Linear DEs & Numerical Methods

Cheatsheet Content

### Euler's Method (Chapter 2.6 Introduction) Euler's method is a numerical technique for approximating solutions to first-order initial-value problems (IVPs) of the form: $$y' = f(x, y), \quad y(x_0) = y_0$$ The approximation is generated iteratively using the formula: $$y_{n+1} = y_n + hf(x_n, y_n)$$ where $x_n = x_0 + nh$ and $h = \Delta x$ is the step size. #### Procedure for Euler's Method 1. **Define the IVP:** Clearly identify $f(x, y)$, the initial point $x_0$, and the initial value $y_0$. 2. **Choose a step size ($h$):** A smaller $h$ generally leads to more accurate approximations but requires a greater number of computational steps. 3. **Iterate:** * Begin with $n=0$. * Calculate the next $y$ value: $y_1 = y_0 + hf(x_0, y_0)$. * Update the $x$ value: $x_1 = x_0 + h$. * Continue this process for $n=1, 2, 3, \dots$ until the desired $x$ value is reached. Each step uses the previously calculated $y_n$ and $x_n$ to find $y_{n+1}$. #### Error Analysis in Euler's Method * **Local Truncation Error:** This is the error introduced in a single step of the method, typically proportional to $O(h^2)$. It represents how much the tangent line approximation deviates from the true solution curve over one step. * **Global Truncation Error:** This is the accumulated error over many steps, typically proportional to $O(h)$. It reflects the total deviation of the numerical solution from the true solution over the entire interval. * **Factors Affecting Error:** The accuracy of the method depends significantly on the chosen step size $h$ and the nature (e.g., nonlinearity, stiffness) of the function $f(x,y)$. * **Accumulation:** Errors accumulate over the interval of integration and can be magnified, especially if the slope $f(x,y)$ changes rapidly. ### Exact Differential Equations (Chapter 2.6) Exact differential equations represent a specific class of first-order DEs that can be solved by finding a potential function. #### General Form An exact differential equation is written in the form: $$M(x,y)dx + N(x,y)dy = 0$$ #### Condition for Exactness The equation is exact if and only if the partial derivative of $M$ with respect to $y$ equals the partial derivative of $N$ with respect to $x$: $$\frac{\partial M}{\partial y} = \frac{\partial N}{\partial x}$$ #### Conceptual Meaning When an equation is exact, it implies that $M(x,y)dx + N(x,y)dy$ is the total differential of some potential function $F(x,y)$. That is, $dF = \frac{\partial F}{\partial x}dx + \frac{\partial F}{\partial y}dy$. This means: * $\frac{\partial F}{\partial x} = M(x,y)$ * $\frac{\partial F}{\partial y} = N(x,y)$ The solution to the differential equation is then implicitly defined by $F(x,y) = C$, where $C$ is a constant. This concept is analogous to: * Conservative vector fields in multivariable calculus. * Energy functions in physics (where total energy is conserved). * Potential energy surfaces in various scientific domains. #### Procedure for Solving Exact Equations 1. **Check for Exactness:** Verify if $\frac{\partial M}{\partial y} = \frac{\partial N}{\partial x}$. If not, the equation is not exact (unless an integrating factor is used). 2. **Integrate M with respect to x:** Assume $\frac{\partial F}{\partial x} = M(x,y)$. Integrate $M(x,y)$ with respect to $x$, treating $y$ as a constant. This will introduce an arbitrary function of $y$, denoted as $h(y)$: $$F(x,y) = \int M(x,y)dx + h(y)$$ 3. **Differentiate F with respect to y and Match to N:** Differentiate the expression for $F(x,y)$ obtained in step 2 with respect to $y$. Set this equal to $N(x,y)$: $$\frac{\partial F}{\partial y} = \frac{\partial}{\partial y}\left(\int M(x,y)dx\right) + h'(y) = N(x,y)$$ 4. **Solve for h(y):** Rearrange the equation from step 3 to solve for $h'(y)$, then integrate $h'(y)$ with respect to $y$ to find $h(y)$. (Note: no integration constant is needed here, as it will be absorbed into the final constant $C$). 