本文从生物进化的角度,将群体中的每只昆虫看成一个神经元,彼此之间通过随机、松散的(软)的连接组成一个神经网络;然后,类似于人工神经网络模拟人类大脑,用“松散”的神经网络来模拟昆虫的群体行为。即建立一个松散的大脑——群体智能的随机结构的神经网络模型。最后,根据所建立的数学模型,分别对“蚂蚁筑巢”和“蜜蜂筑巢”行为进行计算机仿真模拟,以表明上述模型的合理性。 In this paper, from the point of view of biological evolution, each insect in the population is regarded as a neuron, which forms a neural network through random and loose (soft) connections. Then, similar to the artificial neural network to simulate the human brain, “Loose ” neural network to simulate insect group behavior. That is, to build a loose brain-group neural network model with random structure. Finally, according to the established mathematical model, we simulate the behaviors of “nest ” and “bee nest ” respectively to show the rationality of the above model.