Existence and well-posedness for excess demand equilibrium problems (Q6135315)

The authors consider an economic equilibrium problem with \(l\) commodities and \(m\) economic agents dealing with such commodities. The set \(\mathcal{M} =\{1,\ldots ,m\}\) is divided into two subsets \(\mathcal{M}_{s}\) and \( \mathcal{M}_{c}\) which correspond to producers and consumers, respectively. Let \(\mathcal{A}=\{p\in \mathbb{R}_{+}^{l}\mid 0\leq p_{i}\leq \widehat{p} _{i}\}\), where \(\widehat{p}_{i}\) is the maximal price of the \(i\)-th product. The total supply of the \(j\)-th producer, with respect to the price vector \( p\in \mathcal{A}\), is denoted by \(S_{j}(p)=\{y^{\ast }\in \mathcal{K}\mid \left\langle p,y^{\ast }\right\rangle \geq \left\langle p,y\right\rangle \), \( \forall y\in \mathcal{K\}}\), where \(\mathcal{K}\) is a convex, closed and bounded set in \(\mathbb{R}_{+}^{l}\). The total demand of the \(k\)-th consumer is denoted by \(D_{k}(p)=\{x\in \mathcal{K}\mid \left\langle p,x\right\rangle \leq b_{k}\}\), where \(b_{k}\) is the individual budget of the \(k\)-th consumer. The excess demand map \(Z:\mathcal{A}\rightrightarrows \mathbb{R} ^{l}\) defined as: \(Z(p)=D(p)-S(p)\) is proved to be upper semicontinuous, nonempty, compact, and convex valued on \(\mathcal{A}\). The excess demand equilibrium problem is expressed as: Find \(p^{\ast }\in \mathcal{A}\) such that \(Z(p^{\ast })\cap \mathbb{R}_{-}^{l}\neq \emptyset \), where \(\mathbb{R} _{-}^{l}=-\mathbb{R}_{+}^{l}\). Assuming that \(Z\) satisfies Walras's law: for all \(p\in \mathcal{A}\) and \(z\in Z(p)\) \(\left\langle z,p\right\rangle =0\), the authors prove that the solution set of this problem \(Sol(Z,\mathcal{A} )=\{p\in \mathcal{A}\mid Z(p)\cap \mathbb{R}_{-}^{l}\neq \emptyset \}\) is nonempty and compact. For the proof, the authors use Glicksberg's fixed point theorem, the compactness of \(\mathcal{A}\) and the properties of \(Z\). The second main result of the paper proves that, under the same hypothesis, the excess demand equilibrium problem is generalized Levitin-Polyak well-posed in the sense of Painlevé-Kuratowski, that is it has solutions, and every Levitin-Polyak approximating sequence in the sense of Painlevé -Kuratowski has a subsequence converging to a solution. The authors again use the compactness of \(\mathcal{A}\).