Dynamic networking control method for'edge-end 'equipment based on differentiated optimization of physical ID (Identity)

Abstract

The invention relates to an edge-end equipment dynamic networking control method based on differentiated optimization of a real object ID, and the method comprises the following steps: 1, building a differentiated model of the real object ID, and discriminating real object information; step 2, dynamic networking is carried out by using the real object ID obtained in the step 1, and an optimal'edge-end 'equipment dynamic networking mode is obtained; and step 3, according to the actual demand of the power grid and the physical ID, sending an instruction to the equipment through the network to complete the control of the equipment. According to the invention, accurate information transmission between system equipment can be realized, even if the equipment is in an interference environment, the equipment can still operate normally, and the power supply reliability of power grid operation is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of Internet of Things optimization control, and relates to a dynamic networking control method for "edge-end" devices, and in particular to a dynamic networking control method for "edge-end" devices based on differentiated optimization of physical IDs. Background Art

[0002] With the expansion of the grid coverage area and the diversification of grid construction equipment, the current grid has the following main problems: the networking structure cannot be updated in real time, the operating status of the equipment and the handling of abnormal operating status cannot be grasped in real time, and in order to achieve the purpose of tracking and controlling the equipment, a large amount of manpower and material resources are required to conduct daily inspections and professional inspections of the equipment.

[0003] At present, there are many studies on the dynamic networking methods of power grids. For example, considering energy management and optimal scheduling, some existing technologies use the minimum function model to reduce the cost of the power grid, but ignore the communication problem between the network and the device, and do not consider that when the device is in an interference environment, the communication accuracy will be lacking. Especially when the network is actually used to transmit information, it is easy to have deviations in the reception of information between devices, and one signal controls multiple devices, resulting in errors in the transmission of information between dynamic networking and devices, and the existing dynamic networking methods cannot take into account the occurrence of these problems.

[0004] After searching, no public documents of the prior art identical or similar to the present invention were found. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a "edge-to-end" device dynamic networking control method based on differentiated optimization of physical ID, which can realize accurate transmission of information between system devices. Even if the device is in an interference environment, the device can still operate normally. This invention improves the power supply reliability of the power grid operation.

[0006] The present invention solves the practical problem by adopting the following technical solutions:

[0007] A method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs, comprising the following steps:

[0008] Step 1: Establish a differentiated model of physical ID to identify physical information;

[0009] Step 2: Use the physical ID obtained in step 1 to dynamically network and obtain the optimal "edge-end" device dynamic networking mode;

[0010] Step 3: According to the actual needs of the power grid and the physical ID, send instructions to the device through the network to complete the control of the device.

[0011] Moreover, the specific steps of step 1 include:

[0012] Step 1.1 adopts the method of complementing the edge layer and the end layer to collect the basic data of the IoT devices, and builds the attribute library of all IoT devices in the IoT cloud layer based on the collected basic data of the IoT devices, and summarizes and classifies the attributes of the IoT devices in the edge layer and the end layer;

[0013] Step 1.2: Based on the attribute summary and classification results of the edge layer and end layer IoT devices, asset coding and physical coding are performed on the edge layer and end layer IoT devices, and the asset coding and physical coding information are integrated to obtain the uniqueness and rationality coding of the edge layer and end layer IoT device IDs;

[0014] Step 1.3: Use the operation and inspection equipment to read the ID coding information of the edge layer and end layer IoT devices in the IoT obtained in step 1.2, and store the coding information in the on-site operation and inspection equipment;

[0015] Step 1.4 Scan the QR code of the actual equipment on the construction site through non-contact rapid identification technology represented by RFID technology to obtain the verification code of the physical ID;

[0016] Step 1.5 Verify whether the verification code of the physical ID is correct;

[0017] Step 1.6 If the verification code of the physical ID is correct, combine the coding rules in step 1.2 to obtain the physical company code segment, physical category and physical flow number; if the verification code of the physical ID is incorrect, establish a differentiated model of the IoT device and the physical ID based on the different verification codes, and re-scan;

[0018] 1.7 Filter out the physical ID with accurate verification code according to the differentiation model in 1.6, and obtain the physical information according to the encoding rules.

[0019] Moreover, the encoding rules for asset encoding and physical encoding of edge layer and end layer IoT devices in step 1.2 are as follows: the general encoding structure is a binary bit stream, which can be divided into encoding type, header and manager code, object classification code and serial number.

