Maximal Interval of Existence for IVP

We are given an initial value problem (IVP):

\(\dot{x} = f(t, x)\)

\(x(t_0) = x_0\)

where \(f: \mathbb{R}^2 \to \mathbb{R}\) is a locally Lipschitz function, and \((t_0, x_0) \in \mathbb{R}^2\). We are asked about the nature of the maximal interval of existence, denoted by \(J(t_0, x_0)\).

Understanding the Maximal Interval

The maximal interval of existence \(J(t_0, x_0)\) is the largest possible open interval containing \(t_0\) on which a unique solution to the IVP exists. The fact that \(f\) is locally Lipschitz is a crucial condition. According to the Picard-Lindelöf theorem (also known as the Existence and Uniqueness Theorem) for ODEs, if \(f\) is locally Lipschitz in \(x\) uniformly in \(t\) (or on a compact set in \((t, x)\)), a unique solution exists in some open interval around \(t_0\). The maximal interval extends this local existence to the largest possible interval.

Properties of the Maximal Interval

A fundamental result in the theory of ordinary differential equations states that for an IVP with a locally Lipschitz right-hand side function \(f\), the maximal interval of existence \(J(t_0, x_0)\) is always an open interval. Let's analyze why this is the case and consider the given options.

Analyzing the Options

• Option 1: \(J(t_0, x_0) = \mathbb{R}\).

  This is not always true. While some solutions exist for all time (e.g., linear ODEs with bounded coefficients), solutions to general non-linear ODEs can 'blow up' in finite time, meaning the solution goes to infinity as \(t\) approaches a finite value. A classic example is \(\dot{x} = x^2\) with \(x(0)=1\), whose solution \(x(t) = 1/(1-t)\) exists only for \(t \lt 1\).

• Option 2: \(J(t_0, x_0)\) is an open set.

  This statement aligns with the standard theorem regarding the maximal interval of existence for IVPs with locally Lipschitz functions. The maximal interval is defined such that the solution cannot be extended to include either endpoint if those endpoints are finite. This implies the interval must be open.

• Option 3: \(J(t_0, x_0)\) is a closed set.

  This is incorrect. If the maximal interval were closed, say \([a, b]\), and the solution exists up to time \(b\), then if \(f\) is defined and locally Lipschitz at \((b, x(b))\), the existence and uniqueness theorem would guarantee that the solution could be extended to an interval slightly larger than \(b\), contradicting the maximality of \([a, b]\). The solution either blows up or approaches the boundary of the domain of \(f\) at a finite time, preventing extension to a closed interval.

• Option 4: \(J(t_0, x_0)\) could be an empty set.

  This is incorrect. The Picard-Lindelöf theorem guarantees the existence of a unique solution in some open interval around \(t_0\). Since \(t_0\) must be contained in \(J(t_0, x_0)\), the maximal interval of existence is never empty.

Conclusion

Based on the theory of ordinary differential equations, for an initial value problem with a locally Lipschitz function, the maximal interval of existence is always an open interval containing the initial time \(t_0\). This open interval can be bounded or unbounded (in which case it would be \((-\infty, b)\), \((a, \infty)\), or \((-\infty, \infty)\)).