Chaos, Solitons and Fractals 91 (2016) 649–655

Contents lists available at ScienceDirect

Chaos, Solitons and Fractals

Nonlinear Science, and Nonequilibrium and Complex Phenomena
journal homepage: www.elsevier.com/locate/chaos

How the expanded crowd-funding mechanism of some southern rural

areas in China affects cooperative behaviors in threshold public goods
game
Qiao Chen, Tong Chen, Yongjie Wang∗
College of Management and Economics, Tianjin University, Tianjin 300072, PR China

a r t i c l e i n f o a b s t r a c t

Article history: The pursuit of high cooperation rates in public goods games has attracted many researchers. However,
Received 16 July 2016 few researchers attach much weight to the influence of emotions on decision-making, especially on public
Revised 30 August 2016
goods games. From ancient to modern times, publishing the list of cooperators to stimulate cooperation
Accepted 31 August 2016
has been a common phenomenon in some southern rural areas in China. Actually, the published list can
influence individuals’ behaviors by affecting their emotions. Here we extend the method of publishing the
Keywords: list and optimize it by adding a lobbyist mechanism. Through numerical simulations, we find that the role
Publish the list of lobbyists can not be ignored unless the synergy factor is larger than a certain value. Additionally, we
Lobbyist find that publishing the list certainly has a great effect on individual’s cooperative behavior. But whether
Emotion
to publish the list or not and how to publish the list depend on the situation.
Threshold public goods game
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction
At present, the problems we have encountered can be summed
up for the contradiction between the collective interests and indi-
vidual interests. Cooperation means that people have to contribute
personal interests to others, even may help the potential competi-
tors. However, there is no doubt that cooperative behaviors are
universal in reality. Therefore, how to promote cooperation in real
word become an important and challenging issue [1–3]. In recent
years, great progress has been made in the analysis of coopera-
tion from the perspective of evolutionary game theory. The pris-
oner’s dilemma (PDG) is a representative example of non zero-sum
game which describes how two rational individuals choose their
own best strategies, but not the best choices for groups [4]. In real
life, however, most conflicts often occur in muti-agent groups. The
public goods game (PGG) is a standard of experimental economics
which is mainly used for analyzing the multi-person group. In this
model, each individual can secretly make decision that whether
to contribute their private property or not to the public pool.
Then, the payoff of public goods will be divided among players
evenly. Defectors would benefit from this rule. Thus, it will cause
some popular problems, like the tragedy of the commons and the
∗
Corresponding author.
E-mail addresses: 2015209055@tju.edu.cn (Q. Chen), wang_yj@tju.edu.cn (Y.
Wang).
free rider [5–8]. Due to these issues, many researchers are mak-
ing great efforts to explore mechanisms to solve these dilemmas
[9]. Punishment mechanism is proposed by many researchers [10–
18]. Through this method, defectors will be added cost by penalty.
Considering their own profit, they would choose cooperative strat-
egy. The reward mechanism, which is contrary to the punishment
mechanism, is also received widespread attention from researchers
[19–22]. Due to the existence of indirect reciprocity, a lot of peo-
ple choose to cooperate in order to get a better reputation. There-
fore, reputation mechanism is introduced into the PGG to promote
cooperation [15,23–25]. However, the punishment mechanism and
reward mechanism are not available when it’s failure to identify
defectors [26–30]. Besides, many researchers also have studied the
influence of the group’s own attributes (like group diversity, group
size, group heterogeneity, agents’ ability of thought independent,
etc) on the game of public goods [31–38]. What’s more, many
other mechanisms have been explored, e.g., insurance, the religious
belief, the establishment of institutions, the tolerance toward de-
fectors, etc. [1,39–41]. Mechanisms above mostly have tended to
emphasize the operation of analytic processes in guiding cooper-
ative behavior. However, bounded rationality, somatic marker hy-
pothesis (SMH) and many other economic models emphasize that
we can not ignore the role of emotions in decision-making [42,43].
Nowak also has pointed out that each person’s emotion would be
affected by the network of friends and loved ones, etc [44]. Ac-
cording above, it is worth digging much deeper in the cooperative
field of the public goods game.

http://dx.doi.org/10.1016/j.chaos.2016.08.015
0960-0779/© 2016 Elsevier Ltd. All rights reserved.
650 Q. Chen et al. / Chaos, Solitons and Fractals 91 (2016) 649–655
Culture is playing a more and more important role in people’s
life. However, the increasing demand of spirit and culture can not
timely and efficiently to be met in many areas, especially in ru-
ral region. If we can motivate social forces to improve the effective
supply of public cultural products, it will not only light the burden
of government, but also provide lots of products with high quality.
Therefore, how to encourage people to contribute their own inter-
ests to provide basic, even higher quality public cultural products,
is a significant topic.
Here, we propose a model to attempt to solve those problems.
