.^3 - 3*x);
P_4 = (35/8)*x.^4 - (15/4)*x.^2 + (3/8);
Q_0 = atanh(x);
Q_1 = x.*atanh(x) - 1;
Q_2 = (1/2)*(3*x.^2 - 1).*atanh(x) - (3/2)*x;
Q_3 = (1/2)*(5*x.^3 - x).*atanh(x) - (5/2)*x.^2 + (2/3);
Q_4 = ((35/8)*x.^4 - (15/4)*x.^2 + (3/8)).*atanh(x) -(7/8)*(5*x.^3 - 3*x) - (1/3)*x;
figure(1);
plot(x,P_0,'--r',x,P_1,'--b',x,P_2,'--g',x,P_3,'--c',x,P_4,'--m');
legend('P_0','P_1','P_2','P_3','P_4');
figure(2);
plot(x,Q_0,'--r',x,Q_1,'--b',x,Q_2,'--g',x,Q_3,'--c',x,Q_4,'--m');
legend('Q_0','Q_1','Q_2','Q_3','Q_4');

Plots:

The plots for \displaystyle P_n(x) with \displaystyle n = 0,1,2,3,4 are shown below:

File:Prob2-P(x).jpg.

The plots for \displaystyle Q_n(x) with \displaystyle n = 0,1,2,3,4 are shown below:

File:Prob2-Q(x).jpg.

Problem 3: Orthogonality of Legendre Solutions [edit]

From Lecture Slide 37-2.

Given [edit]

From lecture slide 31-3, general expression for Legendre Polynomial is given by,

\displaystyle 
P_n(x) = \sum_{i=0}^{\lfloor n/2 \rfloor}
\frac{1 \cdot 3 \dots \left(2n - 2i - 1\right)}{2^i i! (n-2i)!} (-1)^i x^{n-2i} ,

\displaystyle (Eq. 1)

and from lecture slide 37-1, the general expression for the second solution of Legendre Differential Equation is given as,

\displaystyle 
Q_n(x) = P_n(x) \tanh^{-1}(x) - 2 \sum_{j=1,3,5...}^{J}
\frac{2n-2j+1}{(2n-j+1)j} P_{n-j}(x),

\displaystyle (Eq. 2)

where \displaystyle J is given by,

\displaystyle 
J := 1 + 2 \left\lfloor \frac{n-1}{2} \right\rfloor

\displaystyle (Eq. 3)

Find [edit]

Show that, \displaystyle < P_n, Q_n >  = 0.

Solution [edit]

From lecture slide 33-1, the inner (scalar) product is given by,

\displaystyle 
<G(x),F(x)> = \int_{-1}^{+1} G(x)F(x)dx,

\displaystyle (Eq. 4)

and Equation (4) becomes zero, when \displaystyle G(x) is even and \displaystyle F(x) is odd or when \displaystyle G(x) is odd and \displaystyle F(x) is even.

For \displaystyle P_n(x) and \displaystyle Q_n(x) using Equation (4),

\displaystyle 
<P_n(x),Q_n(x)> = \int_{-1}^{+1} P_n(x)Q_n(x)dx.

\displaystyle (Eq. 5)

From Solution to Problem 1 above, \displaystyle Q_n(x) is odd and \displaystyle P_n(x) is even, when \displaystyle n is even. When \displaystyle n is odd, \displaystyle Q_n(x) is even and \displaystyle P_n(x) is odd.

So, for all values of \displaystyle n , \displaystyle Q_n(x) is opposite to \displaystyle P_n(x) in oddness and evenness, which means,

\displaystyle 
 \int_{-1}^{+1} P_n(x)Q_n(x)dx = 0.

And so,

\displaystyle 
<P_n(x),Q_n(x)> = 0

Problem 4: Integration by Parts: Legendre Solutions [edit]

From lecture slide 37-3.

Given [edit]

From lecture slide 37-2,

\displaystyle
\alpha = \int_{-1}^{+1} L_m \left[ (1-x^2) L'_n \right]' dx

\displaystyle (Eq. 1)

where \displaystyle L_n and \displaystyle L_m are solutions to the Legendre differential equation.

Find [edit]

Show that

\displaystyle
\alpha = -\int_{-1}^{+1} (1-x^2) L'_n L'_m dx,

\displaystyle (Eq. 2)

after integration by parts of \displaystyle \alpha in Equation (1).

Solution [edit]

Let,

\displaystyle
u = L_m \Rightarrow du = L'_m dx,

\displaystyle (Eq. 3)

and,

\displaystyle
dv = \left[ (1-x^2) L'_n \right]' dx.

