CN116366459A - Dual-mode service flow distribution method - Google Patents

Abstract

The invention relates to a dual-mode service flow distribution method, which belongs to the field of user information acquisition systems and comprises the following steps: s1: different utility functions are designed according to the characteristics of different types of services, and utility values of each network attribute to different services are calculated; s2: calculating the weight of each network attribute; s3: calculating the relative closeness of candidate networks by using a TOPSIS algorithm based on relative entropy improvement; s4: and calculating the comprehensive score of the candidate network by combining the relative closeness and the service preference value, and selecting the network with the highest score for service flow distribution.

Description

Dual-mode service flow distribution method

Technical Field

The invention belongs to the field of user information acquisition systems, and relates to a dual-mode service flow distribution method.

Background

In the electricity consumption information acquisition system, a local communication system uses a wired communication technology and a wireless communication technology, wherein the wired communication technology comprises the following steps: narrowband power line carrier communication technology, broadband power line carrier communication technology, and the like; the wireless communication technology comprises the following steps: 230M wireless communication technology, narrowband micropower wireless communication technology, broadband micropower wireless communication technology, etc.; among them, the broadband power line carrier communication technology and the broadband micro-power wireless communication technology can provide higher transmission bandwidth and faster transmission rate, so as to meet the increasing demands of the power consumer electricity consumption information acquisition system, and have become the mainstream technology of development and application.

Broadband power line carrier communication (Broadband Power Line Communication, BPLC) technology, commonly referred to as a BPLC access network, has been widely used for local communication of low-voltage distribution networks to realize automatic meter reading service, and BPLC is particularly suitable for communication between the internet of things and intelligent home systems, and has a wide development prospect. The BPLC communication technology can directly utilize the power line without rewiring, and has the advantages of simple and quick networking, low cost and wide application range. As shown in fig. 1, a typical network topology provided by a national grid company in the technical specification of high-speed carrier communication interconnection of a voltage power line.

Micropower Wireless communication is classified into narrowband micropower Wireless communication and Broadband micropower Wireless (BMPW) communication. The narrow-band micropower wireless communication technology is mature and commercially available, but has low effective communication rate and limited practical application scene because the channel bandwidth is smaller than 100kHz, and the application range of the narrow-band micropower wireless communication technology in a national power grid is limited at present. Accordingly, in order to solve the defects of the narrowband micropower communication technology, the wideband micropower communication technology has developed, and at present, there are various schemes in the developed wideband micropower communication technology, one representative scheme is that a physical layer adopts a hybrid high-order modulation technology based on a chirp technology and a psk technology, a protocol stack adopts a communication protocol similar to a high-speed carrier communication technology, and a point-to-point communication rate of up to 3.84Mbps is realized.

For the electricity consumption information acquisition system, the broadband power line carrier communication and the broadband micro-power wireless communication have similar network architecture, namely: a tree network with a central coordinator (central coordinator, CCO) as a center and a proxy coordinator (proxy coordinator, PCO) as a relay proxy, connecting All Stations (STAs) in a multi-level association, the network level supporting a maximum of 16 levels.

The distribution and utilization network has complex topological structure, various distribution and utilization equipment, large quantity and wide coverage range, complex application scene, various service demands and high transmission reliability requirements, and the communication demands of the intelligent power grid cannot be completely met by singly adopting any communication mode of HPLC and BMP. The dual-mode communication system integrating the HPLC and the BMPW can realize the complementary advantages of the two, eliminate communication blind spots, enlarge communication coverage, meet the urgent requirements of the smart grid on high-performance and high-reliability communication, and is a key content to be researched. And traffic flow allocation is an important link in researching dual-mode communication systems.

The current dual-mode system model is that the HPLC and BMPW modes are independently networked to form two independent tree networks. The two networks have the characteristics of higher HPLC transmission bandwidth and lower time delay, but have great influence on communication quality and reliability due to the load characteristic of the power line; and BMPW networking is flexible, and the reliability is higher as the number of nodes increases and the service transmission paths increase. The design divides the services into control services, data acquisition services, event reporting services and equipment upgrading services, selects network attributes such as time delay, packet loss rate, bandwidth, hop count and the like, and distributes service flows under double modes according to the characteristics of different services. A schematic diagram of the system model is shown in fig. 2.

