The axiomatization of affine oriented matroids reassessed (Q1754395)

The main aim of the paper is to provide a correct proof of Karlander's axiomatization of affine oriented matroids. Before we formulate the main result, we need to recall some basics on oriented matorids (these objects can be considered as a combinatorial abstraction of real hyperplane arrangements). Let \(E\) be a finite set. A signed subset \(X\) of \(E\) is a member of \(\{+,-,0\}^{E}\). Every signed subset \(X\) can be identified with an ordered pair \((X^{+},X^{-})\), according to the sign. Let \(X,Y\) be signed subsets of \(E\). Then \(\bar{X} = X^{+} \cap X^{-}\) is the support of \(X\). The composition \(X \circ Y\) of \(X,Y\) is defined by \((X\circ Y)_{e} = X_{e}\) if \(X_{e} \neq 0\) or \(Y_{e}\) otherwise. The separation set \(S(X,Y)\) of \(X,Y\) is defined by \[ S(X,Y) = (X^{+} \cap Y^{-}) \cup (X^{-} \cap Y^{+}). \] Let \(\mathcal{O}\) be an oriented matroid on \(E\) and fix \(g \in E\). Let \(g^{+}\) be the signed subset \((\{g\},0)\), then the set \(\mathcal{O}_{g^{+}}\) consists of those covectors in \(\mathcal{O}\) whose sign at \(g\) is \(+\). Definition 1. Let \(E\) be a finite set. A set \(\mathcal{W}\) of signed subsets of \(E\) is an affine oriented matroid if and only if there is an oriented matroid \(\mathcal{O}\) on \(E\) such that \(\mathcal{W} = \mathcal{O}_{g^{+}}(E \setminus \{g\})\). Definition 2. Let \(E\) be a finite set. For \(X,Y \in \{+,-,0\}^{E}\) with \(\bar{X} = \bar{Y}\) and \(X \neq Y\) we define {\parindent=8mm \begin{itemize}\item[i)] the \(e\)-th elimination set of \(X\) and \(Y\) for some \(e \in S(X,Y)\) by \[ I_{e}(X,Y) = \{V \in \{+,-,0\}^{E} : \bar{V} \subseteq \bar{X} \setminus \{e\} \text{ and } V_{f} = X_{f} \text{ for all } f \notin S(X,Y)\}. \] \item[ii)] the elimination set of \(X\) and \(Y\) by \[ I(X,Y) = \bigcup_{e \in S(X,Y)} I_{e}(X,Y). \] \end{itemize}} Definition 3. Let \(E\) be a finite set and \(X,Y\) be signed subsets of \(E\). Then \(X+Y\) is given by \((X+Y)_{e} = 0\) if \(e \in S(X,Y)\) or \((X \circ Y)_{e}\) otherwise. Definition 4. Let \(E\) be a finite set and \(\mathcal{W} \subseteq \{+,-,0\}^{E}\). Then {\parindent=8mm \begin{itemize}\item[i)] \(\text{Sym}(\mathcal{W}) = \{V \in \{+,-,0\}^{E} : \pm V \in \mathcal{W}\},\) \item[ii)] \(\text{Asym}(\mathcal{W}) = \{V \in \{+,-,0\}^{E} : V \in \mathcal{W}, -V \notin \mathcal{W} \},\) \item[iii)] \(\mathcal{P}(\mathcal{W}) = \{X + (-Y): X,Y \in \text{Asym}(\mathcal{W}), \bar{X} = \bar{Y} \text{ and } I(X,-Y) \cap \mathcal{W} = I(-X,Y) \cap \mathcal{W} = \emptyset\}.\) \end{itemize}} Now we are ready to present the main result of the paper (the idea is due to Karlander). Main Theorem. A set \(\mathcal{W} \subseteq \{+,-,0\}^{E}\) is an affine oriented matroid if and only if \(\mathcal{W}\) satisfies the following: {\parindent=8mm \begin{itemize}\item[A1)] if \(X,Y \in \mathcal{W}\), then \(X \circ (\pm Y) \in \mathcal{W}\), \item[A2)] if \(X,Y \in \mathcal{W}\) with \(\bar{X} = \bar{Y}\), then \(I_{e}(X,Y) \cap \mathcal{W} \neq \emptyset\) for all \(e \in S(X,Y)\), \item[A3)] \(\mathcal{P}(\mathcal{W}) \circ \mathcal{W} \subseteq \mathcal{W}\). \end{itemize}}