Abstract
In real life, individuals often need to pay some costs to build and maintain long-termed relationships as well as interactions among them. However, previous studies of the repeated public goods game have focused almost exclusively on the cost-free participation. Here we introduce costly participation to the repeated public goods game in a conditional manner, and study the evolution of cooperation in both deterministic and stochastic dynamics for well-mixed populations. In the limit of an infinite population size, the deterministic dynamics can lead to either a stable coexistence between cooperators and defectors or even a complete dominance of cooperators over defectors if the initial frequency of cooperators is larger than some invasion barrier. In general, defectors are always able to resist invasion by cooperators. However, in finite populations, we show that natural selection can favor the emergence of cooperation in the stochastic dynamics. In the limit of weak selection and large populations, we derive a critical condition required for a cooperator to replace a population of defectors with a selective advantage by using several approximation techniques. Theoretical analysis of the critical condition reveals that participation cost determines the impacts of the number of game round in the emergence of cooperation. Interestingly, there exists an intermediate value of the participation threshold of cooperators leading to the optimal condition for a cooperator to invade and fixate in a population of defectors if the participation cost is smaller than a critical value. Numerical calculations confirm that the validity of the analytical approximations extends to much wider ranges of the selection strength as well as of the population size.
Similar content being viewed by others
References
Allen JM, Skeldon AC, Hoyle RB (2018) Social influence preserves cooperative strategies in the conditional cooperator public goods game on a multiplex network. Phys Rev E 98:062305
Altrock PM, Traulsen A (2009) Deterministic evolutionary game dynamics in finite populations. Phys Rev E 80:011909
Axelrod R (1984) The evolution of cooperation. Basic Books, New York
Battu B, Pammi VSC, Srinivasan N (2018) Evolution of cooperation with heterogeneous conditional cooperators. Sci Rep 8:4524
Ezaki T, Horita Y, Takezawa M, Masuda N (2016) Reinforcement learning explains conditional cooperation and its moody cousin. PLoS Comput Biol 12:e1005034
Ginsberg AG, Fu F (2018) Evolution of cooperation in public goods games with stochastic opting-out. Games 10(1):1–27
Grujić J, Cuesta JA, Sánchez A (2012) On the coexistence of cooperators, defectors and conditional cooperators in the multiplayer iterated prisoner’s dilemma. J Theor Biol 300:299–308
Hardin G (1968) The tragedy of the commons. Science 162:1243–1248
Hauert C, De Monte S, Hofbauer J, Sigmund K (2002) Volunteering as Red Queen mechanism for cooperation in public goods game. Science 296:1129–1132
Hilbe C, Wu B, Traulsen A, Nowak MA (2014) Cooperation and control in multiplayer social dilemmas. Proc Natl Acad Sci USA 111:16425–16430
Hilbe C, Wu B, Traulsen A, Nowak MA (2015) Evolutionary performance of zero-determinant strategies in multiplayer games. J Theor Biol 374:115–124
Hilbe C, Chatterjee K, Nowak MA (2018) Partners and rivals in direct reciprocity. Nat Hum Behav 2:469–477
Hofbauer J, Sigmund K (1998) Evolutionary games and population dynamics. Cambridge University Press, Cambridge
Karlin S, Taylor HMA (1975) A first course in stochastic processes, 2nd edn. Academic, London
Kurokawa S, Ihara Y (2009) Emergence of cooperation in public goods game. Proc R Soc Lond B 276:1379–1384
Li A, Broom M, Du J, Wang L (2016) Evolutionary dynamics of general group interactions in structured populations. Phys Rev E 93:022407
Liu L, Chen X, Perc M (2019) Evolutionary dynamics of cooperation in the public goods game with pool exclusion strategies. Nonlinear Dyn 97:749–766
Masuda N (2007) Participation costs dismiss the advantage of heterogeneous networks in evolution of cooperation. Proc R Soc B 274:1815–1821
Maynard Smith J (1982) Evolution and the theory of games. Cambridge University Press, Cambridge
Milinski M, Semmann D, Krambeck HJ (2002) Reputation helps to solve the ’tragedy of the commons’. Nature 415:424–426
