Python code for a simulation framework to evaluate wildlife security games

The python code is used for the development of an evaluation framework for wildlife security games. The framework file includes a class for collecting geographic vector data in a wildlife park (park boundary, environmental features, and size of grid cells), a class for setting the preferences and type of movement for each of the agents (rangers, poachers, and wildlife animals), and a class for defining the rules of the game (action sets, payoff functions, order of moves, and solution concept). Once these objects are created, movements of the agents are simulated along the grid created within the wildlife park, for a specified number of moves, to determine the frequency of poaching and arrest events. For evaluation, games can be compared with each other, and against two baseline games, by using hypothesis tests to identify significant differences. The knp file includes code to create the geographical features for the Kruger National Park (KNP), using shapefiles that are publicly available. However, the framework allows for the input of shapefiles for any wildlife park and the creation of custom features. The gametheory file includes algorithms for various solution concepts, including the maximin strategy, Nash equilibrium, subgame perfect Nash equilibrium, Stackelberg security game, and for the Stackelberg follower game that is proposed in the thesis. The remaining files include the code used for implementing the simulation experiments in Chapter 6 of the thesis. A summary of the framework parameters and sources for public shapefiles are provided in Appendix A of the thesis.