The current dual-mode service flow distribution method cannot meet the service quality requirements of various services, ignores the influence of service characteristics and service preferences on the service flow distribution process, has lower rationality and accuracy of service flow distribution, and has more times of switching the services among different networks.

Disclosure of Invention

In view of the above, the present invention is directed to a dual-mode service flow allocation method combining a multi-attribute decision algorithm and a utility theory.

In order to achieve the above purpose, the present invention provides the following technical solutions:

a dual mode traffic flow allocation method comprising the steps of:

s1: different utility functions are designed according to the characteristics of different types of services, and utility values of each network attribute to different services are calculated;

s2: calculating the weight of each network attribute;

s3: calculating the relative closeness of candidate networks by using a TOPSIS algorithm based on relative entropy improvement;

s4: and calculating the comprehensive score of the candidate network by combining the relative closeness and the service preference value, and selecting the network with the highest score for service flow distribution.

In step S1, the services are divided into four types, namely a control type, a data acquisition type, an event reporting type and an equipment upgrading type; the network attributes comprise time delay, packet loss rate, bandwidth and hop count; wherein:

s11: for the time delay utility functions of four service types, S-shaped function representation is adopted:

where x represents the average delay of the network, u (x) represents the delay utility value, and a and b represent constants.

S12: for the utility functions of the packet loss rates of four service types, the utility functions are expressed by linear functions:

u(x)＝gx+h

wherein x represents the average packet loss rate of the network, u (x) represents the utility value of the packet loss rate, and g and h represent constants.

S13: and for the bandwidth utility functions of the control class service and the event reporting class service, adopting an exponential function to express:

u(x)＝1-e ^-cx or u(x)＝e ^(x)-c

where x represents the average bandwidth of the network, u (x) represents the bandwidth utility value, and c represents a constant.

For bandwidth utility functions of data acquisition type service and equipment upgrading type service, S-shaped function representation is adopted;

s14: and for the hop count utility functions of the four service types, linear functions are adopted for representation.

Further, the calculating the weight of each network attribute in step S2 includes calculating subjective and objective weights of the network attributes by using Fuzzy hierarchy analysis (Fuzzy-Analytic Hierarchy Process, FAHP) and algorithm (Criteria Importance Through Intercrieria Correlation, CRITIC) on index correlation, respectively, modeling based on minimizing the deviation of the subjective and objective attribute weights, calculating the subjective and objective comprehensive weights of the network attributes, and calculating the preference of the service for different candidate networks by using the FAHP algorithm.

Further, the subjective weight of the network attribute is calculated by adopting the FAHP method, and the method specifically comprises the following steps:

step 1: dividing an analysis object into a scheme layer, an index layer and a target layer from bottom to top in sequence; the scheme layer comprises all candidate networks for distributing the service flow, the index layer comprises all network attributes, and the target layer refers to the optimal network for distributing the service flow;

step 2: performing pairwise comparison on the importance of each attribute of the index layer according to the service type; by r _ij Representing element x _i Relative to element x _j Importance degree of (1) and r at the same time _ij The matrix is also a component of a fuzzy consistency matrix, and the consistency of the matrix is judged by the following formula:

where n represents the number of network attributes considered, r _ii Representing element x _i Relative to element x _i Importance degree of r _ji Representing element x _j Relative to element x _i Importance degree of r _ik Representing element x _i Relative to element x _k Importance degree of r _jk Representing element x _j Relative to element x _k Is of importance;

step 3: calculating subjective weight of each network attribute:

thus, the subjective attribute weight vector α= (α) of n networks is obtained ₁ ,α ₂ ,…,α _n )。

Further, the algorithm CRITIC of the index correlation in step S2 calculates an objective weight of the network attribute, and specifically includes the following steps:

step 1: constructing a parameter matrix, namely constructing a decision matrix X= (X) by assuming that m candidate networks and n network attributes are total _ij ) _m×n Composition, x _ij A j-th attribute value representing network i;

step 2: different standardization modes are adopted for the positive indexes and the negative indexes to carry out parameter standardization;

for the forward index:

for negative going index:

step 3: calculating standard deviation of the j-th attribute:

step 4: calculating the information amount of the j-th attribute:

step 5: calculating objective weights of all the attributes:

obtaining objective attribute weight vectors beta= (beta) of n networks ₁ ,β ₂ ,…,β _n )。

