基于AGENT动态加权无标度网络的双语竞争模型

Two Language Competition Model with Dynamic Weighted Scale-free Network Based on Agent

近年来,濒危语言的灭亡速度越来越快,为了寻求濒危语言的保护措施,提出了以同一区域内的双语竞争为研究对象,利用基于agent的无标度网络构建接近实际的语言网络模型。模型以agent为网络节点生成无标度网络,agent的language等于1、2、3分别表示单语1agent、单语2agent和双语3agent。语言网络通过agent之间的短期连接、长期连接、无连接三种关系之间的演化维持网络的动态特性,语言之间的传播竞争只存在于有连接的agent之间,语言的传播竞争分为语言学习和语言遗忘,引入了语言地位S1、S2和agent语言密度σ之间的乘积作为语言学习和语言遗忘的概率。仿真结果表明,提高弱势语言的地位、降低语言网络中agent个体的度和降低语言网络的动态性保证语言网络的稳定性,更有利于保护弱势语言。

The extinction of endangered languages is rapidly increasing in recent years.Based on language competition in the same areaas the object of study,an agent-based scale-free network is proposed to build closer to the actual language of the social network to seek methods to protect endangered languages.The model uses agents as node to build scale-free network and defines language as an attribute of agent.There are monolingual1 agents,monolingual 2agents and bilingual 3agents in the model.The network keeps its dynamicity by the evolution of three relations that long-term linking,short-term linking and no linking between agents.The language competition includes language learning,language forgetting and it only existed in the agents which are linking.The probability of language learning and forgetting is equal to the product of language status（S1,S2）and density of agent language（σ）.The simulations reveals that to improve the status of language,reduce the degree of agents and the dynamicity of the language of network are useful methods in protecting endangered languages.