A method for node dual motivation for mobile opportunistic networks

Abstract

A node double incentive method for a mobile opportunity network is proposed, a framework based on a graph attention neural network is proposed, the reputation of the node is evaluated through the historical interaction behavior, attribute and the like of the node, the node is stimulated to participate in cooperation based on the reputation of the node, and the first incentive is effectively realized; the contract theory is introduced, after the first reputation incentive, in order to make up for the resource consumption of the node in assisting message transmission and the information asymmetry, a personalized contract is designed and the optimal contract is discussed, the second incentive of the node cooperation transmission reward is carried out, the node utility is maximized, the reputation of the node is evaluated based on the graph attention neural network, the contract is combined, the DIGC double incentive mechanism is proposed, the cooperation transmission between nodes is encouraged, the selfishness of the node is reduced, the delivery rate of the mobile opportunity network is improved, and the data transmission delay is reduced.

Description

Technical Field

[0001] This invention relates to the field of mobile opportunity networks, and more particularly to a method for dual-incentivizing nodes in mobile opportunity networks. Background Technology

[0002] With the rapid development of wireless technology and the increasing number of mobile terminals (such as mobile phones and Bluetooth devices), research on opportunistic networks is booming. Opportunistic networks do not require pre-established communication links; data is transmitted through opportunistic mobile contacts between nodes, enabling flexible data exchange. This characteristic introduces a unique "store-and-carry-forward" routing pattern. In data transmission scenarios within opportunistic networks, successful data transmission from the source node to the target node typically requires the cooperation of multiple relay nodes. Therefore, the willingness of relay nodes to cooperate in forwarding data becomes a key factor for successful data transmission. While ideally, each relay node would selflessly assist in data transmission, many challenges exist in reality. Mobile terminals, as network nodes, may exhibit selfishness due to limitations in energy, network bandwidth, and cache space. These nodes may be unwilling to cooperate in forwarding data for other nodes or may selectively forward data based on their own interests, thus affecting the overall network performance. Therefore, existing technologies lack a reliable incentive mechanism to motivate nodes, resulting in low delivery rates and high data transmission latency in current mobile opportunistic networks. Summary of the Invention

[0003] The purpose of this invention is to provide a dual incentive method for nodes in mobile opportunistic networks, which encourages cooperative transmission among nodes and reduces node selfishness.

[0004] The technical solution adopted by this invention to solve the above-mentioned technical problems is: a node dual-excitation method for mobile opportunistic networks, wherein the mobile opportunistic network includes multiple source nodes, relay nodes, and target nodes. The source nodes transmit data to the relay nodes, and then the data is transmitted sequentially through multiple relay nodes to the target node. The method includes the following steps: Step 1: Initial Incentive S101. Node reputation is evaluated using a graph attention neural network. The node features of the mobile opportunistic network are represented in the graph attention layer as follows: Where n is the number of nodes and F is the feature dimension; S102. Establish the attention coefficient between nodes: (1); In equation (1), a is the attention function implemented by a single-layer feedforward neural network. These are learnable parameters; S103. Normalize the attention coefficients using the softmax function: (2); In equation (2), N i Let i be the set of all neighboring nodes with which it has a communication relationship; S104. Expand and calculate the normalized attention coefficients using the non-linear activation function LeakyReLU: (3); S105. Map the nodes to high-dimensional features to obtain the transformed node feature representation. F' is the dimension of the transformed output: (4); In equation (4), σ(.) is the activation function; S106. Use a multi-head attention mechanism to transform nodes and connect features: (5); In equation (5), f is the number of heads in multi-head attention. The characteristic transformation matrix, To calculate the weight coefficient of the f-th attention; S107, Calculate the reputation value R of the node. i (t): (6); S108. Nodes are distinguished according to reputation thresholds, and high-reputation nodes are selected from the mobile opportunity network to achieve initial incentive; (7); In equation (7), τ is the reputation threshold, T1 is the set of high reputation nodes, and T0 is the set of low reputation nodes; Step 2: Store the reputation value of high-reputation nodes on the blockchain, and send multiple data transmission contracts to the relay nodes. The relay nodes select contracts through the utility functions of the source nodes and relay nodes. S201. The source node utility function is the actual revenue the source node receives from the contract minus the reward the source node receives for transmitting data to the relay node. (8); In equation (8), S i G represents the actual benefit of the source node during transmission. i Let α and β be the reward for the source node in the contract, and α and β be the trade-off parameters. S202, The relay node utility function is the reward a relay node receives for selecting a contract and participating in data transmission, minus the resources consumed by the relay node: (9); In equation (9), O i W is the reward for relay nodes during transmission. i For network resources of relay nodes, λ i For type i nodes, p is the unit cost; In equation (9), For the evaluation function used for rewards: ; S203. After obtaining the utility functions of the source node and relay node, the contract is selected according to the following principles: ; ; Data transmission between nodes is performed according to the selected contract, thus realizing the secondary incentive of nodes in the mobile opportunity network.

