CN111125911B - Modeling simulation method of language competition microscopic simulation model based on partitioned double-layer network - Google Patents

Abstract

The invention relates to a modeling simulation method of a language competition microscopic simulation model based on a partitioned double-layer network, which improves a modeling method based on an agent social circle network, constructs a social network with a remote weak connection relation, and can describe the reinforcing trend of the weak relation of the social network caused by regional population flow and communication connection under the urbanization background. Besides considering the influence of the language position and the language population proportion, the language evolution mechanism of different crowds introduces a new mechanism of influence of bilingual on the conversion from monolingual to bilingual and influence of language attitude on language selection behavior. An improved agent language horizontal and vertical propagation model is provided, and the propagation model comprehensively considers factors such as language status, language density, language attitude, influence of bilingual on monolingual, remote connection weight and the like.

Description

Modeling simulation method of language competition microscopic simulation model based on partitioned double-layer network

Technical Field

The invention relates to a modeling simulation method of a language competition microscopic simulation model based on a partitioned double-layer network, belonging to the field of computer simulation.

Background

In the current research, no matter a macro model or a micro model, people mainly consider population proportions of different languages of a language communication network, strength of language status and influence of family language inheritance on language selection and use. The individual language learning behaviors in the social network are considered to be influenced by different types of language densities and language positions of surrounding people, people tend to select the powerful language and the language with a large language population ratio for learning easily, and competition of the two languages is formed. The key point of policy research is to realize the improvement of the proportion of bilingual population by regulating and controlling the position of the vulnerable language, and further protect the vulnerable language by bilingual. Some macroscopic models consider the buffer factors of geographic space and cultural space, and the geographic and cultural isolation is considered to be beneficial to the preservation of weak languages. Recent studies have shown that the language position or situation is unchangeable in a short period of time and is not suitable as a variable regulatory parameter. According to Chen Bao ya, the main factor influencing language competition is the language potential (namely, the language position generally referred to), the language potential of one language is determined by the using population number, the vocabulary amount and the number of owned documents of the language, so that the language potential or the language position is difficult to change in a short time and cannot be used as a policy regulation parameter. That is to say, the language position or language posture is the sum of historical culture, and is difficult to change in a short period, and as a policy regulation measure, the practical implementation is difficult in reality. With the convenience of urbanization and communication, the buffer effect of the geographic space is weakened. Further search for mechanisms for maintaining language harmony and policy regulation measures is imperative.

Disclosure of Invention

The invention provides a modeling simulation method of a language competition microscopic simulation model based on a partitioned double-layer network, which is used for solving the problems of competition and protection of current social endangered languages.

The technical scheme of the invention is as follows: a two-dimensional grid is divided into two areas according to a proportion, agents are respectively established in the two areas, monolingual and bilingual population is initialized, attributes of the monolingual and bilingual population are set, and a local social network and a remote cross-regional social network are respectively established by utilizing a social circle principle and a virtual mirror image principle; in the established network, agents carry out language communication through horizontal propagation and carry out language inheritance through vertical propagation; describing the dynamics of the network according to the introduced daily movement and long-distance migration; after horizontal transmission, vertical transmission, daily movement and long-distance migration, the agent updates the social network again according to the social circle principle and the virtual mirror principle, and utilizes the established social circle network to perform modeling simulation on language competition transmission; and finally, analyzing the influence on language propagation by adjusting the influence of bilingual on monolingual, the attitude of monolingual x on learning y language, the attitude of learning bilingual and the remote connection proportion parameter.

The method comprises the following specific steps:

step1, create agent and initialize monolingual and bilingual populations: creating a two-dimensional space network within the range of 250 x 241, creating nofArea1 agents in the two-dimensional space network, wherein the abscissa of the nofArea is randomly distributed in the interval of [0, (250 × cc) -1 ]; creating nofArea2 agents with the abscissa randomly distributed within the interval [ (249 × cc), 249], wherein cc belongs to [0, 1; setting a language type: monolingual x, monolingual y and bilingualr z, wherein the population proportions of monolingual x, monolingual y and bilingual z in the nofArea1 agents are nof1 x%, nof1 y% and nof1 z%, respectively, and nof1 x% + nof1 y% + nof1 z% ═ 1; the population proportions of monolingual x, monolingual y and bilingual z in nofArea2 agents were assigned to nof2 x%, nof2 y% and nof2 z%, respectively, and nof2 x% + nof2 y% + nof2 z% ═ 1;

step2, agent attribute setting: setting the language status of two languages, the language status of x is S _x The linguistic position of Y is S _y And S is _x +S _y 1 is ═ 1; allocating social radiuses for agents, and setting population distribution proportions of the three social radiuses as Spop%, Mpop% and Bpop%, wherein Spop% + Mpop% + Bpop% is 1; setting the daily movement proportion movescope of agent, the long-distance immigration proportions of two areas, cross score 1% and cross score 2%, the remote connection proportions of two areas, long linkpeope 1% and long linkpeope 2%, the social radius of the remote connection, long distance and time of the remote connection;

step3, generating a social network: the agent establishes a local social network by utilizing the social circle principle according to the social radius: taking the position of each individual agent on the two-dimensional plane as the center of a circle, taking the circle formed by the lengths of the social radiuses as the social circle of each individual agent, and when the length Q of a connecting line of central coordinate points of the two agents is less than or equal to the smaller social radius r of the agents at the two ends of the connecting line, indicating that the two individual agents are both positioned in the social circle of the other party and are mutually known, and establishing a connection relationship between the agents, otherwise, not establishing the connection relationship; constructing a remote cross-regional social network by using the virtual mirror image principle of an agent: constructing a virtual mirror image agent 'with a new social radius in an area II by an agent in an area I according to a set proportion, wherein the agent' takes a circle formed by the lengths of the social radii of the agent 'as a social circle, and aiming at the area where the agent' is located, when the length Q of a connecting line of two agent center coordinate points at least having one virtual mirror image is less than or equal to a smaller social radius r of agents at two ends of the connecting line, the two individual agents are both positioned in the social circle of the other party and are known mutually, a connection relationship is established between the agents, and the agents which establish the connection relationship with the virtual mirror image agent 'are connected with the original agents of the agent' in the area I through edges;

step4, horizontal and vertical propagation of individuals: performing horizontal propagation when the individual satisfies equations (1) to (5), and performing vertical propagation when equations (6) to (10) are satisfied;

P _x→x ＝1-P _x→z (2)

P _y→y ＝1-P _y→z (4)

P _z→z ＝1 (5)

in the formula, P _x→z Representing the probability, P, of a single speaker x converting to a bilingual speaker z _y→z Representing the probability, P, of a single speaker y converting to a bilingual speaker z _x→x Representing the probability, P, of a single speaker x converting to a single speaker x _y→y Representing the probability, P, of a single speaker y being converted to a single speaker y _z→z Representing the probability of converting bilingual z into bilingual z; c. C _h Is the horizontal propagation probability front coefficient, P _xy Representing attitudes of speaker x versus learning y language, P _yx Representing the attitude of the speaker y to the learning x language with a value range of [0, 1%]；S _x Representing the linguistic status of the monolingual x, S _y Representing the linguistic status of monolingual y; n is _b Represents the total number of all individuals within the social radius of agent individuals, n _m Representing the total number of all individuals with which the remote connection has been made, n _bx 、n _by And n _bz Respectively representing the number of monolingual x, monolingual y and bilingual z within the social radius of agent individual, n _mx 、n _my And n _mz Respectively representing the number of monolingual x, monolingual y and bilingual z in the individual with the remote connection relation; k. u represents the influence of bilingual z on monolingual x and monolingual y to bilingual z, and the value range is [0, 1%](ii) a w represents a remote connection weight;

