Abstract
We study nonatomic games in which players’ choices are guided by general preferences. Rather than ones over actions while also under influences of player-action profiles, we let the preferences be over returns received by individual players and let the returns be then linked to all players’ actions. Our modeling choice has rendered otherwise standard analysis quite fruitful. Not only can we establish equilibrium existence results, but we can also derive the upper hemi-continuity of equilibrium-environment sets with respect to the return function and players’ preference profile. Advances concerning pure equilibria can also be made on a framework involving a rich set of players, cruder traits, and an externality midway between semi-anonymity and anonymity.
Similar content being viewed by others
Data availability
No real data set has been used in the derivation of the paper’s results.
References
Aliprantis CD, Border K (1999) Infinite dimensional analysis: a Hitchhiker’s guide. Springer-Verlag, Berlin
Balder EJ (1995) A unifying approach to existence of Nash equilibria. Int J Game Theory 24:79–94
Balder EJ (1999) On the existence of Cournot–Nash equilibria in continuum of games. J Math Econ 32:207–223
Balder EJ (2000) Incompatibility of usual conditions for equilibrium existence in continuum economies without ordered preferences. J Econ Theory 93:110–117
Balder EJ (2002) A unifying pair of Cournot–Nash equilibrium results. J Econ Theory 102:437–470
Borglin A, Keiding H (1976) Existence of equilibrium actions and of equilibrium: a note on the “new’’ existence theorems. J Math Econ 3:313–316
Buckley JJ (1974) Graphs of measurable functions. Proc Am Math Soc 44:78–80
Debreu G (1952) A social equilibrium existence theorem. Proc Natl Acad Sci 38:886–893
Grant S, Meneghel I, Tourky R (2016) Savage games. Theor Econ 11:641–682
He W, Sun X, Sun Y (2017) Modeling infinitely many agents. Theor Econ 12:771–815
Hildenbrand W (1974) Core and equilibria of a large economy. Princeton University Press, Princeton, NJ
Keisler HJ, Sun Y (2009) Why saturated probability spaces are necessary. Adv Math 221:1584–1607
Khan MA, Papageorgiou NS (1987) On Cournot–Nash equilibria in generalized qualitative games with an atomless measure space of agents. Proc Am Math Soc 100:505–510
Khan MA, Rath KP, Sun Y (1997) On the existence of pure strategy equilibrium in games with a continuum of players. J Econ Theory 76:13–46
Khan MA, Rath KP, Sun Y, Yu H (2013) Large games with a bio-social typology. J Econ Theory 148:1122–1149
Khan MA, Rath KP, Yu H, Zhang Y (2013) Large distributional games with traits. Econ Lett 118:502–505
Khan MA, Sun Y (1990) On a reformulation of Cournot–Nash equilibria. J Math Anal Appl 146:442–460
Khan MA, Vohra R (1984) Equilibrium in abstract economies without ordered preferences and with a measure space of agents. J Math Econ 13:133–142
Kim T, Prikry K, Yannelis NC (1989) Equilibria in abstract economies with a measure space of agents and with an infinite dimensional strategy space. J Approx Theory 56:256–266
Loeb P, Sun Y (2006) Purification of measure-valued maps. Ill J Math 50:747–762
Martins-da-Rocha VF, Topuzu M (2008) Cournot–Nash equilibria in continuum games with non-ordered preferences. J Econ Theory 140:314–327
Mas-Colell A (1974) An equilibrium existence theorem without complete or transitive preferences. J Math Econ 1:237–246
