Network topology optimization method based on link cost and node degree constraint

Abstract

The invention discloses a network topology optimization method based on link cost and node degree constraint. The method comprises the following steps: representing a sensor networking topology through a weighted directed graph; the overall transmission link cost, the critical path cost, the transmission node information exchange cost and the network connectivity are comprehensively considered, and a transmission cost-oriented network topology optimization model is constructed; and solving the network topology optimization model through an improved greedy algorithm, and rapidly solving the sensor networking optimization topology. Compared with an existing method, the method has more comprehensive models and higher model solving speed, sensor networking can be rapidly optimized for temporary tasks, and the method is suitable for the fields of military reconnaissance, rail transit and the like with multiple temporary tasks and strict information transmission delay requirements.

Description

Technical Field

[0001] The present invention belongs to the technical field of sensor networking, and in particular relates to a network topology optimization method based on link cost and node degree constraints. Background Art

[0002] With the advent of the information age, the application requirements in the fields of industrial control systems, military reconnaissance systems, rail transit systems, etc. have gradually diversified, the corresponding sensor network topology complexity has increased, the network optimization conditions have increased, and the requirements for sensor network optimization speed have become more stringent. In the context of the information age, how to quickly solve the sensor network topology that meets application requirements has become an urgent problem to be solved in the field of topology optimization technology.

[0003] Different fields have different focuses on the demand for sensor network topology optimization. The industrial control field focuses on improving the network scale and service bandwidth performance; the military reconnaissance and rail transit fields focus on topology optimization speed and critical path transmission delay, which can quickly calculate the network topology that meets the time-sensitive path cost constraints for sudden tasks, and ensure the low-latency and fast transmission of temporary critical information. The traditional network topology optimization method oriented to transmission cost only considers the link cost when designing the model, ignoring the problem of increased exchange cost caused by increased transmission node load in complex business scenarios. Therefore, a network model that comprehensively considers the link cost and the transmission node exchange cost is designed to ensure the low-latency transmission of critical time-sensitive services, which has become a new idea for topology optimization.

[0004] Among the classic solutions to existing network topology optimization models, the greedy algorithm has the fastest model solving speed by selecting the optimal solution in the current state at each step. However, the algorithm does not perform topology optimization from a global perspective during the solving process, and often obtains a local optimal solution. How to improve the greedy algorithm by considering global constraints for diversified network models and further improve the model solving speed has also become a research focus in the field of sensor network topology optimization technology. Summary of the invention

[0005] The technical problem solved by the present invention is: to meet the diverse optimization needs of sensor networking in different fields, to overcome the shortcomings of the existing technology of single design optimization model and slow model solving speed, to provide a network topology optimization method based on link cost and node degree constraints, to design a topology optimization model by comprehensively considering the link transmission cost and the transmission node exchange cost, and to further improve the model solving speed by improving the greedy algorithm.

[0006] The technical solution of the present invention is:

[0007] A network topology optimization method based on link cost and node degree constraints, comprising:

[0008] The sensor network topology is represented by a weighted directed graph;

[0009] A network topology optimization model oriented to transmission cost is constructed by comprehensively considering the overall transmission link cost, critical path cost, transmission node information exchange cost and network connectivity.

[0010] The network topology optimization model is solved by improving the greedy algorithm, and the optimized topology of sensor networking is quickly solved.

[0011] Preferably, the weighted directed graph represents the sensor network topology through G = {V, E}, where V = {v1, v2, ..., v n} is the set of transmission nodes, n is the number of transmission nodes in the sensor network; is a set of directed edges, Represents the transmission node v i With v j The direct link cost value.

[0012] Preferably, the overall transmission link cost is the cost value of all direct links in the sensor networking topology; the critical path cost is the cost value of all direct links contained in the shortest path from the key message source node to the key message destination node in the sensor networking topology; the transmission node information exchange cost is expressed by the number of adjacent transmission nodes of the transmission node, that is, the transmission node connectivity, mapping the transmission node exchange cost; the network connectivity means that any two transmission nodes in the sensor networking topology have a bidirectional reachable path.