[0020] Moreover, the specific method of step 1.5 is:

[0021] Verify whether the code values ​​at each position in the verification code of the physical ID are accurate. The verification formula is: Among them, α i is the code value at the i-th position, W i is the weighted value at the i-th position;

[0022] Moreover, the specific steps of step 2 include:

[0023] (1) Now consider each device as a node in the network. Each node corresponds to a unique ID. The distance between each node is measured. Each node communicates at the maximum transmission power. The actual minimum distance model between nodes is: Among them, d ij is the distance between nodes i and j, is the distance error between nodes i and j.

[0024] (2) For the minimum error The optimization problem is solved by using an improved particle swarm algorithm to find the optimal solution, and then the optimal "edge-end" device dynamic networking method is obtained.

[0025] Moreover, the specific steps of step 2 (2) include:

[0026] ①Use Tent mapping to generate chaotic sequence and particle swarm;

[0027] ②Set the particle swarm size, speed, and position;

[0028] ③Select cluster center;

[0029] ④Set the adjustment rules of inertia weight;

[0030] ⑤ When external interference occurs, update the algorithm rules;

[0031] ⑥ Selection of the global optimal solution: Finally, by comparing the fitness of the cluster centers, gbest is selected as the global optimal solution.

[0032] ⑦ Determination of minimum error: Use the optimal solution in the particle swarm algorithm to obtain the minimum error, and then determine the exact distance between nodes;

[0033] ⑧ Information transmission between network and physical ID: Based on the actual distance between the network and the device, information is transmitted to the physical ID through the network, so that the power grid can be adjusted and dynamically networked in real time, and ultimately the optimal "edge-end" device dynamic networking method is obtained.

[0034] Moreover, the specific method of step ① of step 2 (2) is:

[0035] Randomly generate an initial value in [0,1], record it as y0, and substitute it into the Tent mapping expression Iterate, and when the maximum number of iterations N-1 is reached, the iteration stops and a chaotic sequence is generated, y i ,y i+1 Respectively represent the mapping values ​​of iterations i-1 and i;

[0036] Moreover, the specific method of step ② of step 2 (2) is:

[0037] The size of the particle swarm is set to a total of N particles, the search space is D-dimensional, and the particle swarm velocity and position expressions are: t is the number of particle update iterations; v t 、v t+1 are the velocities of the particle at generation t-1 and generation t respectively; x t 、x t+1 is the position of the particle at generation t-1 and generation t; ω is the inertia weight; c1 and c2 are acceleration factors; r1 and r2 are randomly distributed numbers between [0, 1]; pbest is the individual optimal solution, and gbest is the global optimal solution;

[0038] The specific method of step ③ of step 2 (2) is:

[0039] According to the maximum fitness of the particle position F = f(x i ) as the cluster center, the particle swarm is divided into several subgroups. After the cluster center is determined, after several iterations, a disturbance increment Δ tending to 0 is applied to the particle velocity to observe whether the fitness of the cluster center changes.

[0040] Moreover, the specific method of step ④ of step 2 (2) is:

[0041] Dynamically select cluster centers. When each particle moves toward the cluster center, increase the value of the inertia weight to improve the particle's ability to search for the local optimum. When each cluster center moves toward the global optimal solution, reduce the value of the inertia weight to improve the cluster center's ability to search for the global optimum.

[0042] Moreover, the specific method of step 5 of step 2 (2) is:

[0043] When disturbed by the outside world, the algorithm considers this situation as a predator. The predator is a bird that snatches food from particles and kills particles. Let the predator position be (m1, n1) and the particle position be (m2, n2). m1 and n1 are the components of the predator position in the coordinate axis; m2 and n2 are the components of the particle position in the coordinate axis. The particle position is updated according to the relative distance between the particle and the predator. The relative distance expression is: The particle automatic update expression is: It is represented by the position of the i-th (i=1,2,3,…N) particle in the d-th (d=1,2,3,…,D) dimension of the t-th generation; α is a random number in (0,1]; Q is a standard normal distribution random number; M is the maximum number of iterations; R is the alarm value.

[0044] Advantages and beneficial effects of the present invention:

[0045] 1. The present invention proposes a dynamic networking control method for "edge-end" devices based on differentiated optimization of physical ID. To address the problem of erroneous information transmission from the network to the device, the present invention measures the exact distance between the network and the device and sets an ID for each device. By utilizing the method of the present invention, the problem of erroneous information transmission between dynamic networking and devices can be reduced, and the system can achieve accurate information transmission between the network and the device.

