Complex Agent Network Multi - Language Competition General Model

There are many social concerns about protection of endangered languages; language international trend has been increasing in recent years. It will result in some language is weakened or even extinction, it will result in loss of national cultural wealth. Understanding the evolution trend of the language can be better to adopt effective control methods to curb or even prevent the destruction of the language.This paper proposes a dynamic social network model for the competition among languages with agent based modeling method and social circles theory. The parameters of the topological structure of constructed social network are more close to the actual social network. The individual agents in the network can move, dead and reproduce, so the constructed social network is endowed with dynamic characteristics.The network is divided into mixed residential and fragmentation inhabit according to the characteristics of realistic social language, at the same time to join the diversity of population movement to study the tendency of language competition; Language consists of words and words changes affect the evolution of language.We can summarize the internal mechanism of language change using the words communication rulesFirst of all, taking the trilingual competition as an example, we proposed a method that it is decomposed into three times of bilingual competition. On the basis of it, we presented a universal complex agent network method for multi-lingual competition and policy intervention. The model of learning and forgetting of the agents in the network is proposed. It is used to study the individual language competition in different social environment evolution paths. The agent nodes in the network represents the individuals have learning and forgetting abilities, each agent can be gained the second or third language to become bilingual or trilingual agents, and they can become monolingual or bilingual agent by forgetting one or two languages. The complex evolutionary relationships happen in the network.Secondly, the individual agents in the network can be given space attribute using the simulation modeling methods of multi-lingual competition network. The agents can describe mixed residential and fragmentation inhabit, so the constructed social network is endowed with dynamic characteristics.The agents on the network represent which has the function of learning and forgetting. And each agent can be gained a second or third language to become bilingual or trilingual by learning and can forget to become monolingual or bilingual. The vertical transmission of language is taken into account. The simulating analysis of the proposed model shows that the system parameters of language status, the proportion of people of different languages, the proportion of the mobile population, social radius, living space model and integrated control measures can affect the results of simulation.Finally, taking the trilingual competition as an example, this paper proposes a new universal complex agent network model for multilingual competition. This model is based on social circles theory and bilingual competition model. In order to reflect the vocabulary of the language, a language structure with bit-strings is introduced into the agent which is a node of the network. By the evolution of vocabulary structure of speakers, the model can show the macroscopic process of language evolution. We decompose the competition among three languages for three times of bilingual competition and analyze the formation process and influencing factors of various types of speakers by computational experiments.The results can verify language competition and simulation under the environment of complex agent. We can restore and predict the evolution trend of the language. Our model has a profound impact on the protection of endangered languages.