A network topology generation method for edge computing

Abstract

The embodiment of the present application relates to a kind of network topology generation methods for edge computing, which comprises the following steps: initializing topology set;In topology plane, with service computing providing point as vertex, plane triangulation is made, and minimum spanning tree is calculated;Cross-merging of network topology is realized using mixed integer linear programming algorithm;The network topology of optimal cost is generated.The present application can make edge computing node and terminal equipment collaborative processing task by generating effective network topology scheme, under the condition of guaranteeing optimal cost, share computing resource and carry out cooperative computing task, can effectively improve the utilization of computing resource and the efficiency of task processing.

Description

Technical Field

[0001] This invention relates to the fields of edge computing, distributed computing, and communication technology, and specifically to a network topology generation method for edge computing in the Industrial Internet of Things. Background Technology

[0002] With the continuous development of Industrial Internet of Things (IIoT) technology, the number of IIoT devices in industrial networks is growing at an astonishing rate. Driven by various applications, these devices undertake a wide range of industrial tasks. However, most of these industrial tasks are computationally intensive, requiring significant computing power. Due to inherent limitations, the computing power that industrial equipment can provide is relatively limited, necessitating the transfer of some or all of its required computing power to other peripheral computing devices.

[0003] Leveraging mature communication technologies, computing, storage, and security capabilities can be distributed across nodes at the network edge. This distributes the computing, storage, and communication load from the central node to edge nodes with slightly weaker computing power, achieving low latency, high reliability, and low cost for services, while effectively protecting user privacy at the edge and resolving issues such as transmission latency and network congestion in traditional cloud computing IoT architectures. However, loading all computing tasks onto edge computing nodes results in limited processing power for a single edge computing node, neglecting the computing resources of numerous industrial IoT devices and adjacent edge computing nodes, leading to resource waste.

[0004] To address this issue, an effective solution is to leverage edge computing nodes (including adjacent edge computing nodes) and terminal devices within the edge computing system to collaboratively process tasks. Since both adjacent edge nodes and terminal devices need to perform computing tasks, an efficient network topology scheme needs to be designed to enable them to share computing resources and collaborate on computing tasks at the most cost-effective conditions. Summary of the Invention

[0005] (a) Technical problems to be solved

[0006] To address the aforementioned problems in the prior art, this invention provides a network topology generation method for edge computing, which enables edge computing nodes and terminal devices to collaboratively process tasks, share computing resources and perform cooperative computing tasks while ensuring optimal cost, thereby effectively improving the utilization rate of computing resources and the efficiency of task processing.

[0007] (II) Technical Solution

[0008] To achieve the above objectives, the main technical solutions adopted by the present invention include:

[0009] An embodiment of the present invention provides a network topology generation method for edge computing, comprising the following steps:

[0010] Step S100: Initialize the topology set G set If empty, calculate the overall computational power ∑ of the topology on the topological plane. p∈P g(p) and the total topology computation cost ∑ q∈Q d(q), where g(p) is the computing power that each service computing point p∈P can provide to the outside, and d(q) is the computing power that each service computing demand point q∈Q needs to provide to the outside;

[0011] Step S200: Determine the overall computational capability ∑ of the topology. p∈P Is g(p) greater than or equal to the total topological computation cost ∑? q∈Q Does d(q) hold true, i.e., ∑ p∈P g(p)≥∑ q∈Q Check if d(q) is true. If it is true, proceed to step S300. If it is not true, proceed to step S900.

[0012] Step S300: On the topology plane, calculate the point P for each service. i Perform a planar triangulation around the vertex, dividing the area into planar triangular regions of region R. i In the divided planar triangular region R i Inside, service computing is provided at point P. i Begin by calculating the minimum spanning tree T. i Connect the service computing demand points Q within the triangular region of this plane. i Planar triangle R i The topology in G is denoted as G i Minimum Spanning Tree T i The edge in the middle is denoted as E. i Minimum Spanning Tree T i The node in the denoted 'V' is denoted as V. i V i =(P i Q i ), G i =(V i E i ), and all topologies G i Put into set G set middle;

[0013] Step S400, from set G set The topology G is extracted from two adjacent planar triangular regions. x and G y ,judge If the condition is met, proceed to step S600; otherwise, proceed to step S500.

