Reputation-Based Investment Helps to Optimize Group Behaviors in Spatial Lattice Networks

Hong Ding; Lin Cao; Yizhi Ren; Kim-Kwang Raymond Choo; Benyun Shi

doi:10.1371/journal.pone.0162781

Abstract

Encouraging cooperation among selfish individuals is crucial in many real-world systems, where individuals’ collective behaviors can be analyzed using evolutionary public goods game. Along this line, extensive studies have shown that reputation is an effective mechanism to investigate the evolution of cooperation. In most existing studies, participating individuals in a public goods game are assumed to contribute unconditionally into the public pool, or they can choose partners based on a common reputation standard (e.g., preferences or characters). However, to assign one reputation standard for all individuals is impractical in many real-world deployment. In this paper, we introduce a reputation tolerance mechanism that allows an individual to select its potential partners and decide whether or not to contribute an investment to the public pool based on its tolerance to other individuals’ reputation. Specifically, an individual takes part in a public goods game only if the number of participants with higher reputation exceeds the value of its tolerance. Moreover, in this paper, an individual’s reputation can increase or decrease in a bounded interval based on its historical behaviors. We explore the principle that how the reputation tolerance and conditional investment mechanisms can affect the evolution of cooperation in spatial lattice networks. Our simulation results demonstrate that a larger tolerance value can achieve an environment that promote the cooperation of participants.

Citation: Ding H, Cao L, Ren Y, Choo K-KR, Shi B (2016) Reputation-Based Investment Helps to Optimize Group Behaviors in Spatial Lattice Networks. PLoS ONE 11(9): e0162781. https://doi.org/10.1371/journal.pone.0162781

Editor: Chris T. Bauch, University of Waterloo, CANADA

Received: May 18, 2016; Accepted: August 29, 2016; Published: September 9, 2016

Copyright: © 2016 Ding et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: Data and soucrce code are available at the open online digital repository Figshare https://dx.doi.org/10.6084/m9.figshare.3753390.v1.

Funding: This work was supported by the National Natural Science Foundation of China (Grant Nos. 61100194 and 61272173):YRZ http://www.nsfc.gov.cn/ and National Natural Foundation of China (No.: 81402760):BYS. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

In collaborative and distributed systems, such as Internet of Things (IoT) and Peer to Peer (P2P) networks, autonomous individuals cooperate with each other to accomplish relatively complicated tasks for their reciprocity targets. However, most individuals in such networks are rational, and pursue their selfish interest or gains. In other words, individuals in such systems may only prefer to obtaining services from, rather than providing services to, others [1, 2]. For example, in a P2P network, there exist individuals who are only interested in downloading (shared) resources but are reluctant to contribute any resources in return. Such individuals are known as free-riders, whose behaviors may result in degradation for the system performance, and have negative effect on the quality of experience of other individuals (especially, those individuals who have contributed resources but are unable to reap the associated benefits). If no mitigation mechanisms are taken, more individuals will stop contributing/sharing resources, and the system will suffer from the problem of commons dilemma, also known as the tragedy of the commons. In this case, incentive mechanisms for promoting cooperation among selfish individuals are one useful solution to enhance the smooth operation of such systems.

Evolutionary public goods game (PGG) has been widely used in the literature to model group decision and cooperation dynamics in many fields, such as biology, sociology, physics and computer science [3–5]. A typical N-player PGG involves N individuals, each of which decides whether to cooperate (C) or defect (D) as its strategy. A cooperator will contribute an investment into the public pool which represents for a cost to establish a service, while a defector will not contribute anything. The total investment in the public pool will be multiplied by a multiplication factor r (i.e., 1 < r < N), which represents the eventual benefits upon the completion of the service. Finally, the benefits will be equally distributed among all N players. Defectors in a public goods game play as free riders by sharing the same public goods as cooperators without any contribution. Such behaviors may ultimately lead to the tragedy of the commons.

In order to overcome the commons dilemma, extensive studies has focused on investigating the evolution of cooperation in complex networks, where individuals are forced to participate in all PGGs that centered on itself and its neighbors [3, 6], or optionally decide whether or not to participate a PGG [7], or unidirectionally select their group members [8]. Accordingly, a number of mechanisms have been proposed to promote cooperation behaviors, such as reputation [9], mobility [10, 11], punishment [12], rewards [13] and social diversity [14, 15]. Among them, reputation as a social engine of indirect reciprocity [16, 17], has been widely used to explain the high cooperation levels in human society. Existing research generally assumes that there is a common reputation standard accepted by all individuals, where an individual will be regarded as good (respectively, bad) if its reputation is larger (respectively, smaller) than a standard/threshold value. However, reputation is more or less a subjective value. Different individuals may have different evalation criteria (i.e., tolerance in this paper). Moreover, in a more realistic scenario, the reputation value should be bounded, and can be dynamically updated according to individuals’ historical behaviors in an evolutionary game.

In this paper, we propose a reputation-based investment model that takes into account both bounded reputation and reputation tolerance among individuals during the evolutionary process of a public goods game. By allowing for differences in reputation tolerance, a cooperative individual will contribute investment into the public pool only when the number of individuals with higher reputation is not less its tolerance value. In this case, the situation where all individuals possess the same tolerance value is regarded as a special case in our model, and generally exists in a collaborative environment. We simulate our model in a homogeneous spatial lattice network to demonstrate how the reputation-based model can significantly promote cooperation among selfish individuals. It is interesting to observe that a larger tolerance value can make for a better environment that favors the prevalence of cooperative behaviors. We also explore how the initial conditions (including the reputation distribution and the upper bound of reputation value) can influence the evolutionary rate of cooperation in the spatial public goods game.

The rest of this paper is organized as follows. In the next section, we introduce related studies about the evolution of cooperation. Then, we present our reputation-based model and evaluate the performance in a spatial lattice network setting. Finally, we summarize our findings and discuss potential implications of our proposed model.

Related Literature

In the literature, there are a number of incentive mechanisms designed to promote cooperation behaviors in both collaborative and distributed applications, such as Ad Hoc networks [18], cellular networks [19] and wireless networks [20]. Efficient incentive mechanisms include kin selection [21, 22], direct reciprocity [23], indirect reciprocity [16, 24], group selection [25], punishment [12, 26], and rewards [13, 27]. Reputation has also been identified as one of the typical mechanisms of indirect reciprocity, where the evolution of cooperation relies on individuals’ mutual monitoring and assessments [28]. Specifically, indirect reciprocity describes a kind of phenomenon where two individuals behavior toward each other is based on their past behaviors. Suppose that individual A acted positively to individual B, then individual B would assign a high reputation score for individual A. Meanwhile individual C, who witnessed the positive behavior of individual A towards individual B, would remember individual A’s good performance (i.e., high reputation) and cooperate with individual A with a high possibility. In reality, one’s reputation cannot infinitely increase or decrease due to bounded rationality [29], therefore, the reputation values should be in a reasonable interval (i.e., bounded reputation in this paper).

