Ultra-dense Internet of Things computing unloading method in converter station scene

By adopting a two-layer game theory greedy offloading solution in ultra-intensive IoT networks and synergistically allocating resources, the processing delay and energy consumption problems of resource scarcity and computing-intensive tasks are solved, and the multi-user and dynamic resource changes are realized. Optimal computing and offloading improves the service quality and experience quality of the network.

CN118900433APending Publication Date: 2024-11-05GUIYANG BUREAU OF CHINA SOUTHERN POWER GRID CO LTD EHV TRANSMISSION CO
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410952033.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-16
Publication Date
2024-11-05

AI Technical Summary

Technical Problem

In ultra-intensive IoT networks, conflicts between resource-scarce mobile applications and constrained IoT MDs, especially the processing delay and energy consumption problems of computing-intensive tasks, existing mobile edge computing offload solutions have failed to effectively solve multiple problems. The changes in user scenarios and dynamic computing resources.

Method used

In the multi-user super-intensive edge server scenario, a two-layer game theory greedy offloading scheme is adopted to jointly allocate computing and wireless resources between the Internet of Things MD, MBS and SC, and establish a system model to optimize the unloading scheme and the double-layer Game theory greedy approximate offloading scheme meets wireless channel constraints while minimizing overall computing overhead.

Benefits of technology

It effectively reduces task processing delay and energy consumption, improves the service quality and experience quality of ultra-intensive IoT networks, and adapts to scenarios where multiple users and dynamic computing resources change.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118900433A_ABST
    Figure CN118900433A_ABST
Patent Text Reader

Abstract

The invention provides an ultra-dense Internet of Things computing unloading method in a converter station scene. Relates to the field of Internet of Things, and aims to solve the problems that many mobile applications requested by the Internet of Things are compute-intensive, high energy consumption is needed and resources are limited. The technical method provided by the invention comprises the step of providing an ultra-dense Internet of Things computing unloading method in a converter station scene in a 6G mobile communication satellite ground network to realize effective utilization of resources. In a multi-user ultra-dense edge server scene, a system model is established according to the energy consumption of the Internet of Things and the processing delay of a calculation task, and the optimal MECO problem in the ultra-dense Internet of Things is modeled as a constraint optimization problem, so that the overall calculation overhead is minimized while the given wireless channel constraint is met. An optimal enumeration unloading scheme is provided to solve the optimal MECO problem in the ultra-dense Internet of Things network, and the scheme can realize the optimal solution of the problem. The invention further provides a double-layer game theory greedy approximate unloading scheme with higher calculation efficiency, wherein cooperative calculation unloading is carried out among the Internet of Things MDs, the edge server connected with the MBS and the SCs.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of ultra-dense Internet of Things, to edge computing, and in particular to an ultra-dense Internet of Things computing unloading method in a converter station scenario. Background Art

[0002] In recent years, with the emergence of IoT technologies such as narrowband IoT and smart IoT, many new IoT applications have emerged and attracted a lot of attention. A large number of IoT mobile devices, such as smart phones, smart bracelets, smart cameras, virtual reality, etc., have been widely used in our lives. On the other hand, with the rapid development of 5G communication technology, it is believed that the deployment of ultra-dense 5G will be the trend of future communication networks. The deployment of large-scale IoT MD and ultra-dense 5G cells have jointly promoted the evolution of the Internet of Things to the future ultra-dense IoT network. Compared with the existing IoT, the ultra-dense IoT has put forward diversified requirements for the communication network in terms of quality of service (QoS) and quality of experience (QoE), and the future ultra-dense IoT faces unprecedented challenges.

[0003] Among these challenges, a fundamental one is how to resolve the conflict between resource-hungry mobile applications and resource-constrained IoT MDs. Many mobile applications requested by IoT MDs are computationally intensive and require high energy consumption, such as fingerprint recognition, face recognition, natural language processing, and interactive games. However, restricted by the physical size of these lightweight IoT MDs, they always have limited computing resources and battery life.