[0005] Preferably, step one also includes: S109. Establish a reputation feedback function: ; ; In the above formula, R i (t) represents the reputation value of the node at time t, F i For forwarding behavior, C i (t) is the contribution function, к and η are the reward and penalty factors for forwarding behavior and contribution, respectively, with к and η taking values ​​of (-1, 1), and ζ is the adjustment factor, with a value of (0, 1). C i (t-1) represents the contribution of node at time t-1, Rc i (t) represents the resource consumption of the node at time t; Then the reputation is updated using an update function: ; In the above formula, R i (t-1) represents the node's historical reputation value at time t-1, Rf i (t) represents the reputation feedback value, and γ is the reputation update parameter, with a value of (0, 1).

[0006] According to the above technical solution, the beneficial effects of the present invention are: This invention proposes a framework based on graph attention neural networks. It evaluates node reputation through historical interaction behavior and attributes, using this reputation to incentivize node participation in collaboration and effectively achieve initial incentives. Furthermore, this invention introduces contract theory to address the resource consumption and information asymmetry associated with node-assisted message transmission after the initial reputation incentive. It designs personalized contracts and discusses the optimal contract formulation to provide secondary incentives for node collaborative transmission rewards, maximizing node utility. By combining graph attention neural network-based evaluation of node reputation and contractual incentives, a dual incentive mechanism (DIGC) is proposed to encourage collaborative transmission among nodes, reduce node selfishness, improve delivery rates in mobile opportunity networks, and reduce data transmission latency. Attached Figure Description

[0007] Figure 1 A comparison chart of message delivery rates for mobile opportunity networks; Figure 2 A comparison chart of average message delivery latency for mobile opportunistic networks. Detailed Implementation

[0008] A node dual-incentive method for a mobile opportunistic network, the mobile opportunistic network including multiple source nodes, relay nodes, and target nodes, wherein the source nodes transmit data to relay nodes, and then the data is sequentially transmitted through multiple relay nodes to the target node, the method comprising the following steps: Step 1: Initial Incentive S101. Node reputation is evaluated using a graph attention neural network. The node features of the mobile opportunistic network are represented in the graph attention layer as follows: , where n is the number of nodes and F is the feature dimension.

[0009] S102. Establish the attention coefficient between nodes: (1); In equation (1), a is the attention function implemented by a single-layer feedforward neural network. These are learnable parameters.

[0010] S103. Normalize the attention coefficients using the softmax function: (2); In equation (2), N i Let i be the set of all neighboring nodes with which node i has a communication relationship.

[0011] S104. Expand and calculate the normalized attention coefficients using the non-linear activation function LeakyReLU: (3).

[0012] S105. Map the nodes to high-dimensional features to obtain the transformed node feature representation. F' is the dimension of the transformed output: (4); In equation (4), σ(.) is the activation function.

[0013] S106. Use a multi-head attention mechanism to transform nodes and connect features: (5); In equation (5), f is the number of heads in multi-head attention. The characteristic transformation matrix, To calculate the weight coefficient of the f-th attention.

[0014] S107, Calculate the reputation value R of the node. i (t): (6).

[0015] S108. Nodes are distinguished according to reputation thresholds, and high-reputation nodes are selected from the mobile opportunity network to achieve initial incentive; (7); In equation (7), τ is the reputation threshold, T1 is the set of high reputation nodes, and T0 is the set of low reputation nodes.

[0016] S109. Establish a reputation feedback function: ; ; In the above formula, R i (t) represents the reputation value of the node at time t, F i For forwarding behavior, C i (t) is the contribution function, к and η are the reward and penalty factors for forwarding behavior and contribution, respectively, with к and η taking values ​​of (-1, 1), and ζ is the adjustment factor, with a value of (0, 1). C i (t-1) represents the contribution of node at time t-1, Rc i (t) represents the resource consumption of the node at time t.

[0017] Then the reputation is updated using an update function: ; In the above formula, R i (t-1) represents the node's historical reputation value at time t-1, Rf i (t) represents the reputation feedback value, and γ is the reputation update parameter, with a value of (0, 1).

[0018] Step 2: Store the reputation value of high-reputation nodes on the blockchain, transmit multiple data transmission contracts to the relay nodes, and use the utility functions of the source node and the relay node to enable the relay node to select a contract.