[0046] 2. The present invention proposes a "edge-end" device dynamic networking control method based on differentiated optimization of physical ID. First, the physical information is collected and an attribute library is established in the Internet of Things; the physical QR code is scanned and identified using RFID technology; the physical ID information is verified by comparing the check code; and the physical ID technology can effectively perform dynamic networking and maximize energy efficiency. The present invention uses physical ID to ensure the accuracy of device management, and can accurately identify and control each device to avoid information errors and unnecessary energy consumption.

[0047] 3. The present invention provides an "edge-end" device dynamic networking control method based on differentiated optimization of physical ID. First, a minimum error function is established; secondly, an improved particle swarm algorithm is used to optimize the minimum error, so that the distance between the device and the network is more accurate; a predator variable is introduced into the particle swarm algorithm, through which the measured distance can be still relatively accurate when the device is in an interference environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a schematic diagram of a commonly used RFID label format of the present invention;

[0049] Figure 2 It is a schematic diagram of the physical ID coding structure of the power grid of the present invention.

[0050] Figure 3 It is a flow chart of the particle swarm algorithm of the present invention;

[0051] Figure 4 It is a 2004 switchgear physical ID diagram. DETAILED DESCRIPTION

[0052] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings:

[0053] A method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs, comprising the following steps:

[0054] Step 1: Establish a differentiated model of physical ID to identify physical information;

[0055] The specific steps of step 1 include:

[0056] Step 1.1 adopts the method of complementing the edge layer and the end layer to collect the basic data of the IoT devices, and builds the attribute library of all IoT devices in the IoT cloud layer based on the collected basic data of the IoT devices, and summarizes and classifies the attributes of the IoT devices in the edge layer and the end layer;

[0057] Step 1.2: Based on the attribute summary and classification results of the edge layer and end layer IoT devices, asset coding and physical coding are performed on the edge layer and end layer IoT devices, and the asset coding and physical coding information are integrated to obtain the uniqueness and rationality coding of the edge layer and end layer IoT device IDs;

[0058] The coding rules for asset coding and physical coding of edge layer and end layer IoT devices in step 1.2 are as follows:

[0059] As shown in Table 1, the general coding structure is a binary bit stream, which can be divided into four parts: coding type, version number (header) and the remaining three codes (manager code, object classification code and serial number):

[0060] Table 1 General coding structure

[0061]

[0062] In this embodiment, the header length of the existing data standard is divided into two types, 2 bits and 8 bits. The 2-bit header has a total of 3 values ​​11, 10, and 01 (00 for headers of more than 2 bits), and the 8-bit header has a total of 63 values ​​00000001-00111111 (0000000 for headers of more than 8 bits); the remaining data is stored in various types of information of the physical object by allocating space areas of different sizes.

[0063] Step 1.3 uses the operation and inspection equipment to read the edge layer and end layer IoT device ID coding information (such as device-related parameters, usage methods, etc.) obtained in step 1.2, and stores the coding information in the on-site operation and inspection equipment.

[0064] Step 1.4 Scan the QR code of the actual equipment on the construction site through non-contact rapid identification technology represented by RFID technology to obtain the verification code of the physical ID;

[0065] Step 1.5 Verify whether the verification code of the physical ID is correct.

[0066] The specific method of step 1.5 is:

[0067] Verify whether the code values ​​at each position in the verification code of the physical ID are accurate. The verification formula is: Among them, α i is the code value at the i-th position, W iis the weighted value at the i-th position;

[0068] Step 1.6 If the verification code of the physical ID is correct, combined with the coding rules in step 1.2, the company code segment, physical category and physical flow number of the physical ID are obtained; if the verification code of the physical ID is incorrect, according to the difference in the verification code, a differentiated model of the IoT device and the physical ID is established (indicating that the verification codes of the two are different to avoid repeated scanning the next time), and re-scan;

[0069] 1.7 Filter out the physical ID with accurate verification code according to the differentiation model in 1.6, and obtain the physical information according to the encoding rules.

[0070] Step 2: Use the physical ID obtained in step 1 to perform dynamic networking to obtain the optimal "edge-end" device dynamic networking method.

[0071] According to the current load and energy demand of the power grid, the system can dynamically select and connect the most suitable device. There are different quantities for the same device, but the physical ID can be accurate to a certain device. The device ID can be used to achieve real-time network control of the device to maximize energy efficiency and grid stability. This dynamic scheduling can be carried out based on real-time data to ensure the optimal operation of the power grid under different load conditions.