Further, the comprehensive weight calculation step in step S2 is as follows:

assuming that there are m candidate networks in total, n network attributes, a decision matrix y= (Y) is constructed _ij ) _m×n Composition, y _ij A j-th attribute value representing network i;

defining subjective weight coefficient and objective weight coefficient in the combined weight as t and tau respectively, wherein t+tau is more than or equal to 0, and t+tau=1, and then the combined weight:

w _j ＝tα _j +τβ _j (j＝1,2,…,n)

subjective attribute weight for the jth attribute of the ith network is denoted tα _j y _ij Objective attribute weight is τβ _j y _ij The distance of the subjective and objective attribute weights is expressed as:

wherein Z is _i Representing the weight distance of the objective and subjective attribute of the ith network;

and establishing a weight combination optimization model by taking the minimized subjective and objective attribute weight distance as an objective function:

wherein, minZ represents the minimum subjective and objective attribute weight distance;

and (3) optimizing a model solving process:

establishing Lagrange functions:

wherein lambda represents Lagrange multiplier, to obtain

The three formulas are combined to obtain weight coefficients t and tau, which are respectively:

further, the calculating the preference of the service for different candidate networks by using the FAHP algorithm in step S2 specifically includes the following steps:

step 1: calculating importance ranking of the index layer to the target layer, and obtaining weight relations of different services and network attributes according to the calculated weight of the index layer relative to the target layer;

step 2: the importance ranking of the index layer by the calculation scheme layer is performed, the importance of each candidate network with respect to the network attribute is compared, a fuzzy consistent matrix and corresponding weight are obtained, and then the weight relation between different network attributes and the candidate network is obtained;

step 3: calculating importance ranking of the scheme layer relative to the target layer;

multiplying the weight relation matrix of different services and network attributes by the relation matrix between the network attributes and the candidate networks, and calculating the sequence of a scheme layer relative to a target layer, namely the preference value of the services to the different candidate networks.

Further, the step S3 specifically includes the following steps:

s31: let w be _j Is the weight value of the j-th attribute, and the utility matrix Y= (Y) _ij ) _m×n The weight vector w= [ w ] for each column and the column ₁ ,w ₂ ,…,w _n ] ^T The weights in the matrix are multiplied one by one to obtain a weighted utility matrix Z= (Z) _ij ) _m×n ：

S32: determining an ideal network

And negative ideal network->

S33: calculating the relative entropy distance between each candidate network and the positive ideal network, wherein the relative entropy distance between the candidate network and the positive ideal network is as follows

The relative entropy distance from the negative ideal network is +.>

S34: calculating relative closeness:

further, the step S4 specifically includes:

calculating to obtain a business preference value T _i Combining the business preference value and the relative closeness to be recorded as the comprehensive score R of the candidate network _i ：

In the method, in the process of the invention,

for adjusting the factors, adjusting the proportion of the relative closeness to the business preference value according to business requirements; r is R _i And the candidate networks are ordered according to the magnitude of the comprehensive score as the judging criterion of the optimal network for the service flow distribution, and the optimal network with the large comprehensive score as the service flow distribution is selected.

The invention has the beneficial effects that:

first: the utility function is designed according to the characteristics of different services, characteristics and preferences of the different services are considered, the services can be distributed in the most suitable communication network, the QoS requirements of various services are met, and the overall service quality of the services is improved.

Second,: the attribute weighting method is optimized, and the two methods are combined aiming at the defects of the traditional subjective weighting method and the objective weighting method, and then the subjective and objective comprehensive weight is calculated by modeling with the minimized subjective and objective attribute weight distance, so that the weight setting is more reasonable.

Third,: the method improves the traditional network sequencing method, aims at the problems of abnormal sequencing and low accuracy of the traditional TOPSIS algorithm, improves the relative entropy, improves the accuracy of network selection, reduces the switching times of services among different networks, and reduces the influence caused by the ping-pong effect.

Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objects and other advantages of the invention may be realized and obtained by means of the instrumentalities and combinations particularly pointed out in the specification.

Drawings

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in the following preferred detail with reference to the accompanying drawings, in which:

fig. 1 is a topology of a broadband carrier communication network;

FIG. 2 is a schematic diagram of an HPLC-BMPW dual mode system model;

FIG. 3 is a flow chart of a dual mode service flow allocation method according to the present invention;

fig. 4 is a view of the FAHP hierarchy.

