A method for constructing approximate MDS based on local information in mobile ad hoc networks

Abstract

The application relates to a kind of approximate MDS construction method based on local information in mobile ad hoc network, the whole network node domination set is cut, algorithm is carried out in multiple rounds, each round is divided into four stages: neighbor information perception stage, tentative cutting stage, domination stage and dominated stage;Each node has a serial number to indicate its own ID, and maintains a local variable s to indicate the current state of node;Approximate MDS is constructed using the method disclosed in the application, and MCDS is constructed on this basis, so that the nodes in MCDS act as a virtual backbone network, which can provide routing for nodes when a segment of link in the shortest path fails or a node cannot calculate the shortest path due to network topology not converging.In actual application scenarios, the algorithm disclosed in the application can reduce the occurrence of un-dominated nodes and improve fault tolerance.

Description

Technical Field

[0001] This invention relates to the field of computer network technology, and in particular to an approximate MDS construction method based on local information in mobile ad hoc networks. Background Technology

[0002] Mobile ad hoc networks have been widely used in various fields due to their ease of deployment and high scalability, such as military operations, environmental protection, and emergency disaster relief.

[0003] Adhoc (also known as Peer-to-Peer) refers to a self-organizing wireless network, consisting of only two or more STAs (Stations, Units, and Nodes) without access points (APs). This type of network has a loose structure, allowing all STAs to communicate directly. With the development of adhoc network trunking technology, the number of nodes in adhoc networks is increasing, leading to more complex network topologies and higher communication density. Compared to traditional networks, mobile adhoc networks lack a wired backbone or any centralized network management system for unified management of network routing and global topology. Broadcasting is a crucial method for information exchange among nodes in mobile adhoc networks, and flooding is the most common broadcast mode. However, unrestricted flooding can cause network nodes to forward redundant data packets multiple times, potentially leading to broadcast storms (in the OSPF protocol, flooding is used to send information to all routers when link states change). To address these issues, existing research has proposed selecting key nodes in mobile adhoc networks to form a virtual backbone network. Virtual backbone networks can uniformly forward data packets and calculate and maintain routes, reducing the forwarding of redundant data packets and improving the utilization of network resources. The key to a virtual backbone network lies in how to select critical nodes. The most common approach is to abstract this problem as solving the Minimum Connected Dominating Set (MCDS) problem in graph theory. The nodes in the obtained minimum connected dominating set are the critical nodes that make up the virtual backbone network.

[0004] Regarding the problem of finding the minimum connected dominating set, existing research mainly categorizes algorithms into centralized and distributed algorithms. However, due to the lack of centralized control devices and the highly dynamic nature of the global topology in ad hoc networks, centralized algorithms are clearly difficult to deploy on real-world mobile ad hoc network devices. Most existing distributed MCDS solving algorithms rely on coloring methods to obtain maximal independent sets before connecting them, but these algorithms all assume that node information can be successfully transmitted. In distributed MCDS solving algorithms, a better approach is to first construct the minimum dominating set (MDS) in a distributed manner, and then check the connectivity of each node in the minimum dominating set to obtain the minimum connected dominating set. Assuming the minimum dominating set has been obtained, node connectivity can be derived using algorithms such as disjoint-set data structures; therefore, the key to this approach lies in the distributed construction of the minimum dominating set. See prior art documents: F.Luoand PKSrimani, "New Biology Inspired Anonymous Distributed Algorithms toCompute Dominating and Total Dominating Sets in Network Graphs," 2016IEEEInternational Parallel and Distributed Processing Symposium Workshops (IPDPSW), 2016, pp.517-524

[0005] This invention proposes an approximate MDS construction method in mobile ad hoc networks. In this method, nodes determine whether all neighbors are dominated based on local information, reducing the impact of failed interaction information transmission and improving fault tolerance. Summary of the Invention

[0006] The objective of this invention is to address the aforementioned technical problems by proposing an approximate MDS construction method based on local information in mobile ad hoc networks. Specifically, this invention proposes an approximate MDS construction method based on local information, where the dominating set D of the network nodes in the mobile ad hoc network satisfies the following condition: all nodes in the network either belong to set D or are adjacent to nodes that belong to D. The set D with the fewest such nodes is called the minimum dominating set (MDS).