Actually, providing public culture goods through crowd-funding
has been around for a long time in Chinese history. The most typ-
ical example is the providing of village opera in most southern
China. The annual village opera fees come from village commit-
tee and resident’s donation. The donor’s name will be published.
Publishing the list has effects on cooperation mainly because of its
effects on individuals’ emotions. Thus people’s decisions will be in-
fluenced by the list. In fact, this is a traditional way of providing
public cultural goods in china. Since it is still in use, it indicates
the antiquity of the tradition. Of course, it also has its drawbacks.
Therefore, we want to expand and optimize it in a scientific way
to solve modern social dilemma. Although PGG is considered to be
one of the most common games in the study of social dilemmas,
it can not be suited to every social dilemma [45]. Therefore in this
paper, we use the so-called threshold public goods game (TPGG),
which means public goods will be provided only when the contri-
butions are equal or exceed the required threshold [39,46,47]. But
if the group contribution can not reach the certain threshold, the
contributions will not be returned to the cooperators.
We divide people into three types, namely, the unconditional
cooperator (UC), the unconditional defector (UD) and the slick free
rider (SR). Cooperation is one of the most basic human instincts
[44]. Therefore, we have reason to believe that all people have the
impulse to donate money to build public goods. If the donation
exceeds the threshold, the extra contributions will be a significant
role for much more effective public culture goods [1]. In a sense,
it also may be beneficial to provoking humans’ good nature. And
then we apply the way of publishing the list to modern public
goods games. In our view, the influence of emotions on cooper-
ation behavior can not be neglected. However, modern economic
theory always ignores the influence of emotions [42,43]. Our aim
is to increase cooperation level through the full use of people’s
emotions. Therefore, besides making full use the list, we also set
up a type of person called lobbyist. They will influence individual’s
emotions through words. Therefore in this paper, we focus on ex-
amining the importance of the role of lobbyist, and observing how
to publish the list is the best.
2. The TPGG model with three types of agents
In our TPGG model, we consider the evolutionary public game
on a L × L square lattice with periodic boundary conditions and
von Neumann neighborhood. There are three types of players (UC,
UD, SR) who are all randomly distributed on the square. In order
to express convenience, we use a1 , a2 , a3 represent the propor-
tion of UC, SR, UD, respectively. Whereas every UC contributes an
amount (c = 1) to the public pool, UDs and SRs contribute noth-
ing but exploit cooperators’ efforts. That is, everyone in the group
can enjoy the public goods fairly. Here, we assume that a1 > 0,
because it can conform better with the reality. What’s more, it has
been confirmed that a minimal number of cooperators is required
to elicit the full advantage of group [48]. Please note that lobby-
ists derive from the UCs. That means lobbyists are all UCs, but UCs
are not all lobbyists. Some UCs have the lobbying power because
they have accumulated reputation in the previous cooperation or
have special status. Nevertheless, some of those UCs are willing
to spend their time and energy to act as lobbyists but some are
not. Therefore, the proportion of lobbyists is really small. In ad-
dition, every people in the group all has emotion and rationality.
Different people, however, are different in their ability to control
their emotions. Unlike SRs, UCs and UDs are strong-minded. UCs
are sure that they should cooperate and UDs are totally do not
want to cooperate any more at the first time. Therefore, the pub-
lished list and lobbyist’s emotion words seem have little effect on
them. But it is a fact that individuals, whether humans or animals,
will have different behaviors in different situations [49]. Whether
or not to adopt a different strategy depends both on the personal
success of each individual and on the strategies of neighbors [50].
Therefore, SRs and UDs also may change their strategies. SRs could
also be called conditional cooperators. That is, although SRs have
the impulse to donate money, they are always hesitant for various
reasons. But SR’s emotion will be influenced by the people around
(like a neighbor’s decision, lobbyist’s emotional words), then their
desire may be aroused. This may lead them to make relatively ir-
rational decisions. Because they think the payoff will not be very
low if doing this. This is also very similar to the conformity [51].
In addition, SRs and UDs also may change their decisions based on
the payoff of their neighbors. In other words, SRs are the combi-
nation of conformity-driven players and payoff-driven players. And
UDs are all payoff-driven players. They are good at making deci-
sions according to situations. In addition to exploring the role of
lobbyists, we also need to study whether and how to publish the
list is better. Therefore, we here propose two different models.
Model 1: The whole evolution process is a dynamic circulatory
process. Each stage has two parts: SRs’ emotions are passively in-
fluenced by others and both SRs and UDs try to maximize their
payoff through actively imitating others’ strategies. At the first part
of each stage, UCs’ and UDs’ names will be published. Meanwhile,
a Lobbyist also try their best to persuade a SR to cooperate. Thus,
the total influence to a SR’s emotion from outside can be summed
over as follow:
N (t + 1 ) = (1 + nA − mB + xC − yD + zE )N (t )
Where N denotes the agent’s emotion. A stands for the positive
effect from the cooperated neighbor. B means the negative effect
from defector. C denotes the influence from ex-SRs. D is the neg-
ative effect from SRs. E signifies the positive effect from lobbyists.