Reputation-based models have been extensively studied under the condition of public information, where all individuals share the same reputation information. For example, Nowak et al. introduced an image scoring to quantify individuals’ reputation, where an individual’s image scoring increases when the individual donates to others by comparing their image scoring [24]. However, in more general and realistic situations, one should take into consideration private information, where individuals have different perception about other individuals’ reputation. Uchida showed that private information can lead to mismatches between the opinions of individuals even when they share the same moral compass [30]. In another work, Uchida and Sigmund investigated the competition of the assessment rules for indirect reciprocity, where individuals assess each other according to their past cooperative behaviors [31]. Both public and private information determine how individuals dynamically perceive reputation of each other over time, while private information can better simulate the model of reputation in structured populations. Taking into consideration errors of information dissemination, Wang et al. investigated how different information about others individuals’ contributions affects conditional cooperators’ willingness to cooperate in a one-shot linear PGG [32]. In this paper, we introduce a reputation-based model, where individuals contribute investment based only on the private knowledge about the number of cooperators in the game.

There have also been extensive studies on the role of diversity (e.g., social diversity [3, 6], structural diversity [33–35], and behavioral diversity [36, 37]) in the evolution of cooperation. Spatial reciprocity is identified as one of the most effective means to enhance the cooperation levels, where social behaviors of individuals are modeled by the classical and evolutionary game theory in spatially structured populations [38–41]. For example, in a setting where individuals are coupled in a spatial network with interactions restricted to only their neighbors, cooperators can sustain by forming compact clusters in order to resist the exploitation of defectors [42]. Hartig et al. studied PGGs on square lattice and scale-free networks, and determined that not only the diversity of the number/size of PGGs, but also the diversity of individuals’ donation to every group member help to promote cooperation behaviors [43]. Accordingly, in this paper, the performance of a reputation-based investment model will be simulated in spatial lattice networks.

With respect to individuals’ behavioral diversity, investment heterogeneity is also a critical factor in reputation studies [7, 44]. Different levels of reputation heterogeneity may significantly influence the evolution of cooperation. For example, Helbing et al. investigated the impact of reputation-based investment heterogeneity on the evolution of cooperation in collaborative networks [45]. Tian et al. proposed an age-related preferential selection mechanism to study the impact of aging on the evolution of cooperation in the Prison’s dilemma game [35]. Specifically in this paper, an individual will determine whether or not to contribute investment into the public pool based on the individual’s private and heterogeneous tolerances about their neighbors’ reputation in a spatial lattice network.

Reputation-based Investment Model

In this section, we present a reputation-based investment model in a public goods game, and investigate the evolutionary dynamics of cooperation in a spatial lattice network.

Basic model

In a typical N-player PGG, individuals who cooperate are required to contribute some investment (i.e., cost of cooperation) into a public pool. Each individual knows that the total amount of investment in the public pool will be multiplied by a factor r (1 < r < N), and divided equally among them irrespective of their contributions. If all individuals cooperate, each of them will increase its initial capital by (r − 1)c, where c denotes the cost of cooperation. However, individuals are more likely to defect by exploiting other individuals’ investment without any cost. Obviously, such selfish behavior yields a higher payoff, irrespective of other individuals’ actions, because the investment of each cooperator returns only a fraction r/N < 1 of its investment. Therefore, although the group’s total payoff is maximized when all individuals cooperate, the Nash equilibrium in this game is simply zero contributions by all individuals.

In this paper, we focus on studying an evolutionary PGG in a spatially structured population, where individuals interact only with their immediate neighbors (known as Von Neumann neighbors). Specifically, each individual are confined in a square lattice, and restricted to play PGGs with its four neighbors. Formally, we denote G(V, E) as the square lattice with M nodes, where v_i ∈ V denotes the player i and E_ij ∈ E denotes the interaction between player i and j. Each player has exactly four neighbors in G(V, E). During the evolutionary process, each player i can dynamically change its strategy s_i (s_i ∈ {C, D}). Here, C strategy means a player will cooperate with others, while a D strategy means a player will defect.

The proposed reputation-based investment model is built upon the evolutionary PGG in G(V, E), where each round consists of two main phases, namely, an investment phase and a learning phase. In the investment phase, each player interacts with his four nearest neighbors, and totally participates in five PGGs (one is centered at itself and the other four PGGs are centered at its four neighbors). Different from typical PGGs, here a cooperative player will conditionally contribute investment to the public pool based on other players’ reputation. During each round, all players will receive their accumulated payoffs when all PGGs are finished. In the learning phase, each player has chance to change strategy by imitating one of his/her neighbors. More details will be presented in subsequent sections.

Reputation updating rule

A reputation mechanism is introduced to model such situation in this section. We assume that players do not operate in a full anonymity environment. In other words, players may collect information about their potential interaction partners from their neighboring environment. Further, we assume that the reputations of players are publicly known to individuals who play the same PGG, which is reasonable due to the limitation of individual player’s capability.

In order to efficiently record the strategy each individual adopts, we introduce a reputation tolerance mechanism to help players choose potential partners with high reputation. Initially, players in the network randomly adopt a strategy, either cooperate (C) or defect (D). During the interaction phase, players decide whether or not to contribute to the public pool based on their strategies. Cooperative players contribute conditionally to the public pool, while defective players will never contribute. As for a cooperative player i in a PGG, he/she will first count the number of group members whose reputation are equal or greater than his own in the game. If the number exceeds or equals to his/her pre-determined reputation tolerance T_i, the cooperator will contributes one unit to the public pool. We denote s_i as the strategy of player i, where s_i = 0 represents the cooperative strategy and s_i = 1 represents the defective strategy. Then, player i’s action a_i can be denoted as: (1) where n_{r_i} is the number of other players in the game whose reputation is not less than i’s reputation R_i. Here, a_i = 0 denotes player i will contribute, and a_i = 1 denotes a non-contributing player i.

During the evolutionary process, the reputation of cooperative players who contribute investment to the public pool will be increased by one unit, whilst the reputation of players who benefit without contributing will be reduced by one unit. We assume that the players’ reputation is bounded in a region [0, θ]. In other words, if a player’s reputation exceeds θ, his/her reputation will be capped at θ. Similarly, a player’s reputation will never drop below 0. The updating rule for player i’s reputation at evolutionary round t can be described as follows: (2) where R_i(t) represents player i’s reputation at round t.

