Data service network site selection method in fuzzy environment

Abstract

The invention discloses a data service network site selection method in a fuzzy environment. The method comprises the following steps: acquiring a data service network in the fuzzy environment; the data service network comprises an intelligent community, a data analysis station, a communication base station and a data service center; minimizing data transmission consumption in the data service network as an optimization target of data service network site selection, and setting constraint conditions for the optimization target to form a data service network site selection optimization model; converting the data service network site selection optimization model to obtain a converted optimization model; and determining an optimality condition of the converted optimization model, and solving the converted optimization model by adopting a neurodynamics algorithm to obtain an optimal location of the data service network site selection. According to the method, the optimal data service network site selection position can be effectively found, and the communication load of the data service network in the fuzzy environment is improved.

Description

Technical Field

[0001] The invention relates to the field of data service networks, and in particular to a data service network site selection method under a fuzzy environment. Background Art

[0002] As the demand for data resources in smart communities increases, communication base stations need to be established to store data resources and provide feedback to the data service center on the efficiency of data resource utilization in smart communities. Data resources are sent from the data service center and sent to the smart community through communication base stations. After receiving feedback from the communication base stations, the data service center organizes and updates the data resources so that the smart community can use data resources more effectively.

[0003] In order to control the risks of data transmission and use within a safe range, the data analysis station always monitors and evaluates the transmission and feedback of data. In the data service network, the spatial distance between the smart community, communication base station, data analysis station and data service center is extremely important for saving the transmission consumption of data resources. At present, the actual data information transmission process is carried out in a fuzzy and uncertain environment. Due to the interference of the external environment and the complexity and diversity of the transmission medium, it is impossible to effectively determine the optimal address in the data service network by relying solely on a clear problem model.

[0004] Therefore, a new data service network site selection method is needed to effectively find the optimal data service network site selection location and improve the communication load of the data service network in a fuzzy environment. Summary of the invention

[0005] In view of this, the purpose of the present invention is to overcome the defects in the prior art and provide a data service network site selection method under a fuzzy environment, which can effectively find the optimal data service network site selection location and improve the communication load of the data service network under the fuzzy environment.

[0006] The data service network site selection method under fuzzy environment of the present invention comprises:

[0007] Acquire a data service network in a fuzzy environment; the data service network includes a smart community, a data analysis station, a communication base station and a data service center;

[0008] Minimizing data transmission consumption in the data service network is the optimization goal of data service network site selection, and setting constraints for the optimization goal to form a data service network site selection optimization model;

[0009] Transforming the data service network location optimization model to obtain a transformed optimization model;

[0010] The optimality conditions of the converted optimization model are determined, and the converted optimization model is solved using a neurodynamic algorithm to obtain the optimal location for data service network site selection.

[0011] Furthermore, the data service network location optimization model is determined according to the following formula:

[0012]

[0013]

[0014]

[0015] in, Represents the position coordinates of the entity, Represents the column vector of the position coordinates of other entities. The cost function of the entity is a fuzzy convex mapping And it can be represented by a family of interval-valued functions:

[0016] To portray, and They are the fuzzy cost functions The left and right functions of is a convex set.

[0017] Furthermore, the fuzzy cost function is determined according to the following formula:

[0018]

[0019] in, is the fuzzy transmission cost per unit distance between the th entity and the th entity Since the transmission medium on the same interconnection link is the same, Since the data service center cannot be directly connected to the smart community, hour,

[0020] Furthermore, the data service network location optimization model is transformed, including:

[0021] The dual-objective optimization model is set as follows:

[0022]

[0023]

[0024] in, remember is the Pareto optimal solution set of the bi-objective optimization model. If there exists a fuzzy cost function A certain α-cut of makes a feasible location plan Then the feasible site selection plan x is an optimal site selection plan for the data service network site selection optimization model;

[0025] Convert the bi-objective optimization model to a weighted scalar optimization model:

[0026]

[0027]

[0028] Among them, θ1, θ2 are scalar weights that satisfy θ1+θ2=1 and θ1, θ2≥0; an optimal solution of the weighted scalar optimization model corresponds to a Pareto optimal solution of the dual-objective optimization model;

[0029] For a fixed α∈[0,1], the coupling constraint equation of the data service network location optimization model can be rewritten as:

[0030]

[0031] in,

[0032]

[0033]

[0034] Through the conversion of the weighted scalar optimization model, the converted optimization model is:

[0035]

[0036]

[0037]

[0038] Among them, They are a collection of indicators for data service center, communication base station, smart community and data analysis station.