Nowak MA (2006a) Evolutionary dynamics: exploring the equation of life. Harvard University Press, Cambridge, MA
Nowak MA (2006b) Five rules for the evolution of cooperation. Science 314:1560–1563
Nowak MA, Highfield R (2011) Supercooperators: altruism, evolution, and why we need each other to succeed. Free Press, New York
Nowak MA, Sigmund K (1992) Tit for tat in heterogeneous population. Nature 355:250–253
Nowak MA, Sasaki A, Taylor C, Fudenberg D (2004) Emergence of cooperation and evolutionary stability in finite populations. Nature 428:646–650
Pan L, Hao D, Rong Z, Zhou T (2015) Zero-determinant strategies in iterated public goods game. Sci Rep 5:13096
Peña J, Pestelacci E, Berchtold A, Tomassini M (2010) Participation costs can suppress the evolution of upstream reciprocity. J Theor Biol 273:197–206
Perc M (2011) Success-driven distribution of public goods promotes cooperation but preserves defection. Phys Rev E 84:037102
Pinheiro FL, Vasconcelos VV, Santos FC, Pacheco JM (2014) Evolution of all-or-none strategies in repeated public goods dilemmas. PLoS Comput Biol 10:1–5
Rand DG, Dreber A, Ellingsen T, Fudenberg D, Nowak MA (2009) Positive interactions promote public cooperation. Science 325:1272–1275
Sandholm WH (2010) Population games and evolutionary dynamics. MIT University Press, Cambridge, MA
Santos FC, Santos MD, Pacheco JM (2008) Social diversity promotes the emergence of cooperation in public goods games. Nature 454:213–216
Sasaki T, Uchida S (2013) The evolution of cooperation by social exclusion. Proc R Soc B 280:20122498
Sigmund K, Hauert C, Nowak MA (2001) Reward and punishment. Proc Natl Acad Sci USA 98:10757–10762
Szabó G, Tőke C (1998) Evolutionary prisoner’s dilemma game on a square lattice. Phys Rev E 58:69–73
Szolnoki A, Perc M, Szabó G (2009) Topology-independent impact of noise on cooperation in spatial public goods games. Phys Rev E 80:056109
Tanimoto J, Yamauchi A (2010) Does game participation cost affect the advantage of heterogeneous networks for evolving cooperation? Phys A 389:2284–2289
Traulsen A, Hauert C (2009) Stochastic evolutionary game dynamics. In: Schuster HG (ed) Reviews of nonlinear dynamics and complexity. Wiley-VCH, Berlin, pp 25–62
Traulsen A, Claussen JC, Hauert C (2005) Coevolutionary dynamics: from finite to infinite populations. Phys Rev Lett 95:0238701
Traulsen A, Nowak MA, Pacheco JM (2006) Stochastic dynamics of invasion and fixation. Phys Rev E 74:011909
Trivers RL (1971) The evolution of reciprocal altruism. Q Rev Biol 46:35–57
van Kampen NG (1992) Stochastic processes in physics and chemistry. Elsevier, Amsterdam
Van Segbroeck S, Pacheco JM, Lenaerts T, Santos FC (2012) Emergence of fairness in repeated group interactions. Phys Rev Lett 108:158104
Wang J, Fu F, Wu T, Wang L (2009) Emergence of social cooperation in threshold public good games with collective risk. Phys Rev E 80:016101
Acknowledgements
We thank the editor and two anonymous reviewers for constructive comments, which helped us to improve this work.
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.
X.F.W acknowledges gratefully financial support from the NSFC (Grants No. 61903077), the Shanghai Sailing Program (Grants No. 19YF1402500) as well as the Fundamental Research Funds for the Central Universities (Grants Nos. 2232019D3-56 and 2232018G-09).
Appendices
A Detailed Analysis of Payoff Difference
In “Appendix A”, we study in detail the direction of the deterministic dynamics by analyzing the payoff difference \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \). In particular, we analyze the shape of \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \), assuming \({\bar{T}} \in \left\{ {2, \cdots ,N - 1} \right\} \), and later on followed by the analysis for the degenerate cases \({\bar{T}} = N\) and \({\bar{T}} = 1\).
Analysis for \({\bar{T}} \in \left\{ {2, \cdots ,N - 1} \right\} \). Let the polynomial \({G_1}\left( x \right) \) be
Note that \({G_1}\left( x \right) \) has the same shapes of function with the payoff difference in Eq. (13), and thus determines the non-trivial equilibria of the deterministic dynamics described by Eq. (10). Before presenting formally Proposition 2, we first introduce a necessary Lemma 1.