Detailed Description

Other advantages and effects of the present invention will become apparent to those skilled in the art from the following disclosure, which describes the embodiments of the present invention with reference to specific examples. The invention may be practiced or carried out in other embodiments that depart from the specific details, and the details of the present description may be modified or varied from the spirit and scope of the present invention. It should be noted that the illustrations provided in the following embodiments merely illustrate the basic idea of the present invention by way of illustration, and the following embodiments and features in the embodiments may be combined with each other without conflict.

Wherein the drawings are for illustrative purposes only and are shown in schematic, non-physical, and not intended to limit the invention; for the purpose of better illustrating embodiments of the invention, certain elements of the drawings may be omitted, enlarged or reduced and do not represent the size of the actual product; it will be appreciated by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.

The same or similar reference numbers in the drawings of embodiments of the invention correspond to the same or similar components; in the description of the present invention, it should be understood that, if there are terms such as "upper", "lower", "left", "right", "front", "rear", etc., that indicate an azimuth or a positional relationship based on the azimuth or the positional relationship shown in the drawings, it is only for convenience of describing the present invention and simplifying the description, but not for indicating or suggesting that the referred device or element must have a specific azimuth, be constructed and operated in a specific azimuth, so that the terms describing the positional relationship in the drawings are merely for exemplary illustration and should not be construed as limiting the present invention, and that the specific meaning of the above terms may be understood by those of ordinary skill in the art according to the specific circumstances.

The invention provides a dual-mode service flow distribution method combining a multi-attribute decision algorithm and a utility theory according to actual engineering requirements, wherein a service flow distribution method schematic diagram is shown in fig. 3.

Firstly, dividing the services into four types of control types, data acquisition types, event reporting types and equipment upgrading types, designing different utility functions according to the characteristics of different types of services, calculating utility values of each network attribute to different services, and fully considering the characteristics and QoS requirements of different services; and aiming at different service characteristics, proper utility functions are designed to quantify the satisfaction degree of the service on the network attribute. Common utility function types are: linear functions, exponential functions, logarithmic functions, sigmoid functions, etc., as shown in equations (1) through (5).

S-shaped function:

exponential function:

u(x)＝1-e ^-cx or u(x)＝e ^(x)-c (2)

logarithmic function:

u(x)＝d+eln(x+f) (3)

linear function:

u(x)＝gx+h (4)

linear piecewise function:

time delay utility function: firstly, the control service belongs to real-time service, the delay requirement is strict, and once the service cannot be completed within a specified time limit, system faults and breakdown can be caused; secondly, the event reporting service and the equipment upgrading service have relatively loose time delay requirements, do not require strict time synchronization, and have relatively larger time delay variation range; finally, the data acquisition service is the most relaxed time delay requirement in the four services. Through the analysis, the time delay utility function adopts an S-shaped function, and the time delay utility function settings of various services are shown in table 1.

TABLE 1

Packet loss rate utility function: the packet loss rate refers to the proportion of the number of lost data packets in the transmission process to the total number of data packets. The four services of the control service, the data acquisition service, the event reporting service and the equipment upgrading service have certain tolerance to the packet loss rate, but the satisfaction degree can be continuously reduced along with the increase of the packet loss rate. Because the tolerance of the four services to the packet loss rate is different, the lowest packet loss rate of the event reporting service is assumed to be higher than 1%, the lowest satisfaction is achieved when the packet loss rate of the control service and the data acquisition service is higher than 5%, and the lowest satisfaction is achieved when the packet loss rate of the equipment upgrading service is higher than 10%. Therefore, the utility function of the packet loss rate is expressed by adopting a linear function, and the utility function of the packet loss rate of various services is set as shown in table 2.

TABLE 2

Bandwidth utility function: for equipment upgrading service, a larger transmission bandwidth is required because of larger transmission data volume, the data acquisition service is smaller than the data acquisition service, the bandwidth requirement is relatively loose, the minimum bandwidth threshold value of the two services needs to be met, and meanwhile, the threshold value of the data acquisition service is lower than that of the equipment upgrading service, so that the S-shaped function is suitable to be used as a utility function. The control service and the event reporting service do not need to set the lowest bandwidth threshold value, and the utility value and the bandwidth are in positive correlation. The bandwidth requirement of the equipment upgrading service is assumed to be 1Mbps, and the bandwidth requirement of the data acquisition service is assumed to be 0.5Mbps. The bandwidth utility function settings for the various services are shown in table 3.