[0013] Preferably, the network topology optimization model takes the minimum cost of the overall network transmission link as the optimization goal, and is constructed with the critical path cost, transmission node connectivity, and bidirectional reachability of any two transmission nodes as constraints, as follows:

[0014]

[0015] where V = {v1,v2,…,v n} is the set of transmission nodes, n is the number of transmission nodes in the sensor network; weight parameter Represents the transmission node v i With v j The direct link cost value; Z represents the overall transmission link cost of the sensor network optimization topology; v source Neighboring transmission nodes representing key message source sensors; Represents the v obtained in the sensor network optimization topology source With v i path_delay represents the upper limit of the critical path cost; V key represents the set of adjacent transmission nodes of all key message destination sensor nodes; degree(vi ) represents node v i The connection in-degree or out-degree of the transmission node; degree_limit represents the upper limit of the number of transmission nodes for the transmission node adjacency message input or output. Represents the v obtained by using Dijkstra algorithm in the sensor network optimization topology i With v j The link cost value.

[0016] Preferably, solving the network topology optimization model by using an improved greedy algorithm comprises the following steps:

[0017] Step 3-1: Initialize and solve the network topology optimization model, calculate the path from the key message source transmission node to all other transmission nodes in the network based on the initial sensor network topology, and generate the shortest path spanning tree with the key message source transmission node as the root;

[0018] Step 3-2: Optimize the shortest path spanning tree topology based on the key path cost constraint and the transmission node information exchange cost constraint, delete the direct links that do not meet the degree constraint in the current topology by replacing the initial path, and rebuild the sensor network topology that meets the cost constraint and the degree constraint; the degree constraint refers to the transmission node information exchange cost constraint;

[0019] Step 3-3: Further optimize the sensor network topology based on the network connectivity constraints, and obtain the sensor network optimized topology that meets the constraint requirements and optimization objectives by adding direct links to the initial topology.

[0020] Preferably, the direct link cost is the transmission node v i to v j One-way link delay.

[0021] Preferably, in step 3-2, the initial path is a direct link between transmission nodes in the initial sensor networking topology, and the initial topology in step 3-3 is the initial sensor networking topology.

[0022] Preferably, the implementation method of step 3-1 is as follows:

[0023] Step 3-1-1: Initialize and solve the network topology optimization model, and use the Dijkstra algorithm to quickly generate the shortest path spanning tree for the initial sensor network topology. The spanning tree will take the key message source transmission node as the root, cover the entire network topology, and ensure that the path from the source node to all other nodes is the shortest;

[0024] Step 3-1-2: Check whether the path cost from the source node to the destination node in the shortest path spanning tree meets the critical path cost constraint. If so, output the shortest path spanning tree, otherwise, the topology optimization fails.

[0025] Preferably, the implementation method of step 3-2 is as follows:

[0026] Step 3-2-1: Input the shortest path spanning tree, initial sensor network topology, transmission node information exchange cost constraint and critical path cost constraint, and start network topology adjustment;

[0027] Step 3-2-2: Traverse all network nodes in the spanning tree and find the node M whose degree does not meet the constraint condition;

[0028] Step 3-2-3: Get all directly connected child nodes of node M and store them in queue Q;

[0029] Step 3-2-4: Determine whether there is a node N in the queue Q, so that there is a replacement path from the parent node of M to the node N that does not pass through the point M in the initial sensor network topology. If so, proceed to step 3-2-5. Otherwise, it indicates that the degree constraint of the node M cannot be adjusted, and the network optimization failure result is output;

[0030] Step 3-2-5: Determine whether node N and its subordinate branch trees contain the destination node of the key message. If so, proceed to step 3-2-8. Otherwise, replace the replacement path from the parent node of M to node N in the network spanning tree and proceed to step 3-2-6.

[0031] Step 3-2-6: Determine whether the node M in the network spanning tree after the path replacement satisfies the degree constraint requirement. If so, proceed to step 3-2-7; otherwise, delete the node N from the queue Q and proceed to step 3-2-4;

[0032] Step 3-2-7: Determine whether there are still nodes M that do not meet the degree constraint in the network spanning tree after the path replacement. If so, jump to step 3-2-2. Otherwise, output the latest network spanning tree, indicating that the sensor network topology that meets the critical path cost and node degree constraint requirements is obtained;

[0033] Step 3-2-8: Determine whether the network critical path delay meets the constraint requirements after replacing the replacement path from M's mother node to node N into the network spanning tree. If so, replace the replacement path from M's mother node to node N into the network spanning tree and proceed to step 3-2-6. Otherwise, delete node N from queue Q and proceed to step 3-2-4.

