Foundations of Linear Differential Equations

Linear Differential Equations Introduction Linear differential equations form one of the most important and tractable classes of differential equations. Unlike nonlinear equations, which are often impossible to solve analytically, linear equations have a rich theory with systematic solution methods. Understanding their structure—particularly how solutions relate to vector spaces and operators—is essential for solving practical problems in engineering, physics, and applied mathematics. What Is a Linear Differential Equation? A linear differential equation is one where the unknown function and all of its derivatives appear to the first power, with no products or other nonlinear combinations. More formally, we can write any such equation in the standard form: $$a0(x)y + a1(x)y' + a2(x)y'' + \cdots + an(x)y^{(n)} = b(x)$$ Here, $y$ is the unknown function, and $a0(x), a1(x), \ldots, an(x), b(x)$ are known differentiable functions (which can be constants). The key insight is that $y$ and its derivatives appear linearly—each term is a coefficient function times a derivative of $y$, with no terms like $y \cdot y'$ or $(y')^2$. Why does this matter? Linearity is special because it allows us to use powerful mathematical tools from linear algebra. Solutions to linear equations can be added together, scaled, and analyzed using vector space theory in ways that simply don't work for nonlinear equations. Key Terminology Before solving linear differential equations, we need to classify them using several important terms. Order is the highest derivative that appears in the equation. A first-order equation contains $y'$ but no higher derivatives. A second-order equation contains $y''$, and so on. The order determines fundamental properties like the number of initial conditions needed for a unique solution. Homogeneous versus Non-homogeneous: A homogeneous linear differential equation has $b(x) = 0$ (the right side is zero). The equation looks like: $a0(x)y + a1(x)y' + a2(x)y'' + \cdots + an(x)y^{(n)} = 0$ A non-homogeneous equation has $b(x) \neq 0$ (the right side is nonzero). Given any non-homogeneous equation, the associated homogeneous equation is what you get by replacing $b(x)$ with $0$. This is crucial—we'll use it constantly when solving. Constant coefficients: An equation has constant coefficients when all the coefficient functions are constants. For example, $3y'' + 2y' + 5y = \sin(x)$ has constant coefficients (even though the right side isn't constant). When coefficients are not constants, the equation is more difficult to solve. The Structure of Solutions: Vector Spaces Here's a remarkable fact: the set of all solutions to a homogeneous linear differential equation forms a vector space. This means you can add two solutions together and get another solution, and you can multiply a solution by any constant and get another solution. For an ordinary (single-variable) $n$-th order linear differential equation, this vector space has dimension exactly $n$. This means: There exist exactly $n$ "basic" solutions that are independent of each other Every solution to the homogeneous equation can be written as a linear combination of these $n$ basic solutions: $yh = c1 y1 + c2 y2 + \cdots + cn yn$, where $c1, \ldots, cn$ are arbitrary constants The general solution to a non-homogeneous equation combines this with a particular solution: $$y = \underbrace{c1 y1 + c2 y2 + \cdots + cn yn}{yh \text{ (homogeneous part)}} + \underbrace{yp}{\text{particular solution}}$$ This formula says: any solution to the non-homogeneous equation is the sum of a solution to the homogeneous equation plus one specific solution to the non-homogeneous equation. This decomposition is fundamental—it separates the problem into two parts we can handle separately. Linear Differential Operators To handle linear differential equations more abstractly and powerfully, we introduce the concept of a linear differential operator. A basic differential operator of order $i$, denoted $D^i$, simply applies differentiation $i$ times. For instance, $D$ is differentiation ($Dy = y'$), $D^2$ means apply differentiation twice ($D^2y = y''$), and so on. A linear differential operator $L$ is a sum of basic operators, each multiplied by a coefficient function: $$L = a0(x) + a1(x)D + a2(x)D^2 + \cdots + an(x)D^n$$ where $an(x) \neq 0$ and $n$ is the order of $L$. Using this notation, our original differential equation $a0(x)y + a1(x)y' + \cdots + an(x)y^{(n)} = b(x)$ becomes simply: $$Ly = b(x)$$ This operator notation is elegant because it lets us write the equation compactly and apply algebraic insights. The kernel of $L$ is the set of all functions $y$ such that $Ly = 0$ (the homogeneous equation). The kernel is always a vector space of dimension $n$ (under mild continuity conditions). The functions forming a basis of the kernel are exactly those $n$ independent solutions $y1, y2, \ldots, yn$ we discussed earlier. The general solution formula using operators becomes: For an $n$-th order operator $L$, the general solution to $Ly = b(x)$ is $$y = c1 y1 + c2 y2 + \cdots + cn yn + yp$$ where $y1, \ldots, yn$ form a basis of $\ker(L)$, $c1, \ldots, cn$ are arbitrary constants, and $yp$ is any particular solution satisfying $Lyp = b(x)$. <extrainfo> Note: When the equation involves partial derivatives (where $y$ depends on multiple variables like $y(x,t)$), we have a linear partial differential equation. While the underlying principles remain the same, the analysis of partial differential equations is significantly more complex and is typically covered in advanced courses. </extrainfo>