抄録
A fish school, one of the most familiar animal sights for people, fascinates us with its natural beauty. Though there is not any special fish which leads the entire, a fish school behaves like a single organism. It is said that such a collective movement contributes the adaptability toward changes of the surroundings. Therefore, it seems that a collective movement of creatures is applicable to artificial control systems. It is assumed that an autonomous fish school movement emerges by iteration of mutual interaction among neighbor fish. Through studies in the field of fisheries, ecology, physics, and so on have demonstrated some of their instincts. However, an observation of a fish school movement for many hours is not reported, because it's too difficult to grasp precisely their swift behaviors in natural conditions. In this study produce fish school movements in a computer. Therefore we can grasp precisely the position and the velocity of all fish in every simulation step. It is unique that we adopt Fractal Analyses to understand the feature of the complexity of a fish school movement.
In our Fracta Analysesl, we attempt to obtain the feature of a fish school movement in every time-scale of an observation. At the first step of our analyses, we observe the coarsened trail pattern of a school. At the next step, we quantify the complexity of.coarsened patterns as Fractal Dimensions. In this study, we deal with two types of fish school movements, one is behaviors in the space where the area of their action is not restricted and another is the behavior in a tank where the area of action is restricted by walls of the tank. Depictions of coarsened movements and fractal dimensions of their movements in a tank indicate the complexity influenced by the size of the tank.