[0034] Preferably, the implementation method of step 3-3 is as follows:

[0035] Step 3-3-1: Input the sensor network topology and the initial sensor network topology that meet the critical path cost and degree constraint requirements;

[0036] Step 3-3-2: Check whether there are any two-to-two bidirectional unreachable node pairs in the sensor network topology. If so, jump to step 3-3-3. Otherwise, the sensor network topology has a fully connected characteristic, and output the sensor network optimization topology that meets the constraint requirements and optimization objectives.

[0037] Step 3-3-3: Select any pair of nodes that cannot be connected in the sensor network topology;

[0038] Step 3-3-4: Use the Dijkstra algorithm to find the shortest reachable path between two nodes in the initial sensor network topology;

[0039] Step 3-3-5: If there is a shortest reachable path between two nodes in the initial sensor network topology, jump to step 3-3-6, otherwise it is determined that it is impossible to generate a sensor network optimization topology with full connectivity characteristics, and the network optimization failure result is output;

[0040] Step 3-3-6: Determine whether the network node connectivity exceeds the constraint limit after adding the path to the sensor network topology. If so, jump to step 3-3-7, otherwise jump to step 3-3-8;

[0041] Step 3-3-7: Delete the edges in the path that violate the node degree constraint from the initial sensor network topology, and then jump to step 3-3-4;

[0042] Step 3-3-8: Connect the connection paths in the path that are not in the sensor network topology to the sensor network topology, and then jump to step 3-3-2.

[0043] The beneficial effects of the present invention compared with the prior art are:

[0044] 1) Compared with the existing topology optimization methods with transmission cost as the target, this invention proposes for the first time a topology optimization model with the goal of minimizing the overall transmission link cost of the network, with the critical path cost, transmission node connectivity, and bidirectional reachability of any two points as constraints. It comprehensively considers the link cost and the transmission node exchange cost, and proposes a network topology optimization method based on link cost and node degree constraints.

[0045] 2) Based on the traditional greedy algorithm that selects the optimal topology in the current state at each step, this method generates the shortest path spanning tree in one step based on the global critical path cost constraint during the initial solution of the model, rather than the minimum spanning tree generated by the traditional greedy algorithm. The model solution speed is further improved by improving the greedy algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flow chart of the method of the present invention;

[0047] Figure 2 It is the flowchart of topology model initialization solution;

[0048] Figure 3 It is a flow chart for topology adjustment based on critical path cost constraints and transmission node connectivity constraints;

[0049] Figure 4 It is a flowchart of topology adjustment based on connectivity constraints;

[0050] Figure 5 is the original topology diagram in the embodiment of the present invention;

[0051] Figure 6 is the shortest path spanning tree in the embodiment of the present invention;

[0052] Figure 7 is a sensor network topology diagram satisfying cost constraints and degree constraints in an embodiment of the present invention;

[0053] Figure 8 It is an optimized topology diagram of the sensor network that meets the connectivity constraints in the embodiment of the present invention. DETAILED DESCRIPTION

[0054] The specific embodiments of the present invention are further described in detail below with reference to the accompanying drawings.

[0055] The present invention discloses a network topology optimization method based on link cost and node degree constraints. Compared with the existing topology optimization method with transmission cost as the target, this method not only considers the overall link cost when designing the optimization model, but also guarantees the low-delay transmission of critical time-sensitive services by constraining the critical path delay and the transmission node connectivity for the first time. It is suitable for military reconnaissance, rail transit and other fields with many temporary tasks and strict information transmission delay requirements. For the new model, this method improves on the basis of the traditional greedy algorithm, proposes an improved greedy algorithm to optimize the model with constraints, and finds the minimum transmission cost topology between sensor networks under the premise of satisfying the link cost constraint, transmission node connectivity constraint and connectivity constraint.

