Node excitation method based on cooperative NOMA-MEC security unloading system

Abstract

The invention discloses a node excitation method based on a cooperative NOMA-MEC secure unloading system. The node excitation method comprises the following steps: collecting idle resource states and channel gain information, distributing calculation tasks, and measuring energy consumption of user nodes and cooperative nodes; a Stackelberg game model is adopted to describe a resource interaction relationship between the user nodes and the cooperative nodes, and according to game actions and energy consumption of two game parties, an optimal unloading strategy is made for the user nodes, and an optimal pricing strategy is made for the cooperative nodes, so that Stackelberg balance is achieved; according to the optimal unloading strategy and the optimal pricing strategy, the calculation task is unloaded to an MEC server in a cooperative NOMA mode, calculation of the unloading task is completed, and a result is returned. According to the method, the optimal resource allocation of the user nodes, the selection of the cooperative nodes and the optimal resource allocation are realized at the same time, the excitation and selection problems of a plurality of selfish cooperative nodes in a cooperative NOMA-MEC security unloading system are solved, and the method accords with a real application scene.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless communication, and relates to a node incentive method based on a cooperative NOMA-MEC secure offloading system. Background Art

[0002] In 5G and the upcoming 6G communication networks, with the rapid development of technologies such as the Internet of Things (IoT), the number and computing requirements of network edge devices have increased sharply. Limited by resources such as limited computing, storage, and battery power, mobile devices are difficult to cope with the increasing high-load and low-latency requirements, significantly affecting their efficiency and responsiveness when performing computationally intensive tasks. Mobile Edge Computing (MEC) proposes a new paradigm. By migrating functions such as computing, caching, and network functions to the network edge, it not only effectively meets the mobile device's demand for computing power expansion but also significantly improves the quality of service of mobile services. In addition, Non-Orthogonal Multiple Access (NOMA) technology allows a group of wireless users to transmit data simultaneously on the same frequency channel and further uses Successive Interference Cancellation (SIC) to mitigate interference between users, and is considered one of the key technologies to achieve ultra-high spectral efficiency and large-scale connection. NOMA can effectively improve the coverage range and multi-access capacity of the MEC server, thereby effectively improving the system spectral efficiency.

[0003] However, considering the broadcast nature of the wireless channel, the offloading transmission process may be affected by eavesdropping attacks, which brings new challenges to the design of NOMA-MEC offloading strategies. Physical Layer Security (PLS) technology utilizes the inherent random properties of the wireless channel (such as fading, noise, and interference, etc.) to ensure that eavesdroppers cannot obtain useful information, providing high-quality security for the NOMA-MEC system. Considering the limited battery power of devices, nodes in the MEC network are often selfish and rational and will not provide communication and computing resources for other users gratuitously. In a cooperative NOMA-MEC secure offloading system, factors such as communication and computing costs within the cooperative nodes will affect the benefits, and competition from other nodes outside the cooperative nodes will also affect the benefits. Therefore, an effective reward payment mechanism must be developed to encourage cooperative nodes to participate in the cooperative process of task offloading. Summary of the Invention

[0004] The present invention aims to solve the technical problem of how to provide collaborative nodes with optimal communication and computing capabilities and minimize the computational offloading cost. The present invention provides a node incentive method based on a collaborative NOMA-MEC secure offloading system, and the technical solution adopted is as follows:

[0005] A node incentive method based on a collaborative NOMA-MEC secure offloading system, comprising the following steps:

[0006] S1. Collect idle resource status and channel gain information, allocate computing tasks according to the idle resource status and the channel gain information, and measure the energy consumption of user nodes and collaborative nodes;

[0007] S2. Use the Stackelberg game model to characterize the resource interaction relationship between the user node and the collaborative node, and formulate an optimal offloading strategy for the user node and an optimal pricing strategy for the collaborative node according to the game actions of both sides of the game and the energy consumption, so as to achieve the Stackelberg equilibrium;

[0008] S3. According to the optimal offloading strategy and the optimal pricing strategy, offload the computing tasks to the MEC server in the manner of collaborative NOMA, and complete the calculation of the offloading tasks and the return of the results.