Investment and payoff

In the spatial lattice network G(V, E), each player will participate in five PGGs. In each PGG, the total amount of investment in the public pool is multiplied by the multiplication factor r, and then be distributed equally to all group members irrespective of their strategies. If no cooperator contributes, then no payoff will be distributed. Therefore, the payoff of a player i in each PGG will be given by: (3) where n_c is the number of cooperators who contributes to the public pool in the PGG.

Learning rule

At each evolutionary round, all players will synchronously update the strategies according to a learning rule. For each player i, he/she randomly selects a player j from his/her four neighbors, and then adopt j’s strategy based on the following Fermi function: (4) where P_i represent player i’s accumulated payoffs after participating in five PGGs in his/her neighboring environment. K indicates the amplitude of noise, which quantifies the force of selection about strategy adoptions [46, 47]. If K → ∞ (weak selection), players are less responsive to payoff differences, and a player with high payoff may adopt the strategy of a less successful one. If K → 0, then players reliably switch to the strategy with the higher payoff even if the difference is very small. For 0 < K < ∞, there exists a relatively certain possibility that strategies with less payoffs will be adopted. In general, egoistic players prefer to adopt the strategy of more successful neighbors [48–50].

Algorithm 1: The Main Algorithm

1 Generate a spacial lattice network G of size M

2 Initialize strategies s_i for each individual i

3 Initialize model parameters T, K, θ, r and μ

4 foreach iteration t do

5 Setting the initial payoff of each node as zero

6 foreach node i ∈ G do

7 Play PGGs with its four neighbors

8 if s_i = 1 then

9 a_i = 1, i.e., i does not contribute to the pool

10 else

11 if n_{r_i} ≥ T then

12 a_i = 0, i.e., i contribute to the pool

13 else

14 a_i = 1, i.e., i does not contribute to the pool

15 Calculate the cumulative payoff for each node in the PGGs

16 foreach node i ∈ G do

17 Update s_i based on Eq 4

18 Generate a random value x ∈ [0, 1]

19 if x ≤ μ then

20

21 Reset all nodes’ payoffs to zero

22 Record the frequency of cooperation for each iteration

In order to better simulate real-world situations, we introduce strategic mutations in our model, where a cooperator is likely to become a defector, and vice versa. For simplicity, mutation only happens at the end of each round with a constant rate μ. That is to say, each player i has chance to change his/her strategy as follows: (5) where denotes an opposite strategy of s_i. Player i will change to another strategy with probability μ, or remains unchanged with probability 1 − μ.

In summary, the main procedure of the reputation-based investment model is shown in Algorithm 1.

Experiments and Results

Simulations were carried out on a square lattice network of size M = 100 × 100 with periodic boundary conditions. Initially, cooperators and defectors were randomly distributed among the network with equal probability 50%). The key quantity to characterize the cooperative behavior is the cooperation frequency, which is defined as the ratio of the number of cooperator M_c to the total number of players M at the steady state (i.e. ). We aim to evaluate the effects of reputation tolerance T, multiplication factor r, selectoin strength K, maximum reputation θ and initial distribution of strategies on the evolution of cooperation in a PGG. All simulations are carried out over 1,000 Monte Carlo time steps, and we have tested that steady state can be reached.

Effects of reputation tolerance and multiplication factor

We first investigated the dependence of multiplication factor r with a stationary T. Fig 1 shows the average proportion of cooperators as a function of r ranging from 1 to 6, for different values of T. It can be observed that as cooperation frequency increases, so does r. Here, T is a set of six integers, namely: 0, 1, 2, 3, 4 and 5. T = 0 indicates that a cooperator will contribute even if all remaining group members’ reputation are less than himself (i.e., the player will contribute unconditionally). T = 5 indicates that a cooperator will not contribute to the public pool under any circumstances because it is impossible to have five players with a reputation higher than the particular cooperator in a five-member group. In the situation T = 0, where every cooperator contributes unconditionally, defectors who obtain more payoff easily occupy the lattice; thus, the final probability of cooperators is almost 0 under a moderate multiplication factor r. In T = 5 since no cooperators contribute, the probability of cooperators is influenced merely by mutation. Thus, the probability of cooperators fluctuating is around 0.5, regardless of the size of r.

Download:

Fig 1. Cooperation frequency ρ_c acts as a function of r for different reputation tolerance T.

Each data point results from the average value of the proportion of cooperators for the last 100 rounds after reaching steady state. The other parameter settings are θ = 20, K = 0.5 and μ = 10⁻⁴.

https://doi.org/10.1371/journal.pone.0162781.g001

With the increase in r, there is a phase transition in terms of cooperation frequency for each tolerance value of T = 0, 1, 2, 3, 4. It can be observed that for T = 4, the phase transition will happen when r is small (i.e., around 1.7). As T decreases, the critical value of r increases accordingly. When T = 0, the probability of cooperators frequency does not exceed 0 until r reaches approximately 3.9 and increases slowly to 1. The most prominent scenario is when T = 4, where a cooperator would not contribute unless his/her reputation is the smallest among all players in the PGG; thus, the probability of cooperators undergoing an abrupt increase when r is between 1.6 and 1.9. It can be observed from Fig 1 that the larger T is, the more promptly it leads to the cooperation level reaching a plateau for large values of r, except for T = 5.

To further investigate how the growth of T will influence the frequency of cooperators, we conducted experiments with different values of r. It can be observed from Fig 2 that when T = 4, the network will reach a high level of cooperation (i.e., almost 1) for different values of r. Only sufficiently large r can promote cooperation frequency when T = 1. When T = 2, 3, there exists a plateau of high level of cooperation. The results indicate that in order to resist the exploitation of defectors, it is more important for cooperators to contribute only when others’ reputations are no less than his own.

Download:

Fig 2. Cooperation frequency ρ_c, as a function of T for different values of r.

Each data point resulted from average value of the frequency of the proportion of cooperators for the last 100 rounds after reaching steady state. The other parameter settings are θ = 20, K = 0.5 and μ = 10⁻⁴.

https://doi.org/10.1371/journal.pone.0162781.g002

To understand the process of how cooperation evolves, Fig 3 shows the change of cooperation frequency with respect to a fixed multiplication factor r = 2.5. Cooperation frequency decreases at the earlier generations, regardless of the value of T (with the exception of T = 5, which is an impossible situation). It is because the cooperators are distributed randomly in the lattice initially, and they could hardly survive due to the exploitation of defectors who can achieve high payoffs and whose behaviors will be imitated by others. However, individuals gradually congregate to compact clusters, among which cooperative clusters will result in a plateau of high cooperation frequency when T = 3 and 4. While defectors could exploit all cooperators when T = 2 and 3. The results demonstrate that low tolerance to others’ reputations can benefit the construction of cooperative environment.

