Hi Emine,
if you could tell us concisely the players, strategy sets, and payoff functions of your game, I am sure some of us can help with finding Nash equilibria of your game.
Best wishes,
Christoph
Christoph Kuzmics
https://homepage.uni-graz.at/de/christoph.kuzmics/
Professor of Microeconomics
School of Business, Economics, and Social Sciences
and Complexity of Life (COLIBRI)
University of Graz
blog: https://gametheory.life/
Hi Emine,
player i’s utility function is linear in i’s own contribution, so setting its derivative equal to 0 just gives you the value of \alpha_i where this utility happens to be constant, i.e. independent of i’s own contribution, \alpha_i = (N-\beta) / (\beta(N-1)). For \alpha_i below this value, utility is linearly decreasing in own contribution, so contributing 0 is strictly dominant, while for \alpha_i above this value, utility is linearly increasing in own contribution, and contributing the whole endowment is optimal. So generically, in equilibrium all those who are altruistically enough contribute everything, while the others contribute nothing.
Best, Ulrich
_________________________________________
Univ.Prof. DDr. Ulrich BERGER
Department of Economics
WU Vienna University of Economics and Business
Department Volkswirtschaft
Wirtschaftsuniversität Wien
Welthandelsplatz 1/D4
1020 Wien
https://www.wu.ac.at/en/economics/people/berger-u/
Von: learning-evolutio...@googlegroups.com
<learning-evolutio...@googlegroups.com> Im Auftrag von Emine Özge Yurdakurban
Gesendet: Montag, 19. August 2024 11:13
An: Learning Evolution and Games Forum <learning-evolutio...@googlegroups.com>
Betreff: [LEG_Forum] Need help for solution
External Sender |
--