节点文献

小型生态系统的人工生命个体模型与人工智能模型研究

Study on Artificial Life Model and Artificial Intelligence Model in Small Ecosystem

【作者】 丁小军

【导师】 周春艳;

【作者基本信息】 中南大学 , 计算机软件与理论, 2008, 硕士

【摘要】 人工智能是当前计算机研究的热点,被誉为二十世纪的重大科学技术成就之一,并将在新世纪的网络和经济时代发挥重要作用。作为计算机学科的重要分支,人工智能将渗透到应用计算机技术的各行各业,导致这些行业乃至计算机软件产业本身的变革。然而人工智能目前仍处于初期阶段,本文认为只有生命才能产生高级智能,并且生命若想产生高级智能一定不能离开环境,将一个生命隔离开来,那么它最多只会产生有限的低级智能,因此很有必要模拟一个小型的生态环境来研究生命个体的智能。为此本文研究了复杂系统的自组织规律,并依此进行建模,设定出一些规则,使生物个体能够通过进化和相互间的非线性作用,自组织的形成一个相对稳定的生态系统。在此生态系统中,每个生命个体所处的环境比较稳定,因而对每个生命个体来说,它所观察到的大量事件具有一定重复性,规律性,这样就为它的学习提供了一个保障。本文接下来仍然根据自组织的思想,仿照人类社会商人组织的建立,提出一种智能学习模型,生命个体按照这个模型能够学会一定的时序上的因果逻辑。然后本文根据通用问题求解原理,给出了生命个体的行为选择算法,初步完成了动物智能模拟的理论部分。经实验分析,本文提出的自组织生态系统是比较稳定的,只需要一个很小规模的生物种群就能持续下去,这样就为计算机模拟提供了便利。同时本文进行了实验仿真,模拟了简单的狗的智能,这就说明了本文的模型是有效的。

【Abstract】 Computer artificial intelligence research is the current hot spots, known as the 20th century one of the major scientific and technological achievement, and in the new century the era of network it will play an important role in the economy. As an important branch of computer science, artificial intelligence will penetrate into the application of computer technology industries, resulting in these businesses, and even computer software industry itself changes. But artificial intelligence is still at an early stage. This article holds that only life can have a senior intelligence, and life can’t produce senior intelligence if it is seperated from the environment. A isolated life can only have a limited low-smart. Therefore it is necessary to simulate a small ecological environment to study artificial intelligence. This paper studies the self-organizing phenomenon of complexity system, and using the law of this to evolve a relatively stable ecosystem, in this ecosystem, each individual and the environment in which life is relatively stable, and therefore the large number of incidents which each individual observed will have certain repetitive and regularity, and this will provid a safeguard for study. Then this paper brings forward an algorithm which makes individual lives can learn certain temporal causal logic. And using general principles of problem solving, this paper gives the algorithm for actions selection of the individual, and carries out a simple simulation of intelligent animals.By the experimental analysis, we can see that the self-organizing ecological system presented by this paper is relatively stable, only need a small-scale biological population to continue. That will make the computer simulation convenient. At the same time simulation experiments in this paper, a simple simulation of a dog’s intelligence, which proved this model is effective.

  • 【网络出版投稿人】 中南大学
  • 【网络出版年期】2009年 01期
  • 【分类号】TP18
  • 【下载频次】247
节点文献中: 

本文链接的文献网络图示:

本文的引文网络