We call A, C, E emotion promotion coefficients, call B and D emo-
tion inhibition coefficients at the same time. Different people will
have different affection degree to the SR: the UC’s and UD’s deci-
sions have less influence power than the decision made by SR. This
is because people always like to analyze those who have the same
features as they have, and are used to make some comparisons
with them as well. Actually, this kind of phenomenon is very com-
mon, projection effect and availability bias are more or less related
to this phenomenon. That means everyone can project their own
characteristics to others, more attention will be provided to those
people who have the same characteristics with themselves. There-
fore, the ex-SR who used to be a SR now has become a cooperator
has more influence power relatively (C > A, D > B). Compared to
the published list, lobbyists can pertinently persuade the SRs by
persuasive language. Furthermore, it has been confirmed that peo-
ple always become more generous and benevolent when they feel
that they are under the gaze of others[44]. Therefore, the lobbyists
not only can influence others through words, but also can act as
such an observer. Thus, the lobbying effect is always the best.(E >
A, E > C). Because cooperation is one of the most basic human in-
stincts and people’s nature at birth is good, we here assume that
the emotion promotion coefficients are always larger than emotion
inhibition coefficients.
The lowercase letters (n, m, x, y, z) represent the number of
people in each of the categories in the neighborhood. And every-
 Q. Chen et al. / Chaos, Solitons and Fractals 91 (2016) 649–655
one has an initial rational judgment of whether to cooperate or
not, denote by T. When N > T, SR will make a decision that is rel-
ative irrational to himself. Otherwise, they will keep the original
strategy unchanged. If they don not change strategies, they may
maximize their payoff through imitating. (For instance, many peo-
ple may buy something that they don not really need. That maybe
because the influence of advertisements, or maybe because all his
friends buy them, or maybe people around him persuade him that
it is worth buying. Even though they know they don not have to
buy these things any more).
Still, there are plenty of SRs will adhere their first decisions be-
cause the influence from outside is not enough. However, SRs along
with UDs will also try their best to maximize their utility because
of their selfish nature. Thus, they are likely to change their deci-
sions when they see someone else’s payoff is greater than them.
Therefore, every non-cooperator participates in G = 5 groups cen-
tered on his nearest neighbors and himself. The payoff obtained by
each agent can be calculated as follow:
 
  r ∗ ngc
px = pgx = − Sx
G rationality, so the SRs are in the majority. In Fig. 1(a), it can be
g∈x g∈x
Where r(1 < r < N) denotes an enhancement factor, reflecting the
synergetic cooperation. x stands for the set of TPGG groups in
g
which player x participates. g is one element in the set. nc the total
number of cooperators in TPGG group. G is the total number of the
TPGG groups. Sx stands for the contribution donated by agent x.
After that, player x will imitate a person who has a higher payoff
with the following probability:
1 we change the structure of the population, the result is very inter-
W ( sx ← sy ) =
1 + exp[φ ( px − py )]
Where φ characterizes the amplitude of environment noise. Px − Py
means the difference of payoff between agent x and agent y. There-
fore, based on the above two formulas, every non-cooperator has
the chance to imitate other’s strategies and then his payoff may be
increased.
When the above-mentioned two parts of the stage is com-
pleted, a full stage is finished. At this time, all the individuals never
change the strategies again, so the new list will be published. Then
the whole evolution moves on to the next stage. The next stages
are the same: some SRs are also affected by the neighbours’ deci-
sions and lobbyists’ words, some SRs and UDs decide their own
strategies based on the payoff of their neighbors. And the new
list also would not be published until all the individuals’ decisions
never change. This entire process is always repeated like what ex-
pressed above.
Model 2: Compared model 2 to model 1, the assumptions and
the whole frame are the same. The only difference between them
is the lobbying time. In every stage in model 2, every lobbyist will
try their best to persuade more non-cooperators to cooperate. Then
the new list will be published and then enter the next evolution
stage. That is, compared to model 1, model 2 has more lobbying
time in each step.
3. Numerical simulation results and analysis
In order to explore agents’ behaviors, a great deal of simulations
are performed on L × L square lattices. Main simulation parame-
ter is set to be L = 199. Robustness analysis show that the synergy
factor (r) and the amplitude of environment noise (φ ) have great
impacts to the evolution results. Based on the size of these two
coefficients, this paper will discuss the evolution in two parts.
Part one: This part we firstly discuss the special case. Here,
we set the synergy factor r = 2. φ is set to be 10. The emotion
influence coefficients are set to be A = 0.3, B = 0.1, C = 0.35, D =
651
0.15, E = 0.4 at the first, respectively. Those parameters will be dis-
cussed during the following analysis. It is necessary for us to ex-
plain the reasons for the coefficient setting. Although neighbors
will affect the individual’s emotion, this effect is not very large.