[0056] like Figure 1 As shown, the specific implementation steps of the present invention are as follows:

[0057] Step 1: Represent the initial sensor network topology through a weighted directed graph;

[0058] Step 2: Considering the overall transmission link cost, critical path cost, transmission node information exchange cost and network connectivity, a network topology optimization model oriented to transmission cost is constructed;

[0059] Step 3: Solve the network topology optimization model through the improved greedy algorithm to quickly solve the sensor network optimization topology.

[0060] The improved greedy algorithm described in step 3 of the present invention can quickly solve the network topology optimization model, including the following steps:

[0061] Step 3-1: Initialize and solve the network topology optimization model. Based on the initial sensor network topology, use the Dijkstra algorithm to calculate the path from the key message source transmission node to all other transmission nodes in the network, and quickly generate the shortest path spanning tree with the key message source transmission node as the root;

[0062] Step 3-2: Optimize the shortest path spanning tree topology based on the critical path cost constraint and the transmission node information exchange cost constraint. By replacing the initial path, delete the direct links that do not meet the degree constraint in the current topology, and rebuild the sensor network topology that meets the cost constraint and degree constraint.

[0063] Step 3-3: Further optimize the sensor network topology based on the network connectivity constraints, and obtain the sensor network optimized topology that meets the constraint requirements and optimization objectives by adding direct links to the initial topology.

[0064] Step 3-1 Detailed process as follows Figure 2 As shown, the implementation steps are as follows:

[0065] Step 3-1-1: Enter the initial sensor network topology information;

[0066] Step 3-1-2: Use the Dijkstra algorithm to quickly generate the shortest path spanning tree. This spanning tree will take the key message source transmission node as the root, cover the entire network topology, and ensure that the path from the source node to all other nodes is the shortest;

[0067] Step 3-1-3: Check whether the path cost from the source node to the destination node in the shortest path spanning tree meets the critical path cost constraint. If so, output the shortest path spanning tree for constraint adjustment. Otherwise, the topology optimization fails.

[0068] Step 3-2 Detailed process as follows Figure 3 As shown, the implementation steps are as follows:

[0069] Step 3-2-1: Input the shortest path spanning tree, initial sensor network topology, transmission node information exchange cost constraint and critical path cost constraint, and start network topology adjustment;

[0070] Step 3-2-2: Traverse all network nodes in the spanning tree and find the node M whose degree does not meet the constraint condition;

[0071] Step 3-2-3: Get all directly connected child nodes of node M and store them in queue Q;

[0072] Step 3-2-4: Determine whether there is a node N in the queue Q, so that there is a replacement path from the parent node of M to the node N that does not pass through the point M in the initial sensor network topology. If so, proceed to step 3-2-5. Otherwise, it indicates that the degree constraint of the node M cannot be adjusted, and the network optimization failure result is output;

[0073] Step 3-2-5: Determine whether node N and its subordinate branch trees contain the destination node of the key message. If so, proceed to step 3-2-8. Otherwise, replace the replacement path from the parent node of M to node N in the network spanning tree and proceed to step 3-2-6.

[0074] Step 3-2-6: Determine whether the node M in the network spanning tree after the path replacement satisfies the degree constraint requirement. If so, proceed to step 3-2-7; otherwise, delete the node N from the queue Q and proceed to step 3-2-4;

[0075] Step 3-2-7: Determine whether there are still nodes M that do not meet the degree constraint in the network spanning tree after the path replacement. If so, jump to step 3-2-2. Otherwise, output the latest network spanning tree, indicating that the sensor network topology that meets the critical path cost and node degree constraint requirements is obtained;

[0076] Step 3-2-8: Determine whether the network critical path delay meets the constraint requirements after replacing the replacement path from M's mother node to node N into the network spanning tree. If so, replace the replacement path from M's mother node to node N into the network spanning tree and proceed to step 3-2-6. Otherwise, delete node N from queue Q and proceed to step 3-2-4.

[0077] Step 3-3 Detailed process as follows Figure 4 As shown, the implementation steps are as follows:

[0078] Step 3-3-1: Input the sensor network topology and the initial sensor network topology that meet the critical path cost and degree constraint requirements;

[0079] Step 3-3-2: Check whether there are any two-to-two bidirectional unreachable node pairs in the sensor network topology. If so, jump to step 3-3-3. Otherwise, the sensor network topology has a fully connected characteristic, and output the sensor network optimization topology that meets the constraint requirements and optimization objectives.