Mas-Colell A (1984) On a theorem of Schmeidler. J Math Econ 13:201–206
Podczeck K (2009) On purification of measure-valued maps. Econ Theor 38:399–418
Qiao L, Yu H (2017) On the space of players in idealized limit games. J Econ Theory 153:177–190
Qiao L, Yu H, Zhang Z (2016) On the closed-graph property of the Nash equilibrium correspondence in a large game: a complete characterization. Games Econ Behav 99:89–98
Rath KP (1992) A direct proof of the existence of pure strategy equilibria in games with a continuum of players. Econ Theor 2:427–433
Schmeidler D (1969) Competitive equilibria in markets with a continuum of traders and incomplete preferences. Econometrica 37:578–585
Schmeidler D (1973) Equilibrium points of nonatomic games. J Stat Phys 7:295–300
Shafer W, Sonnenschein H (1975) Equilibrium in abstract economies without ordered preferences. J Math Econ 2:345–348
Wiszniewska-Matyszkiel A (2000) Existence of pure equilibria in games with nonatomic space of players. Topol Methods Nonlinear Anal 16:339–349
Yang J (2011) Asymptotic interpretations for equilibria of nonatomic games. J Math Econ 47:491–499
Yang J (2018) Game-theoretic modeling of players’ ambiguities on external factors. J Math Econ 75:31–56
Yannelis NC (1987) Equilibria in noncooperative models of competition. J Econ Theory 41:96–111
Yannelis NC, Prabhakar ND (1983) Existence of maximal elements and equilibria in linear topological spaces. J Math Econ 12:233–245
Acknowledgements
Section 6 concerning a saturated player space was initially suggested by an anonymous referee in an earlier submission. The author would like to thank him/her. Also, this research benefits from the support of Hong Kong RGC Project No. CityU 11505219.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
1.1 Technical details of Sect. 4
Proof of Proposition 1
For any player \(t\in T\) and closed action subset \(A'\subseteq A\). we have from (11) that
Fix some closed \(A'\). Suppose a sequence \(t_n\) in T converges to t and \([{\mathbb {B}}(\delta |\rho ,\psi )](t_n)\cap A'\ne \emptyset\) for each n. Then by (28), for each n there is some \(a_n\in A'\) so that
Since \(A'\) is closed in A while the latter is compact, \(A'\) is compact as well. We might as well suppose that \(a_n\rightarrow a\) for some \(a\in A'\). Also, \(t_n\rightarrow t\) and that \(\rho \in \mathcal{R}\equiv {\mathscr {C}}(T\times A\times \mathcal{D},R)\) is uniformly continuous in (t, a) because T and A are compact. Hence, for any \(\epsilon >0\),
as long as n is large enough. Since \(\psi \in \mathcal{P}\equiv {\mathscr {C}}(T,\Psi )\), we have \(\psi (t_n)\in (\psi (t))^{\epsilon /2}\) when n is large enough. Therefore, \(\rho (t,A,\delta )\times \rho (t,\{a\},\delta )\subseteq (\psi (t))^{\epsilon }\). Because \(\psi (t)\) is closed,
By (28), the limit point t would satisfy \([\mathbb B(\delta |\rho ,\psi )](t)\cap A'\ne \emptyset\) as well. \(\square\)
Proof of Proposition 2:
Let \((t_n,a_n,\delta _n,\rho _n,\psi _n)_{n=1,2,\ldots }\) be a sequence in \(T\times A\times \mathcal{D}\times \mathcal{R}\times \mathcal{P}\) and let \((t,a,\delta ,\rho ,\psi )\) be a point of the aforementioned product space. Suppose that \((t_n,a_n)\in {\mathbb {B}}(\delta _n|\rho _n,\psi _n)\) for each \(n=1,2,\ldots\) and also that \(t_n\rightarrow t\), \(a_n\rightarrow a\), \(\delta _n\rightarrow \delta\), \(\rho _n\rightarrow \rho\), and \(\psi _n\rightarrow \psi\).