[0039] Furthermore, the optimality condition of the converted optimization model is:

[0040] like is an optimal location scheme for the data service network location optimization model under fuzzy environment if and only if for any and, existence and So that:

[0041]

[0042]

[0043] in, Represents projection to a collection The projection operator.

[0044] Furthermore, the converted optimization model is solved using a neural dynamics algorithm, including:

[0045] Construct a non-autonomous generalized gradient neural dynamics algorithm model based on derivative feedback mechanism:

[0046]

[0047]

[0048]

[0049]

[0050]

[0051] in, represents the index set of neighbor entities connected to the th entity, represents the symbolic function defined in a component-wise manner, Respectively represent the fuzzy set valued function based on a certain -cut set The left and right side functions of ;

[0052] Solve to obtain an equilibrium state of the neurodynamic algorithm model:

[0053] When the neural dynamics algorithm model reaches the equilibrium state, the optimal solution of the converted optimization model is obtained. When the converted optimization model achieves the optimal solution, the heterogeneous locations of each entity in the data service network are set as the optimal locations for data service network site selection.

[0054] The beneficial effects of the present invention are as follows: the present invention discloses a method for site selection of a data service network in a fuzzy environment, which takes into account the influence of external conditions and transmission media on the interconnection and transmission of the data service network, and introduces a non-smooth cost function and a coupling constraint equation based on a cut-off set. In view of the fact that the solution to the fuzzy optimization problem is difficult to describe accurately, the concepts of interval value optimization problem solution, Pareto optimal solution of the dual-objective optimization problem and weighted scalar optimization problem solution are introduced to transform the problem. By constructing a class of generalized gradient neural dynamics algorithms based on the differential inclusion framework, the site selection and positioning problem of the data service network in a fuzzy environment is effectively handled, which reduces the construction and transmission consumption of site selection and positioning in real network scenarios, and is conducive to protecting the privacy information of each data node. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0056] Figure 1 A schematic diagram of the network site selection method flow of the present invention;

[0057] Figure 2 It is a principle block diagram of the generalized nonlinear circuit of the present invention;

[0058] Figure 3 A schematic diagram of the communication topology of a data service network according to an embodiment of the present invention;

[0059] Figure 4 It is a schematic diagram of the evolution results of the location coordinates of the data service center and the communication base station of the present invention;

[0060] Figure 5 A schematic diagram of the detailed outline of the state trajectory for finding the optimal position coordinates of the present invention. DETAILED DESCRIPTION

[0061] The present invention is further described below in conjunction with the accompanying drawings, as shown in the drawings:

[0062] This embodiment discloses a method for selecting a data service network site in a fuzzy environment, including the following steps:

[0063] Acquire a data service network in a fuzzy environment; the data service network includes a smart community, a data analysis station, a communication base station and a data service center;

[0064] Minimizing data transmission consumption in the data service network is the optimization goal of data service network site selection, and setting constraints for the optimization goal to form a data service network site selection optimization model;

[0065] Transforming the data service network location optimization model to obtain a transformed optimization model;

[0066] The optimality conditions of the converted optimization model are determined, and the converted optimization model is solved using a neurodynamic algorithm to obtain the optimal location for data service network site selection.

[0067] In this embodiment, a data service network in a fuzzy environment consisting of a smart community, a data analysis station, a communication base station and a data service center is considered, and a unified method is used to mark each entity (smart community, data analysis station, communication base station and data service center).

[0068] remember They are the indicator sets of data service center, communication base station, smart community and data analysis station, namely

[0069] Smart communities, communication base stations, and data analysis stations are located on the same plane, and their locations can be expressed using plane coordinates. To depict, The data service center is not located on the same plane as the smart community, communication base station, and data analysis station. Its location can be described by plane coordinates and vertical coordinates, that is, Therefore, the data service network location selection problem is a nonlinear heterogeneous network optimization problem.

[0070] In this embodiment, by fixing the location of the data analysis station and the smart community, considering that each data service center is not connected to any smart community, the goal is to find the optimal location for the data service center and the communication base station to minimize the corresponding data transmission consumption. Considering the uncertainty of the data demand of the smart community, the location of all data service centers and communication base stations must satisfy an affine coupling equation constraint based on the α-cut set. At the same time, since the data analysis station maintains real-time monitoring of data traffic, the location of the data service center and the communication base station should satisfy their respective local constraints.