Download:

Fig 3. Cooperation frequency ρ_c with r = 2.5.

Every colored solid line describes the evolution of the frequency of cooperators in 1,000 generations. Of the six reputation tolerance levels, T = 3 and 4 resulted in high cooperation. The other parameter settings are θ = 20, K = 0.5 and μ = 10⁻⁴.

https://doi.org/10.1371/journal.pone.0162781.g003

We now investigated the evolution of individuals’ reputation, payoffs and cooperation frequency during the evolution. As shown in Fig 4, a larger T can significantly promote the evolution of cooperation. Cooperators are unable to escape the exploitation of defectors at the earlier stage of evolution. This resulted in a decrease of cooperation frequency and a moderate growth of the average reputation and payoff (expect for T = 1, where cooperation frequency decreases and the average reputation and payoff remain at a lower level). If the member’s T is large, then cooperators assembled faster than a small T, which manifested as their velocities to achieve stabilization are much swifter. The reason is that a larger T requires strict adherence to the group members that maintains high cooperation level in the entire population. The variation tendency of average reputation and payoff is similar to the frequency of cooperators. The maximum reputation is 20 and the maximum payoff is 2, which can be calculated by . Therefore, the optimal scenario is that all five group members adopt cooperative strategy with a_i = 1 and r = 3 in this experiment. Similar to the results demonstrated in Fig 4, the results presented in Fig 5 are conditioned on r = 2. We observed that the trend of cooperation frequency resembled the trend of average reputation and payoff. The average reputation and payoff have the same tendency as the frequency of cooperators globally. The cooperation frequency achieves a high cooperation level only when T = 4.

Download:

Fig 4. Cooperation frequency ρ_c (the left) corresponding average reputation (the center) and relevant average payoff (the right) of all population when r = 3.

Every colored solid line describes how the parameters evolve for 1,000 generations. The other parameter settings are θ = 20, K = 0.5 and μ = 10⁻⁴.

https://doi.org/10.1371/journal.pone.0162781.g004

Download:

Fig 5. Evolution of cooperation frequency ρ_c (the left), average reputation (the center) and relevant average payoff (the right) of all population when r = 2.

The other parameter settings are θ = 20, K = 0.5 and μ = 10⁻⁴.

https://doi.org/10.1371/journal.pone.0162781.g005

To further illustrate the competition between cooperators and defectors in a spatial lattice, in Fig 6, we evaluated the spatial distribution of defectors and cooperators under different parameter settings. The frequency of cooperators may undergo the exploration of defectors around the 10^th generation. As more cooperators survive from the exploitation, the remaining cooperators gradually formed small clusters. This would allow them to swiftly expand to the entire lattices.

Download:

Fig 6. Snapshots of players’ strategies with parameters r = 2, T = 3 (the first row); r = 3, T = 2 (the second row) and r = 3, T = 3 (the third row).

Blue pixels and red pixels represent the cooperative and defective strategy, respectively. The four columns show the results of the 1^st, 10^th, 30^th and 10000^th generations. The other parameter settings are θ = 20, K = 0.5 and μ = 10⁻⁴.

https://doi.org/10.1371/journal.pone.0162781.g006

Effects of maximum reputation and its initial distribution

In this section, we investigated whether the maximum reputation have any effects on the evolution of cooperation. Recall that in our model, the reputation of a cooperator who contributed to the public pool will be increased by one unit, while the reputation of players who benefit but without contributing will be reduced by one unit. The reputation value ranges from 0 to a maximum value θ. If one player’s reputation exceeds the maximum value, then his reputation will be capped. It can be observed from Figs 7 and 8 that the higher the maximum reputation is, the faster the evolution will converge. However, the values of θ will not affect the final cooperation frequency at the steady state. Different from the results in Fig 7, it can be found from Fig 8 that higher maximum reputation may result in better and more stable cooperation frequency.

Download:

Fig 7. Evolutionary dynamics of cooperation frequency for different values of maximum reputation θ when T = 3 and r = 3.

The other parameter settings are K = 0.5 and μ = 10⁻⁴. As shown in the Figure, the higher the maximum reputation is, the faster the evolution will converge.

https://doi.org/10.1371/journal.pone.0162781.g007

Download:

Fig 8. Evolutionary dynamics of cooperation frequency for different values of maximum reputation θ when T = 3 and r = 2.

The other parameter settings are K = 0.5 and μ = 10⁻⁴.

https://doi.org/10.1371/journal.pone.0162781.g008

In Fig 9, we investigate the impact of different initial settings of reputation values on the evolution of cooperation frequency. Two types of settings are investigated. Firstly, all individuals’ reputation values are set to be zero, and secondly, the reputation values are randomly selected from [0, θ]. It can be observed that for the first setting, it is easier for cooperation frequency to reach a stable state. While for the second setting, it takes slightly more generations to reach a stable state. However, both settings do not affect the final results of cooperation frequency. The reason may be that if the reputation of all individuals is initially zero, then cooperative individuals are more likely to have high reputation. In doing so, it would be much easier to form a cooperative cluster in a spatial lattice network. On the contrary, if reputation values are randomly distributed among individuals, then it is possible for defectors to have a high initial reputation. Thus, it will take several generations to correct the reputation of those undeserving individuals. Therefore, under the second setting, the cooperation frequency is more unstable at the earlier generations.

Download:

Fig 9. Evolutionary dynamics of cooperation frequency under different initial settings.

The two blue lines show the results when r = 3 and T = 3; the two reds lines show the results when r = 3.5 and T = 1; and the two yellow lines shown the results when r = 3 and T = 2. For darker colored lines, all individuals’ initial reputation values are set to be 0, while for lighter colored lines, individuals’ initial reputation values are randomly selected from [0, θ]. The other parameter settings are θ = 20, K = 0.5 and μ = 10⁻⁴.

https://doi.org/10.1371/journal.pone.0162781.g009

Effects of selection force K

We also evaluated the effects of selection force K on the evolution of cooperation. It can be observed from Fig 10 that in the case of K = 0.01 (i.e., strong selection), the state of all defectors can only maintain for a very small value of r. In other words, the phase transition from all defectors to all cooperators will happen when r is small. While for K = 10 (i.e., weak selection), it can be observed that the phase transition of cooperation frequency happens for larger values of r. The reason is that the cooperators are more likely to survive by forming spatial clusters by mutually imitating behaviors of each other under a strong selection. On the other hand, once the cluster of cooperators is formed, the behaviors of cooperators are more easily imitated by defectors with a strong selection.