Fix an \(\epsilon >0\). By \(d_\mathcal{R}(\rho _n,\rho )\rightarrow 0^+\) and \(d_\mathcal{R}\)’s definition by (14),
for any \(a'\) and m would occur when n is large enough. By \(\rho\)’s membership to \(\mathcal{R}\equiv {\mathscr {C}}(T\times A\times \mathcal{D},R)\) whose members are uniformly continuous due to the compactness of \(T\times A\times \mathcal{D}\) and by the convergences of \(t_n\) to t, \(a_n\) to a, and \(\delta _n\) to \(\delta\), on the other hand,
would occur when n is large enough. Hence for n sufficiently large,
and
However, by (11) and the hypothesis that \((t_n,a_n)\in {\mathbb {B}}(\delta _n|\rho _n,\psi _n)\) for any n,
In view of (3), we have from the above three that, for a large enough n,
By \(d_\mathcal{P}(\psi _n,\psi )\rightarrow 0^+\) and \(d_\mathcal{P}\)’s definition by (15), \(d_{\Phi }(\psi _n(t_m),\psi (t_m))<\epsilon /4\) for any m would occur when n is large enough. By \(\psi\)’s membership to \(\mathcal{P}\) and the convergence of \(t_n\) to t, on the other hand, \(d_{\Phi }(\psi (t_n),\psi (t))<\epsilon /4\) would occur when n is large enough. Hence, \(d_{\Phi }(\psi _n(t_n),\psi (t))<\epsilon /2\) would occur when n is large enough. By (2), this implies that
when n is large enough. When we combine the above two, we would certainly get
Since \(\psi (t)\) is closed and \(\epsilon\) is arbitrary, we can conclude from the above that \(\rho (t,A,\delta )\times \rho (t,\{a\},\delta )\subseteq \psi (t)\) and hence by (11), \((t,a)\in {\mathbb {B}}(\delta |\rho ,\psi )\). Thus, \(\mathbb B(\cdot |\cdot ,\cdot )\) is closed and hence closed-valued. It is also upper hemi-continuous and compact-valued as \(T\times A\) is compact. \(\square\)
Proof of Proposition 3:
Recall that \({\mathbb {B}}(\delta |\rho ,\psi )\) has a nonempty intersection with any \(\{t\}\times A\). It can thus be treated as a correspondence from T to A. By Proposition 1, for any closed subset \(A'\) of A,
Proposition 2 has shown \({\mathbb {B}}(\delta |\rho ,\psi )\) to be closed in \(T\times A\). Hence, each section \(\left[ \mathbb B(\delta |\rho ,\psi )\right] (t)\) is closed and therefore compact in the compact set A. By Lemma 18.2 of Aliprantis and Border (1999), the above would guarantee that, for any open subset \(A'\) of A,
By the Kuratowski–Ryll-Nardzewski measurable selection theorem—see, e.g., 18.13 of Aliprantis and Border (1999), there would be a measurable function \(b^*(\delta |\rho ,\psi ):T\rightarrow A\) so that \([b^*(\delta |\rho ,\psi )](t)\in [{\mathbb {B}}(\delta |\rho ,\psi )](t)\) for any \(t\in T\).
Meanwhile, \(b^*(\delta |\rho ,\psi )\)’s graph \(B^*(\delta |\rho ,\psi )\) would belong to \({\mathscr {B}}(T\times A)\) which is also the product \(\sigma\)-field \({\mathscr {B}}(T)\otimes {\mathscr {B}}(A)\); see, e.g., Buckley (1974). We can then define a probability measure \(\delta '(\delta |\rho ,\psi )\in {\mathscr {P}}(T\times A)\) so that for any \(T'\in {\mathscr {B}}(T)\) and \(A'\in {\mathscr {B}}(A)\),
Clearly, \(\delta '(\delta |\rho ,\psi )|_T=\theta\) and hence \(\delta '(\delta |\rho ,\psi )\in \mathcal{D}|_T(\theta )\) by (9). Also,
By Proposition 2, \({\mathbb {B}}(\delta |\rho ,\psi )\) is closed and hence a member of \({\mathscr {B}}(T\times A)\). As \(B^*(\delta |\rho ,\psi )\subseteq {\mathbb {B}}(\delta |\rho ,\psi )\), we have \([\delta '(\delta |\rho ,\psi )]({\mathbb {B}}(\delta |\rho ,\psi ))=1\) as well. In view of (12), these are tantamount to the nonemptiness of \({\mathbb {F}}(\delta |\rho ,\psi )\). As the convex combination of two \(\delta '\)’s satisfying the requirement of (12) would still be in compliance, the latter is also convex. \(\square\)
Proof of Proposition 4:
Let \((\delta '_n,\delta _n,\rho _n,\psi _n)_{n=1,2,\ldots }\) be a sequence in \(\mathcal{D}|_T(\theta )\times \mathcal{D}|_T(\theta )\times \mathcal{R}\times \mathcal{P}\) and let \((\delta ',\delta ,\rho ,\psi )\) be a point of the aforementioned product space. Suppose that \(\delta '_n\in {\mathbb {F}}(\delta _n|\rho _n,\psi _n)\) for each \(n=1,2,\ldots\) and also that \(\delta '_n\rightarrow \delta '\), \(\delta _n\rightarrow \delta\), \(\rho _n\rightarrow \rho\), and \(\psi _n\rightarrow \psi\).