Nicholas A. Christakis and James H. Fowler have found that when
a person is happy, the odd of his neighbor’s happiness will increase
by 34% [52]. Moods will be so, the affective state of decision-
makers will certainly have a similar phenomenon. But the factors
related to decision making are much more complex than mood, so
the influence to decision makers may not be stronger than that to
moods. Although people are bounded rational, but the majority of
the initiative is still in their own hands.
First of all, we need to verify that if using lobbyists’ words and
publishing the list really can influence cooperation. Three control
groups are introduced: method 3 (has lobbyists and only publish
the original list once, evolves as model 1), method 4 (has lobbyists
and not publish the list, evolves as model 1) and method 5 (no
lobbyists and publish the list as method 1, evolves as model 1).
Fig. 1 has showed several different evolution results. What need
to emphasize is that in real life people are selfish and bounded
easily found that, the light blue curve and orange curve are in the
highest position. And the other three curves are all in lower po-
sitions. That means under the influence of lobbyists and the pub-
lished list, the number of cooperators is significantly increased. Be-
sides, compared to the second method, the first method is much
earlier to reach to the evolution equilibrium. The gap between dark
blue curve and other curves also shows that lobbyists play a very
important role in improving the cooperation rate. However, when
esting. We find that when the proportion of UD is relatively larger
than UC, publishing the list seems to be less reasonable. When the
number of UC is a constant (a1 = 0.1), the evolution of the results
are not what we originally thought when we increase the propor-
tion of UD. As shown in Fig. 1(b), with the proportion of UD is rose
to 40%, the tendency of the yellow curve is almost the same like
the light blue curve and the orange curve. When the proportion
of UD is increased to 50% (see Fig. 1(c)), not to publish the list be-
come the best strategy in the long term. Therefore, here, we can be
sure that changes in the population structure will cause a change
in decision-making. That is to say, publishing the list is not always
possible. But from the above analysis, our pre-judgment is that it
is wise to publish the list when the number of UCs is larger than
UDs. What’s more, we observe that the dark blue curve is always
in a low position in these three graphs. Interestingly, when UCs
only account for less percent, the number of lobbyists is relatively
much smaller. Minuscule though the number of lobbyists is, the
role of them should not be underestimated. That means the need
for the introduction of lobbyists is beyond question. But whether
to publish the list or not need to be discussed. Therefore, we think
it is necessary to carry on a more in-depth discussion.
In the analysis above, the emotion inhibition coefficients are
relative small (B = 0.1, D = 0.15). Next we will change the relative
size of the emotional coefficients, and then explore how the evolu-
tion results will change as population structure is changed. We also
want to verify our pre-judgment through doing this. Firstly, we
keep the emotion promotion coefficients unchanged, then B and
D are set to be 0.15 and 0.2, respectively. When a1 = 0.2, a3 = 0.2,
the result is that method 1 and method 2 are also the best choices
(here a1 = a3 )(see Fig. 2 (A1)). However, the simulation experiment
results also show (see Fig. 2 (A2)) that the yellow curve become
higher than the other curves when a1 = 0.1 and a3 = 0.3. This rep-
resents method 4 obviously already is the best decision in a longer
term perspective when a3 = 0.3. This time, UDs are three times as
many as UCs. This also confirms what we have mentioned above:
when the number of UCs is relatively small, publishing the list may
652 Q. Chen et al. / Chaos, Solitons and Fractals 91 (2016) 649–655

Fig. 1. (Color online) Time evolution of the number of cooperators (n) for different population structure. The y1, y2, y3, y4, y5 represent the method 1, method 2, method 3,
method 4, method 5, respectively (the same below). (a) a1 = 0.1, a3 = 0.1, a2 = 0.8. (b) a1 = 0.1, a2 = 0.5, a3 = 0.4 (c) a1 = 0.1, a2 = 0.4, a3 = 0.5.

Fig. 2. (Color online) Time evolution of the number of cooperators (n) for different values of emotion inhibition coefficients. (A1) When A = 0.3, B = 0.15, C = 0.35, D =
0.2, E = 0.4 and a1 = 0.2, a2 = 0.6, a3 = 0.2, method 1 and method 2 are the best choices. (A2) When A = 0.3, B = 0.15, C = 0.35, D = 0.2, E = 0.4 and a1 = 0.1, a2 = 0.6, a3 =
0.3, method 4 is the best choice in the long run. (A3) But when the A = 0.3, B = 0.2, C = 0.35, D = 0.25, E = 0.4, method 3 and method 4 also become the best choices in the
long term even a1 = 0.2, a2 = 0.6, a3 = 0.2. (B) We keep the proportion of UC larger than that of UD, a1 = 0.3, a2 = 0.6, a3 = 0.1. Then we gradually increase B and D. (B1)
A = 0.3, B = 0.1, C = 0.35, D = 0.15, E = 0.4.(B2) A = 0.3, B = 0.15, C = 0.35, D = 0.2, E = 0.4.(B3) A = 0.3, B = 0.2, C = 0.35, D = 0.25, E = 0.4. This time method 1 and method
2 also always the best choices.