Lemma 1
Let \({\bar{T}} \in \left\{ {2, \cdots ,N - 1} \right\} \). The polynomial \({G_1}\left( x \right) \) satisfies following properties:
-
1.1
\({G_1}\left( 0 \right) = 0\);
-
1.2
\({G_1}\left( 1 \right) = \delta - 1\);
-
1.3
If \(c = \bar{c} = \frac{{\left( {{\bar{T}} - 1} \right) \left( {N\delta - 1} \right) }}{{N - 1}}\), \({G_1}\left( x \right) \) decreases monotonically in \(x \in \left( { - \infty , + \infty } \right) \); Else if \(0 \le c < \bar{c}\), \({G_1}\left( x \right) \) attains a maximum at \({\bar{x}} = \frac{{\left( {{\bar{T}} - 1} \right) \left( {\delta {\bar{T}} - c - 1} \right) }}{{\left( {{\bar{T}} - 1} \right) \left( {\delta N - 1} \right) - c\left( {N - 1} \right) }}\). Moreover, \({G_1}'\left( x \right) = 0\) at \(x = {\bar{x}}\), \({G_1}'\left( x \right) > 0\) for \(x \in \left( { - \infty ,{\bar{x}}} \right) \) and \({G_1}'\left( x \right) < 0\) for \(x \in \left( {{\bar{x}}, + \infty } \right) \); Else if \(c > \bar{c}\), \({G_1}\left( x \right) \) attains a minimum at \({\bar{x}} = \frac{{\left( {{\bar{T}} - 1} \right) \left( {\delta {\bar{T}} - c - 1} \right) }}{{\left( {{\bar{T}} - 1} \right) \left( {\delta N - 1} \right) - c\left( {N - 1} \right) }}\). Moreover, \({G_1}'\left( x \right) = 0\) at \(x = {\bar{x}}\), \({G_1}'\left( x \right) < 0\) for \(x \in \left( { - \infty ,{\bar{x}}} \right) \) and \({G_1}'\left( x \right) > 0\) for \(x \in \left( {{\bar{x}}, + \infty } \right) \).
Proof
The first two properties follow immediately from Eq. (20). We prove the third property now. From Eq. (20), we obtain the following expression for the first order derivative \({G_1}'\left( x \right) \):
For \(j \in \left\{ {0, \cdots ,N - 2} \right\} \), the following recurrence relation is satisfied, i.e.,
so that
Following Eq. (23), we prove the third property of Lemma 1. \(\square \)
Based on Lemma 1, we are able to prove the following Proposition 2:
Proposition 2
Let \({\bar{T}} \in \left\{ {2, \cdots ,N - 1} \right\} \). The function of the payoff difference \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) satisfies the following properties:
-
2.1
If \({{\left( {c + 1} \right) } /{{\bar{T}}}}< \delta < 1\) and \(0 \le c < \bar{c}\), there exists a critical number of rounds \(\bar{R} = 1 + {{\left( {1 - \delta } \right) } /{{G_1} (\bar{x})}}\), which determines the behavior of \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \):
-
2.1.1
If \(1 \le R < \bar{R}\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \).
-
2.1.2
If \(R = \bar{R}\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) has a double real root at \({x^*} = \bar{x} \in \left( {0,1} \right) \). \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in [0,{x^*})\) and \(x \in ({x^*},1]\).
-
2.1.3
If \(R > \bar{R}\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) has two real roots, \(x_L^*\) and \(x_R^*\), with \(x_L^* \in \left( {0,\bar{x}} \right) \) and \(x_R^* \in \left( {\bar{x},1} \right) \). \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in [0,{x_L^*})\) and \(x \in ({x_R^*},1]\), whereas \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) > 0\) for \(x \in ({x_L^*},{x_R^*})\).
-
2.1.1
-
2.2
If \({1 /N} < \delta \le {{\left( {c + 1} \right) } /{{\bar{T}}}}\) and \(0 \le c < \bar{c}\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \).
-
2.3
If \(c \ge \bar{c}\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \).
Proof
2.1 If \({{\left( {c + 1} \right) } /{{\bar{T}}}}< \delta < 1\) and \(0 \le c < \bar{c}\), Lemma 1 states that \({G_1}\left( x \right) \) attains a maximum at \(\bar{x} \in \left( {0,1} \right) \).