[0019] S201. The source node utility function is the actual revenue the source node receives from the contract minus the reward the source node receives for transmitting data to the relay node. (8); In equation (8), S i G represents the actual benefit of the source node during transmission. i Let α be the reward for the source node in the contract, and β be the trade-off parameters.

[0020] S202, The relay node utility function is the reward a relay node receives for selecting a contract and participating in data transmission, minus the resources consumed by the relay node: (9); In equation (9), O i W is the reward for relay nodes during transmission. i For network resources of relay nodes, λ i For type i nodes, p is the unit cost; In equation (9), For the evaluation function used for rewards: .

[0021] S203. After obtaining the utility functions of the source node and relay node, the contract is selected according to the following principles: ; ; Data transmission between nodes is performed according to the selected contract, thus realizing the secondary incentive of nodes in the mobile opportunity network.

[0022] Figure 1 A comparison chart of message delivery rates for mobile opportunity networks. Figure 2 A comparison chart of average message delivery latency in mobile opportunistic networks, based on... Figure 1 , 2 As shown, the present invention outperforms the comparative method in both message delivery rate and average latency, demonstrating that the present invention can effectively improve message delivery rate and reduce data transmission latency.

Claims

1. A node dual-incentive method for a mobile opportunistic network, the mobile opportunistic network comprising multiple source nodes, relay nodes, and target nodes, wherein the source nodes transmit data to relay nodes, and then the data is sequentially transmitted through multiple relay nodes to the target node, characterized in that, The method includes the following steps: Step 1: Initial Incentive S101. Node reputation is evaluated using a graph attention neural network. The node features of the mobile opportunistic network are represented in the graph attention layer as follows: Where n is the number of nodes and F is the feature dimension; S102. Establish the attention coefficient between nodes: （1）； In equation (1), a is the attention function implemented by a single-layer feedforward neural network. These are learnable parameters; S103. Normalize the attention coefficients using the softmax function: （2）； In equation (2), N i Let i be the set of all neighboring nodes with which it has a communication relationship. S104. Expand and calculate the normalized attention coefficients using the non-linear activation function LeakyReLU: （3）； S105. Map the nodes to high-dimensional features to obtain the transformed node feature representation. F' is the dimension of the transformed output: （4）； In equation (4), σ(.) is the activation function; S106. Use a multi-head attention mechanism to transform nodes and connect features: （5）； In equation (5), f is the number of heads in multi-head attention. The feature transformation matrix, To calculate the weight coefficient of the f-th attention; S107, Calculate the reputation value R of the node. i (t): （6）； S108. Nodes are distinguished according to reputation thresholds, and high-reputation nodes are selected from the mobile opportunity network to achieve initial incentive; （7）； In equation (7), τ is the reputation threshold, T1 is the set of high reputation nodes, and T0 is the set of low reputation nodes; Step 2: Store the reputation value of high-reputation nodes on the blockchain, and send multiple data transmission contracts to the relay nodes. The relay nodes select contracts through the utility functions of the source nodes and relay nodes. S201, The source node utility function is the actual revenue the source node receives from the contract minus the reward the source node receives for transmitting data to the relay node: （8）； In equation (8), S i G represents the actual benefit of the source node during transmission. i Let α and β be the reward for the source node in the contract, and α and β be the trade-off parameters. S202, The relay node utility function is the reward a relay node receives for selecting a contract and participating in data transmission, minus the resources consumed by the relay node: （9）； In equation (9), O i W is the reward for relay nodes during transmission. i For network resources of relay nodes, λ i For type i nodes, p is the unit cost; In equation (9), For the evaluation function used for rewards: ； S203. After obtaining the utility functions of the source node and relay node, the contract is selected according to the following principles: ； ； Data transmission between nodes is performed according to the selected contract, thus realizing the secondary incentive of nodes in the mobile opportunity network.

2. The node dual-excitation method for mobile opportunistic networks according to claim 1, characterized in that, Step one also includes: S109. Establish a reputation feedback function: ； ； In the above formula, R i (t) represents the reputation value of the node at time t, F i For forwarding behavior, C i (t) is the contribution function, к and η are the reward and penalty factors for forwarding behavior and contribution, respectively, with к and η taking values ​​of (-1, 1), and ζ is the adjustment factor, with a value of (0, 1). C i (t-1) represents the contribution of node at time t-1, Rc i (t) represents the resource consumption of the node at time t; Then the reputation is updated using an update function: ； In the above formula, R i (t-1) represents the node's historical reputation value at time t-1, Rf i (t) represents the reputation feedback value, and γ is the reputation update parameter, with a value of (0, 1).

Images