P _x→x ＝1 (6)

P _y→y ＝1 (7)

P _z→z ＝1-P _z→x -P _z→y (10)

in the formula, P _z→x Representing the probability, P, of a bilingual person z converting to a monolingual person x _z→y Representing the probability of converting a bilingual person z into a monolingual person y; c. C _v Is the vertical propagation probability front coefficient; p _a Representing the attitude of learning bilingual with a value range of [0, 1%]；

Step5, dynamic property of the network: agent needs to judge whether the probability of daily movement is met or not at each time step tick, if yes, the agent moves according to the movement distance, and otherwise, the agent does not move; meanwhile, at each time point, the agent needs to judge whether the probability of long-distance trans-regional migration is met, if yes, the agent moves by adopting a moving distance larger than the social radius of the agent, and otherwise, the agent does not move; after moving, the individual agents regenerate a social network according to the social radius and the social circle connection rule, when the length of the connection between the agents is larger than the social radius of the agent with the smaller social radius in the agents which are connected with each other, the connection is disconnected, otherwise, the connection is kept;

step6, updating the social network: updating the social network again by using a social circle principle and a virtual mirror image principle, and repeatedly executing the steps from Step3 to Step5 along with the time updating until the time Step tick reaches a set value time, and stopping updating the social network;

step7, modeling and simulating the competition propagation of the language by utilizing the established social circle network: respectively plotted on the abscissa at [0, (250 × cc) -1]And [ (249 × cc), 249]Population ratio n of monolingual x, monolingual y and bilingual z in the interval _x 、n _y And n _z A track that evolves over time; drawing all monolingual x and monolingual in two-dimensional space gridPopulation ratio n of y and bilingual z _x 、n _y And n _z A track that evolves over time;

step8, parameter adjustment: and (3) repeatedly executing the steps from Step3 to Step7 by adjusting the influence of bilingual on the monolingual, the attitude of the monolingual x on the learning y language, the attitude of the learning bilingual and the influence of remote connection ratio parameter analysis on language propagation.

The invention has the beneficial effects that: the invention improves a modeling method based on agent social circle network, constructs a social network with remote weak connection relation, and can describe the strengthening trend of the social network weak relation caused by regional population flow and communication connection under the urbanization background. Besides considering the influence of the language position and the language population proportion, the language evolution mechanism of different crowds introduces a new mechanism of influence of bilingual on the conversion from monolingual to bilingual and influence of language attitude on language selection behavior. An improved agent language horizontal and vertical propagation model is provided, and the propagation model comprehensively considers factors such as language status, language density, language attitude, influence of bilingual on monolingual, remote connection weight and the like. A dynamic social network language competition model considering the language attitude and the bilingual cohesion is constructed, and the influence of the language attitude and the bilingual cohesion on the language inheritance and propagation rules under the conditions of population confusion and population convergence is simulated and analyzed. The calculation experiment result shows that: the weak language of the strong relationship social network under the traditional settlement background has slow decay speed and long storage time, but the final evolution trend of the system can not be changed; the measures of cultivating positive language attitude, improving the influence of bilingual on the conversion from monolingual to bilingual and the like can guide the strong and weak two language systems to evolve to a deep bilingual fusion state under the background of strengthening the weak connection of the social network, so as to form a stable nationwide bilingual community.

Drawings

FIG. 1 is a first diagram illustrating the relationship between individuals in a social circle according to the present invention;

FIG. 2 is a second schematic diagram of the relationship between individuals in the social circle of the present invention;

FIG. 3 is a third schematic diagram of the relationship between individuals in the social circle of the present invention;

FIG. 4 is a cross-regional remote social network of the present invention;

FIG. 5 is a schematic diagram of a language evolution model of the present invention;

FIG. 6 is a simulation flow chart of the present invention;

FIG. 7 is a trajectory diagram of the system under bilingual influence change in a living mixed condition according to the present invention;

FIG. 8 is a trajectory diagram of the system under bilingual influence change in the presence of colonization in accordance with the present invention;

FIG. 9 shows the coexistence of languages under the coexistence condition according to the present invention _xy A parameter space schematic diagram;

FIG. 10 shows language coexistence P under the aggregation condition of the present invention _xy A parameter space schematic diagram;

FIG. 11 shows the coexistence of languages under the coexistence condition according to the present invention _a A parameter space schematic diagram;

FIG. 12 is a diagram of language coexistence P under the colonization condition of the present invention _a A parameter space schematic diagram;

FIG. 13 is a diagram of dynamic regulation and analysis under the change of the weak connection ratio of the social network.

Detailed Description

Example 1: as shown in fig. 1-13, a modeling simulation method of a language competition micro simulation model based on a partitioned double-layer network, divides a two-dimensional grid into two areas in proportion, respectively creates agents in the two areas, initializes monolingual and bilingual population, sets the attributes thereof, and respectively establishes a local social network and a remote cross-regional social network by using a social circle principle and a virtual mirror principle; in the established network, agents carry out language communication through horizontal propagation and carry out language inheritance through vertical propagation; describing the dynamics of the network according to the introduced daily movement and long-distance migration; after horizontal transmission, vertical transmission, daily movement and long-distance migration, the agent updates the social network again according to the social circle principle and the virtual mirror principle, and utilizes the established social circle network to perform modeling simulation on language competition transmission; and finally, analyzing the influence on language propagation by adjusting the influence of bilingual on monolingual, the attitude of monolingual x on learning y language, the attitude of learning bilingual and the remote connection proportion parameter.

Further, the method may be set as follows:

step1, create agent and initialize monolingual and bilingual populations: creating a two-dimensional space network within the range of 250 x 241, creating nofArea1 agents in the two-dimensional space network, wherein the abscissa of the nofArea is randomly distributed in the interval of [0, (250 × cc) -1 ]; creating nofArea2 agents with the abscissa randomly distributed within the interval [ (249 × cc), 249], wherein cc belongs to [0, 1; setting a language type: monolingual x, monolingual y and bilingualr z, wherein the population proportions of monolingual x, monolingual y and bilingual z in the nofArea1 agents are nof1 x%, nof1 y% and nof1 z%, respectively, and nof1 x% + nof1 y% + nof1 z% ═ 1; the population proportions of monolingual x, monolingual y and bilingual z in nofArea2 agents were assigned to nof2 x%, nof2 y% and nof2 z%, respectively, and nof2 x% + nof2 y% + nof2 z% ═ 1;

step2, agent attribute setting: setting the language status of two languages, the language status of x is S _x The linguistic position of Y is S _y And S is _x +S _y 1 is ═ 1; allocating social radiuses for agents, and setting population distribution proportions of the three social radiuses as Spop%, Mpop% and Bpop%, wherein Spop% + Mpop% + Bpop% ═ 1; setting the daily movement proportion movescope of agent, the long-distance immigration proportions of two areas, cross score 1% and cross score 2%, the remote connection proportions of two areas, long linkpeope 1% and long linkpeope 2%, the social radius of the remote connection, long distance and time of the remote connection;

