论文部分内容阅读
本文从复杂自适应理论和人工生命基本观点出发,研究一个具有“简单性、局部性、全局性、内聚力、动态性”的复杂自适应系统,并创建包含“智能体、环境、交互规则”三个主要概念以及“适者生存”的淘汰规则的AER多主体模型.该模型中具有局部目标和行为的自主体能通过动态环境达到整体的全局目标,这种涌现的智能能够高效求解问题.实验证明该模型对传统的大规模约束满足问题的测试集,包括n-皇后问题(n=7000)和一些大规模染色问题,均能在较快的时间内找到精确解.同时能在3个时间片内找到精确度(≥94%)很高的近似解.
Based on the theory of complex adaptive theory and the basic concept of artificial life, this paper studies a complex adaptive system with “simpleness, locality, globality, cohesion and dynamism” and creates a complex adaptive system that includes “agent, environment, Rules ”and the“ survival of the fittest ”AER multi-agent model of the model with local goals and behavior of autonomy can achieve a global environment through the overall goal of the emerging intelligence can And solve the problem efficiently.The experiment proves that this model can find the exact solution to the traditional test set of large-scale constraint satisfaction problems, including the n-queens problem (n = 7000) and some large-scale coloring problems. Find approximate solutions with high accuracy (≥94%) in 3 time slices.