[0071] Assume that the cost function of each data service center and communication base station is Part of the cost function comes from the interconnection transmission with the data analysis station, and the other part comes from the interconnection transmission with other entities (data service center, communication base station, smart community). The former part is a deterministic value, and the latter part is a fuzzy value due to the uncertainty of the transmission medium. Considering the impact of external conditions on the interconnection transmission of the entire data service network, the cost function of each data service center and communication base station may be non-smooth.

[0072] Constraint equations and local constraints based on α-cut sets (α∈[0,1]) are established. Affected by the uncertainty of data demand in smart communities, the locations of all data service centers and communication base stations must satisfy an affine coupling constraint, which is described by the α-cut set (α∈[0,1]) equation of fuzzy matrices and fuzzy vectors. In addition, considering that the data analysis station maintains real-time monitoring of data traffic, the location of the data service center and communication base station should satisfy their respective local constraints.

[0073] remember They are the fuzzy matrices based on α-cut sets The left and right matrices of ; They are the fuzzy matrices based on α-cut sets Considering the impact of the uncertainty of smart community data demand on the positioning of data service centers and communication base stations, the following constraint equation is introduced:

[0074]

[0075] remember but represents a coupled constraint set, where

[0076] remember is the local constraint set of the ith entity, then represents a set of constraints that all entities must satisfy. is the set of constraints that the i-th entity needs to satisfy, is the set of constraints that all entities must satisfy.

[0077] Considering the non-smooth fuzzy optimization objectives and corresponding constraints, the following data service network location optimization model is constructed:

[0078]

[0079]

[0080]

[0081] in, represents the position coordinates of the i-th entity, Represents the column vector of the position coordinates of other entities. The cost function of the i-th entity is a fuzzy convex mapping And it can be represented by a family of interval-valued functions:

[0082] To portray, and They are the fuzzy cost functions The left and right functions of is a convex set.

[0083] In this embodiment, during the data monitoring, analysis and evaluation process, a relatively accurate short-term load is defined as L(ψ(x))=L0+ιψ(x), where L0 is the basic load and ι is the aggregation coefficient. ψ(x) represents the following nonlinear aggregation function:

[0084]

[0085] in represents the index set of neighbor entities connected to the jth entity, o j 、p j and q j They respectively represent the number of data service centers, communication base stations and smart communities interconnected with the j-th entity.

[0086] remember is the fuzzy transmission cost per unit distance between the i-th entity and the j-th entity Since the transmission medium on the same interconnection link is the same, Therefore, for each data service center and communication base station, the specific form of the corresponding cost function is as follows:

[0087]

[0088] Note hour, This is because the data service center cannot be directly connected to the smart community. The cost function of an entity is determined not only by its own position, but also by the positions of other entities.

[0089] In this embodiment, considering that the optimal solution or approximate optimal solution of the fuzzy optimization problem is not easy to characterize, and the obtained solution may be a fuzzy value, it is not conducive to scientific research on the actual data service network site selection problem. Therefore, it is necessary to perform appropriate equivalent transformation on the basic data service network site selection optimization model constructed in the fuzzy environment.

[0090] Based on the partial order relation theory of interval values, the minimization problem of fuzzy functions on feasible sets can be transformed into a bi-objective optimization problem. In view of the data service network location optimization model under the fuzzy environment, the following equivalent bi-objective optimization model is constructed for each entity:

[0091]

[0092]

[0093] in remember is the Pareto optimal solution set of the bi-objective optimization model. If there exists a fuzzy cost function A certain α-cut of makes a feasible location plan Then the feasible site selection plan x is the optimal site selection plan of the site selection optimization model.

[0094] Since the Pareto optimal solution set of the dual-objective optimization model evolves into The above conversion method makes the optimal solution of the location optimization model still possible to be a fuzzy value. A universal method for finding the Pareto optimal solution of the bi-objective optimization model is to convert it into the following weighted scalar optimization model:

[0095]

[0096]

[0097] Where θ1, θ2 are scalar weights that satisfy θ1+θ2=1 and θ1, θ2≥0. An optimal solution of the weighted scalar optimization model corresponds to a Pareto optimal solution of the dual-objective optimization model.