[0080] Step 3-3-3: Select any pair of nodes that cannot be connected in the sensor network topology;

[0081] Step 3-3-4: Use the Dijkstra algorithm to find the shortest reachable path between two nodes in the initial sensor network topology;

[0082] Step 3-3-5: If there is a shortest reachable path between two nodes in the initial sensor network topology, jump to step 3-3-6, otherwise it is determined that it is impossible to generate a sensor network optimization topology with full connectivity characteristics, and the network optimization failure result is output;

[0083] Step 3-3-6: Determine whether the network node connectivity exceeds the constraint limit after adding the path to the sensor network topology. If so, jump to step 3-3-7, otherwise jump to step 3-3-8;

[0084] Step 3-3-7: Delete the edges in the path that violate the node degree constraint from the initial sensor network topology, and then jump to step 3-3-4;

[0085] Step 3-3-8: Connect the connection paths in the path that are not in the sensor network topology to the sensor network topology, and then jump to step 3-3-2.

[0086] The improved greedy algorithm of the present invention is based on the traditional greedy algorithm selecting the optimal topology in the current state at each step, and comprehensively considering the global constraints to optimize the model. Step 3-1 comprehensively considers the key path cost constraints to generate the shortest path spanning tree in one step, rather than the minimum spanning tree generated by the traditional greedy algorithm, further improving the model solving speed.

[0087] Example:

[0088] by Figure 5-Figure 8 Taking the sensor network topology optimization process shown in the figure as an example, the specific steps of the embodiment of the present invention are described:

[0089] Step 1: Pass Figure 5 The weighted directed graph shown in the figure represents the initial sensor network topology, V = {v a ,v b ,v c ,v d ,v e ,v f} represents 6 transmission nodes in the network, is a set of directed edges, such as w (a,b) =4 indicates the transmission node v a With v b The direct link cost value is 4;

[0090] Step 2: Considering the overall transmission link cost, critical path cost, transmission node information exchange cost and network connectivity, a network topology optimization model oriented to transmission cost is constructed:

[0091]

[0092] Where V = {v a ,vb ,v c ,v d ,v e ,v f} is a collection of transmission nodes; weight parameter Represents the transmission node v i With v j The direct link cost value; Z represents the overall transmission link cost of the sensor network optimization topology; v c Represents the key message source transmission node; Represents the v obtained by using Dijkstra algorithm in the sensor network optimization topology c With v i The link cost value; 9 represents the upper limit of the critical path cost; v d With v f It is the destination transmission node for all key messages; degree(v i ) represents node v i The connection in-degree or out-degree of the transmission node; 2 represents the upper limit of the number of transmission nodes for the transmission node adjacency message input or output. Represents the v obtained by using Dijkstra algorithm in the sensor network optimization topology i With v j The link cost value.

[0093] Step 3: Initialize and solve the initial sensor network topology, and use the Dijkstra algorithm to calculate the key message source transmission node v c The paths to all other transmission nodes in the network are quickly generated with the key message source transmission node v c The shortest path to the root is generated as Figure 6 As shown;

[0094] Step 4: Topology optimization of the shortest path spanning tree based on the critical path cost constraint and the transmission node connectivity constraint. Figure 6 Transmission Node v c The out-degree is 3, which does not meet the node degree constraint requirement. It is necessary to reconstruct the topology by replacing the initial path:

[0095] Get v in the shortest path spanning tree c All child nodes under {v b ,v d ,v f}, determine whether there is a connection through v in the initial network topology b / v d / v f Other alternative paths can satisfy the degree constraint and the critical path cost constraint requirements, using replace like Figure 7As shown, a sensor network topology that satisfies the degree constraint and the critical path cost constraint is constructed;

[0096] Step 5: Further optimize the sensor network topology based on the reachability constraint. Figure 7 By adding direct links to the initial topology Get as Figure 8 The sensor network optimization topology that meets the constraint requirements and optimization objectives is shown.