\displaystyle (Eq. 4)

\displaystyle
\Rightarrow \int dv = \int\left[ (1-x^2) L'_n \right]' dx \Rightarrow v = \left[ (1-x^2) L'_n \right].

We know,

\displaystyle
\int_{a}^{b} udv = \left[uv \right]_{a}^{b} - \int_{a}^{b} vdu ,

\displaystyle (Eq. 5)

and so, with \displaystyle a = -1 and \displaystyle b = +1

\displaystyle
\alpha = \int_{-1}^1 L_m \left[ (1-x^2) L'_n \right]' dx = \left[ L_m(1-x^2) L'_n\right]_{-1}^{+1} - \int_{-1}^{+1} (1-x^2) L'_n L'_m dx ,

\displaystyle (Eq. 6)

For values of \displaystyle x = -1 and \displaystyle x = +1 ,

\displaystyle
(1-x^2) = 0.

\displaystyle (Eq. 7)

Substituting Equation (7) in first term of Equation (6), first term becomes zero.

\displaystyle
\alpha = \cancelto{0}{\left[ L_m(1-x^2) L'_n\right]_{-1}^{+1}} - \int_{-1}^{+1} (1-x^2) L'_n L'_m dx.

So,

\displaystyle
\alpha = - \int_{-1}^{+1} (1-x^2) L'_n L'_m dx

Problem 5: Attraction of Spheres [edit]

From lecture slide 38-2.

Given [edit]

From lecture slide 38-2,

\displaystyle
(r_{PQ})^2 = \sum_{i=1}^{3} \left(x_Q^i - x_P^i\right)^2,

\displaystyle (Eq. 1)

where,

\displaystyle x_P^1 = x_P,\ x_P^2 = y_P,\ x_P^3 = z_P,\

and,

\displaystyle x_Q^1 = x_Q,\ x_Q^2 = y_Q,\ x_Q^3 = z_Q,\

given by,

\displaystyle
x_P = r_P \cos(\theta_P) \cos(\psi_P);\  
y_P = r_P \cos(\theta_P) \sin(\psi_P);\ 
z_P = r_P \sin(\theta_P),

\displaystyle (Eq. 2)

and,

\displaystyle
x_Q = r_Q \cos(\theta_Q) \cos(\psi_Q);\  
y_Q = r_Q \cos(\theta_Q) \sin(\psi_Q);\ 
z_Q = r_Q \sin(\theta_Q).

\displaystyle (Eq. 3)

Find [edit]

Show that,

\displaystyle
(r_{PQ})^2 = (r_P)^2 + (r_Q)^2 - 2 (r_P) (r_Q) \cos\gamma,

\displaystyle (Eq. 4)

where,

\displaystyle
\cos\gamma = \cos(\theta_Q) \cos(\theta_P) \cos(\psi_Q-\psi_P) + \sin(\theta_Q)\sin(\theta_P)

\displaystyle (Eq. 5)

Solution [edit]

Equation (1) can be written as,

\displaystyle
(r_{PQ})^2 =  \left(x_Q - x_P\right)^2 + \left(y_Q - y_P\right)^2 + \left(z_Q - z_P\right)^2.

\displaystyle (Eq. 6)