step3, generating a social network: the agent establishes a local social network by utilizing the social circle principle according to the social radius: taking the position of each individual agent on the two-dimensional plane as the center of a circle, taking the circle formed by the lengths of the social radiuses as the social circle of each individual agent, and when the length Q of a connecting line of central coordinate points of the two agents is less than or equal to the smaller social radius r of the agents at the two ends of the connecting line, indicating that the two individual agents are both positioned in the social circle of the other party and are mutually known, and establishing a connection relationship between the agents, otherwise, not establishing the connection relationship; constructing a remote cross-regional social network by using the virtual mirror image principle of an agent: constructing a virtual mirror image agent 'with a new social radius in an area II by an agent in an area I according to a set proportion, wherein the agent' takes a circle formed by the lengths of the social radii of the agent 'as a social circle, and aiming at the area where the agent' is located, when the length Q of a connecting line of two agent center coordinate points at least having one virtual mirror image is less than or equal to a smaller social radius r of agents at two ends of the connecting line, the two individual agents are both positioned in the social circle of the other party and are known mutually, a connection relationship is established between the agents, and the agents which establish the connection relationship with the virtual mirror image agent 'are connected with the original agents of the agent' in the area I through edges;

step4, horizontal and vertical propagation of individuals: performing horizontal propagation when the individual satisfies equations (1) to (5), and performing vertical propagation when equations (6) to (10) are satisfied;

P _x→x ＝1-P _x→z (2)

P _y→y ＝1-P _y→z (4)

P _z→z ＝1 (5)

in the formula, P _x→z Representing the probability, P, of a single speaker x converting to a bilingual speaker z _y→z Representing the probability, P, of a single speaker y converting to a bilingual speaker z _x→x Representing the probability, P, of a single speaker x converting to a single speaker x _y→y Representing the probability, P, of a single speaker y being converted to a single speaker y _z→z Representing the probability of converting bilingual z into bilingual z; c. C _h Is the horizontal propagation probability front coefficient, P _xy Representing attitudes of speaker x versus learning y language, P _yx Representing the attitude of the speaker y to the learning x language with a value range of [0, 1%]；S _x Representing the linguistic status of the monolingual x, S _y Representing the linguistic status of monolingual y; n is _b Represents the total number of all individuals within the social radius of agent individuals, n _m Representing the total number of all individuals with which the remote connection has been made, n _bx 、n _by And n _bz Respectively representing the number of monolingual x, monolingual y and bilingual z within the social radius of agent individual, n _mx 、n _my And n _mz Respectively representing the number of monolingual x, monolingual y and bilingual z in the individual with the remote connection relation; k. u represents the influence of bilingual z on monolingual x and monolingual y to bilingual z, and the value range is [0, 1%](ii) a w represents a remote connection weight;

P _x→x ＝1 (6)

P _y→y ＝1 (7)

P _z→z ＝1-P _z→x -P _z→y (10)

in the formula, P _z→x Representing the probability, P, of a bilingual person z converting to a monolingual person x _z→y Representing the probability of converting a bilingual person z into a monolingual person y; c. C _v Is the vertical propagation probability front coefficient; p _a Representing the attitude of learning bilingual with a value range of [0, 1%]；

Step5, dynamic property of the network: agent needs to judge whether the probability of daily movement is met or not at each time step tick, if yes, the agent moves according to the movement distance, and otherwise, the agent does not move; meanwhile, at each time point, the agent needs to judge whether the probability of long-distance trans-regional migration is met, if yes, the agent moves by adopting a moving distance larger than the social radius of the agent, and otherwise, the agent does not move; after moving, the individual agents regenerate a social network according to the social radius and the social circle connection rule, when the length of the connection between the agents is larger than the social radius of the agent with the smaller social radius in the agents which are connected with each other, the connection is disconnected, otherwise, the connection is kept;

step6, updating the social network: updating the social network again by using a social circle principle and a virtual mirror image principle, and repeatedly executing the steps from Step3 to Step5 along with the time updating until the time Step tick reaches a set value time, and stopping updating the social network;

step7, modeling and simulating the competition propagation of the language by utilizing the established social circle network: respectively plotted on the abscissa at [0, (250 × cc) -1]And [ (249 × cc), 249]Population ratio n of monolingual x, monolingual y and bilingual z in the interval _x 、n _y And n _z A track that evolves over time; plotting the population proportion n of all monolingual x, monolingual y and bilingual z in the two-dimensional space grid _x 、n _y And n _z A track that evolves over time;

step8, parameter adjustment: and (3) repeatedly executing the steps from Step3 to Step7 by adjusting the influence of bilingual on the monolingual, the attitude of the monolingual x on the learning y language, the attitude of the learning bilingual and the influence of remote connection ratio parameter analysis on language propagation.

Still further, the following experimental data are given:

the initial parameter design is shown in table 1:

TABLE 1 initial fixed parameters

Parameter(s)	Initial value	Description of the invention
			nofArea1	500	Total number of area one population
nofArea2	500	Total number of population of region two
			cc	0.5	Proportion of area one to the whole area
nof1x％	0.4	Ratio of x monolinguals in region one
			nof1y％	0.6	Ratio of y monolinguals in region one
nof2x％	0.4	Ratio of x monolinguals in region two
			nof2y％	0.6	Ratio of y monolinguals in region two
S x	0.4	Position of X language
			S y	0.6	Position of Y language
S-Reach	13	Radius of small society
			M-Reach	23	Radius of the middle society
B-Reach	33	Radius of large society
			Spop％	55	Population proportion of small social radius
Bpop％	8	Population proportion of large social radius
			k	0.15	Influence degree of bilingual person on x monolingual person
u	0.4	Influence degree of bilingual person on x monolingual person
			w	0.2	Weight of remote connection
P xy	0.65	Attitude of x monolingual to learning y language
			P yx	0.45	Attitude of y monolingual to learning x language
P a	0.7	Attitude to learning bilingual
			c h	0.2	Horizontal propagation probability front coefficient
c v	0.2	Vertical propagation probability front coefficient
			movepeople％	60	Ratio of agent daily movements
movedistance	1～6	Length of daily movement within a region
			crossmovement1％	3	Probability of each agent in the first area migrating to the second area in long distance
crossmovement2％	3	Probability of each agent in the second area migrating to the second area in long distance
			longlinkpeople1％	20	Agent proportion with long-range links (to region 2) in region 1
longlinkpeople2％	10	Agent proportion with long-range links (to region 1) in region 2
			timelonglinks	1	Long distance link every several steps
longdistance	10	Radius of social circle in remote connection
			cross-edge％	75	Probability of success of crossing between regions

First, the number, language type, social radius, location, and other attributes are set for the individuals in the network.