[0072] However, when "edge-to-end" devices communicate, they are easily affected by external interference due to the working environment, which causes a deviation between the actual distance between the devices and the measured distance, resulting in the network and the equipment not being able to operate normally. In order to solve this problem, the dynamic establishment of the network is achieved by reducing the deviation.

[0073] The specific steps of step 2 include:

[0074] (1) Now consider each device as a node in the network. Each node corresponds to a unique ID. The distance between each node is measured. Each node communicates at the maximum transmission power. The actual minimum distance model between nodes is: Among them, d ij is the distance between nodes i and j, is the distance error between nodes i and j.

[0075] (2) For the minimum error The optimization problem is solved by using an improved particle swarm algorithm to find the optimal solution, and then the optimal "edge-end" device dynamic networking method is obtained.

[0076] like Figure 3 As shown, the specific steps of step 2 (2) include:

[0077] (1) Use Tent mapping to generate chaotic sequence and particle swarm: randomly generate an initial value in [0,1] denoted as y0 and substitute it into the Tent mapping expression Iterate, and when the maximum number of iterations N-1 is reached, the iteration stops and a chaotic sequence is generated, y i ,y i+1 Respectively represent the mapping values ​​of iterations i-1 and i;

[0078] (2) Set the particle swarm size, speed, and position as follows: Set the particle swarm size to have a total of N particles, the search space to be D-dimensional, and the particle swarm speed and position expressions to be: t is the number of particle update iterations; v t 、v t+1 are the velocities of the particle at generation t-1 and generation t respectively; x t 、x t+1 is the position of the particle at generation t-1 and generation t; ω is the inertia weight; c1 and c2 are acceleration factors; r1 and r2 are randomly distributed numbers between [0, 1]; pbest is the individual optimal solution, and gbest is the global optimal solution.

[0079] (3) Method for selecting cluster centers: Based on the maximum fitness of the particle position F = f(x i ) as the cluster center, the particle swarm is divided into several subgroups. After the cluster center is determined, after several iterations, a disturbance increment Δ tending to 0 is applied to the particle velocity to observe whether the fitness of the cluster center changes.

[0080] (4) Inertia weight adjustment rule: Dynamically select cluster centers. When each particle moves toward the cluster center, increase the value of the inertia weight to improve the particle's ability to search for the local optimum. When each cluster center moves toward the global optimal solution, reduce the value of the inertia weight to improve the cluster center's ability to search for the global optimum.

[0081] (5) Update rules of the algorithm when there is external interference: When there is external interference, the algorithm considers this situation as a predator. The predator is a bird that steals the particle food and kills the particle. Let the predator position be (m1, n1) and the particle position be (m2, n2). m1 and n1 are the components of the predator position in the coordinate axis; m2 and n2 are the components of the particle position in the coordinate axis. The particle position is updated according to the relative distance between the particle and the predator. The relative distance expression is: The particle automatic update expression is: It is represented by the position of the i-th (i=1,2,3,…N) particle in the d-th (d=1,2,3,…,D) dimension of the t-th generation; α is a random number in (0,1]; Q is a standard normal distribution random number; M is the maximum number of iterations; R is the alarm value.

[0082] (6) Selection of the global optimal solution: Finally, by comparing the fitness of the cluster centers, gbest is selected as the global optimal solution.

[0083] (7) Determination of minimum error: The optimal solution in the particle swarm algorithm is used to obtain the minimum error, and then the accurate distance between nodes is determined;

[0084] (8) Information transmission between network and physical ID: Based on the actual distance between the network and the device, information is transmitted to the physical ID through the network, allowing the power grid to perform real-time adjustment and dynamic networking, ultimately obtaining the optimal "edge-to-end" device dynamic networking method.

[0085] Step 3: According to the actual needs of the power grid, based on the physical ID, send instructions to the device through the network to complete the control of the device to meet the actual needs of the power grid.

[0086] The working principle of the present invention is:

[0087] A method for fast networking and optimizing control of "edge-to-end" devices based on physical ID, where the physical ID code is encoded in Figure 1 The RFID tag format commonly used is shown in the figure. The specific steps are as follows:

[0088] A certain number of bits of physical ID code, digital code and its corresponding QR code are stored inside and outside the chip. In the actual operation and maintenance process, the RFID recognition of the physical ID of the Internet of Things and the scanning recognition of the QR code on the label surface are used in parallel.