5. **Form the General Solution:** Substitute $h(y)$ back into the expression for $F(x,y)$ from step 2. The general solution is then $F(x,y) = C$. #### Applications of Exact Equations Exact equations frequently arise in scientific and engineering contexts, including: * **Thermodynamics:** Describing state functions and energy relationships where the path taken between states does not affect the total change in some quantity (e.g., internal energy). * **Conservative Force Fields:** Analyzing work done by forces (e.g., gravitational, electrostatic) where the work is independent of the path. * **Implicit Geometric Relationships:** Defining families of curves where the slope of the tangent at any point $(x,y)$ is given by the differential equation. ### Integrating Factors (Chapter 2.6) If a first-order differential equation $M(x,y)dx + N(x,y)dy = 0$ is not exact, it may sometimes be made exact by multiplying it by an integrating factor $\mu(x,y)$. #### When to Look for Integrating Factors An integrating factor $\mu(x,y)$ is a function such that when multiplied by the non-exact equation, the resulting equation $[\mu M]dx + [\mu N]dy = 0$ becomes exact. This means: $$\frac{\partial}{\partial y}(\mu M) = \frac{\partial}{\partial x}(\mu N)$$ #### Common Cases for Integrating Factors Two common scenarios allow for finding an integrating factor that depends on only one variable: 1. **Integrating Factor $\mu(x)$ (function of $x$ only):** If the expression $\frac{M_y - N_x}{N}$ is a function of $x$ alone (let's call it $f(x)$), then an integrating factor $\mu(x)$ exists and is given by: $$\mu(x) = e^{\int f(x)dx} = e^{\int \frac{M_y - N_x}{N} dx}$$ 2. **Integrating Factor $\mu(y)$ (function of $y$ only):** If the expression $\frac{N_x - M_y}{M}$ is a function of $y$ alone (let's call it $g(y)$), then an integrating factor $\mu(y)$ exists and is given by: $$\mu(y) = e^{\int g(y)dy} = e^{\int \frac{N_x - M_y}{M} dy}$$ #### Procedure with an Integrating Factor 1. **Check for Non-Exactness:** Verify that $\frac{\partial M}{\partial y} \neq \frac{\partial N}{\partial x}$. 2. **Calculate Test Expressions:** Compute $\frac{M_y - N_x}{N}$ and $\frac{N_x - M_y}{M}$. 3. **Identify Type of Integrating Factor:** * If $\frac{M_y - N_x}{N}$ depends only on $x$, use $\mu(x) = e^{\int \frac{M_y - N_x}{N} dx}$. * If $\frac{N_x - M_y}{M}$ depends only on $y$, use $\mu(y) = e^{\int \frac{N_x - M_y}{M} dy}$. 4. **Multiply and Solve:** Multiply the original non-exact equation by the derived integrating factor $\mu$. The new equation $\mu M dx + \mu N dy = 0$ will now be exact. Solve this new exact equation using the procedure outlined in the "Exact Differential Equations" section. ### First-Order Linear Differential Equations (Chapter 3.1) First-order linear differential equations are a fundamental class of DEs with a well-defined solution method. #### Standard Form A first-order linear differential equation can always be written in the standard form: $$y' + P(x)y = f(x)$$ where $P(x)$ and $f(x)$ are continuous functions on some interval. #### Integrating Factor for Linear DEs The integrating factor $\mu(x)$ for a first-order linear DE in standard form is given by: $$\mu(x) = e^{\int P(x) dx}$$ Multiplying the standard form equation by $\mu(x)$ transforms the left side into the derivative of a product: $$\mu(x)y' + P(x)\mu(x)y = f(x)\mu(x)$$ $$\frac{d}{dx}[\mu(x)y] = f(x)\mu(x)$$ #### General Solution Integrate both sides with respect to $x$: $$\mu(x)y = \int f(x)\mu(x) dx + C$$ Finally, solve for $y$: $$y = \frac{1}{\mu(x)}\left[\int f(x)\mu(x) dx + C\right]$$ #### Structure of the Solution The general solution $y$ can often be understood as a sum of two components: $$y = y_p + y_c$$ * $y_p$: The **particular solution** (or steady-state solution). This part arises from $\frac{1}{\mu(x)}\int