Zhao, Lei (2004) Genetic characteristics of artificial agents in formAL. Masters thesis, Concordia University.
- Accepted Version
This work addresses several issues regarding Artificial Life. It focuses on the behaviours of simulated organisms in such models, e.g. the evolutionary tendencies displayed through reproduction. The scope of Artificial Life is briefly discussed, as well as some arguable issues in this area. An Artificial Life model, called FormAL, is presented. FormAL is a platform capable of running simulations with thousands of organisms (agents), and tracing their evolutionary properties through the changes of their genomes. The design principle of FormAL is to allow the simulated agents maximum freedom possible for their behaviours, with fewest possible 'law's to govern them. The ultimate goal of this design is to give rise to emergent properties of the agents governed with very simple rules. Some interesting results came out of extensive experiments conducted with FormAL. Although there is no explicit fitness function in this work, the selection pressure from competing for the finite energy supply drives the agents to evolve into optimized forms. The test results presented in Chapter 4 and 5 indicate that even in a fairly simple environment, premature convergence can be avoided, and that the mutation mechanism plays a crucial role in evolution. The moment the mutation mechanism is lost, the system stops evolving and the course of evolution reaches a dead-end. This project is still young in its development, and a number of future research directions are discussed at the end of this thesis.
|Divisions:||Concordia University > Faculty of Engineering and Computer Science > Computer Science and Software Engineering|
|Item Type:||Thesis (Masters)|
|Pagination:||xi, 90 leaves : ill. ; 29 cm.|
|Degree Name:||M. Comp. Sc.|
|Thesis Supervisor(s):||Grogono, Peter|
|Deposited By:||Concordia University Libraries|
|Deposited On:||18 Aug 2011 18:10|
|Last Modified:||19 Aug 2011 08:04|
Repository Staff Only: item control page
Downloads per month over past year