AiS Challenge Interim Report

Team Number: 90

School Name: Silver High School

Area of Science: Sociology

Project Title: Cultural Evolution

 

 

Culture is composed of the restrictions imposed on a interactions between individuals in a society. How a society develops a culture is a difficult question, one that has been the subject of much research. Current sociologists say that culture develops to provide closer family ties, happiness and greater possibility of success. Anthropologists say it develops to facilitate survival and reproduction, but these are only educated guesses. Due to long spans of time, it is difficult if not impossible to study exactly why and how a culture develops.

A computer simulation of a society built of agents(individuals) provides can provide a valuable resource to sociologists, giving them the ability to look at simplified interactions over the long term, i.e. thousands of generations, something that would be impossible in real time.

Understanding how culture develops is critical to understanding our past and present and preparing for the future.

To create the simulation involves heavy computational demands, on the order of months to run a short small-scale simulation on a typical desktop computer. Only by using massively parallel machines is it possible to effectively run a simulation. The method for this model is based on genetic algorithms(GA’s), a computational technique based on Darwinian evolution. In essence a GA is a heuristic search, one that is suited to problems in linear programming, artificial intelligence and evolutionary simulations. The simulation will be a search through the complex space of possible optimal cultures. The complex space is a mathematical concept which can be thought of as a high dimensional space. Each point in this space represents a combination of factors including cultural rules and agent characteristics. A GA moves through this space and is subjected to evaluation criteria. By the nature of GA’s and natural selection, the phenotype approaches a global optimum and receives higher and higher evaluation scores This optimum may be a global optimum or an sufficiently high score. The simulation will portray development of a ‘good’ culture and how individuals adapt to the culture. To create a simulation that is as unbiased as possible a good culture is defined using an accepted anthropological viewpoint that culture promotes survival and reproduction.

So far all algorithms have been developed with the help of a mentor, some C++ code has been written, and research regarding the subject has been preformed. The basic C++ code is being added to using the message passing interface(MPI), a library for parallelization, in preparation for execution on theta, A 128 processor supercomputer.


Acknowledgements and Sources I would like to thank Mr. Steve Blake, Dr. David Harris, Mr. John Hancock and Mrs. Peggy Larisch for their support, advice and inspiration throughout the course of the project.

1.Holland, John. Adaptation in Natural and Artificial Systems. 1992, MIT Press
2. Koza, John. Genetic Programming. 1992, MIT Press.
3. Epstien, Joshua and Robert Axtell. Growing Artificial Societies from the Bottom Up. MIT Press.

Team Members

Team Mail

Sponsoring Teacher(s)

Project Mentor(s)