Repeated Roots and Reduction of Order

Core Idea

When the characteristic equation has a repeated root r, one solution is e^rx, but we need a second linearly independent solution. The reduction-of-order method yields y₂ = x·e^rx. The general solution is y = (c₁ + c₂x)e^rx. For higher multiplicities, additional solutions involve higher powers of x. This technique extends beyond repeated roots to finding second solutions from any known solution.

Explainer

Recall from the characteristic equation method that for a second-order linear ODE y'' + py' + qy = 0, you try y = e^rx, substitute in, and get a quadratic equation in r. When that quadratic has two distinct roots r₁ and r₂, you get two independent solutions e^r₁x and e^r₂x and you're done. But a repeated root — when the discriminant is zero and r₁ = r₂ = r — gives only one solution e^rx from the characteristic equation, leaving you one solution short of a complete general solution.

The reduction-of-order method finds the missing second solution. The idea: if you already know one solution y₁ = e^rx, try y₂ = v(x)·y₁ = v(x)·e^rx for some unknown function v(x). Substitute y₂ into the ODE and apply the product rule (which you know) to differentiate. The key algebraic miracle is that the terms involving v(x) itself cancel — because y₁ is already a solution — and you're left with an equation only in v' and v''. Setting w = v' reduces this to a first-order ODE for w, which you can solve. For the repeated-root case, you end up with w'' = 0 (after substitution), so w is constant and v(x) = c₁ + c₂x. This gives y₂ = (c₁ + c₂x)e^rx, and factoring out c₁e^rx (a multiple of y₁) leaves the independent part y₂ = xe^rx.

Geometrically, the two solutions e^rx and xe^rx span a two-dimensional solution space. The factor of x makes them linearly independent — neither is a constant multiple of the other — so their linear combination y = (c₁ + c₂x)e^rx provides the full family of solutions. You can verify independence using the Wronskian (which you'll study next): W(e^rx, xe^rx) = e^2rx ≠ 0, confirming the two solutions are independent.

The reduction-of-order technique is more general than the repeated-root setting. Given *any* known solution y₁ to a second-order linear ODE — whether or not it came from a characteristic equation — you can use the same substitution y₂ = v(x)·y₁ to find a second independent solution. This is particularly valuable for variable-coefficient equations like Euler–Cauchy equations, where you might guess one solution by inspection and need a systematic method to find the second. The pattern of "multiply a known solution by an unknown function, substitute, and watch the equation reduce in order" recurs throughout differential equations and is worth internalizing as a general strategy.