Download:

Fig 10. Cooperation frequency ρ_c as a function of r for different selection force K.

The other parameter settings are T = 3, θ = 20 and μ = 10⁻⁴.

https://doi.org/10.1371/journal.pone.0162781.g010

Conclusion

In this paper, we proposed a reputation-based investment mechanism to investigate the evolution of cooperation in a spatial public goods game. The proposed mechanism has two properties, namely: each individual’s investment behavior is conditional based on other individuals’ reputation and its own reputation tolerance, and the reputation values are bounded.

We then evaluated the performance of the proposed mechanism in a spatial lattice network. The findings demonstrated that individuals’ reputation-based conditional investment behaviors can promote the evolution of cooperation for higher values of reputation tolerance. Specifically, if there are more individuals with strict reputation tolerance, then it will be easier to reach high the proportion of cooperators in the network (even with a small multiplication factor). This is due primarily to the spontaneous emergence of quarantine of defectors, who are eventually surrounded by cooperators and forced into isolated convex. This phenomenon can be observed in the simulation results on spatially structured populations. In addition, we investigated the effects of reputation tolerance, maximum reputation and selection force on the evolution of cooperation. Cooperation frequency initially decreased, as shown in the snapshots of several simulations for many reputation tolerance. In the long run, however, the evolution of cooperation ameliorated due to the formation of extremely robust clusters of cooperators. Moreover, the initial distribution of reputation appeared to have little influence on the evolution of cooperation, and only influencing the speed of convergence, where random reputation distribution converged faster. Our findings in this paper can help understand the role of reputation-based conditional investment in spatial public goods games.

Author Contributions

Conceptualization: HD YZR LC.
Formal analysis: HD YZR BYS KKRC.
Funding acquisition: YZR BYS.
Methodology: YZR KKRC.
Software: LC YZR.
Supervision: HD.
Validation: HD LC YZR BYS.
Writing – original draft: HD LC YZR.
Writing – review & editing: HD LC YZR BYS KKRC.