f(x)\mu(x) dx$ and represents the long-term behavior of the system, often influenced by the forcing term $f(x)$. * $y_c$: The **complementary solution** (or transient solution). This part arises from $\frac{C}{\mu(x)}$ and represents the behavior determined by the initial conditions. In many physical systems, $y_c$ tends to zero as $x \to \infty$ (it "dies out"), making $y_p$ the dominant long-term behavior. ### Growth and Decay Models (Chapter 3.1 Applications) Many natural phenomena are governed by processes where the rate of change of a quantity is proportional to the quantity itself. #### General Equation The fundamental differential equation for growth and decay is: $$\frac{dx}{dt} = kx$$ with an initial condition $x(t_0) = x_0$. #### Solution The solution to this IVP is an exponential function: $$x(t) = x_0e^{k(t-t_0)}$$ If the initial time $t_0 = 0$, the solution simplifies to $x(t) = x_0e^{kt}$. #### Applications * **Exponential Growth ($k > 0$):** * Population growth (e.g., bacteria, animal populations). * Compound interest (money growing in a bank account). * Unrestricted propagation of certain phenomena. * **Exponential Decay ($k ### Newton's Law of Cooling / Warming (Chapter 3.1 Applications) Newton's Law of Cooling describes how the temperature of an object changes over time as it equilibrates with its surroundings. #### Equation The rate of change of the temperature $T(t)$ of an object is proportional to the difference between its own temperature and the ambient (surrounding) temperature $T_m$: $$\frac{dT}{dt} = k(T - T_m)$$ where $k$ is a constant of proportionality. #### Solution This is a first-order linear differential equation. The solution is: $$T(t) = T_m + Ce^{kt}$$ The constant $C$ is determined by the initial temperature $T(t_0) = T_0$. If $t_0 = 0$, then $C = T_0 - T_m$. #### Behavior * If $T_0 > T_m$, the object is cooling, and $k$ will be negative ($k 0$). The temperature $T(t)$ approaches $T_m$ from below. In both cases, the object's temperature asymptotically approaches the ambient temperature $T_m$. #### Applications * **Cooling Coffee/Drinks:** Predicting how long it takes for a beverage to reach a drinkable temperature. * **Forensic Science:** Estimating the time of death based on a body's temperature. * **Heat Transfer Modeling:** Analyzing thermal processes in engineering and environmental science. ### Mixture Problems (Chapter 3.1 Applications) Mixture problems involve tracking the amount of a substance (e.g., salt, pollutant) in a tank or container where liquids are flowing in and out. #### Fundamental Principle The rate of change of the amount of substance $A(t)$ in the tank at time $t$ is given by: $$\frac{dA}{dt} = \text{Rate}_{\text{in}} - \text{Rate}_{\text{out}}$$ #### Components of the Rates * **Rate In:** The rate at which the substance enters the tank. $$\text{Rate}_{\text{in}} = (\text{concentration of inflow}) \times (\text{rate of inflow})$$ * **Rate Out:** The rate at which the substance leaves the tank. $$\text{Rate}_{\text{out}} = (\text{concentration of outflow}) \times (\text{rate of outflow})$$ #### Key Considerations * **Volume:** The instantaneous volume of liquid in the tank, $V(t)$, is crucial for determining the concentration of outflow. * If the inflow rate equals the outflow rate, then $V(t)$ is constant. * If the inflow rate does not equal the outflow rate, then $V(t)$ will be a function of time: $V(t) = V_0 + (\text{inflow rate} - \text{outflow rate})t$. * **Concentration of Outflow:** This is typically the amount of substance in the tank divided by the current volume: $\frac{A(t)}{V(t)}$. #### Resulting Equation Mixture problems almost always lead to a first-order linear differential equation that can be solved using the integrating factor method. #### Applications * **Chemical Tanks:** Modeling reactions and concentrations in industrial processes. * **Environmental