[0014] Step S500, extract the topology G x and G y Replace set G set Proceed to step S400;

[0015] Step S600, Topology G x and G y The cross-merging will result in the merged topology G. cost Put into set G Set ;

[0016] Step S700, determine set G Set If the element contains only one element, proceed to step S800 if true, otherwise proceed to step S400.

[0017] Step S800, remove G Set The unique element topology G cost Connect the service computing demand points that were not covered during triangulation to G. cost The node with the closest distance;

[0018] Step S900: Topology cannot be generated;

[0019] Step S1000, End, Exit.

[0020] In one embodiment of the present invention, in step S600, topology G x and G y Cross-mergers also include:

[0021] Step S610: Initialize and set the loop variable Count to zero, initialize and calculate the iteration threshold RTH, and initialize and set the optimal topology G. cost Empty;

[0022] Step S620: Select an S using a random average distribution method. x ∪S y subset S t Create a new topology G xy Vertex set V xyt =V x ∪V y ∪S t , of which S x For the planar triangular region R x In the triangular region R of the plane y Adjacent boundary nodes, S y For the planar triangular region R y In the triangular region R of the plane x Adjacent boundary nodes;

[0023] Step S630, in the new topology G xy Vertex set V xyt Insert all edges μ(E) in the middle. x ), where G xy =G x ∪G y The μ() operation is defined as follows:

[0024] If vw belongs to E x An edge in E, where v and w are the two vertices of the edge, and μ(vw) is the expression for E. x The topology G x node V x The edge connecting the node closest to v and the node closest to w;

[0025] Step S640, select parameter k, and select μ(E) through parameter k. y )\μ(E x A random subset E′ of ) xy , random subset E′ xy Edges in G are inserted into topology G xy In the context of the equation, “A\B” is defined as {x|x∈A, and} }, where parameter k is a random uniform distribution;

[0026] Step S650, check the generated topology G xy If the nodes are not connected, randomly add edges between the unconnected nodes and their adjacent nodes until the topology is connected.

[0027] Step S660: Call the cost evaluation algorithm Cost(G) to calculate the topology G. xy Cost, determine Cost(G) xy ) < Cost(G cost If the condition is true, proceed to step S670; otherwise, proceed to step S680.

[0028] Step S670, let G cost For G xy ;

[0029] Step S680: Increment the loop variable Count by 1. Determine whether the loop variable Count has reached the iteration threshold RTH. If it has not reached the iteration threshold RTH, proceed to step S620. If it has reached the iteration threshold RTH, proceed to step S690.

[0030] Step S690, G cost Put into set G Set middle.

[0031] In one embodiment of the present invention, the formula for calculating the iteration threshold RTH in step S610 is as follows:

[0032] RTH=N×(N-1) / 2

[0033] Where N is the topology G to be merged x and G y The sum of the number of vertices;

[0034] In one embodiment of the present invention, in step S640, the random subset E′ is... xy Edges in G are inserted into topology G xy In this process, it is possible to introduce a cycle, which needs to be eliminated by deleting other edges on the cycle.

[0035] In one embodiment of the present invention, the cost evaluation algorithm Cost(G) in step S660 further includes:

[0036] The cost evaluation algorithm Cost(G) takes topology G as input and outputs a cost metric for topology G. When G is empty, Cost(G) outputs infinity.

[0037] The formula for calculating Cost(G) in the cost assessment algorithm is ∑ e∈E ∑ c∈C (z(e,c)×len(e)×W(c)), where E is the set of edges in topology G, C is the channel type that edges in set E can choose, len(e) calculates the shortest distance from edge e to the service provider point p∈P in topology G, W(c) calculates the bandwidth of the channel type that edge e can choose, and z(e,c) is a binary variable, z(e,c)∈{0,1}, and has ∑ c∈C z(e,c)=1, It is guaranteed that the optional channel for edge e can only be of one type.

[0038] (III) Beneficial Effects

[0039] The beneficial effects of the present invention are: the network topology generation method for edge computing provided in the embodiments of the present invention can enable edge computing nodes and terminal devices to collaboratively process tasks by generating an effective network topology scheme, sharing computing resources and performing cooperative computing tasks under the condition of ensuring optimal cost, which can effectively improve the utilization rate of computing resources and the efficiency of task processing. Attached Figure Description

[0040] Figure 1 This is a flowchart of a network topology generation method for edge computing provided in an embodiment of the present invention.