[0098] For a fixed α∈[0,1], the coupling constraint equation in the location optimization model can be rewritten as:

[0099]

[0100] in,

[0101]

[0102] Accordingly, through the conversion idea of ​​the weighted scalar optimization model, the basic data service network location optimization model under fuzzy environment can be converted into the following constrained optimization model:

[0103]

[0104]

[0105]

[0106] In this embodiment, the optimality condition of the converted optimization model is:

[0107] like is an optimal location scheme for the location optimization model of basic data service network under fuzzy environment if and only if for any and, existence and So that:

[0108]

[0109] in, Represents projection to a collection The projection operator.

[0110] In this embodiment, based on the established optimality conditions, a non-autonomous generalized gradient neural dynamics algorithm is constructed with the help of non-smooth analysis and differential inclusion theory. Taking the entity as an example, its corresponding differential dynamics model is as follows:

[0111]

[0112] Among them, it can be expressed by the generalized gradient of the cost function and the constraint equation.

[0113] Referring to all entities that need to be located in the data service network, the following nonlinear aggregation function is constructed:

[0114]

[0115] in, Represents the local contribution of the entity to the aggregation function.

[0116] Auxiliary function that introduces a fuzzy value for each entity Correspondingly, the local blur cost function can be described as

[0117] Taking the th entity as an example, the following neural dynamics system model based on the differential inclusion framework and with derivative feedback mechanism is constructed:

[0118]

[0119] in, represents the index set of neighbor entities connected to the th entity, represents the symbolic function defined in a component-wise manner, Respectively represent the fuzzy set valued function based on a certain -cut set The left and right side functions.

[0120] Referring to the above design, It can be seen that:

[0121]

[0122] and It can be regarded as a local approximation of the optimal solution of the site selection optimization model.

[0123] In the designed neural dynamics algorithm, an equilibrium state in the projection dynamics system (12a) corresponds to an optimal location in the location optimization model. The role of the projection operator is to drive the algorithm's neuron state to approach the corresponding local constraint set. The dynamics model of the variables in (12b) and the variables in (12c) drives the algorithm's neuron state to approach the corresponding affine coupling constraint set. The consensus protocol is based on the variables The derivative feedback mechanism designed in (12b) is established The approximation of the aggregation function in (11) helps to reach a consensus agreement. The dynamic model of the variables in (12d) is used to estimate the nonlinear aggregation function.

[0124] Reference Figure 2 As shown in the figure, the neural dynamics model designed by simulation in the generalized nonlinear programming circuit is shown in the figure. i , i , η i It can be represented by the input and output of the circuit system, the generalized gradient and Can be generated by electronic components, α and β can be input by external scalars. Projection operator It can be realized by activation function. The whole neural dynamics network model is composed of a series of analog components such as adder, integrator, multiplier, activation function, etc.

[0125] Define the initial neuron state x of the neural dynamics system (12a)-(12e) i (0). If but is a positive invariant set of the dynamical system (12a); conversely, the neuron state x in the neural dynamical system (12a)-(12e) i (t) will approach the set exponentially

[0126] For the neuron states x of the neurodynamic system (12a)-(12e) i (t), exists and So that:

[0127]

[0128] in,

[0129] Let y i (t) = e t x i (t), then:

[0130]

[0131] Integrating both ends of (15) from 0 to t simultaneously, we obtain:

[0132]

[0133] in, because and is a convex set, so if Then for any t∈(0,∞) we have On the contrary, if Then the neuron state x i (t) exponentially converges to the set

[0134] Aggregation of state vectors of neural dynamical systems (12a)-(12e).

[0135] remember Algorithms (12a)-(12e) can be written in the form of vector aggregation as follows:

[0136]

[0137] in,

[0138]

[0139] L N represents the weighted Laplacian matrix, I r represents the r×r identity matrix.

[0140] similar:

[0141]

[0142] Then we have:

[0143]

[0144] so, The equilibrium state of the neural dynamics algorithm (17) is equivalent to:

[0145]

[0146] Among them, 0 s represents an s-dimensional column vector whose entries are all zero.

[0147] The relationship between the equilibrium state of the neurodynamic algorithm (17) and the optimal solution of the site selection optimization model is established.

[0148] If (x * , * , η * , ρ * ) is an equilibrium state of the neural dynamics algorithm (17), then x * is an optimal solution to the location optimization model. Therefore, if x* is an optimal solution to the location optimization model, then there exists Such that (x*, (*, η*, ρ*) is an equilibrium state of the neural dynamics algorithm (17).