Fix an \(\epsilon >0\). The convergences of \(\delta _n\) to \(\delta\), \(\rho _n\) to \(\rho\), and \(\psi _n\) to \(\psi\), as well as Proposition 2 on the upper hemi-continuity of \(\mathbb B(\cdot |\cdot ,\cdot )\) would together lead to that, when n is large enough,
By (12) and the hypothesis that \(\delta '_n\in \mathbb F(\delta _n|\rho _n,\psi _n)\) for any n,
Putting the above two together, we can get that \(\delta '_n((\mathbb B(\delta |\rho ,\psi ))^{\epsilon /2})=1\) for n sufficiently large. But the nature of the Prokhorov metric and the fact that \(d_\mathcal{D}(\delta '_n,\delta ')\equiv \pi _{T\times A}(\delta '_n,\delta ')\rightarrow 0^+\) would dictate that, as long as n is large enough,
For any \(k=1,2,\ldots\), the above would mean that
Since \({\mathbb {B}}(\delta |\rho ,\psi )\) is closed by Proposition 2, it is equal to \(\bigcap _{l=k}^{+\infty }({\mathbb {B}}(\delta |\rho ,\psi ))^{1/l}\). Therefore, \(\delta '({\mathbb {B}}(\delta |\rho ,\psi ))=1\). By (12), this implies \(\delta '\in \mathbb F(\delta |\rho ,\psi )\). Thus, \({\mathbb {F}}(\cdot |\cdot ,\cdot )\) is closed and hence closed-valued. It is also upper hemi-continuous and compact-valued as \(\mathcal{D}|_T(\theta )\) is compact. \(\square\)
1.2 Technical details of Sect. 5
Proof of Proposition 5:
Let \((\tilde{\delta }'_n,t_n,\tilde{\delta }_n,{\tilde{\rho }}_n,\psi _n)_{n=1,2,\ldots }\) be a sequence in \(\tilde{\mathcal{D}}^{\text{ e }}\times T\times \tilde{\mathcal{D}}\times \tilde{\mathcal{R}}\times \mathcal{P}\) and \(({\tilde{\delta }}',t,{\tilde{\delta }},{\tilde{\rho }},\psi )\) a point of the aforementioned product space. Suppose \(\tilde{\delta }'_n\in \left[ \tilde{\mathbb J}({\tilde{\delta }}_n|{\tilde{\rho }}_n,\psi _n)\right] (t_n)\) for each \(n=1,2,\ldots\) and also \(\tilde{\delta }'_n\rightarrow \tilde{\delta }_n\), \(t_n\rightarrow t\), \(\tilde{\delta }_n\rightarrow \tilde{\delta }\), \({\tilde{\rho }}_n\rightarrow {\tilde{\rho }}\), and \(\psi _n\rightarrow \psi\). As \(\tilde{\mathcal{D}}^{\text{ e }}\) is finite, we may suppose that all the \(\tilde{\delta }'_n\)’s and \(\tilde{\delta }'\) are some common \(e_i\).
Fix an \(\epsilon >0\). Using similar arguments as the ones in the proof of Proposition 2, but this time using \(d_{\tilde{\mathcal{R}}}\)’s definition by (16), we can obtain that, as long as n is large enough,
and
However, by (17) and the hypothesis that \(e_i\in \left[ \tilde{\mathbb J}({\tilde{\delta }}_n|{\tilde{\rho }}_n,\psi _n)\right] (t_n)\) for any n,
In view of (3), we have from the above three that, for a large enough n,
Using the same arguments as in the proof of Proposition 2, we can also obtain
when n is large enough. When we combine the above two, it would happen that
Since \(\psi (t)\) is closed and \(\epsilon\) is arbitrary, we can conclude from the above that \({\tilde{\rho }}(t,A,{\tilde{\delta }})\times {\tilde{\rho }}(t,\{a_i\},{\tilde{\delta }})\subseteq \psi (t)\) and hence by (17), \(e_i\in \left[ \tilde{\mathbb J}({\tilde{\delta }}|{\tilde{\rho }},\psi )\right] (t)\). Thus, \(\left[ \tilde{{\mathbb {J}}}(\cdot |\cdot ,\cdot )\right] (\cdot )\) is closed and hence closed-valued. It is also upper hemi-continuous and compact-valued as \(\tilde{\mathcal{D}}^{\text{ e }}\) is finite. \(\square\)
Proof of Proposition 6:
Because \(\left[ \tilde{\mathbb J}({\tilde{\delta }}|{\tilde{\rho }},\psi )\right] (t)\ne \emptyset\), we know \(\tilde{{\mathbb {F}}}(\tilde{\delta }|\tilde{\rho },\psi )\) of (18) is nonempty. As \(\theta\) is atomless, Theorem D.II.3 of Hildenbrand (1974) would guarantee \(\tilde{\mathbb F}(\tilde{\delta }|\tilde{\rho },\psi )\)’s convexity.