[0009] In an embodiment of the present invention, the step S1 includes measuring the local computing energy consumption and offloading transmission energy consumption of the user node;

[0010] The local computing energy consumption of the user node is expressed as:

[0011]

[0012] In formula (1), η is the effective capacitance coefficient determined by the device chip architecture, C u represents the number of cycles required for the user node's CPU to locally compute 1-bit task, represents the local computing task volume of the user node, f u represents the CPU frequency of the user node;

[0013] The offloading transmission energy consumption of the user node is expressed as:

[0014]

[0015] In formula (2), p ub and respectively represent the transmission powers of the user node to the base station and the i-th collaborative node, and t off is the length of the first time slot.

[0016] In one embodiment of the present invention, the step S1 includes measuring the local computing energy consumption and offloading transmission energy consumption of the cooperative nodes;

[0017] The local computing energy consumption of the i-th cooperative node is expressed as:

[0018]

[0019] In formula (3), represents the number of cycles required for the CPU of the i-th cooperative node to locally compute 1 bit of task, represents its local computing task volume, represents its CPU frequency;

[0020] The offloading transmission energy consumption of the i-th cooperative node is expressed as:

[0021]

[0022] In formula (4), represents the transmission power of the i-th cooperative node to the base station, and t co is the length of the second time slot.

[0023] In one embodiment of the present invention, the step S1 includes:

[0024] The nodes in the system obtain the status of surrounding nodes by broadcasting their own information and receiving the information broadcast by other nodes;

[0025] Collect the idle resource status and channel gain information of the nodes related to itself in the network, and obtain the channel gain and noise power ratio;

[0026] Make a decision on computing task allocation according to the idle resource status and the channel gain and noise power ratio.

[0027] In one embodiment of the present invention, the step S2 includes: in the Stackelberg game, the user node acts as the leader of the game, selects the cooperative nodes, and decides the amount of tasks offloaded to the cooperative nodes; multiple cooperative nodes act as the followers of the game and bear the unit computing task pricing to obtain benefits.

[0028] In one embodiment of the present invention, in the Stackelberg game, the cost function of the leader is defined as the sum of the energy consumption cost of the user node and the remuneration paid to the cooperative nodes, and is expressed as:

[0029]

[0030] In formula (5), represents the task allocation of the user node, Denote power allocation, \(r\) is a unit vector representing the selection of the optimal cooperative node, \(c\) u represents the unit energy consumption cost of the user node, \(u\) i represents the pricing of the \(i\)-th cooperative node for a unit of computing task;

[0031] The goal of the user node is to minimize the cost function and obtain the optimal offloading strategy. The corresponding optimization problem is defined as:

[0032]

[0033] In formula (6), represents the power upper limit of the user node, and respectively represent the secrecy rates from the user node to the base station and the \(i\)-th cooperative node.

[0034] In an embodiment of the present invention, in the Stackelberg game, for the non-convex optimization problem of the leader, it is decomposed into two sub-problems: time slot and task allocation, and power allocation;

[0035] The time slot and task allocation sub-problem is expressed as:

[0036]

[0037] is a convex optimization problem under the condition of a given vector \(p1\), and the optimal solutions of \(l1\) and \(t\) are obtained using the interior point method; off of;

[0038] The power allocation sub-problem is expressed as:

[0039]

[0040] In formulas (7) and (8), \(\gamma0\) and \(\gamma\) 0,i are auxiliary variables representing the worst eavesdropping situation among multiple eavesdroppers;

[0041] is a non-convex optimization problem under the condition of given \(l1\) and \(t\) off and is transformed into a convex optimization problem through the successive convex approximation algorithm, and the sub-optimal solution of \(p1\) is obtained using the interior point method;

[0042] The time slot and task allocation and the power allocation problems are alternately optimized through an iterative algorithm until the objective function converges.

[0043] In an embodiment of the present invention, in the Stackelberg game, the revenue function of the follower is defined as the difference between the total revenue obtained by the cooperative node from the user node and the energy consumption cost of the cooperative node, expressed as:

[0044]

[0045] In formula (9), represents the task allocation of the \(i\)-th cooperative node, and \(c\) h represents the unit energy consumption cost of the \(i\)-th cooperative node;

[0046] The goal of the cooperative node is to maximize the profit function and obtain the optimal pricing strategy. The corresponding optimization problem is defined as:

[0047]

[0048] In formula (10), represents the power upper limit of the \(i\)-th cooperative node, represents the secrecy rate from the \(i\)-th cooperative node to the base station.