2.1.1 Let \(1 \le R < \bar{R}\). We obtain the following inequality by applying Lemma 1:
for \(x \in \left[ {0,1} \right] \).
2.1.2 For \(R = \bar{R}\), we have
To prove that \(\bar{x}\) is a double root, we show that \({\left[ {{f_{{C_{T_C}}}}\left( {\bar{x}} \right) - {f_{{D_{T_D}}}}\left( {\bar{x}} \right) } \right] ^\prime } = 0\) and \({\left[ {{f_{{C_{T_C}}}}\left( {\bar{x}} \right) - {f_{{D_{T_D}}}}\left( {\bar{x}} \right) } \right] ^{\prime \prime }} \ne 0\). From Eq. (23), we obtain \({\left[ {{f_{{C_{T_C}}}}\left( {\bar{x}} \right) - {f_{{D_{T_D}}}}\left( {\bar{x}} \right) } \right] ^\prime } = \left( {R - 1} \right) {G_1}'\left( {\bar{x}} \right) = 0\). Besides, the second order derivative \({\left[ {{f_{{C_{T_C}}}}\left( {\bar{x}} \right) - {f_{{D_{T_D}}}}\left( {\bar{x}} \right) } \right] ^{\prime \prime }}\) at \(x = \bar{x}\) is given by
As \(\delta > {{\left( {c + 1} \right) } /{{\bar{T}}}}\), we have \(c + 1 - \delta {\bar{T}} < 0\) and \(c - \delta \left( {{\bar{T}} - 1} \right)< {{\left( {c + 1} \right) } / {{\bar{T}}}} - 1 < {{\left( {\bar{c} + 1} \right) } /{{\bar{T}}}} - 1\). Since \(\bar{c} = {{\left( {{\bar{T}} - 1} \right) \left( {N\delta - 1} \right) } /{\left( {N - 1} \right) }} < {\bar{T}} - 1\), we further obtain \(c - \delta \left( {{\bar{T}} - 1} \right)< {{\left( {\bar{c} + 1} \right) } /{{\bar{T}}}} - 1 < 0\). Together with \(c + 1 - \delta {\bar{T}} < 0\), we conclude that \({\left[ {{f_{{C_{T_C}}}}\left( {\bar{x}} \right) - {f_{{D_{T_D}}}}\left( {\bar{x}} \right) } \right] ^{\prime \prime }} < 0\). Thus \(\bar{x}\) is a double root for \(x \in \left( { - \infty , + \infty } \right) \).
From Lemma 1, we know that \({G_1}^\prime \left( x \right) > 0\) for \(x \in \left( { - \infty ,\bar{x}} \right] \) and \({G_1}^\prime \left( x \right) < 0\) for \(x \in \left[ {\bar{x}, + \infty } \right) \) when \(0 \le c < \bar{c}\), and thus \({\left[ {{f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) } \right] ^\prime } > 0\) for \(x \in \left( { - \infty ,\bar{x}} \right] \) and \({\left[ {{f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) } \right] ^\prime } < 0\) for \(x \in \left[ {\bar{x}, + \infty } \right) \) when \(0 \le c < \bar{c}\). Hence, \({{f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) }\) increases monotonically in \(\left( { - \infty ,\bar{x}} \right] \) and decreases monotonically in \(\left[ {\bar{x}, + \infty } \right) \). As a result, together with the condition that \(\delta > {{\left( {c + 1} \right) } / {{\bar{T}}}}\) and \({f_{{C_{T_C}}}}\left( 1 \right) - {f_{{D_{T_D}}}}\left( 1 \right) = R\left( {\delta - 1} \right) \le {f_{{C_{T_C}}}}\left( 0 \right) - {f_{{D_{T_D}}}}\left( 0 \right) = \delta - 1 < 0\), we can conclude that \(\bar{x} \in \left( {0,1} \right) \).