Substituting Equations (2) and (3) in (6),

\displaystyle
\begin{align}
\Rightarrow (r_{PQ})^2 =  & \left[r_Q \cos(\theta_Q) \cos(\psi_Q) - r_P \cos(\theta_P) \cos(\psi_P)\right]^2 + \\
                          & \left[r_Q \cos(\theta_Q) \sin(\psi_Q) - r_P \cos(\theta_P) \sin(\psi_P)\right]^2 + \\
                          & \left[r_Q \sin(\theta_Q) - r_P \sin(\theta_P)\right]^2
\end{align}
\displaystyle
\begin{align}
\Rightarrow (r_{PQ})^2 =  & \left[r_Q^2 \cos^2(\theta_Q) \cos^2(\psi_Q) + r_P^2 \cos^2(\theta_P) \cos^2(\psi_P) - 2 r_P r_Q \cos(\theta_Q) \cos(\psi_Q) \cos(\theta_P) \cos(\psi_P) \right] + \\
                          & \left[r_Q^2 \cos^2(\theta_Q) \sin^2(\psi_Q) + r_P^2 \cos^2(\theta_P) \sin^2(\psi_P) - 2 r_P r_Q \cos(\theta_Q) \sin(\psi_Q) \cos(\theta_P) \sin(\psi_P)\right] + \\
                          & \left[r_Q^2 \sin^2(\theta_Q) + r_P^2 \sin^2(\theta_P) - 2 r_P r_Q \sin(\theta_Q) \sin(\theta_P)\right]
\end{align}
\displaystyle
\begin{align}
\Rightarrow (r_{PQ})^2 =  &\ r_Q^2 \cos^2(\theta_Q) \cancelto{1}{\left[ \cos^2(\psi_Q) + \sin^2(\psi_Q) \right]} + r_Q^2 \sin^2(\theta_Q) -  2 r_P r_Q \cos(\theta_Q) \cos(\theta_P) \left [\cos(\psi_Q) \cos(\psi_P) \right] + \\ 
                            &\ r_P^2 \cos^2(\theta_P) \cancelto{1}{\left[ \cos^2(\psi_P) + \sin^2(\psi_P) \right]} + r_P^2 \sin^2(\theta_P) - 2 r_P r_Q \cos(\theta_Q)\cos(\theta_P) \left [\sin(\psi_Q) \sin(\psi_P)\right] - 2 r_P r_Q \sin(\theta_Q) \sin(\theta_P)
\end{align}
\displaystyle
\begin{align}
\Rightarrow (r_{PQ})^2 =  &\ r_Q^2 \cancelto{1}{\left[\cos^2(\theta_Q) + \sin^2(\theta_Q) \right]} -  2 r_P r_Q \cos(\theta_Q) \cos(\theta_P) \underbrace{\left [\cos(\psi_Q) \cos(\psi_P) + \sin(\psi_Q) \sin(\psi_P)\right]}_{=\ \cos(\psi_Q-\psi_P)} + \\ 
                            &\ r_P^2 \cancelto{1}{\left[ \cos^2(\theta_P) + \sin^2(\theta_P) \right]} - 2 r_P r_Q \sin(\theta_Q) \sin(\theta_P)

\end{align}
\displaystyle
\begin{align}
\Rightarrow (r_{PQ})^2 =  &\ r_Q^2 + r_P^2 -  2 r_P r_Q \left[\underbrace{\cos(\theta_Q) \cos(\theta_P) {\cos(\psi_Q-\psi_P)} + \sin(\theta_Q)\sin(\theta_P)}_{=\ \cos\gamma} \right]
\end{align}

And so,

\displaystyle
(r_{PQ})^2 = (r_P)^2 + (r_Q)^2 - 2 (r_P) (r_Q) \cos\gamma

Problem 6: Binomial Series [edit]

From lecture note slide 38-4.

Given [edit]

The binomial series expansion is

\displaystyle
(x+y)^r = \sum_{k=0}^\infty
\begin{pmatrix}
r \\ k
\end{pmatrix}
x^{r-k} y^k

\displaystyle (Eq. 1)

where

\displaystyle
\begin{pmatrix}
r \\ k
\end{pmatrix}
=
\frac{r(r-1) \cdot\cdot\cdot (r-k+1)}{k!}

\displaystyle (Eq. 2)

Find [edit]

Use Eqs. 1 and 2 to show that

\displaystyle
(1-x)^{-1/2} = \sum_{i=0}^\infty \alpha_i x^i

\displaystyle (Eq. 3)

where

\displaystyle
\alpha_i =
\frac{1 \cdot 3 \cdot ... \cdot (2i - 1)}{2 \cdot 4 \cdot ... \cdot (2i)}

\displaystyle (Eq. 4)

Solution [edit]

Using Eq. 3, we can rewrite the LHS of Eq. 1 as

\displaystyle
(1-x)^{-1/2} = \sum_{k=0}^\infty
\begin{pmatrix}
-1/2 \\ k
\end{pmatrix}
1^{-1/2-k} (-x)^k

\displaystyle (Eq. 5)

We can expand this

\displaystyle
(1-x)^{-1/2} = \sum_{k=0}^\infty
\frac{(-1/2)(-3/2)\cdot\cdot\cdot(1/2-k)}{k!}
(-x)^k

\displaystyle (Eq. 6)

The number of factors in the numerator is \displaystyle k, hence we can rewrite this as

\displaystyle
(1-x)^{-1/2} = \sum_{k=0}^\infty
\frac{(1)(3)\cdot\cdot\cdot(2k-1)}{(-2)^k k!}
(-x)^k

\displaystyle (Eq. 7)

Change \displaystyle k \rightarrow i, and cancel two minus signs

\displaystyle
(1-x)^{-1/2} = \sum_{i=0}^\infty
\frac{(1)(3)\cdot\cdot\cdot(2i-1)}{2^i i!}
x^i

