Ordinary Differential Equations:

A linear ODE has the form $a_n(x)y^{(n)} + a_{n-1}(x)y^{(n-1)}+...+ a_1(x)y'+a_0(x)y=g(x)$

First Order Ordinary Differential Equations

Separable ODE’s $y'=f(x)g(y)$ or if either variable is missing.

Non-separable reduced to separable ODEs If $f(tx,ty)=f(x,y)$ then we can reduce the ODE into separable using the substitution $v=\frac y x$ thus $y'=xv'+v$

Exact ODE’s $M_{(x,y)}dx+N_{(x,y)}dy=0$ and there is a function $f(x,y)$ such that $\frac{\partial f}{\partial x}=M$ and $\frac{\partial f}{\partial y} =N$ but we know it is exact when $M_y=N_x$

Non-exact reduced to exact ODEs using the integrating factor $\mu$ such that $M\mu dx+N\mu dy=0$ is exact.

$\mu=e^{\int\rho\space d(x,y)}$ with 2 Cases:

$\rho =\frac{M_y-N_x}{N}$ depends only on X (dx) or $\rho=\frac{N_x-M_y}{M}$ depends only on Y (dy)

Linear ODEs $a(x)y' +b(x)y=c(x)$ could be homogenous or none

write in standard form $y'+P(x)y=g(x)$ then find the integrating factor $\mu=e^{\int P(x) dx}$ the solution is $\mu y=\int \mu g dx$

Bernoulli's equation $y'+p(x)y=g(x)y^r , r\in \R$ Divide by $y^r$ then let $v=y^{1-r}$

Second Order Ordinary Differential Equations

$a(x)y''+b(x)y'+c(x)=g(x)$

Homogenous ODE with Constant Coefficients solution is $y=e^{rx}$ thus we find the characteristic equation $ar^2 +br+c=0$ and based on roots we have three forms of solutions:

Reduction of order; finding a solution from the other solution

$y''+p(x)y'+q(x)y=0$ then $y_2=y_1 \int \frac{e^{\int-p(x)dx}}{y_1^2} dx$

($y_1,y_2$ are a basis for the solution set or a fundamental set of solutions)

Missing y use $v=y' \rightarrow v'=y''$ Missing x use $v=y' \rightarrow y''=v'v$

Euler’s Equation has $y=x^r$ solution

$a_nx^ny^{(n)}+a_{n-1}x^{n-1}y^{(n-1)} +...+ a_1xy'+a_0y=0$

The auxiliary/characteristic equation: $a_2r(r-1)+a_1r+a_0=0$


Existence of Solution of an IVP $y''+p(x)y'+q(x)y=g(x)$ with initial values $\space y'(x_0)=y_0 , \space y(x_o)=y_1$

→ If there is an open interval $I$ in which $p(x),\space q(x),\space g(x)$ are all continuous and contains $x_0$ then the IVP has a unique solution.

The Wronskian of $y_1$ and $y_2$/ two solutions of an ODE $y=C_1y_1+C_2y_2$ iff $W(y_1,y_2) \ne 0$ (linear independence)

$$ W(y_1,y_2)\quad = \quad det\begin{bmatrix} y_1 & y_2 \\ y_1' & y_2' \\ \end{bmatrix} \quad=\quad y_1y_2' -y_2y_1' $$


Undetermined Coefficients

The general solution of $y''+p(x)y'+q(x)y=g(x)$ has the form $y=C_1y_1+C_2y_2+y_p$

where $y_1, y_2$ are a basis for the corresponding homogeneous ODE and $y_p$ is the particular solution depending on $g(x)$

$g(x)$ could be a polynomial, exponential or trigonometric.

Differential Equations - Undetermined Coefficients - tutorial.math.lamar.edu.png

Variation of Parameters

Solving the corresponding homogenous $y''+p(x)y'+q(x)y=0$ then write in standard form $y_p=u_1(x)y_1+u_2(x)y_2$

$u_1'=\frac {w_1} w \quad u_2'=\frac {w_2} w$ (since we use g(x) it needs to be in stand. form)

$$ W(y_1,y_2) = det\begin{bmatrix} y_1 & y_2 \\ y_1' & y_2' \\ \end{bmatrix} \quad \\ W_1= det\begin{bmatrix} 0 & y_2 \\ g(x) & y_2' \\ \end{bmatrix} \quad \quad W_2 = det\begin{bmatrix} y_1 & 0 \\ y_1' & g(x) \\ \end{bmatrix} \quad $$

⇒ in short, with this technique we replace the constants $C_1, \space C_2$ with two functions to satiate the actual non-homogeneous equation

→ don’t forget to compare particular solution with basis and modify accordingly


Ordinary and Singular Points

$p(x)y''+q(x)y'+r(x)y=0$ and the functions have no common factor if $p(x_0)\ne 0$ then $x_0$ is an ordinary point. Otherwise, it is a singular point.

Singular points could be regular (Frobenius) or irregular using limits (if finite then regular) $\lim_{x\to x_0} (x-x_0) \frac {Q(x)}{P(x)}$ and also $\lim_{x\to x_0} (x-x_0)^2 \frac {R(x)}{P(x)}$

Power Series Method

→ To solve linear ODE’s in general

let $y=\sum_{n=0} a_nx^n$ then find as many derivatives as the ODE & sub in it

then unify the powers and indices

Unifying the powers either to the highest or to n if exits.

Unifying the indices to the highest with straightforward substitution and expansion.

Frobenius Method

Solving an ODE $a_2(x)y''+a_1(x)y'+a_0(x)y=0$

near regular singular points, for simplicity we solve about $x_0=0$

The indicial equation $r(r-1) +p(0)r+q(0)=0$

where $p(x)=x\space \frac{a_1(x)}{a_2(x)}$ and $q(x)=x^2 \space \frac{a_0(x)}{a_2(x)}$

If $r_1 = r_2 =r$ then $y=\ln x \sum a_nx^{n+r} + \sum b_nx^{n+r}$

If $r_1 \ne r_2$ and $r_1-r_2\ne$ integer then $y= \sum a_nx^{n+r_1} + \sum b_nx^{n+r_2}$

If $r_1 \ne r_2$ and $r_1-r_2=$ integer then $y=k\ln x \sum a_nx^{n+r_1} + \sum b_nx^{n+r_2}$

When it comes to actually solving the question we find $y_1=x^{r_1} \sum _{n=0} a_n x^n$ then we find $y_2$ using reduction of order.


Laplace Transform

→the integral transform of a given function with real variable $t$ to convert it into a complex function with variable $s$.

The Laplace transform of a function f(t) is given by $\mathcal{L} \{f(t)\} =\int^\infty_0 e^{-st} f(t) dt = F(s)$ provided that the integral exists (convergent)