2.1.3 For \(R > \bar{R}\), we have
As shown in 2.1.2, if \(\delta > {{\left( {c + 1} \right) } /{{\bar{T}}}}\) and \({f_{{C_{T_C}}}}\left( 1 \right) - {f_{{D_{T_D}}}}\left( 1 \right) = R\left( {\delta - 1} \right) \le {f_{{C_{T_C}}}}\left( 0 \right) - {f_{{D_{T_D}}}}\left( 0 \right) = \delta - 1 < 0\), we can obtain that \(\bar{x} \in \left( {0,1} \right) \). On the other hand, since \({f_{{C_{T_C}}}}\left( 0 \right) - {f_{{D_{T_D}}}}\left( 0 \right) = \delta - 1 < 0\) and \({f_{{C_{T_C}}}}\left( {\bar{x}} \right) - {f_{{D_{T_D}}}}\left( {\bar{x}} \right) > 0\), the Intermediate Value Theorem predicts that \({{f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) }\) has a root \(x_L^*\) in \(\left( {0,\bar{x}} \right) \). Similarly, since \({f_{{C_{T_C}}}}\left( {\bar{x}} \right) - {f_{{D_{T_D}}}}\left( {\bar{x}} \right) > 0\) and \({f_{{C_{T_C}}}}\left( 1 \right) - {f_{{D_{T_D}}}}\left( 1 \right) = R\left( {\delta - 1} \right) < 0\), the Intermediate Value Theorem predicts that \({{f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) }\) has a root \(x_R^*\) in \(\left( {\bar{x},1} \right) \).
2.2 If \(0 \le c < \bar{c}\), \({{f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) }\) increases monotonically in \(\left( { - \infty ,\bar{x}} \right] \) and decreases monotonically in \(\left[ {\bar{x}, + \infty } \right) \). On the other hand, for \({1 / N} < \delta \le {{\left( {c + 1} \right) } /{{\bar{T}}}}\), \(\bar{x} \le 0\). Therefore, \({{f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) }\) decreases monotonically for \(x \in \left[ {0,1} \right] \). As a result, together with the condition that \({f_{{C_{T_C}}}}\left( 0 \right) - {f_{{D_{T_D}}}}\left( 0 \right) = \delta - 1 < 0\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \).
2.3 Lemma 1 states that \({G_1}^\prime \left( x \right) < 0\) for \(x \in \left( { - \infty , + \infty } \right) \) when \(c = \bar{c}\), and thus \({\left[ {{f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) } \right] ^\prime } < 0\) for \(x \in \left[ {0,1} \right] \). Hence, \({{f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) }\) decreases monotonically in \(\left[ {0,1} \right] \). As \({f_{{C_{T_C}}}}\left( 0 \right) - {f_{{D_{T_D}}}}\left( 0 \right) = \delta - 1 < 0\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \).
From Lemma 1, we know that \({G_1}^\prime \left( x \right) < 0\) for \(x \in \left( { - \infty ,\bar{x}} \right] \) and \({G_1}^\prime \left( x \right) > 0\) for \(x \in \left[ {\bar{x}, + \infty } \right) \) when \(c > \bar{c}\). This means that \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) attains a maximum either at \(x = 0\) or \(x = 1\) for \(x \in \left[ {0,1} \right] \). As \({f_{{C_{T_C}}}}\left( 0 \right) - {f_{{D_{T_D}}}}\left( 0 \right) = \delta - 1 < 0\) and \({f_{{C_{T_C}}}}\left( 1 \right) - {f_{{D_{T_D}}}}\left( 1 \right) = R\left( {\delta - 1} \right) < 0\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \). This completes the proof. \(\square \)
Analysis for \({\bar{T}} = N\). When \({\bar{T}} = N\), we obtain the following expression for \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \):
From Eq. (28), we can obtain the following Proposition 3 for \({\bar{T}} = N\).
Proposition 3
Let \({\bar{T}} = N\). \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) satisfies the following properties:
-
3.1
If \(0 \le c < {\bar{c}} = \delta N - 1\), there exists a critical number of rounds \({\bar{R}}_1 = 1 + {{\left( {1 - \delta } \right) } /{\left( {\delta N - c - 1} \right) }}\), which determines the behavior of \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \):
-
3.1.1
If \(1 \le R < {\bar{R}}_1\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \).
-
3.1.2
If \(R \ge {\bar{R}}_1\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) has one simple root \({{x_1}^*}\) in \(\left[ {0,1} \right] \). \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in [0,{{x_1}^*})\) and \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) > 0\) for \(x \in ({{x_1}^*},1]\).
-
3.1.1
-
3.2
If \(c \ge {\bar{c}} = \delta N - 1\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \).