[0049] In one embodiment of the present invention, in the Stackelberg game, for the follower sub-game with multiple node competitions, it is decomposed into two independent sub-problems of profit maximization and cost minimization;

[0050] The cost minimization problem is expressed as:

[0051]

[0052] It is transformed into a convex optimization problem through variable substitution and solved using the interior point method;

[0053] The utility function of the \(i\)-th cooperative node is transformed into a convex function with respect to the unit resource price \(u\) i :

[0054]

[0055] In formula (12), represents the minimum total energy consumption in the offloading process of the \(i\)-th cooperative node;

[0056] The profit maximization problem adopts an equal welfare solution to formulate an optimal price selection algorithm;

[0057] In the initial stage of the game, the user node preferentially selects the node with the lowest price as the cooperative node, and designs the social welfare of the cooperative node as:

[0058]

[0059] In formula (13), \(\lambda\) is used to describe the additional reward part except the original reward after the \(i\)-th cooperative node successfully participates in the cooperation, that is, \(\lambda=\max(u\) up \(-u\) min ,0), where \(u\) up represents the highest price that the user node can afford, \(u\)min represents the minimum selling price of unit resources among all cooperative nodes at the initial stage of the game;

[0060] The bargaining process is continuously repeated until in a certain round of iteration, no node is willing to lower the price, then the Stackelberg equilibrium within all nodes is reached.

[0061] In one embodiment of the present invention, the S3 includes:

[0062] In the first time slot, the user offloads the computing task to the MEC server and the cooperative nodes through NOMA;

[0063] In the second time slot, the cooperative nodes offload the computing tasks received in the first time slot to the MEC server;

[0064] After the calculation of the offloaded task is completed, the result is returned.

[0065] Advantages of the present invention:

[0066] The node incentive method based on the cooperative NOMA-MEC secure offloading system of the present invention adopts the Stackelberg game mechanism to simultaneously achieve the optimal resource allocation of user nodes, the selection of cooperative nodes and the optimal resource allocation; and considers the cooperation relationship between user nodes and cooperative nodes in the network, can utilize the first-mover advantage of the leader, reduce the cost of user nodes, encourage the system to make full use of the computing resources of cooperative nodes, and reduce the energy consumption overhead; also considers the competition relationship between multiple cooperative nodes, solves the incentive and selection problems of multiple selfish cooperative nodes in the cooperative NOMA-MEC secure offloading system, physically conforms to the actual application scenario, and can be effectively applied to engineering practice. Description of the Drawings

[0067] Figure 1 is the flowchart of the node incentive method based on the cooperative NOMA-MEC secure offloading system provided by the embodiment of the present invention;

[0068] Figure 2 is the schematic diagram of the system model of the node incentive method based on the cooperative NOMA-MEC secure offloading system provided by the embodiment of the present invention;

[0069] Figure 3 is the simulation result diagram of the node incentive method based on the cooperative NOMA-MEC secure offloading system provided by the embodiment of the present invention. Detailed Embodiment

[0070] The present invention will be described in detail below in conjunction with the drawings and specific embodiments.

[0071] The present invention is directed to a cooperative NOMA-MEC secure offloading system, and provides a node incentive method based on the cooperative NOMA-MEC secure offloading system, aiming to solve the incentive problem of selfish nodes, and at the same time make selections according to metrics such as node pricing, so as to obtain cooperative nodes that can provide optimal communication and computing capabilities, and minimize the cost of computing offloading.

[0072] Refer to the attached Figure 1 , the node incentive method based on the cooperative NOMA-MEC secure offloading system includes the following steps:

[0073] S1. Collect the idle resource status and channel gain information, allocate computing tasks according to the idle resource status and channel gain information, and measure the energy consumption of user nodes and cooperative nodes;

[0074] S2. Use the Stackelberg game model to characterize the resource interaction relationship between user nodes and cooperative nodes, and formulate an optimal offloading strategy for user nodes and an optimal pricing strategy for cooperative nodes according to the game actions and energy consumption of both sides of the game, so as to achieve the Stackelberg equilibrium;

[0075] S3. According to the optimal offloading strategy and the optimal pricing strategy, offload the computing tasks to the MEC server through the cooperative NOMA method, and complete the calculation of the offloading tasks and the return of the results.