Let \(({\tilde{\delta }}_n,{\tilde{\rho }}_n,\psi _n)_{n=1,2,\ldots }\) be a sequence in \(\tilde{\mathcal{D}}\times \tilde{\mathcal{R}}\times \mathcal{P}\) that converges to some \(({\tilde{\delta }},{\tilde{\rho }},\psi )\) in the aforementioned product space. Then, Proposition 5 would lead to
On the other hand, Theorem D.II.6 of Hildenbrand (1974) and (18) would result in
Thus, by combining (18) with the above two, we would obtain
This amounts to asserting that \(\tilde{\mathbb F}(\cdot |\cdot ,\cdot )\) is closed and hence closed-valued. It is also upper hemi-continuous and compact-valued as \(\tilde{\mathcal{D}}\) is compact. \(\square\)
1.3 Technical details of Sect. 6
Proof of Proposition 7:
Define a subset of \(R\times R\times \Psi\):
Note \(\mathcal{I}(a,a',\delta )\equiv \left[ {\hat{\rho }}(\cdot ,a,\delta ),{\hat{\rho }}(\cdot ,a',\delta ),{\hat{\psi }}(\cdot )\right] ^{-1}(\Sigma )\). By the measurabilities of \({\hat{\rho }}(\cdot ,a,\delta )\), \({\hat{\rho }}(\cdot ,a',\delta )\), and \({\hat{\psi }}(\cdot )\), all we have to show is that \(\Sigma\) of (29) belongs to \(\mathscr {B}(R\times R\times \Psi )\).
Indeed, \(\Sigma\) is a closed subset of \(R\times R\times \Psi\). Suppose \((r,r',\psi )\in (R\times R\times \Psi ){\setminus }\Sigma\). Then by (29) it would follow that \((r,r')\in (R\times R){\setminus }\psi\). Since \(\psi\) in \(\Psi\) of (1) is a closed subset of \(R\times R\), there would be some \(\epsilon\)-neighborhood of \((r,r')\) so that within it any \((r_0,r'_0)\) is still outside \(\psi\). Then, by (2) and (3), any \((r_0,r'_0)\) in the \((\epsilon /2)\)-neighborhood of \((r,r')\) would be outside any \(\psi _0\) in the \((\epsilon /2)\)-neighborhood of \(\psi\). Therefore, \((R\times R\times \Psi ){\setminus }\Sigma\) is open and hence \(\Sigma\) is closed. \(\square\)
Proof of Proposition 8:
Due to \(\hat{\rho }(i,\cdot ,\delta )\)’s continuity and A’s compactness, \(\hat{\rho }(i,A,\delta )\) would be compact. Also, \({\hat{\psi }}(i)\) as an irreflexive and transitive preference. Then, by Lemma 2 of Schmeidler (1969) or Proposition 3 of Yang (2018), \(\hat{{\mathbb {B}}}^A(i,\delta |{\hat{\Gamma }})\) of (25) would be nonempty.