Proof
3.1 When \(0 \le c < {\bar{c}} = \delta N - 1\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) increases monotonously with x. The Intermediate Value Theorem tells us that \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) has one single root if and only if \({f_{{C_{T_C}}}}\left( 0 \right) - {f_{{D_{T_D}}}}\left( 0 \right) < 0\) and \({f_{{C_{T_C}}}}\left( 1 \right) - {f_{{D_{T_D}}}}\left( 1 \right) \ge 0\). Because the condition \({f_{{C_{T_C}}}}\left( 0 \right) - {f_{{D_{T_D}}}}\left( 0 \right) = \delta - 1 < 0\) always holds for \(\delta < 1\), the sign of \({f_{{C_{T_C}}}}\left( 1 \right) - {f_{{D_{T_D}}}}\left( 1 \right) \) determines the existence of the root for \(x \in \left[ {0,1} \right] \).
3.1.1 If \(1 \le R < {\bar{R}}_1\), \({f_{{C_{T_C}}}}\left( 1 \right) - {f_{{D_{T_D}}}}\left( 1 \right) < 0\). Therefore, \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \).
3.1.2 If \(R \ge {\bar{R}}_1\), \({f_{{C_{T_C}}}}\left( 1 \right) - {f_{{D_{T_D}}}}\left( 1 \right) \ge 0\). Thus \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) has one simple root \({{x_1}^*}\) in \(\left[ {0,1} \right] \). \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in [0,{{x_1}^*})\) and \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) > 0\) for \(x \in ({{x_1}^*},1]\).
3.2 If \(c \ge {\bar{c}} = \delta N - 1\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) non-increases monotonously with x for \(x \in \left[ {0,1} \right] \). Hence we have \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \le {f_{{C_{T_C}}}}\left( 0 \right) - {f_{{D_{T_D}}}}\left( 0 \right) = \delta - 1 < 0\) for \(x \in \left[ {0,1} \right] \). \(\square \)
Analysis for \({\bar{T}} = 1\). For \({\bar{T}} = 1\), \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) \) is given by
Eq. (29) proves the below Proposition 4 for \({\bar{T}} = 1\).
Proposition 4
Let \({\bar{T}} = 1\). \({f_{{C_{T_C}}}}\left( x \right) - {f_{{D_{T_D}}}}\left( x \right) < 0\) for \(x \in \left[ {0,1} \right] \).
B Analytical Approximation for Accumulated Difference of Payoff
In “Appendix B”, we provide technical details of analytical approximation for the accumulated difference of payoff \({\sum \nolimits _{i = 1}^{M - 1} {\sum \nolimits _{k = 1}^i {\left[ {{f_{{C_{{T_C}}}}}\left( k \right) - {f_{{D_{{T_D}}}}}\left( k \right) } \right] } } }\) in the limit of a large but finite population size \(M \gg N\). When \(M \gg N\), the hypergeometric probabilities in Eq. (15) can be approximated by binomial probabilities:
and
Introducing the notation \(x = {i / M}\) and \({p_{{C_{{T_C}}}}} = {k / M}\) in the continuum, one can find that the sums in Eq. (14), \(\sum \nolimits _{i = 1}^{M - 1} {\sum \nolimits _{k = 1}^i {\left( \bullet \right) } }\), can be estimated as the the integral, \(\int _0^1 {dx\int _0^x {d{p_{{C_{{T_C}}}}}\left( \bullet \right) } }\). Using Eqs. (15), (30) and (31), the accumulated difference of payoff in Eq. (14) can be approximated as
Let \(B\left( {j,N - j} \right) = \int _0^1 {{p_{{C_{{T_C}}}}}^j{{\left( {1 - {p_{{C_{{T_C}}}}}} \right) }^{N - j}}d{p_{{C_{{T_C}}}}}}\), we have
From above recursive equation, one can obtain
Substituting Eq. (34) into Eq. (32) leads to the analytical approximation of the accumulated payoff difference in the limit of \(M \gg N\):
C Proof of Proposition 1
In “Appendix C”, we prove formally Proposition 1.
Proof
- 1.1 :
-
Following Eq. (17), we directly prove the first property of Proposition 1.