[0004] To solve this problem, Mobile Edge Computing Offload (MECO) is a feasible solution. By offloading the computing tasks of IoTMD to edge servers deployed in wireless access infrastructure, including macro base stations, small cells, WiFiAPs, etc., MECO can significantly reduce the processing latency of tasks and the energy consumed by IoT MD, thereby enhancing QoS and QoE in ultra-dense IoT networks.

[0005] It is worth noting that most of the existing MECO studies consider single-layer BS scenarios, and the computational offloading scheme is very simple. In particular, the MD either performs computational tasks locally on the CPU or offloads some computationally intensive tasks to the edge servers of the MBS for execution. Considering the large number of IoT MDs and mobile applications in ultra-dense IoT networks, MBS is prone to congestion, which seriously affects QoS and QoE. In fact, we can deploy lightweight edge servers on small cells closer to the IoT MDs and offload some computationally intensive tasks to these edge servers, thereby reducing the burden on the MBS. Most of the existing work considers simple computational task scenarios, ignoring the situation where multiple computational tasks are randomly requested by the MD and the computational resources on the edge servers change dynamically.

[0006] Although some studies have focused on the computation offloading problem of multiple edge servers in recent years, most of these studies consider simple computation task scenarios, that is, a fixed number of computation tasks are presupposed during a computation offloading period, and each MD has only one computation task to complete. In addition, the situation where MDs randomly request different types of computation tasks and the computing resources of edge servers change dynamically are also ignored in most existing works.

[0007] Based on the traditional mobile edge computing offloading, the present invention proposes a computing offloading method for ultra-dense IoT networks for multi-user ultra-dense edge server scenarios, uses a two-layer game theory greedy offloading scheme as a solution, and collaboratively allocates computing and wireless resources among the IoT MD, the edge server connected to the MBS, and the SC. Summary of the invention

[0008] The technical problem to be solved by the present invention is to provide a method for offloading the computation of an ultra-dense Internet of Things network, which satisfies wireless requirements while minimizing the overall computational overhead of all tasks. The purpose of the present invention is to design an ultra-dense Internet of Things computation offloading method in a converter station scenario to address the problems of high energy consumption and limited resources of dense Internet of Things.

[0009] The technical solution of the present invention is: a method for unloading ultra-dense Internet of Things computing in a converter station scenario, mainly comprising:

[0010] 1. In a multi-user ultra-dense edge server scenario, consider the case where MD randomly requests different types of computing tasks, and consider the case where tasks randomly arrive at the edge server and the computing resources of the edge server change dynamically. A system model is established based on the energy consumption of IoT MD and the processing delay of computing tasks, and the optimal MECO problem in ultra-dense IoT is modeled as a constrained optimization problem, which aims to minimize the overall computing overhead while satisfying the given wireless channel constraints.

[0011] 2. An optimal enumeration offloading scheme is proposed to solve the optimal MECO problem in ultra-dense IoT networks, which can achieve the optimal solution to the problem.

[0012] 3. A two-layer game theory greedy approximate offloading scheme with higher computational efficiency is further proposed, in which collaborative computation offloading is performed between IoT MDs, edge servers connected to MBS, and SCs. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 Flowchart of an ultra-dense IoT computing offloading method in a converter station scenario

[0014] Figure 2Flowchart for constructing local execution model and mobile edge execution model Detailed implementation manners

[0015] The following will describe in detail the specific implementation manners of the present invention in conjunction with the accompanying drawings. It should be understood that the protection scope of the present invention is not limited by the specific implementation manners. According to an embodiment of the present invention, a method for ultra-dense Internet of Things (IoT) computing offloading in a converter station scenario is provided. Figure 1 The flowchart of a method for ultra-dense IoT computing offloading in a converter station scenario proposed is as follows, including:

[0016] Step 101, construct a multi-user ultra-dense edge server scenario.