[0041] Figure 2This is a flowchart of a network topology cross-merging method provided in an embodiment of the present invention. Detailed Implementation

[0042] To better explain and facilitate understanding of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] All technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0044] Figure 1 A flowchart of a network topology generation method for edge computing provided in an embodiment of the present invention is shown below. Figure 1 As shown, the method includes the following steps:

[0045] like Figure 1 As shown, in step S100, the topology set G is initialized. set If empty, calculate the overall computational power ∑ of the topology on the topological plane. p∈P g(p) and the total topology computation cost ∑ q∈Q d(q), where g(p) is the computing power that each service computing point p∈P can provide to the outside, and d(q) is the computing power that each service computing demand point q∈Q needs to provide to the outside.

[0046] like Figure 1 As shown, in step S200, the overall computational capability ∑ of the topology is determined. p∈P Is g(p) greater than or equal to the total topological computation cost ∑? q∈Q Does d(q) hold true, i.e., ∑ p∈P g(p)≥∑ q∈Q Check if d(q) is true. If it is true, proceed to step S300. If it is not true, proceed to step S900.

[0047] like Figure 1 As shown, in step S300, on the topology plane, point P is provided for each service calculation. i Perform a planar triangulation around the vertex, dividing the area into planar triangular regions of region R. i In the divided planar triangular region R i Inside, service computing is provided at point P. i Begin by calculating the minimum spanning tree T. i Connect the service computing demand points Q within the triangular region of this plane. i Planar triangle R iThe topology in G is denoted as G i Minimum Spanning Tree T i The edge in the middle is denoted as E. i Minimum Spanning Tree T i The node in the denoted 'V' is denoted as V. i V i =(P i Q i ), G i =(V i E i ), and all topologies G i Put into set G set middle;

[0048] like Figure 1 As shown, in step S400, from set G set The topology G is extracted from two adjacent planar triangular regions. x and G y ,judge If the condition is met, proceed to step S600; otherwise, proceed to step S500.

[0049] like Figure 1 As shown, in step S500, the extracted topology G x and G y Replace set G set Proceed to step S400;

[0050] like Figure 1 As shown, in step S600, topology G x and G y The cross-merging will result in the merged topology G. cost Put into set G Set ;

[0051] like Figure 1 As shown, in step S700, the set G is determined. Set If the element contains only one element, proceed to step S800 if true, otherwise proceed to step S400.

[0052] like Figure 1 As shown, in step S800, G is taken out. Set The unique element topology G cost Connect the service computing demand points that were not covered during triangulation to G. cost The node with the closest distance;

[0053] like Figure 1 As shown, in step S900, the topology cannot be generated;

[0054] like Figure 1As shown, in step S1000, the process ends and exits.

[0055] like Figure 1 As shown, in the technical solution provided by the embodiments of the present invention, by designing an effective network topology scheme, edge computing nodes and terminal devices can collaboratively process tasks, share computing resources and perform cooperative computing tasks under the condition of optimal cost.

[0056] The following are Figure 1 The specific implementation of each step in the illustrated embodiment is described in detail below:

[0057] In step S100, the topology set G is initialized. set If empty, calculate the overall computational power ∑ of the topology on the topological plane. p∈P g(p) and the total topology computation cost ∑ q∈Q d(q), where g(p) is the computing power that each service computing point p∈P can provide to the outside, and d(q) is the computing power that each service computing demand point q∈Q needs to provide to the outside.

[0058] In one embodiment of the present invention, it is necessary to initialize the topology set G. set Empty, topological set G set Used to store topologies to be cross-merged;

[0059] On the topological plane, calculate the overall computational power ∑ of the topology. p∈P g(p) and the total topology computation cost ∑ q∈Q d(q), where g(p) is the computing power that each service computing point p∈P can provide to the outside, and d(q) is the computing power that each service computing demand point q∈Q needs to provide to the outside;

[0060] Each service computing point p∈P, in addition to completing the computing tasks of its own node, can provide its surplus computing power to service computing demand points on the topology plane.

[0061] Each service computing demand point q∈Q is usually a terminal device in an edge computing system. Due to energy limitations and other issues, it does not have sufficient computing power. The surplus computing power of the service computing provider can make up for its lack of computing power.

[0062] By calculating the overall computational power of the topology ∑ p∈P g(p) and the total topology computation cost ∑ q∈Q d(q) can provide a basis for designing an effective network topology scheme.

