In \cite{Romauera92}, Romaguera pointed out that if $X=\mathbb{R}^+$ and $p : X \times X\to \mathbb{R}^+$ defined by $p(x, y) = \max\{x, y\}$ for all $x, y \in X$ then ${CB}^p(X)=\emptyset$ and the approach used in Theorem \ref{THM201} and elsewhere has a disadvantage that the fixed point theorems for self-mappings may not be derived from it, when ${CB}^p(X)=\emptyset$. To overcome from this problem he introduced the concept of mixed multi-valued mappings and obtained a different version of Nadler's theorem in a partial metric spaces. \begin{definition} Let $(X, p)$ be a partial metric space. A mapping $T : X \to X \cup {CB}^p(X)$ is called a mixed multi-valued mapping on $X$ if $T$ is a multi-valued mapping on $X$ such that for each $x\in X$ …show more content…
\end{theorem} \begin{proof} It can be completed using the proofs of Theorem \ref{THM202} above and Theorem 2.2 \cite{Singh91}. \end{proof} The following example illustrate our results. \begin{example}\label{EXM302} \cite{Jleli90}. Let $X=\{0, 1, 2\}$ and $p : X \times X\to \mathbb{R}^{+}$ defined by \begin{eqnarray*} & &p(0,0)=0,\; p(1,1)=p(2,2)=\dfrac{1}{4},\; p(1,0)=\dfrac{1}{3},\; p(2,0)=\dfrac{3}{5},\; p(2,1)=\dfrac{2}{5},\\ & &\text{and $p(x,y)=p(y,x)$ for all } x,y\in X. \end{eqnarray*} Clearly, $(X, p)$ is a $0$-complete partial metric space. Now, define the mapping $S:X\rightarrow C^{p}(X)$ such that \[ Sx= \begin{cases} \{0\} \text{ if $x \neq 2, $} \\ \{0,1\} \text{ if $x = 2. $} \end{cases} \] It can be easily seen that $S$ asymptotically regular and for all $x, y\in X$, \[ H_p(Sx, Sy)\leq\varphi(\mu(x,y)) \] with $\gamma\geq 1$ and $\varphi(t)=\frac{3t}{4}$. Therefore the assumptions of Theorem \ref{THM202} are fulfilled. \end{example} The following example shows the generality of our results. \begin{example}\label{EXM302} Let $X=\{0, 1, 4\}$ be endowed with the partial metric $p : X \times X\to \mathbb{R}^{+}$ defined by