TABLE 3 Table 3

Hop count utility function: the requirements of four kinds of services, namely control service, data acquisition service, event reporting service and equipment upgrading service, on the hop count are consistent, and the service satisfaction degree can be reduced along with the increase of the hop count. In the project, the maximum number of node hops cannot exceed 15 hops, so the invention sets that the utility value is the lowest when the number of hops exceeds 15 hops. The hop count utility function settings for various services using a linear function as the utility function for hop count are shown in table 4.

TABLE 4 Table 4

Then, the invention improves the traditional attribute weighting method and provides a comprehensive weighting method combining subjective and objective weights. The subjective weighting method determines the network attribute weight according to various service demands and subjective experience of a decision maker, but ignores the actual attribute characteristics of the network; although the objective weighting method can accurately reflect the characteristics of the network attributes, the preference of a decision maker for the attributes cannot be accurately reflected. The invention firstly adopts a Fuzzy-hierarchical analysis method (Fuzzy-Analytic Hierarchy Process, FAHP) and an algorithm (Criteria Importance Through Intercrieria Correlation, CRITIC) related to index correlation to respectively calculate subjective and objective weights of network attributes, calculates subjective and objective comprehensive weights of the network attributes based on modeling of minimized subjective and objective attribute weight deviation, and calculates the preference of the service to different candidate networks by using the FAHP algorithm.

FAHP calculates subjective weights: the FAHP algorithm has a certain improvement compared with the AHP algorithm, particularly in the aspect of consistency, the FAHP algorithm is based on consistency pairwise comparison matrixes, and the consistency of the matrixes is ensured when the matrixes are built. The section calculates subjective weights of network attributes using a fuzzy consistent matrix based FAHP algorithm. The method comprises the following specific steps:

step 1: in order to make the analysis of the relationships between the network attributes clear, it is necessary to divide the analysis object into a solution layer, an index layer, and a target layer in order from bottom to top. The scheme layer comprises all candidate networks for traffic flow distribution, the index layer comprises various network attributes, and the target layer refers to the optimal network for traffic flow distribution. The hierarchical structure is shown in fig. 4.

Step 2: because different communication services have different requirements on network attributes, the importance of each attribute of the index layer needs to be compared pairwise according to the service type. r is (r) _ij Representing element x _i Relative to element x _j Importance degree of (1) and r at the same time _ij Also components of the fuzzy identity matrix, the scale for importance is shown in Table 5. According to the characteristic that the difference value of any two rows of corresponding elements of the fuzzy consistency matrix is a constant value, and meanwhile, the difference value of any row and any other rows of the matrix is specified to be a constant value, the consistency of the matrix can be judged through a formula (6).

Step 3: and (5) calculating the subjective weight of each network attribute by using a formula (7).

Thus, the subjective attribute weight vector α= (α) of n networks is obtained ₁ ,α ₂ ,…,α _n )。

TABLE 5

The invention sets fuzzy consistent matrixes for four kinds of service respectively according to the characteristics of the four kinds of service and QoS requirements as shown in tables 6 to 9, and calculates attribute weights.

TABLE 6

TABLE 7

TABLE 8

TABLE 9

CRITIC calculates objective weights: since the FAHP algorithm is inherently dependent on subjective experience and judgment, it is not possible to obtain a weight of an objective and accurate network attribute. Therefore, the CRITIC algorithm is introduced without referring to subjective willingness, and objective weights of network attributes are obtained through calculation, so that the situation that the weights are set too subjectively is avoided. The method mainly comprises the following steps:

step 1: constructing a parameter matrix

Assuming that there are m candidate networks in total, n network attributes, a decision matrix x= (X) is constructed _ij ) _m×n Composition, x _ij Represents the j-th attribute value of network i.

Step 2: parameter normalization

Because the dimensions of the network attribute parameters are different, the network attribute parameters need to be standardized for facilitating subsequent calculation, and different standardized modes are adopted for the positive indexes and the negative indexes.

