ordinary differential equations
Existence and Uniqueness of ODE Solutions
You should know: first order differential equation
Overview
Before solving an initial value problem y′ = f(x,y), y(x₀) = y₀, it is worth asking whether a solution exists at all, and if so whether it is the only one. The Picard–Lindelöf theorem (also called the Picard existence theorem) answers both questions: if f(x,y) is continuous and Lipschitz continuous in y near (x₀,y₀), then a unique solution exists on some (possibly small) interval around x₀. Continuity alone (Peano's existence theorem) guarantees existence but not uniqueness, and the interval of existence can be much shorter than the interval where f is defined — solutions can 'blow up' in finite time even when f itself is perfectly smooth everywhere.
Intuition
Continuity of f alone only guarantees the direction field is well-defined at every point, which is enough to guarantee SOME curve threads through it (Peano's theorem) — but continuity doesn't prevent two different curves from touching at (x₀,y₀) and splitting apart afterward, as happens when f is 'too steep' in y (like y^{1/3} near y=0). The Lipschitz condition is a bound on how fast f can change with y — it prevents nearby solution curves from ever being forced to converge to exactly the same point and then diverge, ruling out the branching that causes non-uniqueness. Separately, even a perfectly well-behaved (Lipschitz, even polynomial) f can cause a solution to escape to infinity in finite x — existence is only guaranteed locally, near x₀, not necessarily for all x.
Formal Definition
The Picard–Lindelöf theorem gives sufficient conditions for a unique local solution to an initial value problem:
Worked Examples
This is separable: dy/y² = dx.
Apply y(0)=1 to find C, then solve for y.
f(x,y)=y² is smooth (polynomial, hence Lipschitz in y) everywhere, so Picard–Lindelöf guarantees a unique LOCAL solution — but the solution itself has a vertical asymptote at x=1.
Answer: y = 1/(1-x), which exists uniquely only on (-∞, 1) despite f(x,y)=y² being defined and smooth for all real x, y — a finite-time blowup (verified: y'=1/(1-x)^2=y^2 ✓).
Practice Problems
For y′ = y², y(0) = 2, find the solution and its interval of existence.
Is f(x,y) = y² Lipschitz in y on the strip |y| ≤ 5? Is it Lipschitz on all of ℝ?
Explain why y′ = y^{1/3} with y(0) = 1 (not y(0)=0) does NOT exhibit the non-uniqueness seen at y(0)=0.
Common Mistakes
Assuming that because f(x,y) is smooth (infinitely differentiable), a solution to y′=f(x,y) must exist for all x.
Smoothness only gives LOCAL existence and uniqueness; the maximal interval of existence can still be finite, as in y′=y², y(0)=1, which blows up at x=1 despite f(x,y)=y² being a polynomial.
Concluding that failure of the Lipschitz condition at a point means no solution exists there.
Peano's theorem only needs continuity to guarantee existence; a failure of the Lipschitz condition threatens UNIQUENESS, not existence — as seen in y′=y^{1/3}, y(0)=0, which has solutions, just more than one.
Quiz
Summary
- The Picard–Lindelöf theorem guarantees a unique local solution to y′=f(x,y), y(x₀)=y₀ when f is continuous and Lipschitz continuous in y.
- Continuity alone (Peano's theorem) guarantees existence of some solution but not uniqueness.
- Even smooth, globally-defined f can produce solutions that blow up in finite time, so existence is generally only local, not global.
- Failure of the Lipschitz condition (e.g. an unbounded ∂f/∂y) can allow multiple distinct solutions through the same initial point, as in y′=y^{1/3}, y(0)=0.
References
- BookBoyce, W. & DiPrima, R. Elementary Differential Equations, 11th ed., Ch. 2.
Mathematics