[0017] We consider a collaborative mobile edge computing offloading (MECO) scenario in an ultra-dense IoT network, where N co-located IoT mobile devices (MDs) are alternately covered by a macro base station (MBS) and S small cells (SCs). Let N = {1, 2,.., N}, S = {1, 2,..., S} (S ≥ N), and M = {1, 2,..., M} represent the sets of IoT MDs, SCs, and different types of computing tasks (e.g., face recognition, fingerprint recognition, natural language processing, and interactive games, etc.), respectively. For an IoT MD with M types of computing tasks, in an ultra-dense network, by adopting dual connectivity or coordinated multi-point (CoMP) technology, it can communicate with the MBS and access the Internet through a nearby SC simultaneously. We assume that each type of computing task is atomic and cannot be divided. To provide MEC services for MDs, one or more MEC servers are connected to the MBS / SCs through fiber optic links. Without loss of generality, this paper only considers the case where the MEC server is connected to the MBS / SC. Therefore, for any IoT MD i, it can choose at most three offloading strategies to complete its computing tasks, such as local computing on the CPU of the MD, offloading to the MEC server connected to the MBS, and offloading to the MEC server connected to the SC associated with MD i.

[0018] Specifically, for any type-j computing task of MD i, i.e., Ti,j, it can generally be described by two terms, i.e., Ti,j = (mi,j, ci,j), where mi,j is the input data size of task Ti,j and ci,j is the number of CPU cycles required to complete Ti,j. Denote Pi,j as the probability that MD i requests a type-j task at a certain moment, 0 < Pi,j < 1. In addition, we assume that the MBS and each SC have K and L orthogonal radio frequency channels, respectively. Let λi,j ∈ {0, -1, 1, 2,..., S} represent the offloading decision of the j-th type of task at MD i. Assume that during a computing offloading period, the MDs are still covered by the MBS and SCs.

[0019] Step 102, construct a local execution model and a mobile edge execution model.

[0020] Figure 2 A flow chart is constructed for the proposed local execution model and mobile edge execution model of the method, including:

[0021] Step 201, construct a local execution model.

[0022] If MD i decides to execute the jth task locally on its own CPU, that is, when λ i,j = 0, its energy consumption and the processing time of Ti,j can be calculated. Let and denote the computing power of MD i and the energy consumed per CPU cycle, respectively. In this case, the processing time and energy consumption of Ti,j can be calculated as:

[0023]

[0024] Afterwards, we can obtain the corresponding computational cost by using local computation:

[0025]

[0026] Here, ∈[0, 1] are the weight coefficients of the processing time and energy consumption of MD i for making offloading decisions. Obviously, we have + = 1. If, this means that the type-j task of MD i is more sensitive to processing delay, and these applications include face recognition, fingerprint recognition, etc. Otherwise, if MD i is in a very low battery state, we can set even = 1.

[0027] Step 202: construct a mobile edge execution model.

[0028] For the computation offloading scenario in the ultra-dense IoT network, in addition to the local computation on its own CPU, MD i also needs to offload the task Ti,J to the MEC server connected to the MBS. The processing time of type-j task of MD i mainly includes four parts, namely, the transmission time of task Ti,J from MD i to MBS via wireless access, the transmission time of Ti,J from MBS to MEC server via optical fiber link, the queuing time and execution time of task Ti,J on the MEC server. In particular, the processing time of task Ti,J in this case can be calculated as:

[0029]

[0030] Among them, is the uplink data rate of transmitting Ti,J from MD i to MBS through the wireless access channel, and c is the uplink data rate of transmitting Ti,J through the optical fiber link. and are the queuing time and execution time of task Ti,J on the MEC server, respectively. It should be noted that we ignore the time cost of sending the calculation results from the MEC server back to the MD, because for most mobile applications, such as face recognition, interactive games, etc., the calculation results are usually much smaller than the input data volume. In addition, the data rate of the wireless uplink can be calculated by:

[0031]

[0032] Where is the wireless channel bandwidth of MBS and is the channel gain between MD i and MBS.