[0076] The nodes in the system include a single user node and multiple cooperative nodes, as shown in the attached Figure 2 . Obtain the status of surrounding nodes by broadcasting its own information and receiving the information broadcast by other nodes, collect the idle resource status and channel gain information of the nodes related to itself in the network, and obtain the channel gain to noise power ratio.

[0077] The channel gain to noise power ratio of the user node and where, |h ub | 2 and |h uh | 2 respectively represent the signal-to-noise ratios at the base station and the cooperative node, and represent the received noise power. The channel gain to noise power ratio of the cooperative node where, |h hb | 2 represents the signal-to-noise ratio at the base station.

[0078] Make a decision on computing task allocation according to the idle resource status, channel gain, and channel gain-to-noise power ratio. The allocation method is reflected in two optimization problems: minimizing the cost function and maximizing the revenue function. The optimization variables include parameters such as task allocation ratio, time allocation ratio, transmission power, and pricing. The corresponding answers are obtained by solving the problems (interior point method). The more idle resources there are, the more inclined to local computing; the larger the channel gain-to-noise power ratio, the more inclined to offloading. However, due to the differences in values (quantitative), there is a trade-off between the two aspects, which is ultimately expressed by minimizing the cost function and maximizing the revenue function. Of course, additional constraints such as competition and a secrecy rate / time / power are also included. Provide a reference for the optimal offloading decision and optimal pricing strategy in step S2.

[0079] In the present invention, the energy consumption includes local computing energy consumption and offloading transmission energy consumption.

[0080] The local computing energy consumption of the user node is expressed as:

[0081]

[0082] In formula (1), η is the effective capacitance coefficient determined by the device chip architecture, C u represents the number of cycles required for the user node's CPU to locally compute 1 bit of task, represents the local computing task volume of the user node, f u represents the CPU frequency of the user node.

[0083] The offloading transmission energy consumption of the user node is expressed as:

[0084]

[0085] In formula (2), p ub and respectively represent the transmission powers from the user node to the base station and the i-th cooperative node, and t off is the length of the first time slot.

[0086] The local computing energy consumption of the i-th cooperative node is expressed as:

[0087]

[0088] In formula (3), represents the number of cycles required for the i-th cooperative node's CPU to locally compute 1 bit of task, represents its local computing task volume, represents its CPU frequency;

[0089] The offloading transmission energy consumption of the i-th cooperative node is expressed as:

[0090]

[0091] In formula (4), represents the transmission power from the i-th cooperative node to the base station, and t co is the length of the second time slot.

[0092] In the Stackelberg game, the user node acts as the leader of the game and makes decisions first. Based on indicators such as the pricing of each cooperative node, the status of idle resources, the channel gain, and the noise power ratio, it selects cooperative nodes and determines the amount of tasks offloaded to the cooperative nodes. Multiple cooperative nodes act as followers of the game and make decisions second. They observe the task allocation strategy of the user node and accordingly determine the pricing of their own unit resource allocation. In the specific pricing process, among the cooperative nodes, they also observe the strategies of other cooperative nodes and adjust their own prices accordingly. While ensuring that they will not incur losses (the reward ratio is lower than the energy consumption cost), they increase the probability of being selected (or enhance their competitive advantage) to obtain benefits by undertaking the pricing of unit computing tasks to obtain benefits. The present invention characterizes the cooperation and competition relationships among nodes in the network through the Stackelberg game, can utilize the first-mover advantage of the leader, reduce the cost of the user node, encourage the system to make full use of the computing resources of the cooperative nodes, and reduce the energy consumption overhead.

[0093] There is no resource interaction among the cooperative nodes, only competition. Specifically, if other cooperative nodes set lower prices, the nodes with lower prices are more likely to be selected by the user node and obtain rewards, while the currently un-reduced-price nodes have a zero income because they are not selected. Therefore, if other nodes reduce their prices, on the premise of ensuring that the income is greater than the energy consumption cost, this node also needs to reduce accordingly.