[0007] The technical solution of this invention is as follows: A method for constructing an approximate MDS based on local information in a mobile ad hoc network. This method involves pruning the dominant set of all nodes in the network. The algorithm runs in multiple rounds, each round consisting of four stages: a neighbor information perception stage, a tentative pruning stage, a dominating stage, and a dominated stage. Each node has a sequence number representing its ID and maintains and sets up a local variable `s` to represent the node's current state. For node i, `s`... i =0 means that the node does not currently belong to the dominating set D; i =1 means the node currently belongs to the dominating set D; s i =2 indicates that the node's previous time slot belonged to the dominating set D, but this time slot will be pruned; during initialization, all nodes in the network belong to the dominating set D, so s is set to 1 for all nodes; the specific steps of the MDS construction method based on local information are as follows:

[0008] In the neighbor information perception phase, step S1, each node broadcasts its own state information and simultaneously receives and updates the state information of its one-hop neighbors locally; the specific process is as follows:

[0009] Step S11: Each node broadcasts its own state information StateInfo(ID, Onehop, State, Dominator); the state information StateInfo includes the node's ID, the node's one-hop neighbor Onehop, the node's current state State, and the node's dominator number Dominator; during initialization, each node's state is 1, and the dominator number is equal to the node's ID.

[0010] Step S12: After each node receives the StateInfo information from its one-hop neighbor, it maintains and updates its two-hop neighbor table, one-hop neighbor state table, and one-hop neighbor dominator table locally.

[0011] After step S1 is completed, proceed to step S2, which is the exploratory trimming stage;

[0012] Step S21: The node in state 1 first reads its own number of one-hop neighbors N, and then reads them with a constant independent and identically distributed probability. Once enabled, a node immediately sets its state to 2 and broadcasts a QuitRequest message to its one-hop neighbors.

[0013] Step S22: When a node in state 2 receives a QuitRequest broadcast by another node, it immediately abandons broadcasting the QuitRequest information and resets its state to 1. When a node in state 1 receives a QuitRequest broadcast by another node, it immediately sends a QuitPermission message to notify the node that it can exit.

[0014] Step S23: When the exploratory pruning phase time slot ends, if a node in state 2 does not receive the QuitPermission message, it resets its state to 1.

[0015] After step S2 is completed, step S3, the domination phase, begins. All nodes in state 1 broadcast the dominate information to their surroundings. Nodes in state 0 or 2 update their local dominate index after receiving the dominate information. The dominate index is the source node ID of the received dominate information. If a node in state 0 or 2 does not receive the dominate information in the entire S3 time slot, it sets its dominate index to -1.

[0016] The interpretation of the dominance point number is as follows: if node 3 is in state 0 or 2 and receives a Dominate message from node 5, then node 3 knows that it is dominated by node 5 and updates its dominance point number to 5.

[0017] After step S3 is completed, step S4, the dominated phase, begins. In step S3, if the locally maintained dominant point changes, the nodes in states 0 and 2 broadcast the Dominated message to notify the surrounding nodes of their node state and dominant point number.

[0018] At the end of phase S4, all nodes in state 2 traverse their own one-hop neighbor state table and one-hop neighbor dominance table. If the dominance of a node in state 0 among all one-hop neighbors is not -1 and the node's own dominance is not -1, then the node sets its own state to 0. If there is a node in state 0 among the one-hop neighbors (nodes) with a dominance of -1 or the node's own dominance is -1, then the node sets its own state to 1 and resets its own dominance to its own index. At this point, all nodes in state 1 represent the obtained approximate MDS.

[0019] The method disclosed in this invention is used to construct an approximate MDS, and based on this, an MCDS is constructed, allowing the nodes in the MCDS to act as a virtual backbone network. This allows the MCDS to provide routing for nodes when a link in the shortest path fails or when a node cannot calculate the shortest path due to network topology non-convergence. In practical applications, the algorithm disclosed in this invention can reduce the occurrence of uncontrolled nodes and improve fault tolerance.