[0033] We can model the arrival and execution of tasks Ti,J on the MEC server as an M / M / 1 queue. In this case, the service time on the MEC server follows an exponential distribution. Let represent the average arrival rate of computing tasks on the MEC server using the offloading decision representation, which can be expressed as:

[0034]

[0035] Then, the average service time to complete task Ti,J, including the queuing time and the computation execution time, can be expressed as:

[0036]

[0037] The energy consumed in this case can be expressed as:

[0038]

[0039] Finally, the corresponding computational overhead in this case can be obtained as follows:

[0040]

[0041] Step 103, propose the optimal mobile edge computing offloading (OMOU) problem.

[0042] For the MECO scenario in ultra-dense IoT networks, any type-j task of MD i has S+2 offloading decision choices, and one of the tasks with the minimum computational overhead can be easily selected. As more types of computational tasks are offloaded to the MEC server, the wireless uplink data rate will drop sharply due to the wireless interference model, and the average service time to complete task Ti,j will increase synchronously. Therefore, the overall computational overhead obtained will increase significantly. Therefore, there is a trade-off relationship between the overall computational overhead achieved and the number of tasks offloaded to the MEC server. In summary, the problem of collaborative mobile edge computing offloading in ultra-dense IoT networks is to find an optimal computational offloading profile for the tasks of all MDs, minimizing the overall computational overhead while satisfying the wireless channel constraints of MBS and SCs.

[0043] Given the initial information of all IoT MDs and their computational tasks, the computational capabilities of the MEC servers connecting the MBS and SC, the number of wireless channels, and the uplink data rates of the wireless and optical networks, the OMOU problem is to obtain an optimal computational offloading profile for different types of computational tasks of the IoT MDs, while minimizing the overall computational overhead in terms of processing time and energy consumption while satisfying the wireless channel constraints of the MBS and SC.

[0044] OMOU can be expressed mathematically as:

[0045]

[0046] In formula (10), constraint C1 ensures that the total number of tasks offloaded to the MEC server connected to the MBS is less than the number of wireless channels owned by the MBS, and constraint C2 ensures that the total number of tasks offloaded to a certain SC k (1≤k≤S) is not greater than the number of wireless channels owned by SC k. For constraint C3, it indicates that the type-j task of MD i can choose any of the S+2 offloading decision options, such as local calculation on the CPU of MD i, offloading to the MEC server connected to the MBS, offloading to the MEC server connected to the first SC, offloading to the MEC server connected to the second SC, etc.

[0047] To solve the above OMOU problem, all possible offloading decisions for all MDs tasks should be examined, and a globally optimal offloading decision profile should be selected for the tasks, which achieves the overall minimum computational overhead while satisfying the given wireless channel constraints.

[0048] Step 104, construct an optimal enumeration uninstallation solution.

[0049] To solve the OMOU problem, a powerful approach is to enumerate all optional computation offloading decisions for all tasks and select an offloading decision profile for all MDs' tasks that has the overall minimum computational overhead while satisfying the given wireless channel constraints, namely the Optimal Enumeration Offloading Algorithm (OEOA). The specific steps are as follows:

[0050] 1. Enumerate all optional offloading decision combinations of all MD computing tasks;

[0051] 2. Discard the offloading decision combinations that do not satisfy the given wireless channel constraints of MBS and SCs;

[0052] 3. Calculate the total computational cost of the remaining unloading decision combinations and find the overall unloading decision combination S with the minimum total computational cost Ψmin;

[0053] All optional offloading decision choices for all MD computing tasks are enumerated, and the entire solution space is traversed. Therefore, if there is an optimal solution to this problem, it can definitely obtain the optimal solution to the OMOU problem. Considering the very high computational complexity of OEOA, OEOA cannot handle a large number of computing tasks and is only used as a benchmark to verify the effectiveness of our solution below.