[0094] In one embodiment of the present invention, the cost function of the leader is defined as the sum of the energy consumption cost of the user node and the reward paid to the cooperative node, expressed as:

[0095]

[0096] In formula (5), represents the task allocation of the user node, represents the power allocation, r is a unit vector representing the selection of the best cooperative node, c u represents the unit energy consumption cost of the user node, u i represents the pricing of the i-th cooperative node for unit computing tasks.

[0097] The goal of the user node is to minimize the cost function. By solving this optimization problem, the user node can obtain approximate optimal solutions for variables such as the computing tasks, transmission power, time, and secrecy rate of the user node until it performs offloading to obtain the optimal offloading strategy. The corresponding optimization problem is defined as:

[0098]

[0099] In formula (6), represents the power upper limit of the user node, and represent the secrecy rates from the user node to the base station and the i-th cooperative node, respectively.

[0100] For the non-convex optimization problem of the leader, it is decomposed into two sub-problems: time slot and task allocation, and power allocation.

[0101] The time slot and task allocation sub-problem is expressed as:

[0102]

[0103] It is a convex optimization problem under the condition of the given vector p1, and the optimal solutions of l1 and t off are obtained by using the interior point method.

[0104] The power allocation sub-problem is expressed as:

[0105]

[0106] In formulas (7) and (8), γ0 and Υ 0,i are auxiliary variables, representing the worst eavesdropping situation among multiple eavesdroppers.

[0107] Under the condition of the given l1 and t off it is a non-convex optimization problem, which is transformed into a convex optimization problem through the successive convex approximation algorithm, and the sub-optimal solution of p1 is obtained by using the interior point method.

[0108] The time slot and task allocation and power allocation problems are alternately optimized through an iterative algorithm. First, solve the time slot and task allocation sub-problem, fix p1 to solve l1 and t off , and hand the results of l1 and t off to the power allocation sub-problem, fix l1 and t off to solve p1, and continuously repeat this iterative process until the objective function converges, then broadcast the solutions of p1, l1 and t off after convergence to the cooperative nodes.

[0109] Specifically, use the interior point method and the successive convex approximation method respectively, and alternately iterate until convergence. The process is as follows:

[0110] (1) Select the initial vectors t0 and l0, set the iteration number i = 0 of the alternating optimization, the corresponding maximum iteration number D0>0 and the precision

[0111] (2) Use the interior point method to solve the time slot and task allocation sub-problem to obtain the solution t i+1and l i+1 ;

[0112] (3) Select the initial vector p0, set the iteration number j = 0 of the SCA, the corresponding maximum iteration number D1>0 and the precision δ1 var >0;

[0113] (4) Solve the convex approximation problem of the problem power allocation sub-problem by the interior point method to obtain the solution p j+1 ;

[0114] (5) Update p i+1 = p j+1 ;

[0115] (6) If or is satisfied, then obtain the approximate solution p i+1 of the power allocation sub-problem; if not satisfied, then let j = j + 1 and return to step (4);

[0116] (7) If or is satisfied, then obtain the solution t i+1 , p i+1 , l i+1 of the original problem; if not satisfied, then let i = i + 1 and return to step (2).

[0117] This algorithm has low complexity and can quickly converge to the sub-optimal solution.

[0118] In an embodiment of the present invention, the revenue function of the follower is defined as the difference between the total revenue obtained by the cooperation node from the user node and the energy consumption cost of the cooperation node, expressed as:

[0119]

[0120] In formula (9), represents the task allocation of the i-th cooperation node, and c h represents the unit energy consumption cost of the i-th cooperation node.

[0121] The goal of the cooperation node is to maximize the revenue function. By solving this optimization problem, the cooperation node can obtain the optimal solution of the pricing and cost of the i-th node (the cost is expressed as the computing task, transmission power, time, and secrecy rate of the cooperation node, and is converted into ) in the next step. The corresponding optimization problem is defined as:

[0122]

[0123] In formula (10), represents the power upper limit of the i-th cooperation node, Denote the secrecy rate of the \(i\)-th cooperative node to the base station.

[0124] For the follower subgame with multiple nodes competing, it is decomposed into two independent sub-problems of maximizing revenue and minimizing cost.

[0125] The cost minimization problem is expressed as:

[0126]

[0127] It can be transformed into a convex optimization problem through variable substitution and solved using the interior point method. The offloading resource allocation of the cooperative node, namely \(p_2\), \(l_2\) and \(t\), is obtained. co .