\displaystyle (Eq. 8)

The denominator can then be expanded

\displaystyle
(1-x)^{-1/2} = \sum_{i=0}^\infty
\underbrace{\frac{1 \cdot 3\cdot\cdot\cdot(2i-1)}{2 \cdot 4 \cdot \cdot \cdot (2i)}}_{\alpha_i}
x^i

\displaystyle (Eq. 9)

So that

\displaystyle
(1-x)^{-1/2} = \sum_{i=0}^\infty
\alpha_i
x^i

\displaystyle (Eq. 10)

where

\displaystyle
\alpha_i = \frac{1 \cdot 3\cdot\cdot\cdot(2i-1)}{2 \cdot 4 \cdot \cdot \cdot (2i)}

\displaystyle (Eq. 11)

Problem 7: Generating Functon for Legendre Polynomials [edit]

From lecture note slide 39-2.

Given [edit]

We have that

\displaystyle
A(\mu, \rho) :=
1 - 2 \mu \rho + \rho^2

\displaystyle (Eq. 1)

and the binomial expansion

\displaystyle
(1-x)^{-1/2} = \sum_{i=0}^\infty \alpha_i x^i

\displaystyle (Eq. 2)

where

\displaystyle
\alpha_i =
\frac{1 \cdot 3 \cdot ... \cdot (2i - 1)}{2 \cdot 4 \cdot ... \cdot (2i)}

\displaystyle (Eq. 3)

Using Eqs. 1 and 2, we can expand the expression

\displaystyle
[A(\mu,\rho)]^{-1/2}
= \alpha_0
+ \alpha_1(2\mu \rho- \rho^2)
+ \alpha_2(2\mu \rho- \rho^2)^2
+ ...

\displaystyle
= \underbrace{\alpha_0}_{P_0(\mu)}
+ \underbrace{2 \mu \alpha_1}_{P_1(\mu)} \rho
+ \underbrace{(- \alpha_1 + 4 \mu^2 \alpha_2)}_{P_2(\mu)} \rho^2
+ ...

\displaystyle (Eq. 4)

Find [edit]

Continue the expansion given by Eq. 4 to yield \displaystyle P_3, \displaystyle P_4, and \displaystyle P_5, and compare the result to that obtained by Eq. 7 on lecture note slide 31-3

\displaystyle
P_n(x) = \sum_{i=0}^{[n/2]} =
\frac{1 \cdot 3 \cdot ... \cdot (2n - 2i - 1)}{2^i i! (n-2i)!} (-1)^i x^{n-2i}

\displaystyle (Eq. 5)


Solution [edit]

We need the first 6 terms of the expansion

\displaystyle
[A(\mu,\rho)]^{-1/2}
= \alpha_0
+ \alpha_1(2\mu \rho- \rho^2)
+ \alpha_2(2\mu \rho- \rho^2)^2
+ \alpha_3(2\mu \rho- \rho^2)^3
+ \alpha_4(2\mu \rho- \rho^2)^4
+ \alpha_5(2\mu \rho- \rho^2)^5
+ ...

\displaystyle
= \alpha_0

\displaystyle
+ \alpha_1(2\mu \rho - \rho^2)

\displaystyle
+ \alpha_2(4\mu^2 \rho^2 - 4\mu \rho^3 +  \rho^4)

\displaystyle
+ \alpha_3 (8\mu^3 \rho^3 - 12\mu^2 \rho^4 +  6\mu \rho^5 -  \rho^6)

\displaystyle
+ \alpha_4  (16\mu^4 \rho^4 - 32\mu^3 \rho^5 +  24\mu^2 \rho^6 - 8 \mu \rho^7 +  \rho^8)

\displaystyle
+ \alpha_5 (32\mu^5 \rho^5 - 80 \mu^4 \rho^6 +  80 \mu^3 \rho^7 - 40 \mu^2 \rho^8 +  10 \mu \rho^9 -  \rho^{10})
+ ...

\displaystyle
= \alpha_0
+ \alpha_1 \rho

\displaystyle
+ (-\alpha_1 + 4 \alpha_2 \mu^2) \rho^2

\displaystyle
+ (- 4 \alpha_2 \mu + 8 \alpha_3 \mu^3) \rho^3

\displaystyle
+ (\alpha_2 - 12 \alpha_3 \mu^2 + 16 \alpha_4 \mu^4) \rho^4

\displaystyle
+ (6 \alpha_3 \mu - 32 \alpha_4 \mu^3 + 32 \alpha_5 \mu^5) \rho^5
+ ...

