This work is licensed under the Creative Commons Attribution 4.0 Public License.
Introduction
Firstly, Zadeh introduced the concept of fuzzy numbers and fuzzy arithmetic [22]. The major application of fuzzy arithmetic is fuzzy differential equations. Fuzzy differential equations are suitable models to model dynamic systems in which there exist uncertainties or vaguness. Fuzzy differential equations can be examined by several approach, such as Hukuhara differentiability, generalized differentiability, the concept of differential inclusion etc [1], [3], [4], [5], [6, 7, 8], [9], [11], [13, 14], [15] [17], [19], [20].
In this paper is on a fuzzy Sturm-Liouville problem with the eigenvalue parameter in the boundary condition. Important notes are given for the problem. Integral equations are found of the problem.
Preliminaries
Definition 1. [18] A fuzzy number is a function u : $\mathbb{R}\to \left[ 0,1 \right]$satisfying the properties:u is normal, u is convex fuzzy set, u is upper semi-continuous on ℝ, $cl\left\{ x\varepsilon \mathbb{R}\left| u\left( x \right)>0 \right. \right\}$is compact, where cl denotes the closure of a subset.
Let RF denote the space of fuzzy numbers.
Definition 2. [15] Let uɛℝF. The a-level set of u, denoted, ${{\left[ u \right]}^{\alpha }},0<\alpha \le 1,$is
$${{\left[ u \right]}^{\alpha }}=\left\{ x\varepsilon \mathbb{R}\left| u\left( x \right)\ge \alpha \right. \right\}.$$
If a = 0; the support of u is defined
$${{\left[ u \right]}^{0}}=cl\left\{ x\varepsilon \mathbb{R}\left| u\left( x \right)>0 \right. \right\}.$$
The notation, ${{\left[ u \right]}^{\alpha }}=\left[ {{\underline{u}}_{\alpha }},{{{\bar{u}}}_{\alpha }} \right]$denotes explicitly the a-level set of u. We refer to$\underline{u}$and ū as the lower and upper branches of u,respectively.
The following remark shows when $\left[ {{\underline{u}}_{\alpha }},{{{\bar{u}}}_{\alpha }} \right]$ is a valid a-level set.
Remark 1. [10,15] The sufficient and necessary conditions for$\left[ {{\underline{u}}_{\alpha }},{{{\bar{u}}}_{\alpha }} \right]$to define the parametric form of a fuzzy number as follows:
${{\underline{u}}_{\alpha }}$is bounded monotonic increasing (nondecreasing) left-continuous function on (0;1] and right-continuous for a = 0,
${{\bar{u}}_{\alpha }}$is bounded monotonic decreasing (nonincreasing) left-continuous function on (0;1] and right-continuous fora = 0,
Definition 3. [15] For u,vεℝF and λ ∈R, the sum u+v and the product λ u are defined by${{\left[ u+v \right]}^{\alpha }}={{\left[ u \right]}^{\alpha }}+{{\left[ v \right]}^{\alpha }},$${{\left[ \lambda u \right]}^{\alpha }}=\lambda {{\left[ u \right]}^{\alpha }}$where means the usual addition of two intervals (subsets) of ℝ and λ [u]α means the usual product between a scalar and a subset of R.
The metric structure is given by the Hausdorff distance
Definition 4. [18] If A is a symmetric triangular numbers with supports$\left[ \underline{a},\bar{a} \right],$the α–level sets of A is${{\left[ A \right]}^{\alpha }}=$$\left[ \underline{a}+\left( \frac{\bar{a}-\underline{a}}{2} \right)\alpha ,\bar{a}-\left( \frac{\bar{a}-\underline{a}}{2} \right)\alpha \right].$
Definition 5. [16] u,v 2 ℝF, ${{\left[ u \right]}^{\alpha }}=\left[ {{\underline{u}}_{\alpha }},{{{\bar{u}}}_{\alpha }} \right],{{\left[ v \right]}^{\alpha }}=\left[ {{\underline{v}}_{\alpha }},{{{\bar{v}}}_{\alpha }} \right],$the product uv is defined by
$${{\left[ uv \right]}^{\alpha }}={{\left[ u \right]}^{\alpha }}{{\left[ v \right]}^{\alpha }},\forall \alpha \in \left[ 0,1 \right],$$
Definition 6. [15, 21] Let u,v 2 ℝF.If there exists w 2 ℝF such that u = v+w; then w is called the Hukuhara difference of fuzzy numbers u and v;and it is denoted by w = u⊖v:
Definition 7. [2,15] Let f : [a;b]→RF and t0∈ [a,b] :We say that f is Hukuhara differential at t0, if there exists an element f' 0 (t0) 2 ℝF such that for all h > 0 sufficiently small, $\exists f\left( {{t}_{0}}+h \right)\ominus f\left( {{t}_{0}} \right),\,\,f\left( {{t}_{0}} \right)\ominus f\left( {{t}_{0}}-h \right)$and the limits (in the metric D)
Definition 8. [12] If${p}'\left( x \right)=0,\,\,r\left( x \right)=1$and$Ly=p\left( x \right)y''+q\left( x \right)y$in the fuzzy differential equation${{\left( p\left( x \right){y}' \right)}^{\prime }}+q\left( x \right)y+\lambda r\left( x \right)y=0,p\left( x \right),{p}'\left( x \right),q\left( x \right),r\left( x \right),$are continuous functions and positive, the fuzzydifferential equation
$$Ly+\lambda y=0$$
is called a fuzzy Sturm-Liouville equation.