As for the forward direction index,

for a negative-going index of the way,

step 3: calculating standard deviation of jth attribute

Step 4: calculating information amount of jth attribute

For facilitating subsequent calculation, the normalized parameters are still x _ij And (3) representing.

Step 5: calculating objective weight of each attribute

Obtaining objective attribute weight vectors beta= (beta) of n networks ₁ ,β ₂ ,…,β _n )。

Calculating comprehensive weight: and calculating the subjective and objective comprehensive weight by adopting the minimized subjective and objective attribute weight distance modeling. The method mainly comprises the following steps:

assuming that there are m candidate networks in total, n network attributes, a decision matrix y= (Y) is constructed _ij ) _m×n Composition, y _ij Represents the j-th attribute value of network i.

Defining subjective weight coefficient and objective weight coefficient in the combined weight as t and tau respectively, wherein t+tau is more than or equal to 0, and t+tau=1, and then the combined weight:

w _j ＝tα _j +τβ _j (j＝1,2,…,n) (13)

subjective attribute weight for the jth attribute of the ith network is denoted tα _j y _ij Objective attribute weight is τβ _j y _ij The distance of its subjective and objective attribute weight can be expressed as:

wherein Z is _i And representing the objective attribute weight distance of the ith network. Obviously Z _i The smaller the subjective and objective attribute weights of the evaluation object i tend to be uniform. And establishing a weight combination optimization model by taking the minimized subjective and objective attribute weight distance as an objective function:

where minZ represents the minimum subjective and objective attribute weight distance. The following is the optimization model solving process.

Establishing Lagrange functions:

wherein lambda represents Lagrange multiplier, to obtain

The three formulas are combined to obtain weight coefficients t and tau which are respectively

The invention considers the preference degree of different services to candidate networks in the service flow distribution process, and calculates the service preference by using FAHP. The method mainly comprises the following steps:

step 1: calculating importance ranking of index layers to target layers

From the weights of the index layers with respect to the target layers that have been calculated, the weight relationships of different services to network attributes can be obtained as shown in table 10.

Table 10

Step 2: importance ordering of calculation scheme layer to index layer

And comparing the importance of each candidate network with respect to the network attribute to obtain a fuzzy consistency matrix and corresponding weight, and then obtaining the weight relation between different network attributes and the candidate network. The results are shown in tables 11 and 15.

TABLE 11

Table 12

/>

TABLE 13

TABLE 14

TABLE 15

The importance ranking of the scheme layer with respect to the target layer is calculated by multiplying the table 10 matrix and the table 15 matrix in step 1, and the ranking of the scheme layer with respect to the target layer, i.e. the preference values of the service for different candidate networks are shown in table 16.

Table 16

The scheme calculates utility functions, attribute weights and service preference values respectively, and prepares for implementation of the dual-mode service flow distribution method.

The method comprises the following specific implementation steps:

1. assume a total of m candidate networks, nNetwork attribute, constructing a decision matrix x= (X) _ij ) _m×n Composition, x _ij Represents the j-th attribute value of network i.

2. Determining utility functions for different attributes of different traffic types

3. Constructing utility matrix

Substituting the attribute values in the decision matrix of the step 1 into the utility function determined in the step 2, and obtaining a utility matrix Y= (Y) through a formula (21) _ij ) _m×n 。

y _ij ＝y _j (x _ij ) (21)

Wherein y is _j (x) And the utility function corresponding to the attribute j.

4. Calculating a weighted utility matrix

Let w be _j Is the weight value of the j-th attribute, and the utility matrix Y= (Y) obtained in the step 3 is obtained _ij ) _m×n The weight vector w= [ w ] for each column and the column ₁ ,w ₂ ,…,w _n ] ^T The weights in (2) are multiplied one by one to obtain a weighted utility matrix Z= (Z) through a formula (22) _ij ) _m×n 。

5. Determining an ideal network

And negative ideal network->

6. Calculating the relative entropy distance between each candidate network and the positive and negative ideal networks

The TOPSIS algorithm can accurately reflect the gap between various candidate networks and is a classical sorting algorithm. The best network is obtained by calculating Euclidean distances between various candidate networks and positive and negative ideal networks, and the candidate network with the nearest distance to the positive ideal network and the farthest distance to the negative ideal network is obtained. The TOPSIS algorithm is not without drawbacks, however, due to the euclidean distance adopted, ordering anomalies occur when the candidate distances are the same from the positive and negative ideal networks. To ameliorate this problem, the present invention employs relative entropy to calculate the distance between candidate networks.