be not reasonable. And then we keep on making B and D larger,
the results are more interesting. When a1 = 0.2, although the pro-
portion of UDs is also only 20%, method 1 and method 2 have ap-
parently already lost its original advantage (see Fig. 2 (A3)). This
time, the number of UC and UD is the same. That is to say, when
we keep the proportion of UC unchanged and let the emotion in-
hibition coefficient get larger, the critical value of a3 which can in-
fluence the decision-making will get smaller (compare Fig. 2 (A1)
to Fig. 2 (A3)). But we can find that when the number of UC is
larger than UD, the method 1 is always the best choice, no mat-
ter how the coefficients change(see Fig. 2 (B)). Namely, we suspect
that only when there are more UCs than UDs in the group, method
1 and method 2 show their advantages then. Otherwise, whether
to publish the list or not is worth pondering.
To confirm our conjecture, we set the proportion of the number
of SR to 0.7. And then we carry on the data scanning. Results are
presented in Fig. 3. We can notice that when the proportion of UC
is subtracted 0.01 gradually, the results are completely different.
First, we can observe the Fig. 3 (a). Here B = 0.2 and D = 0.25.
With the increasing of UD, great change has happened in the con-
 Q. Chen et al. / Chaos, Solitons and Fractals 91 (2016) 649–655 653

Fig. 3. (Color online) The evolution result of four models for different population structure. We set a2 = 0.7 unchange. The vertical axis represents the ratio of non-
cooperators translate into cooperators. The horizontal axis represents the proportions of the UDs. (a) Here we set A = 0.3, B = 0.2, C = 0.35, D = 0.25, E = 0.4, and then
add the UDs’ proportion 0.01 each time until a3 = 0.3. (b) Set A = 0.3, B = 0.15, C = 0.35, D = 0.2, E = 0.4. (c) Set A = 0.3, B = 0.1, C = 0.35, D = 0.15, E = 0.4.

version rate of non-cooperators. At the first time, the conversion
rate of method 1 and method 2 is higher than method 3 and
method 4. However, when the proportion of UDs increases to a
certain extent, the situation completely reverses. If the emotion
inhibition coefficients are relatively small, method1 and method
2 always the optimum selection (see Fig. 3 (b) and Fig. 3 (c)).
Therefore, we know that whether to publish the list or not base
on the population structure and the relative size of emotion
coefficients.
Part two: The influence of different synergy factor (r) and dif-
ferent environment noise (φ ) on evolution will be discussed in this
part. In order to analyze these results comprehensively, the num-
ber of UDs is set to be larger than UCs. At high synergy factor,
the outcome of the TPGG is significantly different. As evidenced by
the results presented in Fig. 4(a), at r = 3, high level of coopera-
tion will generate as long as the list is published. It is not difficult
to find that the yellow and gray curves always in the down posi-
tions (see Fig. 4(a) and the right side of the Fig. 4(A–C) ). Notably,
the results also show that the amplitude of environment noise (φ )
has nothing to do with the evolution result this time. Therefore, it
is wise to publish the list in spite of the fact that the number of
UDs is larger than that of the UCs. Moreover, we can find that the
two blue curves almost coincide with each other. This means the
influence of lobbyists is no longer obvious. Actually, those are rea-
sonable. When synergy factor is big, if individuals put a little bit
in, they may get a lot out.
However, if the synergy factor is less than some certain value
(see Fig. 4(B) and (C)), things will be different. Firstly, it is ob-
vious that the dark blue curves are in a low position. Therefore,
the role of lobbyists can not be underestimated in such a situa-
tion. Secondly, whether to publish the list or not will depend on
various parameters. Notably, here we can find that the results of
y1 , y2 and y5 change with the change of the amplitude of envi-
ronment noise (φ ). Beyond that, the change amplitude varies with
the amplitude of environment noise (φ ) (see the left side of the
Fig. 4(A–C)). If φ is small (see the left side of the Fig. 4(C)), the
decision result can be easily influenced by changing the synergy
factor. Conversely, the result will be less susceptible to r (see the
left side of the Fig. 4(A)). That is, φ also determines the influential
of the synergy factor. Then if r and φ are determined, the analysis
of this part may return to the part one. Combined the conclusion
in part one, whether to publish the list or not depends on the syn-
ergy factor, the environment noise, the population structure and
the emotion coefficients in such a situation. If the proportion of
UCs is larger than that of UDs, publishing the list is a good choice.