- 1.2 :
-
By simplifying Eq. (17), we have
$$\begin{aligned} \begin{array}{l} {\delta _{th}} = \frac{{M\left( {M + 1} \right) \left( {M - N} \right) N}}{{3\left( {N - {\bar{T}} + 1} \right) \left( {N + {\bar{T}}} \right) }} \times \frac{{\frac{{M - 1}}{{M - N}} - \frac{{N - {\bar{T}} + 2c + 2}}{{N + {\bar{T}}}}}}{{\frac{{M\left( {M + 1} \right) \left( {M - N} \right) N}}{{3\left( {N - {\bar{T}} + 1} \right) \left( {N + {\bar{T}}} \right) }} + \frac{{R - 1}}{{N + 1}}}} + \frac{{N - {\bar{T}} + 2c + 2}}{{N + {\bar{T}}}}. \\ \end{array} \end{aligned}$$(36)From Eq. (36), one can obtain that
- 1.2.1:
-
\({\delta _{th}}\) is decreased with \(R \in \left\{ {1, \cdots , + \infty } \right\} \) if \(c \in \left[ {0,{{\bar{c}}_1}} \right) \);
- 1.2.2:
-
\({\delta _{th}}\) is increased with \(R \in \left\{ {1, \cdots , + \infty } \right\} \) if \(c \in \left( {{{\bar{c}}_1}, + \infty } \right) \);
- 1.2.3:
-
\({\delta _{th}}\) is invariant with \(R \in \left\{ {1, \cdots , + \infty } \right\} \) if \(c = {{\bar{c}}_1}\), where the critical value \({{\bar{c}}_1}\) is given by Eq. (18). This proves the second property of Proposition 1.
- 1.3:
-
If replacing the discrete variable \({\bar{T}} \in \left\{ {1, \cdots ,N} \right\} \) with a continuous variable \(z \in \left[ {1,N} \right] \) in Eq. (17), we obtain the following expression for the first order partial derivative of \({\delta _{th}}\) with respect to z:
$$\begin{aligned} \frac{{\partial {\delta _{th}}}}{{\partial z}} = \frac{{\frac{{{{\left( {R - 1} \right) }^2}\left( {N + c + 1} \right) }}{{2{N^2}{{\left( {N + 1} \right) }^2}}}{G_2}\left( z \right) }}{{{{\left[ {\frac{{M\left( {M + 1} \right) \left( {M - N} \right) }}{6} + \frac{{\left( {R - 1} \right) \left( {N - z + 1} \right) \left( {N + z} \right) }}{{2N\left( {N + 1} \right) }}} \right] }^2}}}, \end{aligned}$$(37)where the polynomial \({{G_2}\left( z \right) }\) is given by
$$\begin{aligned} {G_2}\left( z \right)= & {} - {z^2} + \left[ {\frac{{M\left( {M + 1} \right) N\left( {N + 1} \right) \left( {2M - N - 1} \right) }}{{3\left( {R - 1} \right) \left( {N + c + 1} \right) }} + 2\left( {N + 1} \right) } \right] z \nonumber \\&- {\left( {N + 1} \right) ^2} - \frac{{M\left( {M + 1} \right) N\left( {N + 1} \right) \left[ {2\left( {M - N} \right) \left( {N + c + 1} \right) + 2M - N - 1} \right] }}{{6\left( {R - 1} \right) \left( {N + c + 1} \right) }}.\nonumber \\ \end{aligned}$$(38)From Eq. (38), one can observe that \({G_2}\left( z \right) \) increases with z in the interval \(\left[ {1,N} \right] \). Furthermore,
$$\begin{aligned} {G_2}\left( 1 \right) = - \frac{{M\left( {M + 1} \right) N\left( {N + 1} \right) \left[ {2\left( {M - N - 1} \right) \left( {c + N} \right) + N + 2c + 1} \right] }}{{6\left( {R - 1} \right) \left( {c + N + 1} \right) }} - {N^2} < 0.\nonumber \\ \end{aligned}$$(39)The Intermediate Value Theorem predicts that \({G_2}\left( z \right) \) has one single root in the interval \(\left( {1,N} \right] \) if and only if