In the network, individuals have social communication radiuses with different lengths, and the individuals are used as centers, and the social circle is formed by rounding the social communication radiuses with the lengths as radiuses. When the length Q of the connecting line of the central points of the two bodies is less than or equal to the smaller social radius r of the bodies at the two ends of the connecting line, the bodies establish a connection relationship to generate a connection. As shown in FIG. 1, Q>r, individuals are not connected, so no association occurs, as shown in FIG. 2, Q<r, the two establish a connection so they are neighbors of each other. As shown in fig. 3, as the radius of society increases, the circle of society expands and the number of individuals having a connection relationship gradually increases. Remote cross-regional social networkAs shown in FIG. 4, a person in a certain proportion of the area can generate a connection relationship with all individuals in a certain social radius range of the area two besides generating a connection relationship with people in the social radius of the person, which is equivalent to a virtual mirror image of the individual in the area two in the area one or the other way around. In order to realize the connection, the concept of agent virtual mirror image is introduced, and the cross-regional social connection is realized by combining the method of social circle theory. The method comprises the following steps: regional agent _i There is a fixed location, agent, in zone one _i ' is agent _i The virtual mirror image in the second area has a new social radius, and agent _i ' As an individual in the second area, the social network circle method in the previous section is used to find the individuals with connection relation with the individual, and the individuals and the agent in the first area are combined _i Are connected by edges, and then the agent is obtained _i Remote social connection relationships in region two.

After the social circle network is constructed, individuals in the network are set to carry out horizontal transmission and vertical transmission of languages, and the specific explanation is as follows:

horizontal and vertical propagation of the individual as shown in fig. 5. When the individual satisfies the formulas 1 to 5, horizontal propagation is performed, and when the formulas 6 to 10 are satisfied, vertical propagation is performed.

P _x→x ＝1-P _x→z (2)

P _y→y ＝1-P _y→z (4)

P _z→z ＝1 (5)

In the formula, P _x→z Representing the probability, P, of a single speaker x converting to a bilingual speaker z _y→z Representing the probability, P, of a single speaker y converting to a bilingual speaker z _x→x Representing the probability of a single speaker x converting to a single speaker x,P _y→y representing the probability, P, of a single speaker y being converted to a single speaker y _z→z Representing the probability of converting bilingual z into bilingual z; c. C _h Is the horizontal propagation probability front coefficient, P _xy Representing attitudes of speaker x versus learning y language, P _yx Representing the attitude of the speaker y to the learning x language with a value range of [0, 1%](ii) a The higher the attitude, the more aggressive the learning of another language, and instead the more likely a behavior is to be created that excludes the learning of other languages, S _x Representing the linguistic status of the monolingual x, S _y Representing the linguistic status of monolingual y; n is _b Represents the total number of all individuals within the social radius of agent individuals, n _m Representing the total number of all individuals with which the remote connection has been made, n _bx 、n _by And n _bz Respectively representing the number of monolingual x, monolingual y and bilingual z, n _mx 、n _my And n _mz Respectively representing the number of monolingual x, monolingual y and bilingual z in the individual with the remote connection relation; k. u represents the influence of bilingual z on monolingual x and monolingual y to bilingual z, and the value range is [0, 1%](ii) a A larger value indicates a larger attraction and influence of the bilingual individual; w represents a remote connection weight;

P _x→x ＝1 (6)

P _y→y ＝1 (7)

P _z→z ＝1-P _z→x -P _z→y (10)

in the formula, P _z→x Representing the probability, P, of a bilingual person z converting to a monolingual person x _z→y Representing the probability of converting a bilingual person z into a monolingual person y; c. C _v Is the vertical propagation probability front coefficient; p _a Representing the attitude of learning bilingual with a value range of [0, 1%](ii) a The larger the value is, the larger the representation isThe more active people are on the attitudes of bilingual;

after the communication between the language horizontal transmission and the language vertical transmission is finished, the agents perform daily movement (the movement distance is movedistance and long-distance trans-regional transfer (the movement distance exceeds the social radius of the agents) from one region to another region, after the movement, the individual agents regenerate the social network according to the social radius and the social circle connection rule, when the length of the connection between the agents is longer than the social radius of the agents with the smaller social radius in the agents connected with each other, the connection is disconnected, otherwise, the connection is kept.

FIG. 6 is a simulation flow chart, first dividing a two-dimensional grid into two regions in proportion, creating agents in the two regions and setting their attributes, respectively, constructing a regional two-layer language competition network with remote connection relationship by using the social circle principle and the virtual mirror principle of the agents, wherein the language propagation of the agents in the network comprehensively considers the influence of factors such as individual language status, language density, language attitude, bilingual influence on monolingual and remote connection weight, and describes the network dynamics according to the introduced daily movement and long-distance migration. Modeling and simulating the competition propagation of the language by utilizing the established social circle network, and analyzing the dynamic characteristics of the system by adjusting the influence of bilingual on monolingual language; setting language attitude as a single variable parameter, and analyzing the influence of the change of model parameters on language coexistence possibility; and finally, verifying the influence of policy control on system evolution under the condition of weak connection relation of different proportions by the model through dynamic control.

The first step is as follows: simulating the influence of bilingual influence change on system evolution under the condition of mixed population, wherein the ratio of mixed population initial population is set to realize random distribution of monolingual x and monolingual y in the whole area, the population ratio of monolingual x in the first area and the second area is the same, and the population ratio of monolingual y is also the same, namely nof1 x%, nof2 x%, nof1 y%, nof2 y%. Discussing the final evolution state of the system when the proportion of the mixed population changes, the main simulation parameters of the system are set as follows: language status S of the disadvantaged language x _x 0.4, the language position of the dominant language y is S _y 0.6, monolingual x pairs learning the dominant languageHas a language attitude of P _xy 0.65, the language attitude of monolingual y to learning vulnerable language is P _yx 0.45, the attitude of monolingual to learning bilingual is P _a 0.4. The evolution trend of the system under the mixed population condition is analyzed by taking bilingual influence (k and u) as a change parameter, and experiments show that the evolution state of the system does not change after the time step of system evolution exceeds 2000 steps, so that the evolution end point state of the system is represented by the evolution state of the system at 2000 steps of the system.

(1) When the influence of bilingual on the monolingual persons x and y is not considered, that is, k is 0 and u is 0, as shown in fig. 7(a), the final state of the system depends on the initial population ratio of the monolingual persons x and y, and when the initial population ratio of the monolingual person x is equal to or less than 0.65, the system finally converges to (0,1,0), and only the dominant language y remains in the system. The attitude of the monolingual y for learning the x language is lower than that of the monolingual x for learning the y language, so that the monolingual x can more easily learn the monolingual y to form the bilingual. When the initial proportion range of the monolingual x is 0.65-0.7, the system finally converges to (x, y, z), the system is a three-language coexistence state, the population proportion parameter coexistence interval is small, the condition is difficult to maintain in the actual society, the state is an unstable state, and the end point state of the system evolution under the same initial condition can be shifted to a stable state with a certain probability. When the initial proportion of monolingual x is greater than 0.7, the system eventually converges to (1,0,0), leaving only the disadvantaged language x. The initial population of the disadvantaged language x is large, the agent social circle almost speaks the monolingual language x, in order to communicate more conveniently, the monolingual person y can learn the disadvantaged language x to form a bilingual person in a short time, and the monolingual person y can forget the language of the monolingual person gradually due to low attitude of learning bilingual. Therefore, the vulnerable language x can be made up for the difference in language status and remain in the case where the population ratio is dominant.

(2) When considering that both the bilingual speakers have small influence on the monolingual speakers, i.e., k is 0.15 and u is 0.4, as shown in fig. 7 (b). The final state of the system depends on the initial population proportion of the monolingual x and the monolingual y, when the initial population proportion of the monolingual x is less than or equal to 0.55, the system finally converges to (0,1,0), and only the powerful language y remains in the system. When the initial occupation ratio range of the single speaker x is 0.55-0.65, the final evolution state of the system is a three-language coexistence state. Compared with fig. 8(a), the position of the final equilibrium state of the system is changed, and the population initial proportion coexistence interval is slightly enlarged but still small. The system generates more bilinguals in the evolution process due to the improvement of k and u. When the initial proportion of monolingual x is equal to or greater than 0.65, the system eventually converges to (1,0,0), leaving only the disadvantaged language x.