[0063] In step S200, the overall computational capability ∑ of the topology is determined. p∈P Is g(p) greater than or equal to the total topological computation cost ∑?q∈Q Does d(q) hold true, i.e., ∑ p∈P g(p)≥∑ q∈Q Check if d(q) is true. If it is true, proceed to step S300. If it is not true, proceed to step S900.

[0064] In one embodiment of the present invention, it is necessary to determine the overall computational capability ∑ of the topology. p∈P Is g(p) greater than or equal to the total topological computation cost ∑? q∈Q d(q), that is, to determine ∑ p∈P g(p)≥∑ q∈Q If d(q) is true, it means that the supply of computing power in the topology is greater than the demand. The computing power can be transferred by designing a reasonable network topology. Conversely, if the supply of computing power in the topology is less than the demand, no matter what kind of network topology structure is used, it will not be able to meet the insufficient computing power.

[0065] In step S300, on the topological plane, point P is provided for each service calculation. i Perform a planar triangulation around the vertex, dividing the area into planar triangular regions of region R. i In the divided planar triangular region R i Inside, service computing is provided at point P. i Begin by calculating the minimum spanning tree T. i Connect the service computing demand points Q within the triangular region of this plane. i Planar triangle R i The topology in G is denoted as G i Minimum Spanning Tree T i The edge in the middle is denoted as E. i Minimum Spanning Tree T i The node in the denoted 'V' is denoted as V. i V i =(P i Q i ), G i =(V i E i ), and all topologies G i Put into set G set middle.

[0066] In one embodiment of the invention, on the topology plane, a service computing point P is provided. i By performing Delaunay triangulation on the vertices, the resulting triangles have the properties of empty circles and maximizing the minimum angle, avoiding the generation of narrow and elongated triangles. This is beneficial for the initial topology in the triangular regions to have a more reasonable structure. In particular, it can achieve trees between the divided regions without edge crossing (edge ​​crossing means that the initial topology may not be optimal), which brings convenience to the subsequent traversal of the entire topological plane space.

[0067] On the topological plane, point P is provided for each service computation. i After performing Delaunay triangulation on the vertices, some service computing requirements will still not be covered. In the subsequent step S800, the uncovered service computing requirements need to be connected to the cross-merged topology.

[0068] Service computing provides point P i Begin by calculating the minimum spanning tree T. i Connect the service computing demand points Q within the triangular region of this plane. i This can effectively avoid loops in the topology (preventing broadcast storms) and reduce the distance calculated by len(e) in the cost evaluation algorithm Cost(G).

[0069] In step S400, from set G set The topology G is extracted from two adjacent planar triangular regions. x and G y ,judge If the condition is met, proceed to step S600; otherwise, proceed to step S500.

[0070] In one embodiment of the present invention, it is necessary to determine from set G set Take out the topology G of two adjacent planar triangular regions x and G y Before cross-merging, if the combined computing power supply in the two topologies exceeds the combined demand, the merge will proceed. If not, the merged topologies can be returned to set G. set In the process, it will be merged with other adjacent topologies or adjacent topologies after intersection and merger;

[0071] In one embodiment of the present invention, this step can be performed in parallel using a multi-threaded approach to achieve cross-merging of multiple adjacent topologies. Starting from multiple triangular regions, adjacent topologies are gradually merged. In particular, the greater the gap between the supply and demand of computing power in two adjacent topologies, the more beneficial it is to the subsequent cross-merging of adjacent topologies.

[0072] In step S500, the extracted topology G x and G y Replace set G set Proceed to step S400.

[0073] In one embodiment of the present invention, if the supply of computing power for the topology to be cross-merged is less than the demand for merging, the topology to be cross-merged is returned to set G. set In the process, it will be merged with other adjacent topologies or adjacent topologies after cross-merger.

[0074] In step S600, topology G x and G y The cross-merging will result in the merged topology G. cost Put into set G Set .

[0075] In one embodiment of the present invention, the topology G in this step x and G y There are multiple options for cross-merging; this embodiment takes one network topology cross-merging method as an example. Figure 2 A flowchart of a network topology cross-merging method provided in an embodiment of the present invention includes the following steps:

[0076] like Figure 2 As shown, in step S610, the loop variable Count is initialized and set to zero, the iteration threshold RTH is initialized and calculated, and the optimal topology G is initialized and set. cost Empty;

[0077] like Figure 2 As shown, in step S620, an S is selected using a random average distribution method. x ∪S y subset S t Create a new topology G xy Vertex set V xyt =V x ∪V y ∪S t , of which S x For the planar triangular region R x In the triangular region R of the plane y Adjacent boundary nodes, S y For the planar triangular region R y In the triangular region R of the plane x Adjacent boundary nodes;

[0078] like Figure 2 As shown, in step S630, in the new topology G xy Vertex set V xyt Insert all edges μ(E) in the middle. x ), where G xy =G x ∪G y The μ() operation is defined as follows:

[0079] If vw belongs to E x An edge in E, where v and w are the two vertices of the edge, and μ(vw) is the expression for E. x The topology G x node V x The edge connecting the node closest to v and the node closest to w.