$$\begin{aligned} {G_2}\left( N \right)= & {} \frac{{M\left( {M + 1} \right) N\left( {N + 1} \right) }}{{6\left( {R - 1} \right) \left( {c + N + 1} \right) }}\nonumber \\&\times \left[ { - 2\left( {M - N} \right) c + 2MN - 4M + N + 1} \right] - 1 \ge 0, \end{aligned}$$(40)which leads to the critical value \({{\bar{c}}_2}\) given by Eq. (19). Otherwise, there is no root for \({G_2}\left( z \right) \) in the interval \(\left( {1,N} \right] \). Furthermore, one can observe from Eq. (37) that \({G_2}\left( z \right) \) has the same sign and root with \({{\partial {\delta _{th}}} /{\partial z}}\). Therefore, the optimal discrete solution \({\bar{T}}^* \in \left\{ {1, \cdots ,N} \right\} \) resulting in the minimum \({{\delta _{th}}}\) is given by
$$\begin{aligned} \begin{array}{l} {\bar{T}}^* = \left\{ \begin{array}{l} \left\lfloor {z'} \right\rfloor \; {\mathrm{if}}\; {\delta _{th}}\left( {\left\lfloor {z'} \right\rfloor \; } \right) \le {\delta _{th}}\left( {\left\lceil {z'} \right\rceil \; } \right) \; {\mathrm{and}}\; c \le {{\bar{c}}_2}, \\ \left\lceil {z'} \right\rceil \; {\mathrm{if}}\; {\delta _{th}}\left( {\left\lfloor {z'} \right\rfloor \; } \right) > {\delta _{th}}\left( {\left\lceil {z'} \right\rceil \; } \right) \; {\mathrm{and}}\; c \le {{\bar{c}}_2}, \\ N\; \; {\mathrm{otherwise,}} \\ \end{array} \right. \\ \end{array} \end{aligned}$$(41)where \(z'\) corresponding to the left root of \({G_2}\left( z \right) = 0\) is given by
$$\begin{aligned} z'= & {} \frac{{M\left( {M + 1} \right) \left( {2M - N - 1} \right) N\left( {N + 1} \right) }}{{6\left( {R - 1} \right) \left( {c + N + 1} \right) }} + N + 1 \nonumber \\&- \sqrt{\frac{{M\left( {M + 1} \right) \left( {2M - N - 1} \right) N\left( {N + 1} \right) }}{{6\left( {R - 1} \right) \left( {c + N + 1} \right) }}\left[ {\frac{{M\left( {M + 1} \right) \left( {2M - N - 1} \right) N\left( {N + 1} \right) }}{{6\left( {R - 1} \right) \left( {c + N + 1} \right) }} - \frac{{2\left( {M - N} \right) \left( {c + N + 1} \right) }}{{2M - N - 1}} + 2N + 1} \right] }.\nonumber \\ \end{aligned}$$(42)If \(c \le {{\bar{c}}_2}\), \({\delta _{th}}\) is decreased with \({\bar{T}}\) in the set \(\left\{ {1, \cdots ,{{\bar{T}}^*}} \right\} \) and is increased with \({\bar{T}}\) in the set \(\left\{ {{{\bar{T}}^*}, \cdots ,N} \right\} \) since \({{\partial {\delta _{th}}} /{\partial z}} < 0\) for \(\left[ {1,z'} \right) \) and \({{\partial {\delta _{th}}} /{\partial z}} > 0\) for \(\left( {z',N} \right] \). If \(c > {{\bar{c}}_2}\), \({\delta _{th}}\) is decreased with \({\bar{T}}\) in the set \(\left\{ {1, \cdots ,N} \right\} \) since \({{\partial {\delta _{th}}} /{\partial z}} < 0\) for \(\left[ {1,N} \right] \). Therefore, for any given \({{T_D}}\), one can obtain that the optimal solution \({{T_C}^*}\) leading to the minimum \({\delta _{th}}\) satisfies
- 1.3.1:
-
\({T_C}^* = {\bar{T}}^*\) if \(c \le {{\bar{c}}_2}\);
- 1.3.2:
-
\({T_C}^* = N\) if \(c > {{\bar{c}}_2}\), which proves the third property of Proposition 1. \(\square \)
D Summary of Critical Parameters: Notations and Mathematical Expressions
In “Appendix D”, we describe the critical parameters used in this study for the purpose of clarifying their definitions and mathematical expressions (see Table 3).
Rights and permissions
About this article
Cite this article
Wang, X. Costly Participation and The Evolution of Cooperation in the Repeated Public Goods Game. Dyn Games Appl 11, 161–183 (2021). https://doi.org/10.1007/s13235-020-00352-1
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13235-020-00352-1