(3) When considering that the influence of the bilingual speaker on the monolingual speaker x is small and the influence on the monolingual speaker y is large, that is, k is 0.15 and u is 0.85, as shown in fig. 7(c), the system eventually converges to (1,0,0) regardless of the initial population ratio of the monolingual speaker x and the monolingual speaker y, only the weak language x remains, and the time required to reach the final equilibrium state becomes shorter as the initial ratio of the monolingual speaker x becomes larger. Therefore, under the condition, the influence of bilingualrs on the conversion of the strong speakers into bilingualrs plays a decisive role in system evolution, the bilingualrs in the system increase rapidly, but the bilingualrs are not strong in stability, and are converted into monolingual x when being inherited to offspring, and only x monolingual is left at the end point of the system evolution.

(4) When considering that the influence of the bilingual speaker on the monolingual speaker x is large and the influence of the bilingual speaker on the monolingual speaker y is small, that is, k is 0.75 and u is 0.25, as shown in fig. 7(d), the system finally converges to (0,1,0) regardless of the initial population ratio of the monolingual speaker x and the monolingual speaker y, and only the dominant language y remains, and the time required to reach the final equilibrium state is longer as the initial population ratio of the monolingual speaker x is higher, for the same reason as above.

(5) When considering that the influence of bilingual on both monolingual persons x and y is relatively large, that is, k is 0.55 and u is 0.65, as shown in fig. 7(e), the system finally converges to (0,0,1) regardless of the initial population ratios of monolingual person x and monolingual person y, and the system finally assumes a full bilingual state. Bilingual people have strong attraction to the conversion from two monolingues to bilingual, the probability of converting monolingues to bilingual people is increased, although the stability of bilingual is weak, the whole system is still converted into mainstream from bilingual, the end point of system evolution is stabilized in the stable state of the whole nationality bilingual, and the bilingual agglutination force guiding system plays a leading role in the bilingual state evolution.

The second step is that: the influence of bilingual influence change on system evolution under the condition of gathering is simulated, the independent distribution of monolingual x and monolingual y in an area is realized by setting the proportion of the initial crowd of gathering, agents in an allocation area I only can be weak language x, agents in an allocation area II only can be strong language y, namely nof2 x% is 0, and nof2 y% is 1. cc represents the proportion of the area of the whole two-dimensional space grid occupied by the first region, and the value is between 0 and 1. The size of the area corresponds to the population proportion, and the population density of the area I is kept the same as that of the area II. And (5) discussing the final evolution state of the system when the initial population proportion of the colony is changed. The main simulation parameters of the system are set to be co-mingled, the evolution trend of the system under the condition of aggregation is analyzed by taking bilingual influence (k and u) as change parameters, and experiments show that the evolution state of the system does not change any more after the time step of system evolution exceeds 4000 steps, so that the evolution state of the system at the time of 4000 steps of the system represents the evolution end point state of the system.

(1) When the influence of bilingual on the monolingual persons x and y is not considered, that is, k is 0 and u is 0, as shown in fig. 8(a), the final state of the system depends on the initial population ratio of the monolingual persons x and y, and when the initial population ratio of the monolingual person x is equal to or less than 0.84, the system finally converges to (0,1,0), and only the dominant language y remains in the system. Compared with the figure 7(a), the time required by the system to reach the final state is obviously prolonged, a lot of bilingual crowds do not appear in the evolution process, the dominant language permeates from the boundary of the two areas to the weak language layer by layer, the weak language exists in a bilingual form in a short time, and the dominant language is quickly assimilated into the dominant language. When the initial population proportion of the monolingual x is within the range of 0.84-0.88, the system finally converges to (x, y, z), which is a three-language coexistence state, mainly weak languages are taken as the main, and a small amount of strong languages and bilinguals coexist. When the initial population proportion of monolingual x is higher than 0.88, the system eventually converges to (1,0,0), leaving only the disadvantaged language x. Although the language status of the vulnerable language is low, the vulnerable language can be well maintained when the initial population proportion is high. That is, the language position difference can be compensated and maintained under the condition that the population proportion of the vulnerable language x is dominant.

(2) When considering that the influence of bilingual speakers on monolingual speakers is small, that is, k is 0.15 and u is 0.4, as shown in fig. 8(b), the final state of the system depends on the initial ratio of monolingual speakers x and y, and when the initial population ratio of monolingual speaker x is equal to or less than 0.55, the system finally converges to (0,1,0), and only the dominant language y remains in the system. When the initial population proportion range of the monolingual x is 0.55-0.6, the final evolution state of the system is a three-language coexistence state, and compared with fig. 7(b), the position of the final equilibrium state of the system is changed, and the coexistence region is expanded. The system generates more bilinguals in the evolution process due to the improvement of k and u. When the initial proportion of monolingual x is greater than or equal to 0.6, the system eventually converges to (1,0,0), leaving only the disadvantaged language x.

(3) When considering that the influence of bilingual speaker on the monolingual x is small and the influence on the monolingual y is large, that is, k is 0.15 and u is 0.85, as shown in fig. 8(c), the system eventually converges to (1,0,0) regardless of the initial population ratio of monolingual x and monolingual y, only the weak language x remains, and the time required to reach the final equilibrium state becomes shorter as the initial ratio of monolingual x becomes higher. Under the condition, the influence of bilingualrs on conversion of strong speakers into bilingualrs plays a decisive role in system evolution, bilingualrs in the system increase rapidly, but bilingual stability of the bilingualrs is not strong, the bilingual bilinguals are converted into monolingual x when being inherited to filial generations, and only x monolingual is left at the end point of the system evolution. The general trend is the same as in fig. 7(c), but not as many bilinguals are produced.

(4) When considering that the influence of a bilingual on a monolingual x is large and the influence of a monolingual y is small, that is, k is 0.75 and u is 0.25, as shown in fig. 8(d), the system finally converges to (0,1,0) no matter what the initial population ratio of the monolingual x and the monolingual y is, only the dominant language y remains, the monolingual x will be completely changed into bilingual in the evolution process, then the native language of the monolingual x is slowly forgotten, and finally only the dominant language y is mastered.

The general trend is the same as in fig. 7(d), but not as many bilinguals are produced.

(5) When considering that the influence of bilingual on both monolingual persons x and y is relatively large, that is, k is 0.55 and u is 0.65, as shown in fig. 8(e), the system finally converges to (0,0,1) regardless of the initial population ratios of monolingual person x and monolingual person y, and the system finally assumes a full bilingual state. Bilingual people have strong attraction to the conversion from two monolingues to bilingual, the probability of converting monolingues to bilingual people is increased, although the stability of bilingual is weak, the whole system is still converted into mainstream from bilingual, the end point of system evolution is stabilized in the stable state of the whole nationality bilingual, and the bilingual agglutination force guiding system plays a leading role in the bilingual state evolution.

