A Model of Biological Differentiation in Adaptiogenesis to the Environment

ai(bi, sti).
(1)
The behavioral strategy sti is represented as a function which determines action. Let ACT be a set of finite
2
in Artificial Life VIII, Standish, Abbass, Bedau (eds) (MIT Press) 2002. pp 93–96
In this paper, as a first step of this research, we propose a computationary evolution model of the biological differentiation. Our model is based on the interrelation between physical feature and behavioral strategy. This paper, then, presents a simulated ecology by our model,
in Artificial Life VIII, Standish, Abbass, Bedau (eds)(MIT Press) 2002. pp 93–96
1
A Model of Biological Differentiation in Adaptiogenesis to the பைடு நூலகம்nvironment
Introduction
This paper proposes a computationary evolution model for biological differentiation. A number of evolutionary models for an ecosystem has been proposed during past decade or so (e.g., (Takashina & Watanabe 1994),(Eiben & van Hemert 1999),(Downing 1998)). In the models, some species acquire behavioral and physical features suitable for a certain environment to which the species is exposed. The ecosystem of a few different species is well simulated by the evolutionary computation based on the models (e.g., (Michalewicz 1996),(Miychell 1996),(Kaneko & Yomo 2000)). Our research group (Moriwaki et al. 1996) has presented a computationary evolution model for a pseudo ecosystem including three types of artificial living things: herbivores, carnivores and plants. The model has well simulated food-chain. In the model, an ecological niche of the living things is given as fixed element. On the other hand, natural living things acquire an ecological niche, such as carnivore and herbivore, with the macro-evolution. It is obvious that the emergence of the variation in ecological niche is necessary for the development of an ecosystem. A longrange target of this research is to propose an evolution model for emergence of ecological niche.
actions, and actti be an action in ACT , which is selected by ai at time t. actti is determined by the value of the function for sti, as follows.
actti = sti(ini(t), exi(t)),
A Model of Biological Differentiation
Biological differentiation has been generally observed in the nature. It is said that the evolution of physical feature and behavioral strategy, as the one of many factors, causes the biological differentiation. In this paper, we propose a computationary evolution model of the biological differentiation. The following section gives a description of the definitions of an artificial living thing (called autonomous agent, or simply agent) and its reproductive activity, and also gives a brief description of the common property of reproductive isolation that the living things have.
Takamasa Sawada, Atsuko Mutoh, Shohei Kato and Hidenori Itoh
Department of Intelligence and Computer Science, Nagoya Institute of Technology
Gokiso-cho, Showa-ku, Nagoya 466-8555, Japan. {sawada, atsuko, shohey, itoh}@ics.nitech.ac.jp
Abstract
In this paper, we propose a model of biological differentiation based on physical features and behavioral strategies in order to probe the mysteries of biological differentiation. We, then, implement the model, and perform the Alife simulations of birds’ evolution in an artificial ecology. The ecology consists of two islands, and the atmospheric temperature in each of the islands fluctuates periodically. As a result of the simulations, we have observed that birds, adapting themselves to a change of temperature in their habitat, branch off in two types of biology: migratory bird and resident bird.
Agent in the model
At an unit time, an agent chooses one from finite actions, and then, does it. The sensory information which an agent can perceive consists of a self state (internal state) and an environmental situation (external stimuli). An agent has a behavioral strategy for determining an action by the sensory information. Moreover, the relative merit of a behavioral strategy for an agent depends on the agent’s body, and we call the body element physical feature. Let ai be an agent (i is identifier), and let bi and sti be the physical feature and the behavioral strategy of ai, respectively. ai is, then, characterized as follows.
(2)
where ini(t) and exi(t) are internal state and external stimuli of ai at time t, respectively. This equation means that a behavioral strategy of the agent determines an action by agent’s sensory information. After the action, the internal state of the agent is updated.