[0149] Assuming that (x*, (*, η*, ρ*) is an equilibrium state of the neural dynamics algorithm (17), then from (19c), we can see that there exists Make Here 1 N represents an N-dimensional column vector whose values ​​are all 1. That is, for any two entities i and j, Then from (19b) we can get:

[0150]

[0151] Based on Kronecker product sum We can get:

[0152]

[0153] Combining (21) and (20) we can derive (19b). From (18), we know that there exists So that:

[0154]

[0155] This shows that (9a) holds and

[0156] in, Therefore, it can be seen that x* is an optimal solution to the site selection optimization model.

[0157] Assuming that x* is an optimal solution of the location optimization model, then for any entity in the data service network, (9a) and (9b) hold.

[0158] Based on (9a), (19a) can be established. (19c) can be obtained by Get. Introduce a column vector From (9b), we can see that:

[0159]

[0160] Kronecker Exportable:

[0161] This means About Space Performing orthogonal decomposition, we can know Existence So that (19b) holds, where and Respectively represent matrices The kernel and image of . For a time-invariant undirected connected network, if Then in a limited time there is And the upper bound of convergence time is

[0162] Therefore, for any two entities i and j, Thus (19d) is established. Therefore, (x * , * , η * , ρ * ) is an equilibrium state of the neural dynamics algorithm (17).

[0163] Among them, the boundedness analysis of the state trajectory of the neural dynamics algorithm (17) is as follows:

[0164] Starting from any initial state (x(0), ζ(0), η(0)), the state trajectory of the algorithm (x(t), ζ(t), η(t)) is bounded.

[0165] Construct the following Lyapunov function:

[0166] V(x,ζ,η)=V1(x)+V2(ζ,η) (24)

[0167]

[0168] because It is a fuzzy convex mapping, which can be derived from the properties of convex functions:

[0169]

[0170] This shows that V(x, ζ, η) is globally positive semidefinite, and V(x, ζ, η) = 0 if and only if (x, ζ, η) = (x * , * , η * ). At the same time, It can be seen that the Lyapunov function V is a radially unbounded function of x, ζ, and η.

[0171] Based on the construction of V1 and V2, it can be calculated that the Lyapunov function V is the generalized gradient of x, ζ, η, that is,

[0172] Therefore, the state trajectory of the neural dynamics algorithm (17) can be derived as:

[0173]

[0174] According to the convexity of the fuzzy objective function and the weighted Laplacian matrix L N The positive definiteness of can be derived as:

[0175]

[0176] Therefore, the state trajectory of the algorithm (x(t), ((t), η(t)) is bounded and globally existent.

[0177] In addition, the convergence analysis of the state solution of the neural dynamics algorithm (17) is as follows:

[0178] Based on the state trajectory (x, ζ, η) of the algorithm, the following set-valued function is constructed:

[0179]

[0180] At the same time, the following collection is introduced:

[0181]

[0182] Therefore, if any state Then there is inf μ′ Υ(x,ζ,η)=0.

[0183] Further analysis of the state (x′, ζ′, η′) shows that the neuron state x′ corresponds to an optimal solution in the site selection optimization model.

[0184] Since the state trajectory of the algorithm (x(t), ζ(t), η(t)) is bounded for any t∈[0,∞), there exists a sequence (x(t τ )},{ζ(t τ )}, {η(t τ )} and its corresponding limit point So that:

[0185]

[0186] Among them, {t τ} is a strictly monotonically increasing sequence. So, is the ω-limit point about (x, ζ, η).

[0187] Based on Lebesgue measure theory, we know that the limit point It corresponds to an optimal solution in the site selection optimization model. Referring to (24)-(26), x * , * , η * Replace with New Lyapunov functions U, U1, and U2 can be obtained. By calculating the derivative of the Lyapunov function U with respect to time t and further analyzing it, we can obtain:

[0188]

[0189] Therefore, the state solution of the neurodynamic algorithm converges asymptotically to the optimal position of the entity in the data service network under the fuzzy environment.

[0190] The non-autonomous generalized gradient neural dynamics algorithm with derivative feedback mechanism designed by the present invention is used for the location selection problem of data service network in fuzzy environment to verify the feasibility and effectiveness of the algorithm model:

[0191] Taking a data service network with 1 data service center, 2 communication base stations, 4 smart communities, and 1 data analysis station as an example, the corresponding network topology is as follows: Figure 3 shown.