[0020] Beneficial Effects: Traditional source routing calculations for shortest paths in mobile ad hoc networks often require the initiating node to possess global topology information. In practical applications of existing mobile ad hoc networks, nodes sometimes fail to successfully transmit their information due to interference or channel conflicts. This makes it difficult for some undominated nodes not in the MDS to successfully broadcast their undominated state information to regain dominance. To address this issue, the method disclosed in this invention involves nodes exchanging information with their one-hop neighbors, including the dominance point information of those one-hop neighbors. Nodes maintain the dominance point information of their one-hop neighbors to ensure that all neighbors are in a dominant or subdominated state before leaving the MDS. In practical applications, the algorithm disclosed in this invention can reduce the occurrence of undominated nodes and improve fault tolerance.

[0021] The method disclosed in this invention constructs an approximate MDS, and based on this, an MCDS is constructed, allowing the nodes in the MCDS to act as a virtual backbone network. This allows the MCDS to provide routing for nodes when a link in the shortest path fails or when a node cannot calculate the shortest path due to network topology non-convergence. The method disclosed in this invention is similar to the method proposed by F. Luo and PKSrimani in existing literature. However, the methods mentioned above, like many MDS construction methods based on coloring algorithms, consider the scenario of an ideal channel where each node can successfully transmit its information. Attached Figure Description

[0022] Figure 1 The diagram shown is the overall flowchart of the present invention for constructing an approximate MDS.

[0023] Figure 2 The figure shows the node-specific state machine of the time slot in the exploratory pruning phase;

[0024] Figure 3 The diagram shows the specific steps of the control phase time slot;

[0025] Figure 4 The diagram shows the specific state machine flow of a node in the dominated phase time slot;

[0026] Figure 5 The original network topology diagram is shown below;

[0027] Figure 6 The diagram shown is a node state diagram after the trial trimming phase of a certain embodiment has ended;

[0028] Figure 7 The figure shown is an approximate MDS diagram of several nodes in a certain embodiment;

[0029] Figure 8 The diagram shows the network topology of an embodiment of the present invention. Detailed Implementation

[0030] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0031] like Figure 1 As shown, a general flowchart of the present invention for constructing an approximate MDS is given, wherein the specific steps of each time slot are described as follows:

[0032] Step S1: Each node broadcasts its own state information and simultaneously receives and updates the state information of its one-hop neighbors locally. The specific process is as follows:

[0033] In step S11, each node broadcasts its own state information, StateInfo(ID, Onehop, State, Dominator). StateInfo contains the node's ID, its one-hop neighbor Onehop, its current state State, and its dominator number Dominator. When a node initializes online, its state is 1, and its dominator number is equal to its node ID.

[0034] In step S12, after each node receives the StateInfo information from its one-hop neighbor, it maintains and updates its two-hop neighbor table, one-hop neighbor state table, and one-hop neighbor dominator table locally.

[0035] like Figure 2 The diagram shows the specific state machine of a node during the exploratory pruning phase. Condition 1 indicates that a QuitRequest broadcast by another node has been received, and condition 2 indicates that a QuitPermission message has not been received. Details are as follows:

[0036] Step S2: After step S1 is completed, the trial cutting stage begins.

[0037] Step S21: The node in state 1 first reads its own number of one-hop neighbors N, and then reads them with a constant independent and identically distributed probability. Once enabled, a node immediately sets its state to 2 and broadcasts a QuitRequest message to its one-hop neighbors. This message serves as a tentative notification to its one-hop neighbors that it wishes to leave the global dominant node set D.

[0038] In step S22, when a node in state 2 receives a QuitRequest broadcast by another node, it immediately abandons broadcasting the QuitRequest information and resets its state to 1, because if two nodes on the same link leave the dominant set D, D may become a non-dominated set. When a node in state 1 receives a QuitRequest broadcast by another node, it immediately sends a QuitPermission message to notify the node that it can leave.

[0039] Step S23: When the exploratory pruning phase time slot ends, if a node in state 2 does not receive the QuitPermission message, it resets its state to 1, because this indicates that there are likely no dominant set nodes around the node.

[0040] like Figure 3 The specific steps of the dominance phase time slot are shown below:

[0041] Step S3: After step S2, the domination phase begins. All nodes in state 1 broadcast the dominate message to their surroundings. Nodes in state 0 or 2, upon receiving the dominate message, update their local dominate point number, which is the source node ID of the received dominate message. If a node in state 0 or 2 does not receive a dominate message during the entire S3 time slot, it sets its dominate point number to -1.