Definition 9. [12]${{\left[ y\left( x,{{\lambda }_{0}} \right) \right]}^{\alpha }}=\left[ \underline{y}\left( x,{{\lambda }_{0}} \right),\bar{y}\left( x,{{\lambda }_{0}} \right) \right]\ne 0,$we say that λ = λ0is eigenvalue of (2.1) if the fuzzy differential equation (2.1) has the nontrivial solutions$\underline{y}\left( x,{{\lambda }_{0}} \right)\ne 0,\bar{y}\left( x,{{\lambda }_{0}} \right)\ne 0.$
where q(x) is continuous function and positive on [-1,1] , λ > 0 and β; γ > 0.
Let u1 (x;λ) and u2 (x;λ) be linearly independent solutions of the classical differential equation τu+λu=0. Then, the general solution of the fuzzy differential equation (3.1) is
Also, since u1 (x;λ) and u2 (x;λ) are linearly independent solutions of the classical differential equation $\tau u+\lambda u=0,$the solution of the equation is
φ(x;λ) be the solution of the classical differential equation τu+lu = 0 satisfying the conditions $u\left( -1 \right)=$$1,{u}'\left( -1 \right)=-1.$Using boundary conditions, we have
Again,χ (x,λ) be the solution of the classical differential equation τu+lu = 0 satisfying the conditions $u\left( 1 \right)=\gamma ,u\left( 1 \right)=-\lambda \beta .$Similarly, χ(x;λ) is obtained as
is the solution of the equation (3.1) satisfying the conditions (3.5), where $\left[ {{c}_{1}}\left( \alpha \right),{{c}_{2}}\left( \alpha \right) \right]={{\left[ 1 \right]}^{\alpha }}.$We take ${{\left[ 1 \right]}^{\alpha }}=\left[ \alpha ,2-\alpha \right].$From here,
are obtained. Computing the value $\varphi \left( x,\lambda \right){\chi }'\left( x,\lambda \right)-\chi \left( x,\lambda \right){\varphi }'\left( x,\lambda \right),$we have
Considering the 3.7), the value is ${{\underline{c}}_{1\alpha }}\left( \lambda \right){{\underline{c}}_{4\alpha }}\left( \lambda \right)-{{\underline{c}}_{2\alpha }}\left( \lambda \right){{\underline{c}}_{3\alpha }}\left( \lambda \right)$
Theorem 1. The Wronskian functions$W\left( {{\underline{\varphi }}_{\alpha }},{{\underline{\chi }}_{\alpha }} \right)$(x;λ) and W$\left( {{{\bar{\varphi }}}_{\alpha }},{{{\bar{\chi }}}_{\alpha }} \right)$(x;λ) are independent of variable x for x ∈ (–1;1), where functions$\underline{{{\varphi }_{\alpha }}},\underline{{{\chi }_{\alpha }}},{{\bar{\varphi }}_{\alpha }},{{\bar{\chi }}_{\alpha }}$are the solution of the fuzzy boundary value problem (3.1)-(3.3).