In summary, in the case of small k and u, the time required for the population to evolve to the final state is longer, the equilibrium position is changed compared with the mixed population, in the case of large k and u, the final state of the system is the same whether the population is mixed or not, and the population is equivalent to the mixed population due to the large influence of the bilingual. This indicates that the colonisation is beneficial to the preservation of the vulnerable language and prolongs its existence time, but does not twist the situation of the final extinction.

The third step: setting P _xy Being a single variable parameter, P _xy The attitude of a monolingual x learning y language is represented, and main simulation parameters of the system are set as follows: s _x ＝0.4、S _y ＝0.6、n _x ＝450、n _y ＝550、P _yx ＝0.45、P _a When the ratio is 0.4, nx refers to the ratio of the weak language monolingus, and ny refers to the ratio of the strong language monolingus; through calculation experiment, P corresponding to system language coexistence state under the condition of mixed population 2000 steps is researched _xy The parameter intervals are analyzed below in the case of different k and u.

(1) When the influence of bilingual on monolingual person x and y is not considered, i.e., k is 0 and u is 0, as shown in fig. 9(a), when 0 ≦ P _xy When the number is less than or equal to 0.25, the enthusiasm of the monolingual x for learning the y language is very low, and the y language is hardly learned; a monolingual y has a higher enthusiasm for learning x language, and will learn xThe language is changed into bilingual person z, and because the attitude to study bilingual is low, the monolingual person y can forget the language of the monolingual person gradually, and finally the system is that the population of the weak language x can occupy the whole population. When it is 0.25<P _xy <At 0.3, as the enthusiasm of monolingual x for learning y language is increased, some monolingual x will become bilingual z in a short time, and the bilingual has a short existence time in the system due to low attitude to learn bilingual, so the system is finally in a coexistence state of monolingual x and monolingual y. When P is present _xy And when the language attitude of the monolingual x for learning the y language is not particularly low, the system finally wins the dominant language y.

(2) When considering that the influence of the bilingual speaker on the monolingual speaker is small, i.e., k is 0.15 and u is 0.45, as shown in fig. 9(b), when 0 ≦ P _xy When the number of the weak language x is less than or equal to 0.45, the population of the final weak language x of the system occupies the whole population, compared with the graph (a) in fig. 9, the final weak language winning interval of the system is larger, and the weak language can be completely maintained, because the influence of the bilingual on the monolingual person y is larger than that of the monolingual person x. When it is 0.45<P _xy <At 0.5, the system is eventually co-existent in monolingual x and bilingual z. When P is present _xy When the number is more than or equal to 0.5, the system is a coexistence state of the dominant language y and the bilingual z. Compared with fig. 9(a), the system is no longer the dominant language wins, and the weak language is saved in bilingual form.

(3) When considering that the influence of bilingual on both monolingual persons x and y is relatively large, that is, k is 0.85 and u is 0.55, as shown in fig. 9(c), P is 0 ≦ P _xy And when the language attitude of the monolingual x for learning the y language is extremely low, the weak language can be won. When P is more than 0.1 _xy Below 0.2, the system ends up as a co-existence of the disadvantaged language x and bilingual language z. When P is present _xy Above 0.2, the system ends up in a full bilingual state. The bilingual has strong attraction to the conversion from two monolingues to bilingual, the probability of the conversion from monolingues to bilinguals is increased, and although the bilingual stability is weak, the whole system is still mainly converted from bilingualThe flow and bilingual agglutination force guiding system plays a leading role in the bilingual state evolution.

FIG. 9 shows a lateral contrast when k and u are 0, with P _xy The system has only two states, all the strong languages or all the weak languages, and the situation of all bilinguals can not occur. As k and u are increased, the system starts to have all bilingual situations, and the larger k and u are, the higher the proportion of bilingual population of the system is finally, and the existing interval becomes larger.

The fourth step: p corresponding to system language coexistence state for representing spatial distribution of different language crowds _xy Under the condition that the main simulation parameters of the system are not changed, the influence of the interval is calculated and tested to research the P corresponding to the system language coexistence state under the condition of 2000 steps of aggregation _xy The parameter intervals are analyzed below in the case of different k and u.

(1) When the influence of bilingual on monolingual person x and y is not considered, i.e., k is 0 and u is 0, as shown in fig. 10(a), P is 0 ≦ P _xy When the number is less than or equal to 0.1, the enthusiasm of the monolingual x for learning the y language is very low, and finally the population of the vulnerable language x occupies the whole population. When 0.1<P _xy <When 0.7, the attitude of learning the y language along with the monolingual x is improved, the system is finally in a coexistence state of three languages, and the ratio of the monolingual y in the coexistence state is the highest; when P is present _xy When the language position and the initial population proportion of the monolingual y are dominant, the attitude of the monolingual x for learning the y language is very positive, and the system is finally superior to the powerful language y when the language position and the initial population proportion of the monolingual y are more than or equal to 0.7. Compared with fig. 9(a), the interval in which languages coexist is significantly wider, which is more advantageous for maintaining vulnerable languages.

(2) When considering that the influence of the bilingual speaker on the monolingual speaker is small, i.e., k is 0.15 and u is 0.45, as shown in fig. 10(b), when 0 ≦ P _xy When the number is less than or equal to 0.4, the situation of learning the x language by the monolingual y is higher, and finally, the population of the vulnerable language x occupies the whole population. When P is present _xy The system is a coexistence state of the dominant language y and the bilingual z when the number is more than 0.4. Compared with FIG. 9(b), monolingual y is more dominant than bilingual z in the co-existence state, and with P _xy The monolingual y in the coexisting state is more.

(3) When considering that the influence of bilingual on both monolingual persons x and y is relatively large, that is, k is 0.85 and u is 0.55, as shown in fig. 10(c), P is 0 ≦ P _xy When the number is less than or equal to 0.1, the enthusiasm of the monolingual x for learning the y language is very low, and finally the population of the vulnerable language x occupies the whole population. When P is more than 0.1 _xy When the number is less than 0.2, the system is finally in a coexisting state of the weak language x and the bilingual z; when P is present _xy And when the number is more than or equal to 0.2, the system is finally in a full bilingual state. Compared with fig. 9(c), the evolution state of the system in the step 2000 is almost the same finally, the progress of the system evolution is accelerated due to the large values of k and u, namely, the influence of the bilingual is large, and the vulnerable language can be well maintained under the condition of spatial distribution (living together and gathering) of people with different languages.

As is apparent from fig. 10(a), (b), the system is still in dynamic change and has not yet fully evolved into a final state; the final evolved state-space diagram and the clutter are almost the same as shown in fig. 9. Therefore, aggregation can prolong the retention time of the disadvantaged language, but cannot change the final evolution state of the system, and about 4000 steps are required for reaching the final evolution state.

The fifth step: when the parent is bilingual individual, the attitude change of the people to the learning bilingual can influence the probability that the child inherits the bilingual of the parent, and P is set _a For a single variable parameter, the main simulation parameters of the system are set as follows: s _x ＝0.4、S _y ＝0.6、n _x ＝450、n _y ＝550、P _xy ＝0.6、P _yx When the system language coexistence status is 0.45, P corresponding to the system language coexistence status under the condition of the mixed population 2000 steps is researched through calculation experiments _a The parameter intervals are analyzed below in the case of different k and u.