[0042] like Figure 4 The diagram shows the specific state machine of a node in the dominated phase time slot. Condition 3 is that the dominator of a node with state 0 among all its one-hop neighbors is not -1 and its own dominator is not -1. Condition 4 is that there is a node with state 0 and dominator -1 among its one-hop neighbors, or that node itself has a dominator of -1. A detailed explanation follows:

[0043] In step S4, nodes in states 0 and 2, if their locally maintained dominator has changed in step S3, broadcast a Dominated message to notify surrounding nodes of their node state and dominator index. At the end of phase S4, all nodes in state 2 traverse their own one-hop neighbor state table and one-hop neighbor dominator table. If the dominator of any node in state 0 among all one-hop neighbors is not -1 and their own dominator is not -1, then they set their own state to 0. If there is a node in state 0 among their one-hop neighbors with a dominator of -1, or if that node itself has a dominator of -1, then they set their own state to 1 and reset their own dominator index to their own index. At this point, all nodes in state 1 represent the obtained approximate MDS.

[0044] The above four steps are illustrated below with a more specific example. The original network topology is as follows: Figure 5 As shown. Node IDs are displayed next to the nodes. Solid lines represent connections between nodes. Node states are indicated by color: black represents state 1, a diagonal line represents state 2, and white represents state 0. During initialization, each node broadcasts its own state information and simultaneously receives and updates the state information of its one-hop neighbors locally. All nodes are initially in state 1.

[0045] During the exploratory pruning phase, nodes in state 1 first read their own number of one-hop neighbors N, and then read them with a constant independent and identically distributed probability. Once enabled, a node immediately sets its state to 2 and broadcasts a QuitRequest message to its one-hop neighbors. If a node has only one neighbor, it will always broadcast the message and immediately become state 2, such as nodes 1, 3, 6, and 12. If its only adjacent node is also enabled, these nodes will become state 1 due to condition 1, and then immediately broadcast a QuitRequest to their neighboring nodes, causing them to revert to state 1. The node states after a trial pruning phase are as follows: Figure 6 As shown.

[0046] During the domination phase, all nodes in state 1 broadcast the domination message "Dominate" to their surroundings. Nodes in state 0 or 2 update their local dominance point number upon receiving the Dominate message. If a node in state 0 or 2 does not receive the Dominate message during the entire S3 time slot, it sets its dominance point number to -1.

[0047] During the dominated phase, nodes in states 0 and 2 broadcast a Dominated message to neighboring nodes to notify them of their node state and dominated index if their locally maintained dominated index changes. At the end of the dominated phase, all nodes in state 2 traverse their one-hop neighbor state table and one-hop neighbor dominated index table. If the dominated index of any node in state 0 among all one-hop neighbors is not -1 and their own dominated index is not -1, then they set their own state to 0. If there is a node in state 0 among their one-hop neighbors with a dominated index of -1, or if that node itself has a dominated index of -1, then they set their own state to 1 and reset their own dominated index to their own index. At this point, all nodes in state 1 represent the obtained approximate MDS, such as... Figure 7 As shown, nodes 2, 5, 7, 8, 10, and 11 are the approximate MDS we are looking for.

[0048] To reflect the fault tolerance of the algorithm, consider the following: Figure 8 The simplified network topology shown depicts the states of each network node before a certain round of neighbor time-slot awareness. Figure 8As shown in the diagram, node 1, in the previous round, failed to receive the dominance information broadcast by node 2 due to interference and set its dominance point to -1. In this round, during the neighbor information awareness phase, node 2 detects that node 1's dominance point is -1 and its state is 0. Assuming that node 2 wants to exit the MDS during the exploratory pruning phase and successfully receives the QuitPermission information from node 3, even if node 2 still fails to receive the Dominated information from node 1 in this round, it will still change from state 2 to state 1, thus avoiding an error in the construction of the MDS.

