Evolving game playing strategies

This proposal examines the use of genetic algorithms to evolve complex and dynamic strategies for playing games. The system would be supplied with a selection of simple strategies and would then combine them into more complex strategies that would be created using a Genetic Algorithm. The same process would allow the development of strategies that react to individual players and to evolve as the players own style of play adapts.

Using Genetic Algorithms to crack codes

This proposal would examine strategies for cracking codes using Genetic Algorithms. There are many different codes and the aim would be to develop a strategy for cracking each of them as well as ¡°Meta-strategy¡± for cracking a code where the cracker does not know which system was used to encode the message. The work could also examine the effects of various aspects of the process (e.g. if using frequency of letters or groups of letters what effect does the size or nature of the source document have on the efficacy of the system).