| Co-Investigator(Kenkyū-buntansha) |
NAKAMURA Tsuyoshi Nagoya Institute of Technology, Faculty of Engineering, Assistante, 工学部, 助手 (90303693)
INUZUKA Nobuhiro Nagoya Institute of Technology, Faculty of Engineering, Associate Professor, 工学部, 助教授 (10221780)
|
|---|
| Research Abstract |
An environment isprepared for the multi-agents self-evolution in the computer. This environment is called artificial life. In this environment, multi-agents movements and actions look like the living things.
Multi-agents live ecological and move around the virtual space in the computer. We experimented and evaluated the pattern of the agents in the software environment is developed by the genetic algorithm. Agents get information from the current environments, and decide next movement and action and then go into the new environment. Agents do this incrementally. Here, agents are evolved in genetic. We developed the crossover, delete and mutation technology in our genetic programming methods.
In general agents decide one move or action from the many choices by the binary-decision-diagram method or classifier System on the research of the artificial life. In this research binary-decision diagram method are analyzed and evaluated in the evolution of the agents. Genes are defined by states o
… More
f the binary decision diagram. In the classifier system, genes are expressed by the sequences of 0/1 digits. When agents receive the current information of environments as input, they decide an output branch from 2 branches. These judgment methods are evolved in genetic. In this research, we developed the model of the migrating birds and non-migrating birds ecological-differentiation is boned from the same agent-programs. This differentiation is caused by the initial direction of the evolution. In this case, the directions of the evolution are depends on the agents-weight and efficiency of the energy-consumption. Next model of the evolution is meme. Meme is the medium of the culture-transfer. Agents are able to evolve using it and also genetic ways. Our research is how meme effect on the food-chain system among some species of the agents. For example, human is born with the ability of conversation from the parents. However he is not yet decided which is his mother language. After he was borne, he decides his mother language. Then, genetic ability means the conversational potential, and meme decide which language he speaks. The group of the agents has the language -culture. In our research, a model is proposed the meal-culture among some species of agents. We experimented and evaluated the efficiency of the meal-culture of the agents. For example, relations of the efficiency of the molding meme, number of the generations and the population of agents. Moreover, what information in the meme is essential is analyzed. Less
|