The relative entropy distance between the candidate network and the positive ideal network is

The relative entropy distance from the negative ideal network is +.>

7. Calculating relative closeness

8. Ranking candidate networks in combination with relative closeness and traffic preference values

Calculating from formula (28) to obtain service preference value T _i Combining the business preference value and the relative closeness to be recorded as the comprehensive score R of the candidate network _i 。

In the method, in the process of the invention,

and adjusting the ratio of the relative closeness to the business preference value according to the business requirement for adjusting the factor. Herein +.>

0.5, R _i And the candidate networks are ordered according to the magnitude of the comprehensive score as the judging criterion of the optimal network for the service flow distribution, and the optimal network with the large comprehensive score as the service flow distribution is selected.

Finally, it is noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made thereto without departing from the spirit and scope of the present invention, which is intended to be covered by the claims of the present invention.

Claims (9)

1. A dual-mode service flow distribution method is characterized in that: the method comprises the following steps:

s1: different utility functions are designed according to the characteristics of different types of services, and utility values of each network attribute to different services are calculated;

s2: calculating the weight of each network attribute;

s3: calculating the relative closeness of candidate networks by using a TOPSIS algorithm based on relative entropy improvement;

s4: and calculating the comprehensive score of the candidate network by combining the relative closeness and the service preference value, and selecting the network with the highest score for service flow distribution.

2. The dual mode traffic flow assignment method as recited in claim 1, wherein: in step S1, the service is divided into four types, namely a control type, a data acquisition type, an event reporting type and an equipment upgrading type; the network attributes comprise time delay, packet loss rate, bandwidth and hop count; wherein:

s11: for the time delay utility functions of four service types, S-shaped function representation is adopted:

wherein x represents the average delay of the network, u (x) represents the delay utility value, and a and b represent constants;

s12: for the utility functions of the packet loss rates of four service types, the utility functions are expressed by linear functions:

u(x)＝gx+h

wherein x represents the average packet loss rate of the network, u (x) represents the utility value of the packet loss rate, and g and h represent constants;

s13: and for the bandwidth utility functions of the control class service and the event reporting class service, adopting an exponential function to express:

u(x)＝1-e ^-cx or u(x)＝e ^(x)-c

wherein x represents the average bandwidth of the network, u (x) represents the bandwidth utility value, and c represents a constant;

for bandwidth utility functions of data acquisition type service and equipment upgrading type service, S-shaped function representation is adopted;

s14: and for the hop count utility functions of the four service types, linear functions are adopted for representation.

3. The dual mode traffic flow assignment method as recited in claim 1, wherein: and step S2, calculating the weight of each network attribute, namely respectively calculating the subjective weight and the objective weight of the network attribute by adopting a fuzzy analytic hierarchy process FAHP and an algorithm CRITIC related to index correlation, modeling based on the minimized subjective and objective attribute weight deviation degree, solving the subjective and objective comprehensive weight of the network attribute, and calculating the preference of the service to different candidate networks by utilizing the FAHP algorithm.

4. The dual mode traffic flow assignment method according to claim 3, wherein: the subjective weight of the network attribute is calculated by adopting a fuzzy analytic hierarchy process FAHP, and the method specifically comprises the following steps of:

step 1: dividing an analysis object into a scheme layer, an index layer and a target layer from bottom to top in sequence; the scheme layer comprises all candidate networks for distributing the service flow, the index layer comprises all network attributes, and the target layer refers to the optimal network for distributing the service flow;

step 2: performing pairwise comparison on the importance of each attribute of the index layer according to the service type; by r _ij Representing element x _i Relative to element x _j Importance degree of (1) and r at the same time _ij The matrix is also a component of a fuzzy consistency matrix, and the consistency of the matrix is judged by the following formula:

where n represents the number of network attributes considered, r _ii Representing element x _i Relative to element x _i Importance degree of r _ji Representing element x _j Relative to element x _i Importance degree of r _ik Representing element x _i Relative to element x _k Importance degree of r _jk Representing element x _j Relative to element x _k Is of importance;

step 3: calculating subjective weight of each network attribute:

thus, the subjective attribute weight vector α= (α) of n networks is obtained ₁ ,α ₂ ,…,α _n )。

