采用人工生命方法模拟七星瓢虫捕食行为进化-生态学杂志.PDFVIP

2001, 20( 1) : 65- 69 Chinese Journal of Ecology * 王 俊 李松岗 ( 北京大学生命科学学院, 100871) Simualtionof Lady Beetle s Preying Behavior Evolution with Neural Network and Genetic Algorithms. Wang Jun, Li Songgang( College of L if e Sciences, Peking University, 1000871) . Chinese Journal of Ecology , 2001, 20( 1) : 65- 69. This paper escribes a computat ional mo el of searching behavior of la y beetle ( Coccinella sep temp unctata L.) by using Neural Network an Genetic Algorithms. It also escribes several Neural Network structures to interpret how the Neural Network controls the behavior of la y beetle, as well as the evolution of the computational mo el. In the nature, Alphi s pattern is clumpe , an the la y beetle s foo - foraging strategy is: w hen the pre ator oes not touch the prey, the pre ator s behavior is searching in a large area; but if the pre ator fin s a prey, it behavior sw itch to searching in the neighborhoo area aroun the prey. In this paper, several Neural NetWork structures have been use to control the in ivi u als foo - searching behavior. An a form of Genetic Algorithms have been use for the Neural Network s Learning.After simulation, the mo el s foo - foraging strategy is similar to the nature. It is not very necessary for the la y beetle to have the ability of fin ing the Alphi in a short istance. In fact the Alphi almost can not escape after the la y beetle fin s it. So the most important factor for improving the searching efficiency is the ability of memory. It coinci es with the la y bee tle s foo - foraging behavior in the nature. ey words:art ificial life, neural network, genetic algorithms, evolution. : Q958. 1 : A : 2001) 20 70 Burks[ 1] ! !∀[ 4] , ,