[0089] Encode the physical assets of the IoT. From the perspective of efficient management, the composition of the physical ID code of the IoT is shown in Figure 2 .pass Figure 2 Coding can help us understand that material coding is a digital ID compiled for production supplies, non-production materials and inventory spare parts based on their intrinsic relationships; equipment coding is based on the production management system, with equipment as the application object, to improve the level of sophistication of production activity management.

[0090] If the internal code of the chip is inconsistent with the QR code, it will directly cause confusion in the business information associated with the physical ID, thereby affecting the effectiveness of the subsequent tracing of related information throughout the life cycle of the asset. Therefore, the physical ID information will be integrated and shared below.

[0091] After obtaining the ID information of the physical IoT object, the identification technology is used to integrate and share the information of the physical IoT object, and then the physical IoT object is finely controlled and managed, and its related information or data is accurately located.

[0092] In the particle swarm algorithm, after the cluster center is determined, a perturbation increment Δ is given to the particle velocity that tends to 0 after each iteration, and the order of magnitude of Δ is 10 -3 , the particle velocity update expression is v and V are the velocities before and after the particle update respectively.

[0093] When each particle goes to the cluster center, the value of the inertia weight is increased, and ω=0.8 is taken to improve the particle's ability to search for the local optimum. When each cluster center goes to the global optimal solution, the value of the inertia weight is reduced, and ω=0.4 is taken to improve the cluster center's ability to search for the global optimum.

[0094] The present invention is further described below through specific examples:

[0095] Taking a switch in a substation as an example, after collecting information from the Internet of Things, the main information obtained is as follows: Table equipment information

[0096] serial number Device Information 1 Equipment source: State Grid Corporation of China, a power station 2 Substation voltage level: 220kV 3 Equipment application voltage level: 110kV 4 Equipment use: spare 5 Device name: 2004 switch

[0097] According to the coding principle, the device information of the switch is encoded, and the physical ID is unique. Through the coding technology, the data characteristics and information of the device can be fully represented. For the device information of the 2004 switch in the table, the physical ID of the device is as follows Figure 4 shown.

[0098] When the power grid needs to control the switch to disconnect, it needs to send instructions to it according to the physical ID. Using infrared remote sensing technology, the distance between the power grid and the switch can be measured as 1800m. When environmental interference occurs, the particle swarm algorithm starts to run. The scale N = 50, the spatial dimension is 10, and the maximum number of iterations is 100. c1 = c2 = 2, r1 = 0.2, r2 = 0.8. The particles start to update.

[0099] Through calculation, we know that When the grid is finally disconnected through the network control switch, the distance needs to be set to 1900m.

[0100] It should be emphasized that the embodiments of the present invention are illustrative rather than restrictive. Therefore, the present invention includes but is not limited to the embodiments described in the specific implementation modes. Any other implementation modes derived by those skilled in the art based on the technical solutions of the present invention also fall within the scope of protection of the present invention.

Claims

1. A method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical ID, characterized in that: The following steps are involved: Step 1: Establish a differentiated model of physical ID to identify physical information; Step 2: Use the physical ID obtained in step 1 to dynamically network and obtain the optimal "edge-end" device dynamic networking mode; Step 3: According to the actual needs of the power grid and the physical ID, send instructions to the device through the network to complete the control of the device.

2. According to claim 1, a method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs is characterized by: The specific steps of step 1 include: Step 1.1 adopts the method of complementing the edge layer and the end layer to collect the basic data of the IoT devices, and builds the attribute library of all IoT devices in the IoT cloud layer based on the collected basic data of the IoT devices, and summarizes and classifies the attributes of the IoT devices in the edge layer and the end layer; Step 1.2: Based on the attribute summary and classification results of the edge layer and end layer IoT devices, asset coding and physical coding are performed on the edge layer and end layer IoT devices, and the asset coding and physical coding information are integrated to obtain the uniqueness and rationality coding of the edge layer and end layer IoT device IDs; Step 1.3: Use the operation and inspection equipment to read the ID coding information of the edge layer and end layer IoT devices in the IoT obtained in step 1.2, and store the coding information in the on-site operation and inspection equipment; Step 1.4 Scan the QR code of the actual equipment on the construction site through non-contact rapid identification technology represented by RFID technology to obtain the verification code of the physical ID; Step 1.5 Verify whether the verification code of the physical ID is correct; Step 1.6 If the verification code of the physical ID is correct, combine the coding rules in step 1.2 to obtain the physical company code segment, physical category and physical flow number; if the verification code of the physical ID is incorrect, establish a differentiated model of the IoT device and the physical ID based on the different verification codes, and re-scan; 1.7 Filter out the physical ID with accurate verification code according to the differentiation model in 1.6, and obtain the physical information according to the encoding rules.

