Solutions to a class of nonlinear differential equations of fractional order, Electronic Journal of Qualitative Theory of Differential Equations

In this paper we investigate the formulation of a class of boundary value problems of fractional order with the Riemann-Liouville fractional derivative and integral-type boundary conditions. The existence of solutions is established by applying a fixed point theorem of Krasnosel’skiuo and Zabreiko for asymptotically linear mappings.

In this section we introduce basic facts of fractional calculus involving the Riemann-Liouville fractional differential operator and list several recent and classical results dealing with initial and boundary value problems of fractional order.In the present section we also formulate the boundary value problems and establish the equivalence of those to certain integral equations.The existence of solutions of the integral equations is shown in Section 2 by applying a fixed point theorem of Krasnosel'skiȋ and Zabreiko [10].
The study of primarily initial value problems for differential equations of fractional order with various types of integro-differential operators such as Riemann-Liouville is extensive.It includes several well-known monographs [8,13,14,15] and papers [2,3,4,5,6,9,11,12].Boundary value problems for fractional order have received less attention than initial value problems [1].In this work we obtain several existence results for a Riemann-Liouville differential equation fractional order 1 < α < 2. We point out that if the solutions are sought in the class of continuous functions, one of the conditions loses its meaning as a boundary condition and should instead be interpreted as a well-posedness condition.In this regard the paper clears certain misconceptions about formulation of boundary value problems for differential equations of fractional order.
We study the Riemann-Liouville integro-differential equation EJQTDE, 2009 No. 20, p. 3 of fractional order 1 < α < 2. We assume throughout the note that We seek solutions of (4) in the class of continuous functions satisfying the boundary condition where β > −α.In addition, we assume that γ > 1 − α and impose the condition to which we refer as the well-posedness condition.The condition (6) plays the role of the second boundary condition.
By a solution of the boundary value problem ( 4)-( 6) we understand a function satisfying the equation ( 4) and the conditions ( 5) and (6).
we have by the semigroup property (3) that In particular, the right side of the above equation vanishes at t = 0 thus justifying the use of ( 6) and the applicability of the term "well-posedness" condition to the scope of this work. Similarly, EJQTDE, 2009 No. 20, p. 4 It follows from ( 5) that and Replacing g with the inhomogeneous term of (4), we obtain that if u ∈ C[0, 1] is a solution of the problem ( 4)-( 6), then u ∈ C[0, 1] is the solution of the integral equation The converse is also true in view of (2).
Since the solvability of ( 4)-( 6) is equivalent the existence of a solution of the integral equation ( 7), we will seek a fixed point of the integral mapping, for t ∈ [0, 1], Note that the parameter γ is absent from the integral equation ( 7) simply because the shape of the mapping above is dictated by the class of admissible functions rather than the condition (6) at t = 0 (see, e. g., [1] with u(0) = 0 which is the case γ = 0 in this note).
Let X = C[0, 1] be endowed with the sup-norm denoted by • .It is clear that T : X → X is a completely continuous mapping.The existence of a fixed point of T will follow from the Krasnosel'skiȋ-Zabreiko fixed point theorem [10]: EJQTDE, 2009 No. 20, p. 5 Theorem 1.1 Let X be a Banach space.Assume that T : X → X is completely continuous mapping and L : X → X is a bounded linear mapping such that 1 is not an eigenvalue of L and Then T has a fixed point in X.
For applications of Theorem 1.1 we refer the reader to [7] and the references therein.
Although the inhomogeneous term in the above mentioned work was considered in the form φ(t)g(u(t)), we observe that there is no need in separating the variables so that a slightly more general scenario f (t, u(t)) is considered in the present note.

The existence result
The first existence result is obtained for the problem (4)- (6).
In addition to the hypothesis (H 1 ) we impose the following conditions: Theorem 2.1 Let the assumptions (H 1 )-(H 3 ) be satisfied.Assume that Then the boundary value problem ( 4)-( 6) has a nontrivial solution.
EJQTDE, 2009 No. 20, p. 6 Proof.Define a bounded linear mapping, L : X → X, by together with Lu = u with u = 0 lead to a contradiction.Thus λ = 1 is not an eigenvalue of L.
Let > 0. There exists an A > 0 such that (11), we also have u .
We see that the condition ( 9) is verified.Thus, by Theorem 1.1, T has a fixed point in X which is a solution of the boundary value problem ( 4)- (6).It follows from (H 3 ) that the boundary value problem does not possess the trivial solution.