[0128] Transform the utility function of the \(i\)-th cooperative node into a convex function with respect to the unit resource pricing \(u\) i :

[0129]

[0130] In formula (12), denotes the minimum total energy consumption in the offloading process of the \(i\)-th cooperative node.

[0131] Since there are multiple followers in the subgame, its optimization variable \(u\) i not only depends on the offloading decision of the user node, but is also closely related to other cooperative nodes. Therefore, the revenue maximization problem adopts an equal welfare solution to formulate an optimal price selection algorithm.

[0132] In the initial stage of the game, the user node preferentially selects the node with the lowest price as the cooperative node, and designs the social welfare of the cooperative node as:

[0133]

[0134] In formula (13), \(\lambda\) is used to describe the additional reward part except the original reward after the \(i\)-th cooperative node successfully participates in the cooperation, that is, \(\lambda=\max(u\) up -u min ,0), where \(u\) up represents the highest price that the user node can afford, and \(u\) min represents the minimum value of the unit resource selling price among all cooperative nodes in the initial stage of the game.

[0135] The specific process of the revenue maximization problem is as follows:

[0136] (1) Set the number of cooperative nodes \(m\), the price ceiling \(u\) i of the user node, initialize the iteration number \(k = 0\), and the maximum iteration number \(D_0\geq0\);

[0137] (2) Each collaborative node randomly determines its unit resource price u i and it is necessary to satisfy the constraint

[0138] (3) Traverse all collaborative nodes and record the lowest price u min ;

[0139] (4) If u min ≤u up then the user node selects the collaborative node with the price of u min and broadcasts the value of u min to all nodes; otherwise, exit the algorithm;

[0140] (5) Traverse each collaborative node, calculate the offloading task allocation and revenue of the node, and adjust the node price to maximize its welfare;

[0141] (6) If there is a node that reduces its price, then select the node with the lowest welfare from all nodes as the collaborative node, let k = k + 1, and return to step (5); otherwise, exit the algorithm.

[0142] Continuously repeat the bargaining process until in a certain round of iteration, no node is willing to reduce the price, then the Stackelberg equilibrium within all nodes is reached.

[0143] According to the equal welfare solution algorithm, in the initial stage of the game, the user node preferentially selects the node with the lowest price as the collaborative node, while other nodes with higher prices need to enter the bargaining process. All collaborative nodes need to appropriately reduce the price to improve their competitive advantages in order to increase the probability of being selected by the user node. Continuously repeat the bargaining process until in a certain round of iteration, no node is willing to reduce the price, then the Stackelberg equilibrium within all nodes is reached, ensuring that no node can obtain a better utility by unilaterally changing its strategy.

[0144] Broadcast the optimal price u i of the equal welfare solution algorithm to the user nodes, make offloading decisions again, and continuously repeat this game process until convergence. At this time, neither the user node nor the offloading node can achieve a better utility by changing their own strategies, reaching the Stackelberg equilibrium, and use the converged p1, l1, t off , p2, l2, t co and u i as the execution basis for step S3.

[0145] Through the game actions of both sides of the game according to the rules, the final Stackelberg equilibrium is obtained, that is, the optimal offloading strategy of the user nodes and the optimal pricing strategy of the cooperative nodes obtained in step S2, and the resource allocation for computing task offloading is performed accordingly. The Stackelberg equilibrium of this game exists and is unique, with low complexity and fast convergence.

[0146] According to the optimal offloading strategy, including the computing task allocation ratio, time allocation ratio, transmission power, etc. of the system, and the optimal pricing strategy, the computing tasks are offloaded to the MEC server through the cooperative NOMA method. When the transmission power is sent to the user node for computing task offloading transmission, it is the power required for information transmission in the wireless channel. The time allocation ratio refers to the allocation ratio of the two stages of the game / the offloading stages of the user and the cooperative node during execution. In the first time slot, during the offloading process of the user node in the game, the user offloads the computing task to the MEC server and the cooperative node through NOMA; in the second time slot, during the offloading process of the cooperative node in the game, the cooperative node offloads the computing tasks received in the first time slot to the MEC server, and after the computing of the offloaded tasks is completed, the results are returned.