5. The dual mode traffic flow assignment method according to claim 3, wherein: the algorithm CRITIC of the index correlation in step S2 calculates an objective weight of the network attribute, and specifically includes the following steps:

step 1: constructing a parameter matrix, namely constructing a decision matrix X= (X) by assuming that m candidate networks and n network attributes are total _ij ) _m×n Composition, x _ij A j-th attribute value representing network i;

step 2: different standardization modes are adopted for the positive indexes and the negative indexes to carry out parameter standardization;

for the forward index:

for negative going index:

step 3: calculating standard deviation of the j-th attribute:

step 4: calculating the information amount of the j-th attribute:

step 5: calculating objective weights of all the attributes:

obtaining objective attribute weight vectors beta= (beta) of n networks ₁ ,β ₂ ,…,β _n )。

6. The dual mode traffic flow assignment method according to claim 3, wherein: the comprehensive weight calculation step in the step S2 is as follows:

assuming that there are m candidate networks in total, n network attributes, a decision matrix y= (Y) is constructed _ij ) _m×n Composition, y _ij A j-th attribute value representing network i;

defining subjective weight coefficient and objective weight coefficient in the combined weight as t and tau respectively, wherein t+tau is more than or equal to 0, and t+tau=1, and then the combined weight:

w _j ＝tα _j +τβ _j (j＝1,2,…,n)

subjective attribute weight for the jth attribute of the ith network is denoted tα _j y _ij Objective attribute weight is τβ _j y _ij The distance of the subjective and objective attribute weights is expressed as:

wherein Z is _i Representing the weight distance of the objective and subjective attribute of the ith network;

and establishing a weight combination optimization model by taking the minimized subjective and objective attribute weight distance as an objective function:

wherein, minZ represents the minimum subjective and objective attribute weight distance;

and (3) optimizing a model solving process:

establishing Lagrange functions:

wherein lambda represents Lagrange multiplier, to obtain

The three formulas are combined to obtain weight coefficients t and tau, which are respectively:

7. the dual mode traffic flow assignment method according to claim 3, wherein: the calculating the preference of the service for different candidate networks by using the FAHP algorithm in step S2 specifically includes the following steps:

step 1: calculating importance ranking of the index layer to the target layer, and obtaining weight relations of different services and network attributes according to the calculated weight of the index layer relative to the target layer;

step 2: the importance ranking of the index layer by the calculation scheme layer is performed, the importance of each candidate network with respect to the network attribute is compared, a fuzzy consistent matrix and corresponding weight are obtained, and then the weight relation between different network attributes and the candidate network is obtained;

step 3: calculating importance ranking of the scheme layer relative to the target layer;

multiplying the weight relation matrix of different services and network attributes by the relation matrix between the network attributes and the candidate networks, and calculating the sequence of a scheme layer relative to a target layer, namely the preference value of the services to the different candidate networks.

8. The dual mode traffic flow assignment method as recited in claim 6, wherein: the step S3 specifically comprises the following steps:

s31: let w be _j Is the weight value of the j-th attribute, and the utility matrix Y= (Y) _ij ) _m×n The weight vector w= [ w ] for each column and the column ₁ ,w ₂ ,…,w _n ] ^T The weights in the matrix are multiplied one by one to obtain a weighted utility matrix Z= (Z) _ij ) _m×n ：

S32: determining an ideal network

And negative ideal network->

S33: calculating the relative entropy distance between each candidate network and the positive ideal network, wherein the relative entropy distance between the candidate network and the positive ideal network is as follows

The relative entropy distance from the negative ideal network is +.>

S34: calculating relative closeness:

9. the dual mode traffic flow assignment method as recited in claim 1, wherein: the step S4 specifically comprises the following steps:

calculating to obtain a business preference value T _i Combining the business preference value and the relative closeness to be recorded as the comprehensive score R of the candidate network _i ：

In the method, in the process of the invention,

for adjusting the factors, adjusting the proportion of the relative closeness to the business preference value according to business requirements; r is R _i And the candidate networks are ordered according to the magnitude of the comprehensive score as the judging criterion of the optimal network for the service flow distribution, and the optimal network with the large comprehensive score as the service flow distribution is selected.

Images