[0080] like Figure 2 As shown, in step S640, parameter k is selected, and μ(E) is selected through parameter k. y )\μ(E x A random subset E′ of ) xy , random subset E′ xy Edges in G are inserted into topology G xy In the context of the equation, “A\B” is defined as {x|x∈A, and} }, where parameter k is a random uniform distribution;

[0081] like Figure 2 As shown, in step S650, the generated topology G is checked. xy If the nodes are not connected, randomly add edges between the unconnected nodes and their adjacent nodes until the topology is connected.

[0082] like Figure 2 As shown, in step S660, the cost evaluation algorithm Cost(G) is called to calculate the topology G. xy Cost, determine Cost(G) xy ) < Cost(G cost If the condition is true, proceed to step S670; otherwise, proceed to step S680.

[0083] like Figure 2 As shown, in step S670, let G... cost For G xy ;

[0084] like Figure 2 As shown, in step S680, the loop variable Count is incremented by 1, and it is determined whether the loop variable Count has reached the iteration threshold RTH. If it has not reached the iteration threshold RTH, proceed to step S620; if it has reached the iteration threshold RTH, proceed to step S690.

[0085] like Figure 2 As shown, in step S690, G is... cost Put into set G Set middle.

[0086] In one embodiment of the present invention, the formula for calculating the iteration threshold RTH in step S610 is as follows:

[0087] RTH=N×(N-1) / 2

[0088] Where N is the topology G to be merged x and G y The sum of the number of vertices;

[0089] In one embodiment of the present invention, in step S630, the newly created topology G is utilized. xy Vertex set V xyt Using a mapping method, first insert all elements belonging to topology G. x The edge;

[0090] In one embodiment of the present invention, in step S640, the required cardinality k is randomly and uniformly selected to achieve the insertion of μ(E). y )\μ(E x A random subset E′ of ) xy ;

[0091] In one embodiment of the present invention, in step S640, the random subset E′ is... xy Edges in G are inserted into topology G xy In this process, it is possible to introduce a cycle, which needs to be eliminated by deleting other edges on the cycle.

[0092] In one embodiment of the present invention, the insertion method adopted in steps S630, S640, and S650 aims to maintain what may be the optimal local substructure, with the idea being to obtain G. y Add a subtree to G x During this process, delete G x To avoid introducing potential cycles, all edges of G need to be removed by deleting other edges on the cycle. x It is connected, therefore the generated topology is also connected.

[0093] In one embodiment of the present invention, the cost evaluation algorithm Cost(G) in step S660 further includes:

[0094] The cost evaluation algorithm Cost(G) takes topology G as input and outputs a cost metric for topology G. When G is empty, Cost(G) outputs infinity.

[0095] The formula for calculating Cost(G) in the cost assessment algorithm is ∑ e∈E ∑ c∈C (z(e,c)×len(e)×W(c)), where E is the set of edges in topology G, C is the channel type that edges in set E can choose, len(e) calculates the shortest distance from edge e to the service provider point p∈P in topology G, W(c) calculates the bandwidth of the channel type that edge e can choose, and z(e,c) is a binary variable, z(e,c)∈{0,1}, and has ∑ c∈C z(e,c)=1, Ensure that edge e can have only one type of optional channel;

[0096] The cost evaluation algorithm Cost(G) processes the selection of the optimal topology according to mixed integer linear programming. Among the selectable cross-merge topologies, it solves for the optimal solution and optimal value of cost under linear constraints.

[0097] In step S700, the set G is determined. Set If the element contains only one element, proceed to step S800 if true, otherwise proceed to step S400.

[0098] In one embodiment of the present invention, this step involves determining the set G. Set To determine whether a set G contains only one element, we can ascertain whether the cross-merge has been completed. Set If the merging process contains only one element, it indicates that the cross-merging has been completed. If it has not been completed, proceed to step S400 to continue merging.