Pollution:** Tracking pollutants in lakes, rivers, or enclosed spaces. * **Pharmaceutical Mixing:** Ensuring correct drug concentrations in solutions. ### Series Circuit Problems (Chapter 3.1 Applications) Differential equations are essential for analyzing the behavior of electrical circuits, specifically LR and RC series circuits. #### LR-Series Circuit A series circuit containing an inductor ($L$), a resistor ($R$), and an electromotive force ($E(t)$). According to Kirchhoff's Voltage Law, the sum of voltage drops across each component equals the applied voltage: $$L\frac{di}{dt} + Ri = E(t)$$ * $i(t)$: Current in the circuit (dependent variable). * $L$: Inductance (constant). * $R$: Resistance (constant). * $E(t)$: Applied electromotive force (voltage source, a function of time). This is a first-order linear differential equation for the current $i(t)$. * **Application:** Models current buildup in circuits, inductance effects, and transient current decay. #### RC-Series Circuit A series circuit containing a resistor ($R$), a capacitor ($C$), and an electromotive force ($E(t)$). According to Kirchhoff's Voltage Law: $$R\frac{dq}{dt} + \frac{1}{C}q = E(t)$$ * $q(t)$: Charge on the capacitor (dependent variable). * $R$: Resistance (constant). * $C$: Capacitance (constant). * $E(t)$: Applied electromotive force (voltage source, a function of time). This is a first-order linear differential equation for the charge $q(t)$. The current $i(t)$ is related to the charge by $i(t) = \frac{dq}{dt}$. * **Applications:** Modeling charging and discharging of capacitors, timing circuits, and filter circuits. ### Second-Order Linear Equations with Constant Coefficients (Chapter 3.2) This section introduces the foundational concepts for solving homogeneous second-order linear differential equations with constant coefficients. These equations are crucial for modeling physical systems involving oscillations and damping. #### General Form A homogeneous second-order linear differential equation with constant coefficients takes the form: $$ay'' + by' + cy = 0$$ where $a, b, c$ are constants, and $a \neq 0$. #### Characteristic Equation To solve this type of equation, we assume a solution of the form $y = e^{rx}$. Substituting this into the differential equation yields the **characteristic equation**: $$ar^2 + br + c = 0$$ This is a quadratic equation whose roots $r_1, r_2$ determine the form of the general solution. The nature of these roots depends on the discriminant $\Delta = b^2 - 4ac$. #### Case 1: Distinct Real Roots ($\Delta > 0$) If $b^2 - 4ac > 0$, there are two distinct real roots, $r_1$ and $r_2$. The general solution is: $$y = C_1e^{r_1x} + C_2e^{r_2x}$$ #### Case 2: Repeated Real Roots ($\Delta = 0$) If $b^2 - 4ac = 0$, there is one repeated real root, $r_1 = r_2 = -\frac{b}{2a}$. The general solution is: $$y = C_1e^{r_1x} + C_2xe^{r_1x}$$ #### Case 3: Complex Conjugate Roots ($\Delta ### Applications: Mechanical Vibrations (Chapter 3.2) Second-order linear differential equations with constant coefficients are the cornerstone for modeling mass-spring systems, which are fundamental to understanding vibrations and oscillations. #### Mass-Spring System Equation Consider a mass $m$ attached to a spring with spring constant $k$, and possibly subjected to a damping force proportional to velocity, with damping coefficient $c$. Applying Newton's second law ($F=ma$) leads to the equation: $$m y'' + c y' + k y = 0$$ * $m$: mass (always positive). * $c$: damping coefficient (non-negative). * $k$: spring constant (always positive). * $y(t)$: displacement of the mass from its equilibrium position at time $t$. This equation is directly analogous to the $ay'' + by' + cy = 0$ form, with $a=m$, $b=c$, and $c=k$. The characteristic equation is $mr^2 + cr + k = 0$. The nature of the roots depends on the discriminant $\Delta = c^2 - 4mk$. #### Case 1: Overdamped