References

1. Kamvar SD, Schlosser MT, Garcia-Molina H. The eigentrust algorithm for reputation management in P2P networks. The 12th international conference on World Wide Web. ACM. 2003;640–651.
2. Zhou R, Hwang K. Powertrust: A robust and scalable reputation system for trusted peer-to-peer computing. IEEE Transactions on Parallel and Distributed Systems. 2007;18(4):460–473.
- View Article
- Google Scholar
3. Santos FC, Santos MD, Pacheco JM. Social diversity promotes the emergence of cooperation in public goods games. Nature. 2007;454(7201):213–216.
- View Article
- Google Scholar
4. Hauert C, De Monte S, Hofbauer J. Volunteering as red queen mechanism for cooperation in public goods games. Science. 2002;296(5570):1129–1132. pmid:12004134
- View Article
- PubMed/NCBI
- Google Scholar
5. Semmann D, Krambeck HJ, Milinski M. Volunteering leads to rock-paper-scissors dynamics in a public goods game. Nature. 2003;425(6956):390–393. pmid:14508487
- View Article
- PubMed/NCBI
- Google Scholar
6. Perc M, Szolnoki A. Social diversity and promotion of cooperation in the spatial prisoner’s dilemma game. Physical Review E. 2008;77(1):011904.
- View Article
- Google Scholar
7. Hauert C, De Monte S, Hofbauer J. Replicator dynamics for optional public good games. Journal of Theoretical Biology. 2002;218(2):187–194. pmid:12381291
- View Article
- PubMed/NCBI
- Google Scholar
8. Wu T, Fu F, Wang L. Partner selections in public goods games with constant group size. Physical Review E. 2009;80(2):026121.
- View Article
- Google Scholar
9. Cuesta JA, Gracia-Lázaro C, Ferrer A. Reputation drives cooperative behaviour and network formation in human groups. Scientific Reports. 2015;5.
- View Article
- Google Scholar
10. Helbing D, Yu W. The outbreak of cooperation among success-driven individuals under noisy conditions. PNAS. 2009;106:3680–3685. pmid:19237576
- View Article
- PubMed/NCBI
- Google Scholar
11. Cong R, Wu B, Qiu Y, Wang L. Evolution of coopertion driven by reputation-based migarion. PLoS ONE. 2012;7(5):e35776. pmid:22615739
- View Article
- PubMed/NCBI
- Google Scholar
12. Hauert C. Replicator dynamics of reward & reputation in public goods games. Journal of Theoretical Biology. 2010;267:22–28. pmid:20708018
- View Article
- PubMed/NCBI
- Google Scholar
13. Helbing D, Attila S, Matjaž P, György S. Punish, but not too hard: how costly punishment spreads in the spatial public goods game. New Journal of Physics. 2010;12(8):083005.
- View Article
- Google Scholar
14. Xia CY, Meng XK, Wang Z. Heterogeneous coupling between interdependent lattices promotes the cooperation in the prisoner’s dilemma game. PLoS ONE. 2015;10(6):e0129542. pmid:26102082
- View Article
- PubMed/NCBI
- Google Scholar
15. Xia CY, Meloni S, Perc M. Dynamic instability of cooperation due to diverse activity patterns in evolutionary social dilemmas. EPL. 2015;109(5):58002.
- View Article
- Google Scholar
16. Nowak MA, Sigmund K. Evolution of indirect reciprocity. Nature. 2005;437(7063):1291–1298. pmid:16251955
- View Article
- PubMed/NCBI
- Google Scholar
17. Nowak MA. Five rules for the evolution of cooperation. Science. 2006;314(7063):1560–1563. pmid:17158317
- View Article
- PubMed/NCBI
- Google Scholar
18. Li Z, Shen H. Game-Theoretic analysis of cooperation incentive strategies in mobile ad hoc networks. IEEE Transactions on Mobile Computing. 2012;11(8):1287–1303.
- View Article
- Google Scholar
19. Wu D, Zhou L, Cai Y. Energy-aware dynamic cooperative strategy selection for relay-assisted cellular networks: an evolutionary game approach. IEEE Transactions on Vehicular Technology. 2014;63(9):4659–4669.
- View Article
- Google Scholar
20. Zhang W, Ismail M. Cooperation in wireless communication networks. IEEE Wireless Communications. 2012;19(2):10–20.
- View Article
- Google Scholar
21. Hamilton WD. The genetical evolution of social behaviour. Journal of Theoretical Biology. 1964;17–52. pmid:5875340
- View Article
- PubMed/NCBI
- Google Scholar
22. Som BA,Joel SB, Amy TS. Evolution of Cooperation: Combining kin selection and reciprocal altruism into matrix games with social dilemmas. PLOS ONE. 2013,8(5):e63761.
- View Article
- Google Scholar
23. Trivers RL. The evolution of reciprocal altruism. Quarterly Review of Biology. 1971;35–57.
- View Article
- Google Scholar
24. Nowak MA, Sigmund K. Evolution of indirect reciprocity by image scoring. Nature. 1998;393(6685):573–577. pmid:9634232
- View Article
- PubMed/NCBI
- Google Scholar
25. Jonathan NP, Charles JG. Site-specific group selection drives locally adapted group compositions. Nature. 2014;514:359–362.
- View Article
- Google Scholar
26. Szolnoki A, Szabo G, Perc M. Rhase diagrams for the spatial public goods game with pool punishment. Physical Review E. 2011;83(3):856–875.
- View Article
- Google Scholar
27. Szolnoki A, Perc M. Reward and cooperation in the spatial public goods game. Europhysics Letters. 2010;92(3):2333–2358.
- View Article
- Google Scholar
28. Wang Z, Wang L, Yin ZY, Xia CY. Inferring reputation promotes the evolution of cooperation in spatial social dilemma games. PLoS ONE. 2012;7(7):E.e40218.
- View Article
- Google Scholar
29. Chen X, Schick A, Doebeli M, Blachford A, Wang L. Reputation-based conditional interaction supports cooperation in well-mixed Prisoner’s dilemmas. PLoS ONE. 2012;7(5):e36260. pmid:22615761
- View Article
- PubMed/NCBI
- Google Scholar
30. Uchida S. Effect of private information on indirect reciprocity. Physical Review E. 2010;82(3):036111.
- View Article
- Google Scholar
31. Uchida S, Sigmund K. The competition of assessment rules for indirect reciprocity. Journal of Theoretical Biology. 2010;263(1):13–19. pmid:19962390
- View Article
- PubMed/NCBI
- Google Scholar
32. Wang X, Chen X, Jia G. Reputation-based mutual selection rule promotes cooperation in spatial threshold public goods games. Chaos Solitons & Fractals. 2013,56(4):181–187.
- View Article
- Google Scholar
33. Santos FC, Pinheiro FL, Lenaerts T, Pacheco JM. The role of diversity in the evolutio of cooperation. Journal of Theoretical Biology. 2012;299:88–96. pmid:21930134
- View Article
- PubMed/NCBI
- Google Scholar
34. Jin Q, Wang L, Xia CY, Wang Z. Spontaneous symmetry breaking in interdependent networked game. Scientific Reports. 2014;4:4095. pmid:24526076
- View Article
- PubMed/NCBI
- Google Scholar
35. Wang Z, Wang L, Szolnoki A, Perc M. Evolutionary games on multilayer networks: A colloquium. The European Physical Journal B. 2015;88:124.
- View Article
- Google Scholar
36. Tian LL, Li MC, Lu K. The influence of age-driven investment on cooperation in spatial public goods games. Chaos, Solitons & Fractals. 2013;54(9):65–70.
- View Article
- Google Scholar
37. Morsky B, Bauch CT. Truncation selection and payoff distributions applied to the replicator equation. Journal of Theoretical Biology. 2016;404:383–390. pmid:27343031
- View Article
- PubMed/NCBI
- Google Scholar
38. Wang Y, Vasilakos AV, Ma J. VPEF: A simple and effective incentive mechanism in community-based autonomous networks. IEEE Transactions on Network and Service Management. 2015;106(1):75–86.
- View Article
- Google Scholar
39. Shutters ST, Hales D. Altruism displays a harmonic signature in structured societies. Journal of Artificial Societies & Social Simulation. 2015;18(3):1–14.
- View Article
- Google Scholar
40. Szabo G, Hauert C. Phase transitions and volunteering in spatial public goods games. Physical Review Letters. 2002;89(11):1401–1412.
- View Article
- Google Scholar
41. Nowak MA, May RM. Evolutionary games and spatial chaos. Nature. 1992;359(6398):826–829.
- View Article
- Google Scholar
42. Sicardi EA, Fort H, Vainstein MH. Random mobility and spatial structure often enhance cooperation. Journal of Theoretical Biology. 2009;256(2):240–246. pmid:18952109
- View Article
- PubMed/NCBI
- Google Scholar
43. Hartig B, Irlenbusch B, Kölle F. Conditioning on what? Heterogeneous contributions and conditional cooperation. Journal of Behavioral and Experimental Economics. 2015,55:48–64.
- View Article
- Google Scholar
44. Fu F, Hauert C, Nowak MA. Reputation-based partner choice promotes cooperation in social networks. Physical Review E. 2008,78(2):026117.
- View Article
- Google Scholar
45. Helbing D, Szolnoki A, Perc M. Punish, but not too hard: How costly punishment spreads in the spatial public goods game. New Journal of Physics. 2010,12(29):3208–3214.
- View Article
- Google Scholar
46. Wu B, García J, Hauert C, Traulsen A. Extrapolating weak selection in evolutionary games. PLoS Computational Biology. 2013,9:e1003381. pmid:24339769
- View Article
- PubMed/NCBI
- Google Scholar
47. Fu F, Rosenbloom DI, Wang L, Nowak MA. Imitation dynamics of vaccination behaviour on social networks. Proceedings of the Royal Society B. 2011,278:42–49. pmid:20667876
- View Article
- PubMed/NCBI
- Google Scholar
48. Blume LE. The statistical mechanics of strategic interaction. Games and Economic Behavior. 1993, 5:387–424.
- View Article
- Google Scholar
49. Szabó G, Hauert C. Phase transition and volunteering in spatial public goods games. Physical Review Letters. 2002, 89:118101. pmid:12225171
- View Article
- PubMed/NCBI
- Google Scholar
50. Wu B, Altrock PM, Wang L, Traulsen A. Universality of weak selection. Physical Review E. 2010, 82:046106.
- View Article
- Google Scholar

[ref1] 1. Kamvar SD, Schlosser MT, Garcia-Molina H. The eigentrust algorithm for reputation management in P2P networks. The 12th international conference on World Wide Web. ACM. 2003;640–651.

