Ideal equilibria in noncooperative multicriteria games. (Q1403163)

The authors consider finite noncooperative multicriteria games in the form of a tuple \(G\) = \(\langle N, (X_i)_{i\in N},(u_i)_{i\in N}\rangle\) with a finite set \(N\) of players, where for \(i\in N\), \(X_i\) is a finite set of Player \(i\)'s pure strategies, and \(u_i = (u_{ik})_{k=1}^{r(i)}\) is a vector payoff function of Player \(i\) defined on \(\prod_{j\in N}X_j\) with values in \(R^{\,r(i)}\) (\(r(i)\) are finite). Let \(\Gamma\) denote the mixed extension of game \(G\) obtained in the standard way, with \(\Delta (X_i)\) as the set of Player \(i\)'s mixed strategies. For such games \(\Gamma\) a new equilibrium, called \textit{ideal equilibrium} is introduced. By definition, a strategy profile \(x\in \prod_{j\in N} \Delta (X_i)\) is an ideal equilibrium if for every Player \(i\in N\) and each \(\tilde{x}_i \in \Delta (X_i)\), \(u_i(x_i,x_{-i}) \geq u_i(\tilde{x}_i,x_{-i})\). Two main results have been shown in the paper. For \(i\in N\), let \(\Delta_{r(i)}\) be the set of nonnegative weights \(\lambda_i = (\lambda_{ik})_{k=1}^{r(i)}\) with \(\sum_{k=1}^{r(i)} \lambda_{ik} =1\). For a vector of weights \(\lambda = (\lambda_i)_{i\in N} \in \prod_{i\in N} \Delta_{r(i)}\), let \(\lambda\)-\textit{weighted game} be defined as \(G_{\lambda}\) = \(\langle N, (X_i)_{i\in N},(v_i)_{i\in N}\rangle \), where for all \(i\in N\) and \(x\in \prod_{j\in N} X_i\), \(v_i(x) = \sum_{k=1}^{r(i)}\lambda_{ik}u_{ik}(x)\). The first result says that the set \(IE(G)\) of mixed ideal equilibria of game \(G\) coincides (if nonempty) with the set \(\bigcap_{\lambda\in \Lambda}NE(G_{\lambda})\), where \(NE(G_{\lambda})\) denotes the set of (mixed) Nash equilibria of game \(G_{\lambda}\) and \(\Lambda\) is a known finite set with cardinality at most \(\max_{i\in N}r(i)\). The second main result gives an axiomatic characterization of the set \(IE(G)\). It is based on a consistency axiom. The first result is illustrated by an example of a multicriteria game with all \(r(i)=2\) which was constructed with the help of an \textit{ordinal potential game} and for which it is shown that it has a pure-strategy ideal equilibrium.