[0097] Based on the traditional link-cost topology optimization model, the present invention proposes a topology optimization model that comprehensively considers the overall transmission link cost, critical path cost, transmission node information exchange cost and network connectivity, and improves the solution speed of the network topology optimization model by improving the greedy algorithm. The network topology optimization model solution includes the following steps: S1, initializing and solving the initial sensor network topology, and outputting the shortest path spanning tree; S2, reconstructing the topology based on the optimization model constraints, and outputting the sensor network topology that meets the cost constraint and degree constraint; S3, adjusting the connectivity, and outputting the sensor network optimization topology that meets the connectivity constraint. Compared with the existing methods, the present invention has a more comprehensive model design and a faster model solution speed, and can quickly optimize the sensor network for temporary tasks. It is suitable for military reconnaissance, rail transportation and other fields with many temporary tasks and strict information transmission delay requirements.

[0098] The contents not described in detail in the specification of the present invention belong to the common knowledge of the professionals in this field.

Claims

1. A network topology optimization method based on link cost and node degree constraints, characterized in that: include: The sensor network topology is represented by a weighted directed graph; Taking into account the overall transmission link cost, critical path cost, transmission node information exchange cost and network connectivity, a network topology optimization model oriented to transmission cost is constructed. The network topology optimization model is solved by improving the greedy algorithm, and the optimized topology of sensor networking is quickly solved.

2. The network topology optimization method based on link cost and node degree constraints according to claim 1, characterized in that: The weighted directed graph represents the sensor network topology through G = {V, E}, where V = {v1, v2, ..., v n } is the set of transmission nodes, n is the number of transmission nodes in the sensor network; is a set of directed edges, Represents the transmission node v i With v j The direct link cost value.

3. The network topology optimization method based on link cost and node degree constraints according to claim 1, characterized in that: The overall transmission link cost is the cost value of all direct links in the sensor network topology; the critical path cost is the cost value of all direct links contained in the shortest path from the key message source node to the key message destination node in the sensor network topology; The transmission node information exchange cost is expressed by the number of adjacent transmission nodes of the transmission node, that is, the transmission node connectivity, which maps the transmission node exchange cost; the network connectivity is that any two transmission nodes in the sensor networking topology have a bidirectional reachable path.

4. The network topology optimization method based on link cost and node degree constraints according to claim 1, characterized in that: The network topology optimization model takes the minimum cost of the overall network transmission link as the optimization goal, and is constructed with the critical path cost, transmission node connectivity, and bidirectional reachability of any two transmission nodes as constraints. The specific conditions are as follows: where V = {v1,v2,…,v n } is the set of transmission nodes, n is the number of transmission nodes in the sensor network; weight parameter Represents the transmission node v i The direct link cost with vj; Z represents the overall transmission link cost of the sensor network optimization topology; v source Neighboring transmission nodes representing key message source sensors; Represents the v obtained in the sensor network optimization topology source With v i path_delay represents the upper limit of the critical path cost; V key represents the set of adjacent transmission nodes of all key message destination sensor nodes; degree(v i ) represents node v i The connection in-degree or out-degree of the transmission node; degree_limit represents the upper limit of the number of transmission nodes for the transmission node adjacency message input or output. Represents the v obtained by using Dijkstra algorithm in the sensor network optimization topology i With v j The link cost value.

5. The network topology optimization method based on link cost and node degree constraints according to claim 1, characterized in that: The method of solving the network topology optimization model by improving the greedy algorithm includes the following steps: Step 3-1: Initialize and solve the network topology optimization model, calculate the path from the key message source transmission node to all other transmission nodes in the network based on the initial sensor network topology, and generate the shortest path spanning tree with the key message source transmission node as the root; Step 3-2: Optimize the shortest path spanning tree topology based on the key path cost constraint and the transmission node information exchange cost constraint, delete the direct links that do not meet the degree constraint in the current topology by replacing the initial path, and rebuild the sensor network topology that meets the cost constraint and the degree constraint; the degree constraint refers to the transmission node information exchange cost constraint; Step 3-3: Further optimize the sensor network topology based on the network connectivity constraints, and obtain the sensor network optimized topology that meets the constraint requirements and optimization objectives by adding direct links to the initial topology.

6. A network topology optimization method based on link cost and node degree constraints according to any one of claims 2 to 4, characterized in that: The direct link cost is the transmission node v i One-way link delay to vj.