[0147] To further prove the results of the present invention, simulation verification was carried out. The simulation results are referred to in the appendix Figure 3 , where the comparison schemes include:

[0148] Active interference Stackelberg scheme: The user node schedules an interference node from multiple nodes to send interference signals to hinder eavesdropping. The user node pays a reward for the interference power paid to the interference node to encourage its participation in cooperation, and constructs the optimal power and pricing decisions of the system through the Stackelberg game model.

[0149] Cooperative NOMA-NE scheme: Adopting the same offloading mode as this scheme, the user node decides the offloading ratio, and the cooperative node conducts a price competition. The difference is that there is no prior and posterior relationship in the decision-making between the user node and the cooperative node, that is, all nodes make decisions simultaneously according to the known information to optimize their own utility.

[0150] The simulation results show that compared with different offloading modes, the user cost of the proposed cooperative NOMA-Stackelberg scheme is significantly lower than that of the active interference Stackelberg scheme; compared with different pricing modes, the user cost of the proposed cooperative NOMA-Stackelberg scheme is significantly lower than that of the cooperative NOMA-NE scheme; therefore, the proposed scheme can encourage user nodes to perform more offloading to make full use of the idle resources of cooperative nodes, and at the same time, using reward incentives and node competition can ensure that cooperative nodes actively participate in cooperation while reducing the user node cost.

[0151] The present invention adopts the Stackelberg game mechanism to simultaneously achieve the optimal resource allocation of user nodes, the selection of cooperative nodes and the optimal resource allocation, and takes into account the fairness of scheduling among cooperative nodes. Compared with the prior art solutions, the present invention can simultaneously achieve lower costs of user nodes and higher benefits of cooperative nodes, solve the incentive and selection problems of multiple selfish cooperative nodes, physically conform to the realistic application scenarios, and can be effectively applied to engineering practice.

[0152] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, any modifications, equivalent replacements and improvements made within the spirit and principle of the present invention shall be covered by the protection scope of the present invention.

Claims

1. A node incentive method based on a collaborative NOMA-MEC security offloading system, characterized in that: The steps include: S1. Collect idle resource status and channel gain information, allocate computing tasks according to the idle resource status and the channel gain information, and measure the energy consumption of user nodes and collaborative nodes; S2. Using the Stackelberg game model to describe the resource interaction relationship between the user node and the collaboration node, according to the game actions of both parties and the energy consumption, an optimal unloading strategy is formulated for the user node, and an optimal pricing strategy is formulated for the collaboration node to achieve Stackelberg equilibrium; S3. According to the optimal offloading strategy and the optimal pricing strategy, the computing task is offloaded to the MEC server through collaborative NOMA to complete the calculation of the offloaded task and return the result.

2. According to claim 1, a node incentive method based on a collaborative NOMA-MEC security offloading system is characterized in that: The step S1 includes measuring the local computing energy consumption and the offloading transmission energy consumption of the user node; The local computing energy consumption of the user node is expressed as: In formula (1), η is the effective capacitance coefficient determined by the device chip architecture, C u Indicates the number of cycles required for the user node's CPU to locally calculate a 1-bit task. represents the local computing task of the user node, f u Indicates the CPU frequency of the user node; The offload transmission energy consumption of the user node is expressed as: In formula (2), p ub and denote the transmission power from the user node to the base station and the i-th cooperative node, respectively, and t off is the length of the first time slot.

3. According to a node incentive method based on a collaborative NOMA-MEC security offloading system according to claim 1, it is characterized in that: The step S1 includes measuring the local computing energy consumption and offloading transmission energy consumption of the cooperation node; The local computing energy consumption of the i-th collaborative node is expressed as: In formula (3), represents the number of cycles required for the CPU of the i-th collaborative node to locally calculate a 1-bit task, represents the amount of local computing tasks, Indicates its CPU frequency; The offloading transmission energy consumption of the i-th cooperative node is expressed as: In formula (4), represents the transmission power from the ith cooperative node to the base station, t co is the length of the second time slot.

4. According to claim 1, a node incentive method based on a collaborative NOMA-MEC security offloading system is characterized in that: The step S1 comprises: Nodes in the system obtain the status of surrounding nodes by broadcasting their own information and receiving information broadcast by other nodes; Collect the idle resource status and channel gain information of the nodes related to itself in the network, and obtain the channel gain and noise power ratio; A decision on computing task allocation is made according to the idle resource status and the channel gain to noise power ratio.