Proof. Derivating of equations W$\left( {{\underline{\varphi }}_{\alpha }},{{\underline{\chi }}_{\alpha }} \right)$(x;λ) and W$\left( {{{\bar{\varphi }}}_{\alpha }},{{{\bar{\chi }}}_{\alpha }} \right)$(x;λ) according to variable x and using the functions ${{\left[ \varphi \left( x,\lambda \right) \right]}^{\alpha }},{{\left[ \chi \left( x,\lambda \right) \right]}^{\alpha }}$are the solutions of the equation (3.1)
Theorem 2. The eigenvalues of the fuzzy boundary value problem (3.1)-(3.3) if and only if are consist of the zeros of functions${{\underline{W}}_{\alpha }}\left( \lambda \right)$and${{\overline{W}}_{\alpha }}\left( \lambda \right).$
Proof. Let be λ = λ0 is the eigenvalue. We show that ${{\underline{W}}_{\alpha }}\left( {{\lambda }_{0}} \right)=0$and ${{\overline{W}}_{\alpha }}\left( {{\lambda }_{0}} \right)=0.$We assume that ${{\underline{W}}_{\alpha }}\left( {{\lambda }_{0}} \right)\ne 0\,\text{or}\,{{\overline{W}}_{\alpha }}\left( {{\lambda }_{0}} \right)\ne 0\,.$Let be ${{\underline{W}}_{\alpha }}\left( {{\lambda }_{0}} \right)\ne 0.$Then, the functions ${{\underline{\varphi }}_{\alpha }}\left( x,{{\lambda }_{0}} \right)$and ${{\underline{\chi }}_{\alpha }}\left( x,{{\lambda }_{0}} \right)$are linearly independent. So, the general solution of the equation (3.1)
Using the boundary condition (3.2) and using the solution function ${{\left[ \varphi \left( x,{{\lambda }_{0}} \right) \right]}^{\alpha }}=\left[ {{\underline{\varphi }}_{\alpha }}\left( x,{{\lambda }_{0}} \right),{{{\bar{\varphi }}}_{\alpha }}\left( x,{{\lambda }_{0}} \right) \right]$satisfies the boundary condition (3.2),
From this, since ${{\underline{W}}_{\alpha }}\left( {{\lambda }_{0}} \right)\ne 0,{{b}_{\alpha }}\left( {{\lambda }_{0}} \right)=0$is obtained. Similarly, using the boundary condition (3.3), we obtained ${{a}_{\alpha }}\left( {{\lambda }_{0}} \right)=0.$Thus, ${{\underline{u}}_{\alpha }}\left( x.{{\lambda }_{0}} \right)=0,{{\lambda }_{0}}$is not an eigenvalue. That is, we have a contradiction. Similarly, if$\,{{\overline{W}}_{\alpha }}\left( {{\lambda }_{0}} \right)\ne 0\,,{{\bar{u}}_{\alpha }}\left( x,{{\lambda }_{0}} \right)=0\,$is obtained. λ0 is not an eigenvalue.
Let λ0 be zero of${{\underline{\text{W}}}_{\alpha }}\left( \lambda \right)$and${{\overline{W}}_{\alpha }}\left( \lambda \right).$Then,
That is, the functions ${{\underline{\varphi }}_{\alpha }},{{\underline{\chi }}_{\alpha }}$and ${{\bar{\varphi }}_{\alpha }},{{\bar{\chi }}_{\alpha }}$are linearly dependent. Also, since ${{\left[ \chi \left( x,\lambda \right) \right]}^{\alpha }}$satisfies the boundary condition (3.3), ${{\underline{\chi }}_{\alpha }}\left( x,{{\lambda }_{0}} \right)$and ${{\bar{\chi }}_{\alpha }}\left( x,{{\lambda }_{0}} \right)$satisfy the boundary condition (3.3). In addition, from (3.14) the functions ${{\underline{\varphi }}_{\alpha }}$x;λ0) and ${{\bar{\varphi }}_{\alpha }}\left( x,{{\lambda }_{0}} \right)$satisfy the boundary condition (3.3). So, ${{\left[ \varphi \left( x,{{\lambda }_{0}} \right) \right]}^{\alpha }}$satisfies the boundary condition (3.3). Hence, [φ(x;λ0)]a is the solution of the boundary value problem (3.1)-(3.3) for λ = λ0. Thus, λ = λ0 is the eigenvalue. The proof is complete.
Lemma 1. Let λ = s2. The lower and the upper solutions${{\underline{\varphi }}_{\alpha }}\left( x,\lambda \right),{{\bar{\varphi }}_{\alpha }}\left( x,\lambda \right)$satisfy the following integral equations for k=0 and k=1:
Substituing the identity $q\left( y \right){{\underline{\varphi }}_{\alpha }}\left( y,\lambda \right)=-\lambda {{\underline{\varphi }}_{\alpha }}\left( y,\lambda \right)-{{\underline{\varphi }}^{\prime \prime }}_{a}\left( y,\lambda \right)$in the right side of (3.15)
Similarly ${{\bar{\varphi }}_{\alpha }}\left( x,\lambda \right)$is found. Derivating in these equations according to x, the derivative equations are obtained.
Lemma 2. Let λ = s2. The lower and the upper solutions${{\underline{\chi }}_{\alpha }}\left( x,\lambda \right),{{\bar{\chi }}_{\alpha }}\left( x,\lambda \right)$satisfy the following integral equations for k=0 and k=1:
Proof. Substituing the identity $q\left( y \right){{\underline{\chi }}_{\alpha }}\left( y,\lambda \right)=-\lambda {{\underline{\chi }}_{\alpha }}\left( y,\lambda \right)-{{\underline{\chi }}^{\prime \prime }}_{a}\left( y,\lambda \right)$in the right side of (3.17), integrating by parts twice and using (3.5) yields (3.17) for k=0. Similarly, the equation (3.18) is found for k=0. Derivating in these equations according to x, the derivative equations are obtained.