Necessary and sufficient criteria for the uniqueness of solutions to the IVPs of scalar autonomous ODEs (Q1888592)

Some necessary and sufficient criteria are established for existence and uniqueness of solutions to the initial value problem \[ x'= f(x),\quad x(t_0)= x_0,\tag{1} \] with \(f\in C(\mathbb{R},\mathbb{R})\). These criteria are given related to the zeros of \(f\), that is, the set \(\mathbb{E}\) of equilibria of (1): \(\mathbb{E}= \{x\in\mathbb{R}: f(x)= 0\}\). The solution \(x(t,t_0,x_0)\) is said to be upper (lower) unique at the point \((t_0, x_0)\) if there exists an \(\varepsilon> 0\) such that in the interval \([t_0-\varepsilon, t_0+\varepsilon]\), \(x(t, t_0,x_0)\) coincides with the maximum (minimum) solution of (1) passing through \((t_0, x_0)\). If the solution \(x(t,t_0, x_0)\) is both upper unique and lower unique at \((t_0, x_0)\), then it is said to be unique at \((t_0,x_0)\). Otherwise, (1) is said to be nonunique at \((t_0,x_0)\). If \((t_0,x_0)\in \mathbb{R}\times (\mathbb{R}\setminus\mathbb{E})\), then (1) has a unique solution through \((t_0, x_0)\). If \(x_0\) is an isolated zero of \(\mathbb{E}\), then the solution \(x(t)\equiv x_0\) is unique at \((t_0, x_0)\) if and only if the following conditions hold simultaneously: \[ (2)\qquad \Biggl|\int_{x_0}{dx\over f(x)}\Biggr|=+\infty,\qquad (3)\qquad \Biggl| \int^{x_0} {dx\over f(x)}\Biggr|=+\infty. \] A point \(x_0\in\mathbb{E}\) is said to be a zero keeping sign of \(f\), if there exists an \(\varepsilon> 0\) such that \(f(x)\) does not change sign for any \(x\in (x_0-\varepsilon, x_0+\varepsilon)\). Otherwise, \(x_0\) is called zero changing sign of \(f\). If \(x_0\) is a zero changing sign of \(f\), then the solution \(x(t)\equiv x_0\) of (1) is unique. In the case when \(x_0\) is a zero keeping sign of \(f\), the following results are stated: (i) If there exists a \(\delta> 0\) such that \(m\{x\in [x_0- \delta, x_0+\delta]: f(x)= 0\}= 0\), then the solution \(x(t)\equiv x_0\) of (1) is unique if and only if (2) and (3) hold simultaneously. (ii) If there exists a \(\delta> 0\) such that for any \(\varepsilon (0,\delta)\), \(m\{x\in [x_0- \varepsilon, x_0+ \varepsilon]: f(x)= 0\}> 0\), then the solution \(x(t)\equiv x_0\) of (1) is unique. (iii) If there exists a \(\delta> 0\) such that \(m\{x\in [x_0+ \delta]: f(x)= 0\}= 0\) and, for any \(\varepsilon\in (0,\delta)\), \(m\{x\in [x_0- \varepsilon, x_0]: f(x)= 0\}> 0\), then the solution \(x(t)\equiv x_0\) of (1) is unique if and only if (2) holds. (iv) If there exists a \(\delta> 0\) such \(m\{x\in [x_0- \delta, x_0]: f(x)= 0\}= 0\) and, for any \(\varepsilon\in (0,\delta)\), \(m\{x\in [x_0, x_0+ \varepsilon]: f(x)= 0\}> 0\), then the solution \(x(t)\equiv x_0\) of (1) is unique if and only if (3) holds.