7. The network topology optimization method based on link cost and node degree constraints according to claim 5 is characterized in that: In step 3-2, the initial path is a direct link between transmission nodes in the initial sensor networking topology, and the initial topology in step 3-3 is the initial sensor networking topology.

8. The network topology optimization method based on link cost and node degree constraints according to claim 5 is characterized in that: The implementation method of step 3-1 is as follows: Step 3-1-1: Initialize and solve the network topology optimization model, and use the Dijkstra algorithm to quickly generate the shortest path spanning tree for the initial sensor network topology. The spanning tree will take the key message source transmission node as the root, cover the entire network topology, and ensure that the path from the source node to all other nodes is the shortest; Step 3-1-2: Check whether the path cost from the source node to the destination node in the shortest path spanning tree meets the critical path cost constraint. If so, output the shortest path spanning tree, otherwise, the topology optimization fails.

9. A network topology optimization method based on link cost and node degree constraints according to claim 5, characterized in that: The implementation method of step 3-2 is as follows: Step 3-2-1: Input the shortest path spanning tree, initial sensor network topology, transmission node information exchange cost constraint and critical path cost constraint, and start network topology adjustment; Step 3-2-2: Traverse all network nodes in the spanning tree and find the node M whose degree does not meet the constraint condition; Step 3-2-3: Get all directly connected child nodes of node M and store them in queue Q; Step 3-2-4: Determine whether there is a node N in the queue Q, so that there is a replacement path from the parent node of M to the node N that does not pass through the point M in the initial sensor network topology. If so, proceed to step 3-2-5. Otherwise, it indicates that the degree constraint of the node M cannot be adjusted, and the network optimization failure result is output; Step 3-2-5: Determine whether node N and its subordinate branch trees contain the destination node of the key message. If so, proceed to step 3-2-8. Otherwise, replace the replacement path from the parent node of M to node N in the network spanning tree and proceed to step 3-2-6. Step 3-2-6: Determine whether the node M in the network spanning tree after the path replacement satisfies the degree constraint requirement. If so, proceed to step 3-2-7; otherwise, delete the node N from the queue Q and proceed to step 3-2-4; Step 3-2-7: Determine whether there are still nodes M that do not meet the degree constraint in the network spanning tree after the path replacement. If so, jump to step 3-2-2. Otherwise, output the latest network spanning tree, indicating that the sensor network topology that meets the critical path cost and node degree constraint requirements is obtained; Step 3-2-8: Determine whether the network critical path delay meets the constraint requirements after replacing the replacement path from M's mother node to node N into the network spanning tree. If so, replace the replacement path from M's mother node to node N into the network spanning tree and proceed to step 3-2-6. Otherwise, delete node N from queue Q and proceed to step 3-2-4.

10. A network topology optimization method based on link cost and node degree constraints according to claim 5, characterized in that: The implementation method of step 3-3 is as follows: Step 3-3-1: Input the sensor network topology and the initial sensor network topology that meet the critical path cost and degree constraint requirements; Step 3-3-2: Check whether there are any two-to-two bidirectional unreachable node pairs in the sensor network topology. If so, jump to step 3-3-3. Otherwise, the sensor network topology has a fully connected characteristic, and output the sensor network optimization topology that meets the constraint requirements and optimization objectives. Step 3-3-3: Select any pair of nodes that cannot be connected in the sensor network topology; Step 3-3-4: Use Dijkstra algorithm to find the shortest reachable path between two nodes in the initial sensor network topology; Step 3-3-5: If there is a shortest reachable path between two nodes in the initial sensor network topology, jump to step 3-3-6, otherwise it is determined that it is impossible to generate a sensor network optimization topology with full connectivity characteristics, and the network optimization failure result is output; Step 3-3-6: Determine whether the network node connectivity exceeds the constraint limit after adding the path to the sensor network topology. If so, jump to step 3-3-7, otherwise jump to step 3-3-8; Step 3-3-7: Delete the edges in the path that violate the node degree constraint from the initial sensor network topology, and then jump to step 3-3-4; Step 3-3-8: Connect the connection paths in the path that are not in the sensor network topology to the sensor network topology, and then jump to step 3-3-2.

Images