[ref2] 2. Zhou R, Hwang K. Powertrust: A robust and scalable reputation system for trusted peer-to-peer computing. IEEE Transactions on Parallel and Distributed Systems. 2007;18(4):460–473.
View Article
Google Scholar

[3] View Article

[4] Google Scholar

[ref3] 3. Santos FC, Santos MD, Pacheco JM. Social diversity promotes the emergence of cooperation in public goods games. Nature. 2007;454(7201):213–216.
View Article
Google Scholar

[6] View Article

[7] Google Scholar

[ref4] 4. Hauert C, De Monte S, Hofbauer J. Volunteering as red queen mechanism for cooperation in public goods games. Science. 2002;296(5570):1129–1132. pmid:12004134
View Article
PubMed/NCBI
Google Scholar

[9] View Article

[10] PubMed/NCBI

[11] Google Scholar

[ref5] 5. Semmann D, Krambeck HJ, Milinski M. Volunteering leads to rock-paper-scissors dynamics in a public goods game. Nature. 2003;425(6956):390–393. pmid:14508487
View Article
PubMed/NCBI
Google Scholar

[13] View Article

[14] PubMed/NCBI

[15] Google Scholar

[ref6] 6. Perc M, Szolnoki A. Social diversity and promotion of cooperation in the spatial prisoner’s dilemma game. Physical Review E. 2008;77(1):011904.
View Article
Google Scholar

[17] View Article

[18] Google Scholar

[ref7] 7. Hauert C, De Monte S, Hofbauer J. Replicator dynamics for optional public good games. Journal of Theoretical Biology. 2002;218(2):187–194. pmid:12381291
View Article
PubMed/NCBI
Google Scholar

[20] View Article

[21] PubMed/NCBI

[22] Google Scholar

[ref8] 8. Wu T, Fu F, Wang L. Partner selections in public goods games with constant group size. Physical Review E. 2009;80(2):026121.
View Article
Google Scholar

[24] View Article

[25] Google Scholar

[ref9] 9. Cuesta JA, Gracia-Lázaro C, Ferrer A. Reputation drives cooperative behaviour and network formation in human groups. Scientific Reports. 2015;5.
View Article
Google Scholar

[27] View Article

[28] Google Scholar

[ref10] 10. Helbing D, Yu W. The outbreak of cooperation among success-driven individuals under noisy conditions. PNAS. 2009;106:3680–3685. pmid:19237576
View Article
PubMed/NCBI
Google Scholar

[30] View Article

[31] PubMed/NCBI

[32] Google Scholar

[ref11] 11. Cong R, Wu B, Qiu Y, Wang L. Evolution of coopertion driven by reputation-based migarion. PLoS ONE. 2012;7(5):e35776. pmid:22615739
View Article
PubMed/NCBI
Google Scholar

[34] View Article

[35] PubMed/NCBI

[36] Google Scholar

[ref12] 12. Hauert C. Replicator dynamics of reward & reputation in public goods games. Journal of Theoretical Biology. 2010;267:22–28. pmid:20708018
View Article
PubMed/NCBI
Google Scholar

[38] View Article

[39] PubMed/NCBI

[40] Google Scholar

[ref13] 13. Helbing D, Attila S, Matjaž P, György S. Punish, but not too hard: how costly punishment spreads in the spatial public goods game. New Journal of Physics. 2010;12(8):083005.
View Article
Google Scholar

[42] View Article

[43] Google Scholar

[ref14] 14. Xia CY, Meng XK, Wang Z. Heterogeneous coupling between interdependent lattices promotes the cooperation in the prisoner’s dilemma game. PLoS ONE. 2015;10(6):e0129542. pmid:26102082
View Article
PubMed/NCBI
Google Scholar

[45] View Article

[46] PubMed/NCBI

[47] Google Scholar

[ref15] 15. Xia CY, Meloni S, Perc M. Dynamic instability of cooperation due to diverse activity patterns in evolutionary social dilemmas. EPL. 2015;109(5):58002.
View Article
Google Scholar

[49] View Article

[50] Google Scholar

[ref16] 16. Nowak MA, Sigmund K. Evolution of indirect reciprocity. Nature. 2005;437(7063):1291–1298. pmid:16251955
View Article
PubMed/NCBI
Google Scholar

[52] View Article

[53] PubMed/NCBI

[54] Google Scholar

[ref17] 17. Nowak MA. Five rules for the evolution of cooperation. Science. 2006;314(7063):1560–1563. pmid:17158317
View Article
PubMed/NCBI
Google Scholar

[56] View Article

[57] PubMed/NCBI

[58] Google Scholar

[ref18] 18. Li Z, Shen H. Game-Theoretic analysis of cooperation incentive strategies in mobile ad hoc networks. IEEE Transactions on Mobile Computing. 2012;11(8):1287–1303.
View Article
Google Scholar

[60] View Article

[61] Google Scholar

[ref19] 19. Wu D, Zhou L, Cai Y. Energy-aware dynamic cooperative strategy selection for relay-assisted cellular networks: an evolutionary game approach. IEEE Transactions on Vehicular Technology. 2014;63(9):4659–4669.
View Article
Google Scholar

[63] View Article

[64] Google Scholar

[ref20] 20. Zhang W, Ismail M. Cooperation in wireless communication networks. IEEE Wireless Communications. 2012;19(2):10–20.
View Article
Google Scholar

[66] View Article

[67] Google Scholar

[ref21] 21. Hamilton WD. The genetical evolution of social behaviour. Journal of Theoretical Biology. 1964;17–52. pmid:5875340
View Article
PubMed/NCBI
Google Scholar

[69] View Article

[70] PubMed/NCBI

[71] Google Scholar

[ref22] 22. Som BA,Joel SB, Amy TS. Evolution of Cooperation: Combining kin selection and reciprocal altruism into matrix games with social dilemmas. PLOS ONE. 2013,8(5):e63761.
View Article
Google Scholar

[73] View Article

[74] Google Scholar

[ref23] 23. Trivers RL. The evolution of reciprocal altruism. Quarterly Review of Biology. 1971;35–57.
View Article
Google Scholar

[76] View Article

[77] Google Scholar

[ref24] 24. Nowak MA, Sigmund K. Evolution of indirect reciprocity by image scoring. Nature. 1998;393(6685):573–577. pmid:9634232
View Article
PubMed/NCBI
Google Scholar

[79] View Article

[80] PubMed/NCBI

[81] Google Scholar

[ref25] 25. Jonathan NP, Charles JG. Site-specific group selection drives locally adapted group compositions. Nature. 2014;514:359–362.
View Article
Google Scholar

[83] View Article

[84] Google Scholar

[ref26] 26. Szolnoki A, Szabo G, Perc M. Rhase diagrams for the spatial public goods game with pool punishment. Physical Review E. 2011;83(3):856–875.
View Article
Google Scholar

[86] View Article

[87] Google Scholar

[ref27] 27. Szolnoki A, Perc M. Reward and cooperation in the spatial public goods game. Europhysics Letters. 2010;92(3):2333–2358.
View Article
Google Scholar