But if there are more UDs than UCs in the group, depth analysis
is needed (This is what we have discussed in part one). There-
fore, this is also the reason we set a1 < a3 . Let us take r = 1.5
for an example. Although the emotion coefficients and the pop-
ulation structure are the same, the result is totally different (see
the bottom row of the Fig. 4 ). The result in Fig. 4(C) tells us that
publishing the list can improve the level of cooperation. But when
φ = 10 0 0, keeping the list confidential is the wiser way to guar-
antee the cooperation rate. Once r and φ are determined, how to
make decision depends on the population and the relative size of
the emotion coefficients. It has already been discussed in the part
one.
In the preceding paragraphs, the evolution results of our model
have been analyzed from a long-term perspective. But compar-
ing all these above data, we can find an interesting phenomenon.
Method 1 always be better than method 2 in the early stage
of evolution. Besides, comparing method 3 to the method 4, the
method 3 always has a better result in the earlier stage (see
Figs. 1 and 2. In the early stage, the position of the light blue curve
is higher than that of the orange curve. Likewise, the position of
the gray curve is higher than that of the yellow curve). Thus, if the
method 1 and method 2 are the reasonable choices for the long
term, then the method 1 should be chose in the short run. Other-
654 Q. Chen et al. / Chaos, Solitons and Fractals 91 (2016) 649–655
Fig. 4. (Color online) Evolution of five models for different environment noise (φ ) (top row) and different synergy factor (r) (bottom row). Here a1 = 0.1, a2 = 0.5, a3 = 0.5
and A = 0.3, B = 0.1, C = 0.35, D = 0.15, E = 0.4.
wise, we should select the method 3 in the short term (We have
verified that it is also true for the other cases).
4. Conclusion
In our daily life, public cultural goods are becoming more and
more popular among the masses. And more and more people want
to pursue high quality public cultural goods as well. In reality, for
government, it is difficult to provide plenty of high quality public
cultural goods for all people. Therefore, it is important to encour-
age everyone to contribute his own strength to build better qual-
ity public cultural goods. Providing public culture goods through
crowd-funding are very common in some southern rural areas in
China from ancient to modern times. Moreover, almost all of deci-
sion makers like to publish the list to stimulate cooperation. Since
it has survived to this day, it must have its own vitality. Therefore,
we use the modern and scientific methods to inherit and expand it.
We introduce three types of people: UC, SR and UD. UCs contribute
their own strength unconditionally. Besides, some UCs will act as
lobbyists to persuade SRs to cooperate. UDs firmly do not want to
contribute their own strength at the first time. But they are all
payoff-driven players. SRs’ emotions will be influenced by others
when making decisions. They are the combination of conformity-
driven players and payoff-driven players. Therefore, we want to
try our best to make more SRs become cooperators by influencing
their emotions. Then more payoff-driven players may choose to co-
operate. Therefore, we can get more contributions and the quality
of the public goods can be raised.
To sum up the above arguments, the synergy factor (r) is the
most influential parameter in our model. Here, we firstly discuss
the role of lobbyists. If r is larger than some certain value, the role
of lobbyists can be ignored. This time, decision makers need not
make the effort to encourage more UCs to act as lobbyists. But if r
is smaller than that certain value, more attention should be paid to
lobbyists. Through our study, we find that the lobbyists are playing
a very significant role in the crowd this time. A lobbyist is some-
one who is able to persuade more waverers to become cooperators.
Therefore, in a sense, lobbyists have a degree of collective influ-
ence. Just as Attila Szolnoki discovered, players with high collec-
tive influence can efficiently transfer their own strategies to others
[53]. Accordingly, we must fully use the role of lobbyists. There-
fore, we should encourage more capable people to act as lobbyists,
if possible. As mentioned above, even the proportion of lobbyists
is small, they can play a very important role in promoting the co-
operation. Unfortunately, there are only few capable people who
are willing to spend time and energy to engage in this work. To
attract more capable UCs, we advise that lobbyists should be re-
warded. The small amount of investment in lobbyists can bring a
lot of cooperators, why not do it? However, whether to publish the
list or not depend on various factors.
As analyzed in section three, there can be little doubt that de-
cision makers should publish the list if r is large. But if the syn-
ergy factor r is really small, we should consider the population
structure, the synergy factor (r), the relative sizes of emotion co-
efficients and the amplitude of environment noise (φ ) at the same
time. If the number of UCs is larger than that of UDs, publishing
the list can greatly improve the level of cooperation. Otherwise,
whether to publish the list or not depends on the other three fac-
tors (r, φ and the relative sizes of emotion coefficients). Firstly, the
influential of the synergy factor has close correlation to the ampli-
tude of environment noise (φ ). In addition, the influential trends
to increasing along with the environment noise decreased. If the
synergy factor (r) and the environment noise (φ ) are determined,
how to make decisions depend on the relative size of the emo-
tion coefficients and the population structure. When the emotion
inhibition coefficients are relatively large, decision makers should
deliberate over whether to publish the list or not even when the
number of UCs and UDs is nearly equal. Meanwhile, if the emo-
tion inhibition coefficients are relatively small, publishing the list
can be seen as the bad decision only when the number of UDs is
much larger than that of UCs. Those critical points depend on the
established environment noise and synergy factor.