[0192] The neural dynamics algorithm (17) is used to find the optimal addressing scheme for the data service network. The algorithm-related parameters are set to,,, and the relevant parameters of each entity in the data service network are shown in Table 1.

[0193] Table 1

[0194]

[0195] Figure 4 (a) Figure 4 (c) Figure 4 (e) shows the evolution of the location coordinates of the data service center and the communication base station driven by the neural dynamics algorithm (17) based on different cut-set and weight combinations. Figure 4 (b) Figure 4 (d) Figure 4 (f) shows the actual optimal location of each entity, where the pentagon is the data service center, the diamond is the communication base station, the triangle is the data analysis station, and the square is the smart community.

[0196] for Figure 4 of Figure 5 (a), (b), and (c) show the search details of the data service center and the communication base station for their optimal location coordinates driven by the neural dynamics algorithm designed by the present invention, which is an optimal combination of the heterogeneous location distribution of each entity in the data service network. It can be seen that the non-autonomous neural dynamics algorithm with derivative feedback mechanism proposed by the present invention can effectively find a set of optimal location solutions for smart communities, data analysis stations, communication base stations, and data service centers.

[0197] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A method for selecting a data service network site in a fuzzy environment, characterized in that: The steps include: Acquire a data service network in a fuzzy environment; the data service network includes a smart community, a data analysis station, a communication base station and a data service center; Minimizing data transmission consumption in the data service network is the optimization goal of data service network site selection, and setting constraints for the optimization goal to form a data service network site selection optimization model; Transforming the data service network location optimization model to obtain a transformed optimization model; The optimality conditions of the converted optimization model are determined, and the converted optimization model is solved using a neurodynamic algorithm to obtain the optimal location for data service network site selection.

2. The method for selecting a data service network site under a fuzzy environment according to claim 1, characterized in that: The data service network location optimization model is determined according to the following formula: in, represents the position coordinates of the i-th entity, Represents the column vector of the position coordinates of other entities. The cost function of the i-th entity is a fuzzy convex mapping And it can be represented by a family of interval-valued functions: To portray, and They are the fuzzy cost functions The left and right functions of is a convex set.

3. The method for selecting a data service network site under a fuzzy environment according to claim 2, characterized in that: The fuzzy cost function is determined according to the following formula: in, is the fuzzy transmission cost per unit distance between the i-th entity and the j-th entity Since the transmission medium on the same interconnection link is the same, Since the data service center cannot be directly connected to the smart community, hour, 4. The method for selecting a data service network site under a fuzzy environment according to claim 1, characterized in that: Transform the data service network site selection optimization model, including: The dual-objective optimization model is set as follows: in, remember is the Pareto optimal solution set of the bi-objective optimization model. If there exists a fuzzy cost function A certain α-cut of makes a feasible location plan Then the feasible site selection plan x is an optimal site selection plan for the data service network site selection optimization model; Convert the bi-objective optimization model to a weighted scalar optimization model: Among them, θ1, θ2 are scalar weights that satisfy θ1+θ2=1 and θ1, θ2≥0; an optimal solution of the weighted scalar optimization model corresponds to a Pareto optimal solution of the dual-objective optimization model; For a fixed α∈[0,1], the coupling constraint equation of the data service network location optimization model can be rewritten as: in, Through the conversion of the weighted scalar optimization model, the converted optimization model is: Among them, They are a collection of indicators for data service center, communication base station, smart community and data analysis station.

5. The method for selecting a data service network site under a fuzzy environment according to claim 1, characterized in that: The optimality condition of the converted optimization model is: like is an optimal location scheme for the data service network location optimization model under fuzzy environment if and only if for any and α∈[0,1], there exists and So that: in, Represents projection to a collection The projection operator.

6. The method for selecting a data service network site under a fuzzy environment according to claim 1, characterized in that: The converted optimization model is solved using a neural dynamics algorithm, including: Construct a non-autonomous generalized gradient neural dynamics algorithm model based on derivative feedback mechanism: in, represents the index set of neighbor entities connected to the ith entity, sgn(·) represents a symbolic function defined in a component-wise manner, They represent the fuzzy set-valued functions based on a certain α-cut set. The left and right side functions of ; Solve to obtain an equilibrium state of the neurodynamic algorithm model: When the neural dynamics algorithm model reaches this equilibrium state, the obtained x * As the optimal solution of the converted optimization model, when the converted optimization model obtains the optimal solution, the heterogeneous locations of each entity in the data service network are set as the optimal locations for data service network site selection.

Images