Optimization Method for Combining Computation Offloading and Resource Allocation in a Mobile Edge Computing Network Based on a Hierarchical Game

Abstract

To provide a method of optimizing coupling of computation offloading and resource allocation in a mobile edge computing network based on a hierarchical game.SOLUTION: A method includes: establishing a mobile edge computing computation offloading and a resource allocation model in a multi-user multi-task scenario; performing computation offloading and resource allocation by a multilayer game algorithm; and obtaining an optimal computation offloading plan and resource allocation plan while minimizing the weighted sum of latency and energy consumption. The method includes: determining whether the current task needs computation offloading on the basis of computation resource size necessary for the task and computation capability on the user; determining whether the task is to be divided on the basis of task attribute; making a decision as to whether to perform complete computation offloading or partial computation offloading; and allocating a mobile edge server to offloaded task or subdivided tasks on the basis of the decision on the offloading.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to the technical field of computation offloading and resource allocation in mobile edge, and in particular to an optimization method for the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games. [Background technology]

[0002] Mobile Edge Computing (MEC) is introduced in 5G networks for computing applications in 5G networks. By deploying computing servers at the edge of the network, users can offload computing tasks to the edge servers of the network. By deploying edge servers, users can achieve better performance, avoid sending and storing user data in the core network, and protect users' personal information to a certain extent, and the system can reduce the burden on the core network. However, edge networks cannot guarantee the offload performance of users due to limited resources, a large number of users, and complex network conditions. Therefore, it is necessary to design an efficient computation offloading decision and resource allocation strategy to ensure the service quality for users and system efficiency.

[0003] In real applications, the optimization of the computation offloading decision and the resource allocation can be more consistent with the actual requirements by independently optimizing the computation offloading decision and the resource allocation, but the optimization of the combination of the computation offloading decision and the resource allocation is correspondingly more complicated. Since the resource-constrained mobile device needs to complete the task using the computation resources on the nearby server such as the remote cloud server or the edge server, the method of the efficient and orderly combination of the computation offloading of the mobile device is particularly important. Most of the existing resource allocation problems adopt the time delay or the energy consumption as the evaluation index of the mobile edge computing network. For example, Patent Document 1 (CN116506896A) and Patent Document 2 (CN110062026A) are the combination methods of the mobile edge computing and the resource allocation based on the typical hierarchical game model. In this model, the first layer game is played between the user side and the cluster head MEC side, and each device tries to maximize its own utility, and the second layer game is played between the cluster head MEC side and the follower MEC side, and the resource allocation is combined according to the utility maximization principle based on the available resources. However, the methods disclosed in Patent Documents 1 and 2 do not consider issues such as collaboration and competition between priority-sensitive tasks and servers during the resource allocation process. Taking this situation into account, it is necessary to invent an optimization method for combining multi-task / multi-user computation offloading decision-making and resource allocation based on a hierarchical game when considering collaboration and competition between priority-sensitive tasks and servers. Summary of the Invention [Problem to be solved by the invention]

[0004] The present invention has been made in consideration of the above problems, and aims to provide an optimization method for combining computation offloading and resource allocation in a mobile edge computing network based on hierarchical games, and to solve the problem that resource allocation methods in the existing computation offloading process do not fully consider computation offloading and resource allocation of multi-task, multi-user hierarchical games. [Means for solving the problem]

[0005] The technical means adopted by the present invention are as follows: According to one aspect of the present invention, a method for optimizing a combination of computation offloading and resource allocation in a mobile edge computing network based on a hierarchical game includes the following steps: S1. Establish a mobile edge computing computation offloading and resource allocation model in multi-user and multi-task scenarios; S2. According to the model, a multi-layer game algorithm is used to perform computation offloading and resource allocation, and an optimal computation offloading scheme and resource allocation scheme are obtained under the condition of minimizing the weighted sum of time delay and energy consumption; In the S1, the computation offloading includes: determining whether the current task needs to be offloaded according to the computation resource size required by the task and the computation capability of the user side; determining whether the task needs to be divided according to the task attributes; and determining whether to perform complete computation offloading or partial computation offloading as an offloading decision; In S1, the resource allocation includes allocating a mobile edge server to each offload task or divided subtask based on the offload decision.

[0006] Furthermore, when complete computation offloading is performed, the task granularity is 1 and there is no need to prioritize the tasks, whereas when partial computation offloading is performed, the task granularity is greater than 0 and less than 1, and the method includes a step of dividing the task according to the task granularity and prioritizing the divided subtasks.