Additionally, we observe that decision results should be recon-
sidered in the short term. Firstly, it is wise to ask every lobby-
ist only to persuade one non-cooperator to cooperate during each
of the stages in the short run. Secondly, if publishing the list reg-
ularly is more reasonable for the long run, the short-term deci-
sion should be in agreement with that long-range decision. Other-
wise, we should choose to only publish the original list only once.
This paper also shows that we can use some traditional methods
to solve modern social dilemma. But it is necessary to take the
essence and discard the dross. Now after improving, decision mak-
ers are no longer working in the dark. Instead, they can choose the
best possible solution depending on the situation.
 Q. Chen et al. / Chaos, Solitons and Fractals 91 (2016) 649–655
Acknowledgments
This work was funded by the National Natural Science Founda-
tion of China (NSFC) under Grant No. 71272148 and by Ph.D. Pro-
grams Foundation of Ministry of Education of China under Grant
No. 20120 032110 039.
References
[1] Zhang J, Zhang C, Cao M. How insurance affects altruistic provision in thresh-
old public goods games. Sci Rep 2015;5.
[2] Axelrod R, Dion D. The further evolution of cooperation. Science
1988;242(4884):1385.
[3] Axelrod R. The evolution of cooperation. J Policy Anal Manag
1984;40(1):167–86.
[4] Weibull JW. Evolutionary game theory. Southern Econ J 1995;22(88):4349.
[5] Rand DG, Nowak MA. Positive interactions promote public cooperation. Science
2009;325(5945). 1272–5.
[6] Fehr E, Gchter S. Altruistic punishment in humans. Nature
2003;415(6868):137–40.
[7] Traulsen A, Hauert C, De Silva H, Nowak MA, Sigmund K. Exploration dynamics
in evolutionary games. Proc Natl Acad Sci USA 2009;106(3):709–12.
[8] Peng D, Yang H, Wang W, Chen G, Wang B. Promotion of cooperation induced
by nonuniform payoff allocation in spatial public goods game. Eur Phys J B
2010;73(3):455–9.
[9] Guan J, Wu Z, Wang Y. Effects of inhomogeneous activity of players and noise
on cooperation in spatial public goods games. Phys Rev E 2007;76:56101.
[10] Hauert C, Traulsen A, Brandt H, Nowak MA, Sigmund K. Via freedom to coer-
cion: the emergence of costly punishment. Science 2007;316(316). 1905–7.
[11] GäChter S, Renner E, Sefton M. The long-run benefits of punishment.. Science
2008;322(5907). 1510–.
[12] Clutton-Brock TH, Parker GA. Punishment in animal societies. Nature
1995;373(6511):209–16.
[13] Helbing D, Szolnoki A, Perc M, Szabo G. Punish, but not too hard: how
costly punishment spreads in the spatial public goods game. New J Phys
2010;12(8):3208–14.
[14] Fehr E, Gachter S. Cooperation and punishment in public goods experiments.
Am Econ Rev 20 0 0;90(4):980–94.
[15] Brandt H, Hauert C, Sigmund K. Punishment and reputation in spatial public
goods games. Proc Biol Sci 2003;270(1519):1099–104.
[16] Yang H, Rong Z. Mutual punishment promotes cooperation in the spatial public
goods game. Chaos Soliton Fract 2015;77:230–4.
[17] Gao J, Li Z, Cong R, Wang L. Tolerance-based punishment in continuous public
goods game. Phys A Stat Mech Appl 2012;391(16):4111–20.
[18] Nikiforakis N. Punishment and counter-punishment in public good games: can
we really govern ourselves? J Public Econ 2008;92(1–2):91–112.
[19] Choi JK, Ahn TK. Strategic reward and altruistic punishment support coopera-
tion in a public goods game experiment. J Econ Psychol 2013;35(1):17–30.
[20] Szolnoki A, Perc M. Reward and cooperation in the spatial public goods game.
Europhys Lett (EPL) 2010a;92(3):38003.
[21] Hauert C. Replicator dynamics of reward & reputation in public goods games.
J Theor Biol 2010;267(1):22–8.
[22] Santos MD. The evolution of anti-social rewarding and its countermeasures in
public goods games. Proc R Soc B 2015;282(1798):20141994.
[23] Nowak MA, Sigmund K. Evolution of indirect reciprocity. Nature
2005;437(7063):1291.
[24] Brandt HW, Sigmund K. Indirect reciprocity, image scoring, and moral hazard.
Proc Natl Acad Sci USA 2005;102(7):2666–70.