System ($\Delta > 0 \implies c^2 - 4mk > 0$) * **Roots:** Two distinct real roots ($r_1, r_2$, both negative). * **Solution:** $y(t) = C_1e^{r_1t} + C_2e^{r_2t}$. * **Behavior:** The system returns to equilibrium slowly without oscillating. The damping force is strong enough to prevent any oscillation. #### Case 2: Critically Damped System ($\Delta = 0 \implies c^2 - 4mk = 0$) * **Roots:** One repeated real root ($r_1 = r_2 = -c/(2m)$, negative). * **Solution:** $y(t) = C_1e^{r_1t} + C_2te^{r_1t}$. * **Behavior:** The system returns to equilibrium as quickly as possible without oscillating. This is the optimal damping for many applications (e.g., shock absorbers). #### Case 3: Underdamped System ($\Delta ### Nonlinear Models (Beyond Linear, but important for context) While Chapters 2.6, 3.1, and 3.2 focus on linear equations, understanding the basic behavior of some nonlinear models provides a broader perspective on differential equations. #### I. Logistic Equation Models population growth in an environment with limited resources, leading to a carrying capacity. $$\frac{dP}{dt} = kP\left(1 - \frac{P}{K}\right)$$ * $P(t)$: Population at time $t$. * $k$: Intrinsic growth rate (constant). * $K$: Carrying capacity (maximum sustainable population, constant). This can also be written in the form $\frac{dP}{dt} = P(a - bP)$, where $a=k$ and $b=k/K$. ##### Behavior * **Equilibrium Solutions:** * $P=0$ (unstable equilibrium): If the population is very small, it will grow. * $P=K$ (stable equilibrium): The population will tend towards the carrying capacity. * **Inflection Point:** The population growth rate is maximized when $P = K/2$. This is the point where the growth curve changes from concave up to concave down. * **S-shaped Curve:** Solutions typically exhibit an S-shaped curve, starting with exponential-like growth, then slowing down as it approaches $K$ asymptotically. ##### Applications * Population dynamics with limited food or space. * Spread of diseases (initial phase). * Technology adoption curves. * Models for resource-limited growth in various fields. #### II. Harvesting Models A modification of the logistic equation that accounts for the removal (harvesting) or addition (restocking) of individuals from the population. $$\frac{dP}{dt} = kP\left(1 - \frac{P}{K}\right) - h$$ where $h$ is a constant harvesting rate. ##### Behavior * The introduction of $h$ can significantly change the equilibrium points and stability. * If $h$ is too large, it can lead to the extinction of the population, even if it started above the carrying capacity. * These models are crucial for sustainable resource management (e.g., fishing quotas, wildlife management). #### III. Doomsday / Extinction Models These nonlinear models can predict scenarios where a quantity either grows unbounded in finite time (doomsday) or goes extinct in finite time, rather than asymptotically. * **Example (Doomsday):** $\frac{dP}{dt} = aP^2$. The solution $P(t) = \frac{P_0}{1 - aP_0t}$ shows that $P(t) \to \infty$ as $t \to \frac{1}{aP_0}$, a finite time. * **Example (Extinction):** $\frac{dP}{dt} = -aP^2$. The solution $P(t) = \frac{P_0}{1 + aP_0t}$ shows that $P(t) \to 0$ as $t \to \infty$, but other similar models can lead to extinction in finite time. ##### Applications * Runaway population growth in abstract models. * Modeling certain economic bubbles or positive feedback systems. #### IV. Competition Models These involve systems of coupled nonlinear differential equations that describe the interactions between two or more species or entities competing for resources. * **Example:** For two competing species $P$ and $Q$: $$\frac{dP}{dt} = P(a_1 - b_1P - c_1Q)$$ $$\frac{dQ}{dt} = Q(a_2 - b_2Q - c_2P)$$ * **Applications:** Predator-prey dynamics, ecological competition, market competition between companies. These models are typically analyzed using phase plane analysis.