\displaystyle (Eq. 6)

Using Eq. 3, we have that

\displaystyle
\alpha_0 = 1 \quad
\alpha_1 = \frac{1}{2} \quad
\alpha_2 = \frac{3}{8} \quad
\alpha_3 = \frac{15}{48} \quad
\alpha_4 = \frac{105}{384} \quad
\alpha_5 = \frac{945}{3840} \quad

we can substitute this into Eq. 6

\displaystyle
= 1
+ \frac{1}{2} \rho

\displaystyle
+ \left(-\frac{1}{2} + \frac{3}{2} \mu^2\right) \rho^2

\displaystyle
+ \left(- \frac{3}{2} \mu + \frac{5}{2} \mu^3\right) \rho^3

\displaystyle
+ \left(\frac{3}{8} - \frac{15}{4} \mu^2 + \frac{35}{8} \mu^4\right) \rho^4

\displaystyle
+ \left(\frac{15}{8} \mu - \frac{35}{4} \mu^3 + \frac{63}{8} \mu^5\right) \rho^5
+ ...

\displaystyle (Eq. 7)

From the above, we can extract

\displaystyle
P_3(\mu) = - \frac{3}{2} \mu + \frac{5}{2} \mu^3

\displaystyle (Eq. 8)

\displaystyle
P_4(\mu) = \frac{3}{8} - \frac{15}{4} \mu^2 + \frac{35}{8} \mu^4

\displaystyle (Eq. 9)

\displaystyle
P_5(\mu) =
\frac{15}{8} \mu - \frac{35}{4} \mu^3 + \frac{63}{8} \mu^5

\displaystyle (Eq. 10)

Using Eq. 5, we get

\displaystyle
P_3(x) = \sum_{i=0}^1
\frac{1 \cdot 3 \cdot ... \cdot (2n - 2i - 1)}{2^i i! (n-2i)!} (-1)^i x^{n-2i}

\displaystyle (Eq. 11)

\displaystyle
P_3(x) =
\frac{1 \cdot 3 \cdot 5}{ 3!}  x^3
-
\frac{1 \cdot 3 }{2} x

\displaystyle (Eq. 12)

\displaystyle
P_3(x) =
\frac{5}{ 2}  x^3
-
\frac{3}{2} x

\displaystyle (Eq. 13)

\displaystyle
P_4(x) = \sum_{i=0}^2
\frac{1 \cdot 3 \cdot ... \cdot (2n - 2i - 1)}{2^i i! (n-2i)!} (-1)^i x^{n-2i}

\displaystyle (Eq. 14)

\displaystyle
P_4(x) =
\frac{1 \cdot 3 \cdot ... \cdot 7}{4!}  x^4
-
\frac{1 \cdot 3 \cdot 5}{4} x^2
+
\frac{1 \cdot 3 }{8}

\displaystyle (Eq. 15)

\displaystyle
P_4(x) =
\frac{35}{8}  x^4
-
\frac{15}{4} x^2
+
\frac{3}{8}

\displaystyle (Eq. 16)

\displaystyle
P_5(x) = \sum_{i=0}^2
\frac{1 \cdot 3 \cdot ... \cdot (2n - 2i - 1)}{2^i i! (n-2i)!} (-1)^i x^{n-2i}

\displaystyle (Eq. 17)

\displaystyle
P_5(x) =
\frac{1 \cdot 3 \cdot ... \cdot 9}{5!} x^5
-
\frac{1 \cdot 3 \cdot ... \cdot 7}{2 \cdot 3!} x^3
+
\frac{1 \cdot 3 \cdot 5}{8} x

\displaystyle (Eq. 18)

\displaystyle
P_5(x) =
\frac{63}{8} x^5
-
\frac{35}{4} x^3
+
\frac{15}{8} x

\displaystyle (Eq. 19)

We have confirmed that the methods are equivalent, at least for terms up to \displaystyle P_5.

Notes and References [edit]

  1. King et al., "Differential Equations (Linear, Nonlinear, Ordinary, Partial)", Cambridge University Press, 2003

Contributing Team Members [edit]

Egm6321.f09.Team1.andy 12:03, 29 November 2009 (UTC)

Egm6321.f09.Team1.sallstrom 18:48, 2 December 2009 (UTC)

Egm6321.f09.Team1.vasquez 05:09, 5 December 2009 (UTC)

Egm6321.f09.Team1.AH 15:59, 8 December 2009 (UTC)