[25] Wang X, Chen X, Gao J, Wang L. Reputation-based mutual selection rule pro-
motes cooperation in spatial threshold public goods games. Chaos Soliton Fract
2013;56:181–7.
655
[26] Hauert C, De Monte S, Hofbauer J, Sigmund K. Volunteering as red
queen mechanism for cooperation in public goods games. Science
2002;296(5570):1129–32.
[27] Semmann D, Krambeck H, Milinski M. Volunteering leads to rock-paper-scis-
sors dynamics in a public goods game. Nature 2003;425(6956):390.
[28] Szabo G, Hauert C. Phase transitions and volunteering in spatial public goods
games. Phys Rev Lett 2002;89(11):118101.
[29] Bagnoli M, Mckee M. Voluntary contribution games: efficient private provision
of public goods. Econ Inq 1991;29(2):351–66.
[30] Rivas MF, Sutter M. The benefits of voluntary leadership in experimental pub-
lic goods games. Econ Lett 2011;112(2):176–8.
[31] Santos FC, Santos MD, Pacheco JM. Social diversity promotes the emergence of
cooperation in public goods games. Nature 2008;454(7201):213.
[32] Perc M. Does strong heterogeneity promote cooperation by group interactions.
New J Phys 2011;13(12):123027.
[33] Gomezgardenes J, Vilone D, Sanchez AM. Disentangling social and group het-
erogeneities: public goods games on complex networks. Europhys Lett (EPL)
2011;95(6):68003.
[34] Shi D, Zhuang Y, Wang B. Group diversity promotes cooperation in the spatial
public goods game. Europhys Lett (EPL) 2010;90(5):58003.
[35] Szolnoki A, Perc M. Group-size effects on the evolution of cooperation in the
spatial public goods game. Phys Rev E 2011;84:47102.
[36] Zhu C, Sun S, Wang L, Ding S, Wang J, Xia C. Promotion of cooperation due to
diversity of players in the spatial public goods game with increasing neighbor-
hood size. Phys A-Stat Mech Appl 2014;406(C):145–54.
[37] Cardillo A, Meloni S, Gomezgardenes J, Moreno Y. Velocity-enhanced coopera-
tion of moving agents playing public goods games. Phys Rev E 2012;85:67101.
[38] Wang Y, Chen T. Heuristics guide cooperative behaviors in public goods game.
Phys A Stat Mech Appl 2015;439:59–65.
[39] Zhang J, Zhang C. The evolution of altruism in spatial threshold pub-
lic goods games via an insurance mechanism. J Stat Mech Theory Exp
2015;2015(5):P05001.
[40] Owen AL, Videras J. Culture and public goods: the case of religion and
the voluntary provision of environmental quality. J Environ Econ Manag
2007;54(2):162–80.
[41] Szolnoki A, Chen X. Benefits of tolerance in public goods games. Phys Rev E
2015;92(4). 42813–42813.
[42] De Martino B, Kumaran O, Seymour B, Dolan RJ. Frames, biases, and rational
decision-making in the human brain. Science 2006;313(5787):684–7.
[43] Bechara A, Damasio AR. The somatic marker hypothesis: a neural theory of
economic decision. Games Econ Behav 2005;52(2):336–72.
[44] Nowak MA, Highfield R, (Mensch) E. Supercooperators. Canongate Books Ltd
2011.
[45] Wang J, Fu F, Wu T, Wang L. Emergence of social cooperation in threshold
public goods games with collective risk. Phys Rev E Stat Nonlinear Soft Matter
Phys 2009;80(2):1274–87.
[46] Marks M, Croson R. Alternative rebate rules in the provision of a threshold
public good: an experimental investigation. J Public Econ 1998;67(2):195–220.
[47] Cadsby CB, Maynes E. Gender and free riding in a threshold public goods
game: Experimental evidence. J Econ Behav Organ 1998;34(4):603–20.
[48] Szolnoki A, Perc M. Impact of critical mass on the evolution of cooperation
in spatial public goods games. Phys Rev E Stat Nonlinear Soft Matter Phys
2010;81(5 Pt 2):561–78.
[49] Szolnoki A, Perc M. Conditional strategies and the evolution of cooperation in
spatial public goods games. Phys Rev E 2012;85(2 Pt 2):287–300.
[50] Szolnoki A, Wang Z, B, Perc M. Wisdom of groups promotes cooperation in
evolutionary social dilemmas. Sci Rep 2012;2(8):576.
[51] Szolnoki A, Perc M. Conformity enhances network reciprocity in evolutionary
social dilemmas. J R Soc Interface 2015;12(103):20141299.
[52] Christakis NA, Fowler JH. Connected: the surprising power of our social net-
works and how they shape our lives; 2009.
[53] Szolnoki A, Perc M. Collective influence in evolutionary social dilemmas. Epl
2016;113(5):58004.