Claims

1. A method for constructing an approximate MDS in a mobile ad hoc network based on local information, characterized in that, The algorithm is performed in multiple rounds, each round being divided into four stages: a neighbor information perception stage, a tentative pruning stage, a domination stage, and a dominated stage. Each node has a serial number representing its own ID and maintains a local variable s representing the current state of the node. For node i, s i =0 represents that the node currently does not belong to the domination set D; s i =1 represents that the node currently belongs to the domination set D; s i =2 represents that the last time slot of the node belongs to the domination set D, but this time slot is to be pruned. In the initialization, all nodes in the network belong to the domination set D, so s is set to 1. The specific steps of the MDS construction method based on local information are as follows: The neighbor information sensing stage, step S1, each node broadcasts its own state information, while receiving and locally updating the state information of one-hop neighbors; the specific process is as follows: Step S11, each node broadcasts its own state information StateInfo; the state information StateInfo includes node serial number ID, node one-hop neighbor Onehop, node current state State, and the node's dominator serial number Dominator; when initialized, the state of each node is 1, and the dominator serial number is equal to the node serial number; Step S12, after each node receives the StateInfo information from its one-hop neighbors, it locally maintains and updates its two-hop neighbor table, one-hop neighbor state table, and one-hop neighbor dominator table; After step S1 ends, step S2, the tentative pruning stage, is entered; The time slot node specific state machine of the tentative pruning stage, condition 1 represents that QuitRequest broadcasted by other nodes is received, and condition 2 represents that QuitPermission information is not received: Step S21, the node in state 1 first reads its one-hop neighbor number N, and broadcasts information QuitRequest to its one-hop neighbors with a constant independent and identically distributed probability is enabled; once a node is enabled, it immediately sets its state to 2 and broadcasts information QuitRequest to its one-hop neighbors; Step S22, when a node in state 2 receives QuitRequest broadcasted by other nodes, it immediately gives up broadcasting QuitRequest information and re-sets its own state to 1; when a node in state 1 receives QuitRequest broadcasted by other nodes, it immediately sends a QuitPermission information to notify the node to exit; Step S23, when the tentative pruning stage time slot ends, if a node in state 2 does not receive QuitPermission information, it re-sets its own state to 1; After step S2 ends, step S3, the domination stage, is entered; all nodes in state 1 broadcast domination information Dominate to the surroundings, and nodes in state 0 or 2 update their local dominator serial number after receiving Dominate information, with the dominator serial number being the source node ID of the received Dominate information; if a node in state 0 or 2 does not receive Dominate information during the entire S3 time slot, it sets its own dominator serial number to -1; After step S3 ends, step S4, the dominated stage, is entered; nodes in state 0 and 2 broadcast Dominated information to notify the surrounding nodes of their node state and dominator serial number if the local maintained dominator changes during step S3; At the end of S4 stage, all nodes in state 2 traverse their one-hop neighbor state table and one-hop neighbor dominator table; if all one-hop neighbors in state 0 have a dominator other than -1 and the node's own dominator is not -1, the node's own state is set to 0; if there is a node in state 0 with a dominator of -1 among the one-hop neighbor nodes or the node's own dominator is -1, the node's own state is set to 1 and the node's own dominator is re-set to its own serial number; thus, all nodes in state 1 are the approximate MDS.

2. The approximate MDS construction method of claim 1, wherein the matrix A is subject to The node state machine of the domination phase time slot, condition 3 is that the domination point of all one-hop neighbors in state 0 is not -1 and the domination point of the node itself is not -1, condition 4 is that there is a node in the one-hop neighbors in state 0 and the domination point is -1 or the domination point of the node itself is -1: During initialization, each node broadcasts its own state information, receives and locally updates the state information of one-hop neighbors; all node states are 1; In the domination phase, all nodes in state 1 broadcast the domination information Dominate to the surroundings, and the nodes in state 0 or 2 update their local domination point sequence number after receiving the Dominate information; if the nodes in state 0 or 2 do not receive the Dominate information in the whole S3 time slot, the domination point sequence number of the nodes is set to -1; In the dominated phase, the nodes in state 0 and 2 broadcast the Dominated information to notify the node states and domination point sequence numbers of the nodes in the surroundings if the local maintained domination point changes in the domination phase; At the end of the dominated phase, all nodes in state 2 traverse the one-hop neighbor state table and the one-hop neighbor domination point table, if the domination point of all one-hop neighbors in state 0 is not -1 and the domination point of the node itself is not -1, the state of the node is set to 0; if there is a node in the one-hop neighbors in state 0 and the domination point is -1 or the domination point of the node itself is -1, the state of the node is set to 1 and the domination point of the node is reset to the sequence number of the node itself; thus, all nodes in state 1 are the approximate MDS.

Images