[0054] Step 105: construct a greedy approximate unloading scheme based on game theory.

[0055] Since OEOA is not feasible in practical applications with a large number of computing tasks, we further propose a two-layer approximation algorithm to solve the OMOU problem, namely, the game-theoretic greedy approximate offloading algorithm (GT-GAOA). In particular, the outer layer of GT-GAOA uses a random game theory strategy to find an overall optimal offloading decision contour for all MDs, and the inner layer uses a greedy approximation method to find a local optimal offloading strategy for different types of computing tasks of MD i.

[0056] Game theory is widely used as an effective method to solve decision-making problems among multiple game players with different goals. In a finite player game, each rational player will respond to the behavior of other players in the previous step and make a locally optimal decision. After several steps, these players self-organize into a mutually balanced state, namely the Nash equilibrium, in which no player can further reduce his cost by unilaterally changing his strategy.

[0057] In order to solve the OMOU problem, we regard the IoT MD as a rational player, and our problem can be transformed into a multi-user computing offloading game. The OMOU problem is modeled as a strategy game, where the set N of MDs is the set of players in the game, is the strategy set of player i in the game, and is the cost function that player i needs to minimize. After converting the OMOU problem into a multi-user computing offloading game problem, the OMOU problem can be solved.

[0058] All MDs first calculate their local optimal offloading decision profiles in future time slots by calling, that is, obtain the optimal associated SC k of any MD i and the computation offloading decision list of all its tasks with the minimum computation offloading overhead. Then, the MD decides whether their offloading profile should be updated. If the answer is yes, then these MDs will compete for the opportunity to update their offloading decision profiles, otherwise vice versa. It is worth noting that only one MD can get the update opportunity in each time slot, such as MD j. Then, MD j updates its offloading decision profile and broadcasts the number of tasks to be offloaded to the edge servers connecting MBS and SC. At the same time, the offloading decision profiles of other MDs remain unchanged. According to the Nash equilibrium existence theorem, our multi-user computation offloading game, as a finite player game, in which each MD can choose a pure strategy from a finite set of offloading strategies, has a Nash equilibrium.

[0059] Therefore, after a finite number of time slots, when no MD needs to update its offloading decision profile, the algorithm converges and obtains the approximately optimal offloading decision profiles for all MD tasks.

Claims

1. A method for offloading ultra-dense IoT computing in a converter station scenario, characterized in that: The following steps are involved: Step 101, constructing a multi-user ultra-dense edge server scenario; Step 102, constructing a local execution model and a mobile edge execution model; Step 201, constructing a local execution model; Step 202, constructing a mobile edge execution model; Step 103, propose an optimal mobile edge computing offloading (OMOU) problem; Step 104, constructing an optimal enumeration uninstallation solution; Step 105: construct a greedy approximate unloading scheme based on game theory.

2. The ultra-dense Internet of Things computing offloading method in a converter station scenario according to claim 1 is characterized in that: In the multi-user ultra-dense edge server scenario, consider the case where the MD randomly requests different types of computing tasks, and consider the case where the tasks randomly arrive at the edge server and the computing resources of the edge server change dynamically. A system model is established based on the energy consumption of IoT MD and the processing delay of computing tasks, and the optimal MECO problem in ultra-dense IoT is modeled as a constrained optimization problem, which aims to minimize the overall computing overhead while satisfying the given wireless channel constraints.

3. The ultra-dense Internet of Things computing offloading method in a converter station scenario according to claim 1 is characterized in that: An optimal enumeration offloading scheme is proposed to solve the optimal MECO problem in ultra-dense IoT networks, which can achieve the optimal solution to the problem.

4. The method for offloading ultra-dense Internet of Things computing in a converter station scenario according to claim 1 is characterized in that: A two-layer game-theoretic greedy approximate offloading scheme with higher computational efficiency is further proposed, in which collaborative computation offloading is performed between IoT MDs, edge servers connected to MBSs, and SCs.