5. According to claim 1, a node incentive method based on a collaborative NOMA-MEC security offloading system is characterized in that: The step S2 includes: in the Stackelberg game, the user node acts as a leader of the game to select the collaboration node and decide the amount of tasks to be offloaded to the collaboration node; and the plurality of collaboration nodes act as followers of the game to undertake unit computing task pricing to obtain benefits.

6. A node incentive method based on a collaborative NOMA-MEC security offloading system according to claim 5, characterized in that: In the Stackelberg game, the leader's cost function is defined as the sum of the energy consumption cost of the user node and the reward paid to the cooperative node, expressed as: In formula (5), represents the task allocation of user nodes, represents power allocation, r is a unit vector representing the selection of the best cooperative node, c u Represents the unit energy consumption cost of the user node, u i Represents the pricing of the i-th collaborative node for a unit computing task; The goal of the user node is to minimize the cost function and obtain the optimal unloading strategy. The corresponding optimization problem is defined as: In formula (6), represents the power upper limit of the user node, and They represent the confidentiality rates from the user node to the base station and the i-th cooperative node respectively.

7. A node incentive method based on a collaborative NOMA-MEC security offloading system according to claim 6, characterized in that: In the Stackelberg game, the non-convex optimization problem for the leader is decomposed into two sub-problems: time slot and task allocation and power allocation; The time slot and task allocation subproblem is expressed as: Given the vector p1, it is a convex optimization problem. The interior point method is used to find l1 and t. off The optimal solution of The power allocation sub-problem is expressed as: In formula (7) and formula (8), γ0 and γ 0,i is an auxiliary variable, indicating the worst eavesdropping situation among multiple eavesdroppers; Given l1 and t off Under the condition of , it is a non-convex optimization problem. Through the successive convex approximation algorithm, it is transformed into a convex optimization problem, and the suboptimal solution of p1 is obtained using the interior point method. The time slot and task allocation and the power allocation problem are alternately optimized through an iterative algorithm until the objective function converges.

8. According to claim 5, a node incentive method based on a collaborative NOMA-MEC security offloading system is characterized in that: In the Stackelberg game, the follower's profit function is defined as the difference between the total profit obtained by the cooperation node from the user node and the energy consumption cost of the cooperation node, which is expressed as: In formula (9), represents the task allocation of the i-th collaborative node, c h represents the unit energy consumption cost of the i-th collaborative node; The goal of the collaboration node is to maximize the revenue function and obtain the optimal pricing strategy. The corresponding optimization problem is defined as: In formula (10), represents the power upper limit of the ith cooperative node, represents the confidentiality rate from the i-th cooperative node to the base station.

9. A node incentive method based on a collaborative NOMA-MEC security offloading system according to claim 8, characterized in that: In the Stackelberg game, the follower subgame with multiple nodes competing is decomposed into two independently solved sub-problems: benefit maximization and cost minimization. The cost minimization problem is expressed as: Transform it into a convex optimization problem by variable substitution and solve it using the interior point method; Transform the utility function of the i-th collaboration node into the unit resource pricing u i Convex function of : In formula (12), represents the minimum total energy consumption of the unloading process of the i-th cooperative node; The revenue maximization problem described above adopts an equal welfare solution to formulate an optimal price selection algorithm; In the initial stage of the game, user nodes give priority to selecting nodes with the lowest price as cooperative nodes, and design the social welfare of the cooperative nodes as follows: In formula (13), λ is used to describe the additional reward after the i-th collaborative node successfully participates in the collaboration, in addition to the original reward, that is, λ = max(u up -u min ,0), where u up Represents the highest price that the user node can bear, u min Indicates the minimum price per unit resource among all cooperative nodes at the initial stage of the game; The bargaining process is repeated continuously until no node is willing to lower the price in a certain round of iterations, and the Stackelberg equilibrium is reached within all nodes.

10. A node incentive method based on a collaborative NOMA-MEC security offloading system according to claim 9, characterized in that: The S3 includes: In the first time slot, the user offloads the computing task to the MEC server and the collaboration node through NOMA; In the second time slot, the collaboration node offloads the computing task received in the first time slot to the MEC server; After the calculation of the offloaded task is completed, the result is returned.

Images