For upper-continuity, let \((a_n,\delta _n)_{n=1,2,\ldots }\) be a sequence in \(A\times \mathcal{D}\) and let \((a,\delta )\) be a point of the aforementioned product space. Suppose that \(a_n\in \hat{\mathbb B}^A(i,\delta _n|{\hat{\Gamma }})\) for each \(n=1,2,\ldots\) and also that \(a_n\rightarrow a\) and \(\delta _n\rightarrow \delta\). Fix an \(\epsilon >0\). By \({\hat{\rho }}(i)\)’s membership to \(\hat{\mathcal{R}}\equiv {\mathscr {C}}(A\times \mathcal{D},R)\),
would occur when n is large enough. Then, for n sufficiently large,
However, by (25) and the hypothesis that \(a_n\in \hat{{\mathbb {B}}}^A(i,\delta _n|{\hat{\Gamma }})\) for any n,
In view of (3), we have from the above that, for a large enough n,
Since \({\hat{\psi }}(i)\) is closed and \(\epsilon\) is arbitrary, we can conclude from the above that \({\hat{\rho }}(i,A,\delta )\times {\hat{\rho }}(i,\{a\},\delta )\subseteq {\hat{\psi }}(i)\) and hence by (25), \(a\in \hat{\mathbb B}^A(i,\delta |{\hat{\Gamma }})\) as well. Thus, \(\hat{\mathbb B}^A(i,\cdot |{\hat{\Gamma }})\) is closed and hence closed-valued. It is also upper hemi-continuous and compact-valued as A is compact. \(\square\)
Proof of Proposition 9:
Since A is compact and hence separable, we can suppose some \(\{a_n:\;n=1,2,\ldots \}\) to be dense in it. As \(A'\) is a closed subset of A, we can further suppose some \(\{a'_m:\;m=1,2,\ldots \}\) to be dense in the former. Let
which is a member of \({\mathscr {I}}\) due to Proposition 7. Define
Just because each \({\hat{\rho }}(i,\cdot ,\delta )\) is continuous, each \(\hat{\psi }(i)\) is closed, and the \(a_n\)’s form a dense subset of A, we have from the above two definitions that
By (25), we have from its own definition that \({\hat{I}}(A',\delta |\hat{\Gamma })\) is equal to
As \(\hat{\rho }(i,\cdot ,\delta )\) is continuous, \(\{a'_m:\;m=1,2,\ldots \}\) is dense in \(A'\), and \(\psi\) is closed,
According to (30), \({\hat{I}}(A',\delta |\hat{\Gamma })\) can be further equated to
Combining the above three, we may see that \({\hat{I}}(A',\delta |\hat{\Gamma })\in {\mathscr {I}}\). \(\square\)
Proof of Proposition 10:
By Proposition 9 and Lemma 18.2 of Aliprantis and Border (1999) which allows the closed \(A'\) in the former to be replaced by an open one, we can apply the Kuratowski–Ryll-Nardzewski measurable selection theorem (18.13 of Aliprantis and Border (1999)) to demonstrate the existence of some \({\hat{b}}^A(\cdot ,\delta |\hat{\Gamma })\in \hat{\mathcal{A}}\equiv \mathscr {M}(I,A)\) such that each \({\hat{b}}^A(i,\delta |\hat{\Gamma })\in \hat{{\mathbb {B}}}(i,\delta |{\hat{\Gamma }})\). Let each \({\hat{b}}(i,\delta |\hat{\Gamma })\) be \(\left( {\hat{\theta }}(i),{\hat{b}}^A(i,\delta |\hat{\Gamma })\right)\). Then, \({\hat{b}}(\cdot ,\delta |\hat{\Gamma })\) would be a measurable selection of the correspondence \(\hat{{\mathbb {B}}}(\cdot ,\delta |{\hat{\Gamma }})\) defined by (26). Therefore, \(\hat{\mathbb F}(\delta |{\hat{\Gamma }})\) of (27) is nonempty. According to Proposition D(P1) of Khan et al. (2013a), the latter is also convex.
By (26) and Proposition 8, we also have \(\hat{{\mathbb {B}}}(i,\cdot |{\hat{\Gamma }})\) being upper hemi-continuous and hence closed-valued at each \(i\in I\). In addition, (26), Proposition 9, and Lemma 18.2 of Aliprantis and Border (1999) would together lead to the measurability of the correspondence \(\hat{{\mathbb {B}}}(\cdot ,\delta |\hat{\Gamma })\) in both the normal and weak senses. Then, Propositions D(P2) and D(P3) of Khan et al. (2013a) would result in the upper hemi-continuity of \(\hat{{\mathbb {F}}}(\cdot |{\hat{\Gamma }})\) defined by (27). At every \(\delta \in \mathcal{D}|_T(\lambda \circ {\hat{\theta }}^{-1})\), the set \(\hat{{\mathbb {F}}}(\delta |{\hat{\Gamma }})\) is thus closed. It is also compact since \(\mathcal{D}|_T(\lambda \circ {\hat{\theta }}^{-1})\) is. \(\square\)
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Yang, J. Nonatomic game with general preferences over returns. Int J Game Theory 52, 861–889 (2023). https://doi.org/10.1007/s00182-023-00843-6
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00182-023-00843-6