Final project for "Stochastic Simulation Method and Application".
This project is known as "基于记忆效应的空间囚徒困境中系统合作的演化" in Chinese. The repo consists of all workflow for running simulation and generating plots.
Two agents game with each other with their own strategies. Output CSV-like format with Python. Figures plotted with R language.
The design of this model is inspired by "Anderson, J. and Schooler, L. (1991). Reflections of the environment in memory. Psychological Science, 2(6):396."
Agents are occupied with their own position on a grid. They game with their
neighbours. In matrix_result folder, we use C++ to run this simulation.
The C++ program exports agent status in every epoch, and we use d3.js to visualize the grid at any moment.
Design of this model is inspired by:
- Wang J, Liu L N, Dong E Z, et al. An improved fitness evaluation mechanism with memory in spatial prisoner's dilemma game on regular lattices[J]. Communications in Theoretical Physics, 2013, 59(3): 257.
- Szabó G, Tőke C. Evolutionary prisoner’s dilemma game on a square lattice[J]. Physical Review E, 1998, 58(1): 69.
Simulation of 1v1 agent & Simulation on grid
(cooperative agent proportion vs simulation epoch)
Animation of evolution on 50x50 grid
The simulation program and the visualization program is licensed under MIT.
Model in this project is designed by Y. Su and R.Q. Lyu.
Also thank J.Y. Xiang for model evaluation and running simulation.