Answer to Question #175524 in Differential Equations for Arsene Dongmo

Question #175524

1) Based on the finite difference method, what is the advanced scheme of the first, second and third derivatives with second order accuracy?


2) Consider the following equation:


d3y/dt3+2d2y/dt2+4dy/dt+8y=te2t


Where y(0) = 0 and y(3) = 28.2

By using the results from the session (1), solve the equation for t E [0 ; 3], with delta(t) = 0.3



1
Expert's answer
2021-04-29T16:25:00-0400

1). We remind that the finite difference method is a class of numerical techniques for solving differential equations by approximating derivatives with finite differences. In order to achieve the second order accuracy, we consider the Taylor series expansion:

"f(x_0+h)=f(x_0)+\\frac{f'(x_0)}{1!}h+\\frac{f''(x_0)}{2!}h^2+\\frac{f'''(x_0)}{3!}h^3+R_n(x),"

where "R_n(x)" denotes the reminder term. Another expansion is:

"f(x_0-h)=f(x_0)-\\frac{f'(x_0)}{1!}h+\\frac{f''(x_0)}{2!}h^2-\\frac{f'''(x_0)}{3!}h^3+\\tilde{R}_n(x),"

where "\\tilde{R}_n(x)" is another reminder term. From the latter we find: "f(x_0+h)-f(x_0-h)=2f'(x_0)h+2h^3\\frac{f'''(x_0)}{3!}\n+R_n(x)-\\tilde{R}_n(x)"

and: "f(x_0+h)+f(x_0-h)=2f(x_0)+f''(x_0)h^2+R_n(x)+\\tilde{R}_n(x)"

The latter yields: "f'(x_0)=\\frac{f(x_0+h)-f(x_0-h)}{2h}-h^2\\frac{f'''(x_0)}{3!}\n-\\frac{(R_n(x)-\\tilde{R}_n(x))}{2h}"; "f''(x_0)=\\frac{f(x_0+h)-f(x_0-h)-2f(x_0)}{h^2}-\\frac{(R_n(x)+\\tilde{R}_n(x))}{h^2}".

Thus, the approximate expressions are "f'(x_0) \\approx\\frac{f(x_0+h)-f(x_0-h)}{2h}" and "f''(x_0)\\approx\\frac{f(x_0+h)+f(x_0-h)-2f(x_0)}{h^2}".

We express "f'''(x_0)" as: "f'''(x_0)\\approx\\frac{f''(x_0+h)-f''(x_0-h)}{2h}\\approx\\frac{1}{2h}\\left(\\frac{f(x_0+2h)+f(x_0)-2f(x_0+h)}{h^2}-\\frac{f(x_0)+f(x_0-2h)-2f(x_0-h)}{h^2}\\right)=\\frac{f(x_0+2h)-2f(x_0+h)+2f(x_0-h)-f(x_0-2h)}{2h^3}"

2). We can substitute the approximations directly and receive the system for the values "f(x_0-2h)", "f(x_0-h)", "f(x_0)", "f(x_0+h)", "f(x_0+2h)" at different points. On the other hand, we can solve the system directly. Namely, we substitute the expression "y=ce^{\\lambda x}", "c,\\lambda\\in{\\mathbb{C}}" into the homogenuous equation and receive: "ce^{\\lambda x}(\\lambda^3+2\\lambda^2+4\\lambda+8)=0". The roots are: "\\lambda=-2,2I,-2I". Thus, the general solution of the homogenuous system is: "y=c_1e^{-2x}+c_2e^{2Ix}+c_3e^{-2Ix}", where "c_1,c_2,c_3\\in{\\mathbb{C}}" . We check, whether the following function "y_p=\\alpha xe^{2x}+\\beta e^{2x}", "\\alpha,\\beta\\in{\\mathbb{C}}", is a solution of the initial equation. The derivatives has the form: "y_p'=2\\alpha x e^{2x}+(\\alpha+2\\beta) e^{2x}", "y_p''=4\\alpha x e^{2x}+(4\\alpha+4\\beta) e^{2x}", "y_p'''=8\\alpha x e^{2x}+(12\\alpha+8\\beta) e^{2x}". We substitute it into equation and receive: "(32\\alpha x+(12\\alpha+8\\alpha+4\\alpha+8\\beta+8\\beta+16\\beta))e^{2x}=(32\\alpha x+(24\\alpha+32\\beta))e^{2x}=xe^{2x}"

From the latter we receive: "\\alpha=\\frac{1}{32}", "\\beta=-\\frac{3}{4}\\alpha=-\\frac{3}{128}" Thus, "y_p=\\frac{1}{128}(4x-3)e^{2x}". The solution (without initial conditions) has the form: "y=c_1e^{-2x}+c_2e^{2Ix}+c_3e^{-2Ix}+\\frac{1}{128}(4x-3)e^{2x}". We set "y(0)=0" and "y(3)=28.2" and receive a system for coefficients "c_1,c_2,c_3." Thus, we get the solution of the original system with initial conditions.


Need a fast expert's response?

Submit order

and get a quick answer at the best price

for any assignment or question with DETAILED EXPLANATIONS!

Comments

No comments. Be the first!

Leave a comment

LATEST TUTORIALS
New on Blog
APPROVED BY CLIENTS