Nonoscillatory solutions of higher-order fractional differential equations (Q2141036): Difference between revisions

The authors study the asymptotic behaviour of the non-oscillatory solutions of forced fractional differential equations in the form \[ CD^{\alpha}_c y(t) + f_1(t, x(t)) = b(t) + k(t)x^{\beta}(t) + f_2(t, x(t)),\tag{1} \] where \[ y = \left(a (x^\prime)^\beta \right)^{(n-1)}, \qquad n \in\mathbb{N}, \] \(\beta \) is the ratio of two odd positive integers, \(CD^{\alpha}_c\) is the Caputo fractional derivative for \(\alpha \in (0, 1)\) and \(c > 1\), \(a, k \in C([c,\infty), (0,\infty))\), \(b \in C([c,\infty),\mathbb{R})\), \(f_1, f_2 \in C([c,\infty) \times \mathbb{R} ,\mathbb{R})\) and there exist \(g_1, g_2 \in C([c,\infty), (0,\infty))\) and \(\lambda_1 > \lambda_2 > 0\) such that \[ xf_1(t, x) \ge g_1(t)|x|^{\lambda_1+1}, \qquad 0 \le xf_2(t, x) \le g_2(t)|x|^{\lambda_2+1}. \] Let \[ q = \frac{p}{p-1} \text{ for } p > 1, \qquad g(t) = \left(\frac{g_2^{\lambda_1} (t)}{g_1^{\lambda_2} (t)} \right)^{\frac{1}{\lambda_1 - \lambda_2}}, \qquad A(t,c) = \int\limits_c^t a^{- \frac{1}{\beta}} (s) \, \mathrm{d}s.\] The main two results read as follows. Theorem 1. Let \(p > 1\) and \(\alpha \in (0, 1)\) satisfy \(p(\alpha - 1) + 1 > 0\). If \[ \int\limits_c^\infty k^q(s)s^{(n-1)q} A^{\beta q} (s, c) \,\mathrm{d}s < \infty \] and if \[ \lim\limits_{t \to \infty} \int\limits_c^t (t - s)^{\alpha-1} |b(s)| \,\mathrm{d}s < \infty, \qquad \lim\limits_{t \to \infty} \int\limits_c^t (t - s)^{\alpha -1} g(s) \,\mathrm{d}s < \infty, \tag{2}\] then any non-oscillatory solution \(x\) of (1) satisfies \[ \limsup\limits_{t \to \infty} \frac{|x(t)|}{ \left( t^{n - 1 } \mathrm{e}^{t} \right)^{\frac{ 1}{\beta}} A(t, c)} < \infty.\] Theorem 2. Let \(\beta = 1\). Let \(p > 1\), \(\alpha \in (0, 1)\), \( S > 0\), and \(\sigma > 1\) satisfy \[p(\alpha - 1) + 1 > 0, \qquad \frac{t^{n-1}}{ a(t)} \le S \mathrm{e}^{- \sigma t}. \] If \[ \int\limits_c^\infty k^{q} (s) \mathrm{e}^{- q s} \,\mathrm{d}s < \infty \] and if the inequalities in (2) hold, then any non-oscillatory solution of (1) is bounded.