(1) When the influence of bilingual on monolingual person x and y is not considered, i.e., k is 0 and u is 0, as shown in fig. 11(a), when 0 ≦ P _a When the language is less than or equal to 0.9, the final dominant language of the system can occupy all population, along with the improvement of the attitude of people to learning bilingual, the probability of the bilingual individuals of the parent to the bilingual of the child is increased, but because the bilingual does not influence the monolingual, the weak language is influenced by the position of the dominant language y and the peopleThe influence of the mouth density is assimilated by the strong language. When 0.9<P _a <At 0.95, the system is finally a co-existence of monolingual y and bilingual z. When P is present _a When the bilingual learning rate is more than or equal to 0.95, namely the attitude of people for learning bilinguals is very positive, the probability that the bilingual individuals of the parent are inherited to the bilingual of the child is very high, and finally, the bilingual population of the system occupies the whole population.

(2) When considering that the influence of the bilingual speaker on the monolingual speaker is small, i.e., k is 0.15 and u is 0.45, as shown in fig. 11(b), when 0 ≦ P _a And when the learning attitude is less than or equal to 0.3, the system finally enables the monolingual person y to occupy the whole population under the condition of not being positive to the bilingual learning attitude. The single speakers can learn each other to become bilingual speakers, the probability of bilingual speakers is very low when the bilingual speakers are inherited to children, the bilingual speakers hardly exist for a long time, and the weak languages are gradually assimilated by the strong languages. When P is more than 0.3 _a Below 0.45, the system is a co-existence of the dominant language y and bilingual z. When P is present _a When the language learning tendency is more than or equal to 0.45, the probability that the parent bilingual individual is inherited to the child bilingual is higher along with the improvement of the attitude of people on learning bilingual, and finally the bilingual population of the system occupies the whole population. Compared with fig. 11(a), the final bilingual winning interval in the system is larger as k and u increase.

(3) When considering that the influence of bilingual on both monolingual persons x and y is relatively large, that is, k is 0.85 and u is 0.55, as shown in fig. 11(c), P is 0 ≦ P _a When the number is less than or equal to 0.25, the system finally occupies the whole population by the dominant language y population, and the interval is very small. When P is more than 0.25 _a Below 0.3, the system is eventually a co-existence of the dominant language y and bilingual z. When P is present _a At more than or equal to 0.3, the system is finally in a full bilingual state, because bilingualrs have strong attraction to the conversion from two monolingues to bilinguals, and even if the attitude to the bilinguals is not particularly high, the whole system is still converted into a mainstream from the bilinguals.

FIG. 11 is a horizontal comparison, in the case that people are active in learning bilingual attitudes, the larger k and u are, the larger the interval of the system which is finally all bilinguals is; when people are not active in learning bilingual attitudes, no matter how large k and u are, the system finally wins the dominant language y.

And a sixth step: to body inP corresponding to system language coexistence state by space distribution of people with different languages _a Under the condition that the main simulation parameters of the system are not changed, the influence of the interval is calculated and tested to research the P corresponding to the system language coexistence state under the condition of 2000 steps of aggregation _a The parameter intervals are analyzed below in the case of different k and u.

(1) When the influence of bilingual on monolingual person x and y is not considered, i.e., k is 0 and u is 0, as shown in fig. 12(a), P is 0 ≦ P _a At ≦ 0.5, the system will eventually be dominated by all people. When P is present _a When the number is more than 0.5, the system is finally in a coexisting state of monolingual y and bilingual z, and the proportion of bilingual z in the coexisting state is higher as the attitude of people on learning bilingual is improved. Compared with fig. 11(a), the final evolution state of the system does not have the situation of all bilinguals, and the coexistence interval of the languages is obviously widened, which is more beneficial to the maintenance of the weak languages.

(2) When considering that the influence of the bilingual speaker on the monolingual speaker is small, i.e., k is 0.15 and u is 0.45, as shown in fig. 12(b), when 0 ≦ P _a When the number is less than or equal to 0.7, the system is a coexistence state of three languages, compared with fig. 11(b), the situation that the dominant language is won does not occur in the final evolution state of the system, the situation of language coexistence occurs, and the language coexistence interval is very wide. When P is present _a When the situation is more than 0.7, with the improvement of the positive attitude of people on learning bilingual, the system finally wins the bilingual, the interval of the bilingual wins becomes narrow, and higher bilingual attitude is needed compared with the mixed living situation.

(3) When considering that the influence of bilingual on both monolingual persons x and y is relatively large, that is, k is 0.85 and u is 0.55, as shown in fig. 12(c), P is 0 ≦ P _a At ≦ 0.25, the system will eventually be dominated by all people. When P is more than 0.25 _a Below 0.4, the system ends up in a co-existence state of the dominant language y and bilingual z. When P is present _a And when the number is more than or equal to 0.4, the system is finally in a full bilingual state. Compared with the figure 11(c), the evolution state of the system in the 2000 step is almost the same finally, the influence of bilingual is great due to the great k and u, the process of the system evolution is accelerated, and the spatial distribution (mixed population and congregation population) of the weak language in different language populations is accelerated) Can be well maintained in all cases.

As is evident from fig. 12(a), (b), the system is still in dynamic change and has not yet fully evolved into a final state; the state space diagram and the mixed population which are finally evolved by the clustering are almost the same, as shown in fig. 11, and only the time is prolonged, and about 4000 steps are needed. The settlement is an effective measure for language protection, and although the tendency of weak language death cannot be finally changed, the decay process is obviously prolonged, and more time and opportunities are provided for people to protect the language.

The seventh step: verifying the influence of policy regulation and control on system evolution under the condition of weak connection relation of different proportions of the model, analyzing by taking the graph 8(a) as a reference, and discussing the condition that n is the number of the elements _x ＝0.4、n _y With an initial population ratio of 0.6, when the ratio of both the disadvantaged speaker x and the bilingual speaker z is less than thirty percent, the influence of the bilingual speaker is increased (k is 0.5 and u is 0.4) and the language attitude (P) is changed _xy ＝0.5，P _a 0.5) to see the change of the future evolution trend of the system, as shown in fig. 13 below. The experiment shows that: (1) under the condition of no policy regulation, the weak language and bilingual in the system will die quickly, and as the proportion of weak connection in the social network increases (from 10% to 30%), the speed of die-out is increased, and finally the time of die-out is advanced from about 1200 steps to 800 steps. This means that when the cross-regional remote flow and communication link become normal, the social relationship range of people is enlarged and the strength is weakened, which is not beneficial to the preservation of the vulnerable language. (2) After policy control is performed by increasing the influence of bilingualrs and changing the language attitude (i.e., k is 0.5, u is 0.4, and P is _xy ＝0.5，P _a 0.5), when the system is operated to 2000 steps, the total ratio of the monolingual x and the bilingual z is 21% under the condition that the weak link ratio in the social network is 10%, and the total ratio of the monolingual x and the bilingual z is increased to 35% when the weak link ratio is increased to 30%. With the increase of the connection proportion in the weak social network, the total proportion of the monolingual x and the bilingual z in the final evolution state of the system can be increased, which shows that the regulation and control measures are easier to achieve effect under the condition of high weak connection proportion. To sum upThe policy regulation and control measures can play a role in protecting the vulnerable language in the social network with weak connection, so that the vulnerable language can be stored in a bilingual form, and the trend of the future vulnerable language extinction can be prevented and improved.

While the present invention has been described in detail with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, and various changes can be made without departing from the spirit of the present invention within the knowledge of those skilled in the art.