[89] View Article

[90] Google Scholar

[ref28] 28. Wang Z, Wang L, Yin ZY, Xia CY. Inferring reputation promotes the evolution of cooperation in spatial social dilemma games. PLoS ONE. 2012;7(7):E.e40218.
View Article
Google Scholar

[92] View Article

[93] Google Scholar

[ref29] 29. Chen X, Schick A, Doebeli M, Blachford A, Wang L. Reputation-based conditional interaction supports cooperation in well-mixed Prisoner’s dilemmas. PLoS ONE. 2012;7(5):e36260. pmid:22615761
View Article
PubMed/NCBI
Google Scholar

[95] View Article

[96] PubMed/NCBI

[97] Google Scholar

[ref30] 30. Uchida S. Effect of private information on indirect reciprocity. Physical Review E. 2010;82(3):036111.
View Article
Google Scholar

[99] View Article

[100] Google Scholar

[ref31] 31. Uchida S, Sigmund K. The competition of assessment rules for indirect reciprocity. Journal of Theoretical Biology. 2010;263(1):13–19. pmid:19962390
View Article
PubMed/NCBI
Google Scholar

[102] View Article

[103] PubMed/NCBI

[104] Google Scholar

[ref32] 32. Wang X, Chen X, Jia G. Reputation-based mutual selection rule promotes cooperation in spatial threshold public goods games. Chaos Solitons & Fractals. 2013,56(4):181–187.
View Article
Google Scholar

[106] View Article

[107] Google Scholar

[ref33] 33. Santos FC, Pinheiro FL, Lenaerts T, Pacheco JM. The role of diversity in the evolutio of cooperation. Journal of Theoretical Biology. 2012;299:88–96. pmid:21930134
View Article
PubMed/NCBI
Google Scholar

[109] View Article

[110] PubMed/NCBI

[111] Google Scholar

[ref34] 34. Jin Q, Wang L, Xia CY, Wang Z. Spontaneous symmetry breaking in interdependent networked game. Scientific Reports. 2014;4:4095. pmid:24526076
View Article
PubMed/NCBI
Google Scholar

[113] View Article

[114] PubMed/NCBI

[115] Google Scholar

[ref35] 35. Wang Z, Wang L, Szolnoki A, Perc M. Evolutionary games on multilayer networks: A colloquium. The European Physical Journal B. 2015;88:124.
View Article
Google Scholar

[117] View Article

[118] Google Scholar

[ref36] 36. Tian LL, Li MC, Lu K. The influence of age-driven investment on cooperation in spatial public goods games. Chaos, Solitons & Fractals. 2013;54(9):65–70.
View Article
Google Scholar

[120] View Article

[121] Google Scholar

[ref37] 37. Morsky B, Bauch CT. Truncation selection and payoff distributions applied to the replicator equation. Journal of Theoretical Biology. 2016;404:383–390. pmid:27343031
View Article
PubMed/NCBI
Google Scholar

[123] View Article

[124] PubMed/NCBI

[125] Google Scholar

[ref38] 38. Wang Y, Vasilakos AV, Ma J. VPEF: A simple and effective incentive mechanism in community-based autonomous networks. IEEE Transactions on Network and Service Management. 2015;106(1):75–86.
View Article
Google Scholar

[127] View Article

[128] Google Scholar

[ref39] 39. Shutters ST, Hales D. Altruism displays a harmonic signature in structured societies. Journal of Artificial Societies & Social Simulation. 2015;18(3):1–14.
View Article
Google Scholar

[130] View Article

[131] Google Scholar

[ref40] 40. Szabo G, Hauert C. Phase transitions and volunteering in spatial public goods games. Physical Review Letters. 2002;89(11):1401–1412.
View Article
Google Scholar

[133] View Article

[134] Google Scholar

[ref41] 41. Nowak MA, May RM. Evolutionary games and spatial chaos. Nature. 1992;359(6398):826–829.
View Article
Google Scholar

[136] View Article

[137] Google Scholar

[ref42] 42. Sicardi EA, Fort H, Vainstein MH. Random mobility and spatial structure often enhance cooperation. Journal of Theoretical Biology. 2009;256(2):240–246. pmid:18952109
View Article
PubMed/NCBI
Google Scholar

[139] View Article

[140] PubMed/NCBI

[141] Google Scholar

[ref43] 43. Hartig B, Irlenbusch B, Kölle F. Conditioning on what? Heterogeneous contributions and conditional cooperation. Journal of Behavioral and Experimental Economics. 2015,55:48–64.
View Article
Google Scholar

[143] View Article

[144] Google Scholar

[ref44] 44. Fu F, Hauert C, Nowak MA. Reputation-based partner choice promotes cooperation in social networks. Physical Review E. 2008,78(2):026117.
View Article
Google Scholar

[146] View Article

[147] Google Scholar

[ref45] 45. Helbing D, Szolnoki A, Perc M. Punish, but not too hard: How costly punishment spreads in the spatial public goods game. New Journal of Physics. 2010,12(29):3208–3214.
View Article
Google Scholar

[149] View Article

[150] Google Scholar

[ref46] 46. Wu B, García J, Hauert C, Traulsen A. Extrapolating weak selection in evolutionary games. PLoS Computational Biology. 2013,9:e1003381. pmid:24339769
View Article
PubMed/NCBI
Google Scholar

[152] View Article

[153] PubMed/NCBI

[154] Google Scholar

[ref47] 47. Fu F, Rosenbloom DI, Wang L, Nowak MA. Imitation dynamics of vaccination behaviour on social networks. Proceedings of the Royal Society B. 2011,278:42–49. pmid:20667876
View Article
PubMed/NCBI
Google Scholar

[156] View Article

[157] PubMed/NCBI

[158] Google Scholar

[ref48] 48. Blume LE. The statistical mechanics of strategic interaction. Games and Economic Behavior. 1993, 5:387–424.
View Article
Google Scholar

[160] View Article

[161] Google Scholar

[ref49] 49. Szabó G, Hauert C. Phase transition and volunteering in spatial public goods games. Physical Review Letters. 2002, 89:118101. pmid:12225171
View Article
PubMed/NCBI
Google Scholar

[163] View Article

[164] PubMed/NCBI

[165] Google Scholar

[ref50] 50. Wu B, Altrock PM, Wang L, Traulsen A. Universality of weak selection. Physical Review E. 2010, 82:046106.
View Article
Google Scholar

[167] View Article

[168] Google Scholar

Figures

Abstract

Introduction

Related Literature

Reputation-based Investment Model

Basic model

Reputation updating rule

Investment and payoff

Learning rule

Experiments and Results

Effects of reputation tolerance and multiplication factor

Effects of maximum reputation and its initial distribution

Effects of selection force K

Conclusion

Author Contributions

References