Evolution of trust in the N-player trust game with transformation incentive mechanism
Data files
Jan 15, 2025 version files 107.22 KB
-
Fig1data.xlsx
105.83 KB
-
README.md
1.38 KB
Abstract
Trust game is commonly used to study the evolution of trust among unrelated individuals. It offers valuable insights into human interactions in a variety of disciplines, including economics, sociology, and psychology. Previous research has revealed that reward and punishment systems can effectively promote the evolution of trust. However, these investigations overlook the gaming environment, leaving unresolved the optimal conditions for employing distinct incentives to facilitate trust level effectively. To bridge this gap, we introduce a transformation incentive mechanism in an N-player trust game, where trustees are given different forms of incentives depending on the number of trustees in the group. Using the Markov decision process approach, our research shows that as incentives increase, the level of trust rises continuously, eventually reaching a high level of coexistence between investors and trustworthy trustees. Specifically, in the case of smaller incentives, rewarding trustworthy trustees is more effective. Conversely, with larger incentives, punishing untrustworthy trustees is of greater efficacy. Furthermore, we find that moderate incentives positively influence the average payoff within the group.
README: A graphics code to plot phase diagrams
https://doi.org/10.5061/dryad.t4b8gtjc4
The dataset shows the data in Figure 1 of the related research paper. Figure 1 is of the type of a phase diagram, which depicts the evolutionary dynamics of investors, trustworthy trustees, and untrustworthy trustees, without incentive in a finite population. The dataset contains data on the computation of stationary distributions of Figure 1.
Description of the data and file structure
These phase diagrams in our main text can be obtained through the provided Matlab code.
Code/Software
Matlab
Description: The main code file is "threemain722", its function is to calculate the stationary distribution and the gradient of selection which are the main components of the phase diagram. In addition, the code file "nchoosek_optimized" is an improved version of Matlab's own function "nchoosek", which is used to compute combinatorial numbers and is slightly faster than "nchoosek". Finally, the code file "quiverc" performs a colour mapping function on the gradient of selection. Furthermore, the code file "quiverc" is referenced as follows:
Bertrand Dano. Quiverc (https://www.mathworks.com/matlabcentral/fileexchange/3225-quiverc), MATLAB Central File Exchange.
Methods
We use MATLAB to numerically compute the evolutionary dynamics of the system.