[0099] In step S800, G is taken out. Set The unique element topology G cost Connect the service computing demand points that were not covered during triangulation to G. cost The node with the closest distance.

[0100] In one embodiment of the present invention, this step ensures that the service computing demand points that are not covered are connected to the topology.

[0101] In step S900, the topology cannot be generated.

[0102] In one embodiment of the present invention, this step is used to provide alarm information when the overall computational capability of the topology is less than the overall computational consumption of the topology.

[0103] In step S1000, the process ends and exits.

[0104] In one embodiment of the present invention, this step implements a unique exit point for a network topology generation method oriented towards edge computing.

[0105] In one embodiment of the present invention, this step is the sole exit point for deploying the optimization algorithm in this distributed service unit.

[0106] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the following claims.

[0107] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A network topology generation method for edge computing, characterized in that, include: Step S100: Initialize the topology set If empty, calculate the overall computational power of the topology on the topological plane. Overall computational cost of topology ,in, Provide points for each service computing Capable of providing computing power to external sources. Calculate the demand point for each service Requires externally provided computing power; Step S200: Determine the overall computing power of the topology. Is it greater than or equal to the total computational cost of the topology? Whether it is true or not, that is If the condition is met, proceed to step S300; otherwise, proceed to step S900. Step S300, on the topology plane, compute the service provision point for each service. Perform a planar triangulation on the vertex, and the resulting planar triangular regions are: In the divided planar triangular region Inside, service computing points Begin by calculating the minimum spanning tree. Connect the service computing demand points within the triangular region of this plane. Planar triangle Topological notation in Minimum Spanning Tree The edge notation in Minimum Spanning Tree The nodes in the are denoted as , , All topologies Add to collection middle; Step S400, from set Extract the topology of two adjacent planar triangular regions. and ,judge If the condition is met, proceed to step S600; otherwise, proceed to step S500. Step S500, extract the topology and Replace Proceed to step S400; Step S600, Topology and topology Cross-merging will result in a merged topology. Add to collection ; Step S700, determine the set If the element contains only one element, proceed to step S800 if true, otherwise proceed to step S400. Step S800, remove Unique element topology Connect the service computing demand points that were not covered during triangulation to The node with the closest distance; Step S900: Topology cannot be generated; Step S1000: End, Exit.

2. The network topology generation method for edge computing as described in claim 1, characterized in that, Step S600 includes: Step S610: Initialize and set the loop variable Count to zero, initialize and calculate the iteration threshold RTH, and initialize and set the optimal topology. Empty; Step S620: Select a random average distribution method. subset of Create a new topology vertex set ,in, For the triangular region in the plane Middle and planar triangular regions Adjacent boundary nodes, For the triangular region in the plane Middle and planar triangular regions Adjacent boundary nodes; Step S630, in the new topology vertex set Insert all edges in the middle. ,in, ,in, The operation is defined as: If vw belongs to In the context of an edge, v and w are the two vertices of that edge. yes Topology nodes The edge connecting the node closest to v and the node closest to w; Step S640, select parameter k, select through parameter k random subset , random subset Edges inserted into the topology In, among them, Defined as The parameter k is a random uniform distribution; Step S650: Check the generated topology If the nodes are not connected, randomly add edges between the unconnected nodes and their adjacent nodes until the topology is connected. Step S660: Invoke the cost assessment algorithm. Computational topology Cost, judgment If the condition is met, proceed to step S670; otherwise, proceed to step S680. Step S670, set for ; Step S680: Increment the loop variable Count by 1. Determine whether the loop variable Count has reached the iteration threshold RTH. If it has not reached the iteration threshold RTH, proceed to step S620. If it has reached the iteration threshold RTH, proceed to step S690. Step S690, will Add to collection middle.

3. The network topology generation method for edge computing as described in claim 2, characterized in that, The formula for calculating the iteration threshold RTH is as follows: Where N is the topology to be merged. and The number of vertices.

4. The network topology generation method for edge computing as described in claim 2, characterized in that, Cost evaluation algorithm The input is the topology G, and the output is the cost metric of the topology G. When G is empty... The output is infinity; cost evaluation algorithm The calculation formula is: Where E is the set of edges of topology G, and C is the channel type that the edges in set E can choose. Calculate the edge e to reach the service computing provider point in topology G. The shortest distance, Calculate the bandwidth of the channel type that edge e can select. It is a binary variable. And there are .

Images