3. According to claim 2, a method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs is characterized by: The encoding rules for asset encoding and physical encoding of edge layer and end layer IoT devices in step 1.2 are as follows: the general encoding structure is a binary bit stream, which can be divided into encoding type, header and manager code, object classification code and serial number.

4. According to claim 2, a method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs is characterized by: The specific method of step 1.5 is: Verify whether the code values ​​at each position in the verification code of the physical ID are accurate. The verification formula is: Among them, α i is the code value at the i-th position, W i is the weighted value at the i-th position.

5. According to claim 1, a method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs is characterized by: The specific steps of step 2 include: (1) Now consider each device as a node in the network. Each node corresponds to a unique ID. The distance between each node is measured. Each node communicates at the maximum transmission power. The actual minimum distance model between nodes is: Among them, d ij is the distance between nodes i and j, is the distance error between nodes i and j; (2) For the minimum error The optimization problem is solved by using an improved particle swarm algorithm to find the optimal solution, and then the optimal "edge-end" device dynamic networking method is obtained.

6. According to claim 1, a method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs is characterized by: The specific steps of step (2) of step 2 include: ①Use Tent mapping to generate chaotic sequence and particle swarm; ②Set the particle swarm size, speed, and position; ③Select cluster center; ④Set the adjustment rules of inertia weight; ⑤ When external interference occurs, update the algorithm rules; ⑥ Selection of the global optimal solution: Finally, by comparing the fitness of the cluster centers, gbest is selected as the global optimal solution; ⑦ Determination of minimum error: Use the optimal solution in the particle swarm algorithm to obtain the minimum error, and then determine the exact distance between nodes; ⑧ Information transmission between network and physical ID: Based on the actual distance between the network and the device, information is transmitted to the physical ID through the network, so that the power grid can be adjusted and dynamically networked in real time, and ultimately the optimal "edge-end" device dynamic networking method is obtained.

7. According to claim 6, a method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs is characterized by: The specific method of step ① of step 2 (2) is: Randomly generate an initial value in [0,1], record it as y0, and substitute it into the Tent mapping expression Iterate, and when the maximum number of iterations N-1 is reached, the iteration stops and a chaotic sequence is generated, y i ,y i+1 Respectively represent the mapping values ​​of iteration i-1 and i; The specific method of step 2 (2) of step 2 is: The size of the particle swarm is set to a total of N particles, the search space is D-dimensional, and the particle swarm velocity and position expressions are: t is the number of particle update iterations; v t 、v t+1 are the velocities of the particle at generation t-1 and generation t respectively; x t 、x t+1 is the position of the particle at generation t-1 and generation t; ω is the inertia weight; c1 and c2 are acceleration factors; r1 and r2 are randomly distributed numbers between [0, 1]; pbest is the individual optimal solution, and gbest is the global optimal solution.

8. According to claim 6, a method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs is characterized by: The specific method of step ③ of step 2 (2) is: According to the maximum fitness of the particle position F = f(x i ) as the cluster center, the particle swarm is divided into several subgroups. After the cluster center is determined, after several iterations, a disturbance increment Δ tending to 0 is applied to the particle velocity to observe whether the fitness of the cluster center changes.

9. According to claim 6, a method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs is characterized by: The specific method of step ④ of step 2 (2) is as follows: Dynamically select cluster centers. When each particle moves toward the cluster center, increase the value of the inertia weight to improve the particle's ability to search for the local optimum. When each cluster center moves toward the global optimal solution, reduce the value of the inertia weight to improve the cluster center's ability to search for the global optimum.

10. According to claim 6, a method for controlling dynamic networking of "edge-end" devices based on differentiated optimization of physical IDs is characterized by: The specific method of step ⑤ of step 2 (2) is as follows: when disturbed by the outside world, the algorithm regards this situation as a predator, and the predator is a bird that snatches particle food and kills particles. Assume the predator position (m1, n1), the particle position (m2, n2), m1, n1 are the components of the predator position in the coordinate axis; m2, n2 are the components of the particle position in the coordinate axis; update the particle position according to the relative distance between the particle and the predator, and the relative distance expression is: The particle automatic update expression is: It is represented by the position of the i-th (i=1,2,3,…N) particle in the d-th (d=1,2,3,…,D) dimension of the t-th generation; α is a random number in (0,1]; Q is a standard normal distribution random number; M is the maximum number of iterations; R is the alarm value.

Images