Claims (2)

1. A language competition microscopic simulation model modeling simulation method based on a partition double-layer network is characterized by comprising the following steps: dividing the two-dimensional grid into two areas according to a proportion, respectively creating agents in the two areas, initializing monolingual and bilingual population, setting the attributes of the monolingual and bilingual population, and respectively establishing a local social network and a remote cross-regional social network by utilizing a social circle principle and a virtual mirror image principle; in the established network, agents carry out language communication through horizontal propagation and carry out language inheritance through vertical propagation; describing the dynamics of the network according to the introduced daily movement and long-distance migration; after horizontal transmission, vertical transmission, daily movement and long-distance migration, the agent updates the social network again according to the social circle principle and the virtual mirror principle, and utilizes the established social circle network to perform modeling simulation on language competition transmission; finally, the influence of bilingual on monolingual influence, the attitude of monolingual x on learning y language, the attitude of learning bilingual and the influence of remote connection proportion parameter analysis on language propagation are adjusted;

and (3) generating a social network: the agent establishes a local social network by utilizing the social circle principle according to the social radius: taking the position of each individual agent on the two-dimensional plane as the center of a circle, taking the circle formed by the lengths of the social radiuses as the social circle of each individual agent, and when the length Q of a connecting line of central coordinate points of the two agents is less than or equal to the smaller social radius r of the agents at the two ends of the connecting line, indicating that the two individual agents are both positioned in the social circle of the other party and are mutually known, and establishing a connection relationship between the agents, otherwise, not establishing the connection relationship; constructing a remote cross-regional social network by using the virtual mirror image principle of an agent: constructing a virtual mirror image agent 'with a new social radius in an area II by an agent in an area I according to a set proportion, wherein the agent' takes a circle formed by the lengths of the social radii of the agent 'as a social circle, and aiming at the area where the agent' is located, when the length Q of a connecting line of two agent center coordinate points at least having one virtual mirror image is less than or equal to a smaller social radius r of agents at two ends of the connecting line, the two individual agents are both positioned in the social circle of the other party and are known mutually, a connection relationship is established between the agents, and the agents which establish the connection relationship with the virtual mirror image agent 'are connected with the original agents of the agent' in the area I through edges;

horizontal and vertical propagation of individuals: performing horizontal propagation when the individual satisfies equations (1) to (5), and performing vertical propagation when equations (6) to (10) are satisfied;

P _x→x ＝1-P _x→z (2)

P _y→y ＝1-P _y→z (4)

P _z→z ＝1 (5)

in the formula, P _x→z Representing the probability, P, of a single speaker x converting to a bilingual speaker z _y→z Representing the probability, P, of a single speaker y converting to a bilingual speaker z _x→x Representing the probability, P, of a single speaker x converting to a single speaker x _y→y Representing the probability, P, of a single speaker y being converted to a single speaker y _z→z Representing the probability of converting bilingual z into bilingual z; c. C _h Is the horizontal propagation probability front coefficient, P _xy Representing attitudes of speaker x versus learning y language, P _yx Representing the attitude of the single speaker y to learning x language with a value range of [0,1]；S _x Representing the linguistic status of the monolingual x, S _y Representing the linguistic status of monolingual y; n is _b Represents the total number of all individuals within the social radius of agent individuals, n _m Representing the total number of all individuals with which the remote connection has been made, n _bx 、n _by And n _bz Respectively representing the number of monolingual x, monolingual y and bilingual z within the social radius of agent individual, n _mx 、n _my And n _mz Respectively representing the number of monolingual x, monolingual y and bilingual z in the individual with the remote connection relation; k. u represents the influence of bilingual z on monolingual x and monolingual y to bilingual z, and the value range is [0, 1%](ii) a w represents a remote connection weight;

P _x→x ＝1 (6)

P _y→y ＝1 (7)

P _z→z ＝1-P _z→x -P _z→y (10)

in the formula, P _z→x Representing the probability, P, of a bilingual person z converting to a monolingual person x _z→y Representing the probability of converting a bilingual person z into a monolingual person y; c. C _v Is the vertical propagation probability front coefficient; p _a Representing the attitude to learning bilingual with a value range of [0, 1%]。

2. The modeling simulation method of the language competition micro simulation model based on the partitioned two-layer network according to claim 1, wherein: the method comprises the following specific steps:

step1, create agent and initialize monolingual and bilingual populations: creating a two-dimensional space network within the range of 250 x 241, creating nofArea1 agents in the two-dimensional space network, wherein the abscissa of the nofArea is randomly distributed in the interval of [0, (250 × cc) -1 ]; creating 2 agents with abscissa randomly distributed within the interval [ (249 × cc), 249], wherein cc belongs to [0,1 ]; setting a language type: monolingual x, monolingual y and bilingualr z, wherein the population proportions of monolingual x, monolingual y and bilingual z in the nofArea1 agents are nof1 x%, nof1 y% and nof1 z%, respectively, and nof1 x% + nof1 y% + nof1 z% ═ 1; the population proportions of monolingual x, monolingual y and bilingual z in nofArea2 agents were assigned to nof2 x%, nof2 y% and nof2 z%, respectively, and nof2 x% + nof2 y% + nof2 z% ═ 1;

step2, agent attribute setting: setting the language status of two languages, the language status of x is S _x The linguistic position of Y is S _y And S is _x +S _y 1 is ═ 1; allocating social radiuses for agents, and setting population distribution proportions of the three social radiuses as Spop%, Mpop% and Bpop%, wherein Spop% + Mpop% + Bpop% is 1; setting the daily movement proportion movescope of agent, the long-distance immigration proportions of two areas, cross score 1% and cross score 2%, the remote connection proportions of two areas, long linkpeope 1% and long linkpeope 2%, the social radius of the remote connection, long distance and time of the remote connection;

step3, generating the social network;

step4, performing horizontal propagation and vertical propagation of the individual;

step5, dynamic property of the network: agent needs to judge whether the probability of daily movement is met or not at each time step tick, if yes, the agent moves according to the movement distance, and otherwise, the agent does not move; meanwhile, at each time point, the agent needs to judge whether the probability of long-distance trans-regional migration is met, if yes, the agent moves by adopting a moving distance larger than the social radius of the agent, and otherwise, the agent does not move; after moving, the individual agents regenerate a social network according to the social radius and the social circle connection rule, when the length of the connection between the agents is larger than the social radius of the agent with the smaller social radius in the agents which are connected with each other, the connection is disconnected, otherwise, the connection is kept;

step6, updating the social network: updating the social network again by using a social circle principle and a virtual mirror image principle, and repeatedly executing the steps from Step3 to Step5 along with the time updating until the time Step tick reaches a set value time, and stopping updating the social network;

step7, modeling and simulating the competition propagation of the language by utilizing the established social circle network: respectively plotted on the abscissa at [0, (250 × cc) -1]And [ (249 × cc), 249]Population ratio n of monolingual x, monolingual y and bilingual z in the interval _x 、n _y And n _z A track that evolves over time; plotting the population proportion n of all monolingual x, monolingual y and bilingual z in the two-dimensional space grid _x 、n _y And n _z A track that evolves over time;

step8, parameter adjustment: and (3) repeatedly executing the steps from Step3 to Step7 by adjusting the influence of bilingual on the monolingual, the attitude of the monolingual x on the learning y language, the attitude of the learning bilingual and the influence of remote connection ratio parameter analysis on language propagation.

Images