[0007] Furthermore, performing computation offloading and resource allocation through a multi-layer game algorithm includes: the multi-layer game algorithm uses a three-layer game resource allocation model, which includes three layers: the user side, the cluster head MEC side, and the lower MEC side; the user side, the cluster head MEC side, and the lower MEC side play a resource game to allocate computation offloading tasks and minimize the weighted sum of time delay and energy consumption; performing calculation on the lower MEC side; allocating resources according to the game result; offloading calculation to the corresponding MEC side based on the allocation result; and returning the task processing result to the user side according to the obtained calculation result to complete the calculation of the offloaded task.

[0008] Furthermore, the three-tier game resource allocation method includes: according to different tasks to be offloaded, the user side, the cluster head MEC side, and the lower MEC side play a non-cooperative game to solve a Nash equilibrium state of the offloading decision; when the Nash equilibrium state is obtained, resource allocation is performed according to the offloading decision in the Nash equilibrium state, and the corresponding MEC side performs corresponding offloading calculation and returns the result to the user side.

[0009] Furthermore, the step of solving the Nash equilibrium state includes: The user side, the cluster head MEC side, and the lower MEC side satisfy the requirements of a three-tiered Stackelberg game, in which the cluster head MEC side is the first leader of the game, the lower MEC side is the second leader of the game, and the user side is the follower of the game, and the requirements of the three-tiered Stackelberg game are as follows:

number

number

number

number

number

[0010] In addition, the corresponding MEC performs the corresponding offloading calculation and returns the result to the user side, specifically, Local execution: When a user does the work locally, the time delay of the local computation is given by

number

number

number

number

number

number

number

number

number

[0011] Furthermore, the entire system satisfies the following constraint requirements:

number

[0012] The present invention has the following beneficial effects compared with the prior art:

[0013] The present invention deploys a multi-server, multi-user, cluster-splitting, edge, computation offloading system, and allocates system resources through a three-tier game, making the allocation effect more efficient and having obvious effects on reducing time delay and energy consumption.

[0014] By combining game theory and evolutionary algorithms, the present invention further rationalizes resource allocation and further optimizes offloading decisions, thereby effectively improving the computing performance of the system. [Brief description of the drawings]

[0015] In order to more clearly explain the technical means in the embodiments of the present invention or the prior art, the accompanying drawings necessary for the description of the embodiments or the prior art will be briefly introduced below. It goes without saying that the drawings below are some embodiments of the present invention, and those skilled in the art can further obtain other drawings from these drawings without creative efforts.

[0016] [Figure 1] FIG. 2 is a frame diagram of the overall model of the present invention. [Diagram 2] FIG. 2 is a model diagram of computation offloading according to the present invention. [Diagram 3] FIG. 1 is a model diagram of the cluster separation game relationship of the present invention. [Figure 4] A comparison of algorithm results when the number of tasks is 50. [Diagram 5] A comparison of algorithm results when the number of servers is 20. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] In order to clarify the purpose, technical means and advantages of the embodiments of the present invention, the technical means of the embodiments of the present invention will be described clearly and completely below with reference to the drawings in the embodiments of the present invention, and it goes without saying that the described embodiments are not all the embodiments but only some of the embodiments of the present invention. Any other embodiments that a person skilled in the art can obtain based on the embodiments of the present invention without any creative effort shall be included in the scope of protection of the present invention.

[0018] It should be noted that terms such as "first" and "second" in the present specification, claims and drawings are used to distinguish between similar objects, and are not used to describe a particular order or priority. It is understood that the embodiments of the present invention described herein can be performed according to orders other than those illustrated herein, and therefore the data thus employed may be interchanged where appropriate. Moreover, the terms "comprise" and "have", as well as any variations thereof, refer to a non-exclusive inclusion. For example, a process, method, system, product or apparatus that includes a series of steps or units need not be limited to those steps or units that are expressly recited, but may further include other steps or units that are not expressly recited or that make the process, method, product or apparatus unique.

[0019] The present invention proposes a technical solution of a method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games, the flow diagram of which is shown in FIG.

[0020] S1. Establish a mobile edge computing computation offloading and resource allocation model in multi-user and multi-task scenarios.

[0021] S11, division of offloading decision: According to the computational resource size required by the task and the computational power of the user side, judge whether the current task needs to perform computation offloading, and judge whether the task needs to be divided according to the task attributes, and further decide whether to perform full computation offloading or partial computation offloading. If full computation offloading is performed, the granularity is 1, and no priority is required. If partial computation offloading is performed, the granularity is greater than 0 and less than 1, and the task needs to be divided according to different granularities of the task.

[0022] When performing partial computation offloading according to different granularity of each task, partial computation offloading assigns priority to the subtasks obtained by dividing each upper task. The present invention adopts a method of combining an integer and a decimal with the first decimal place after zero to divide the task, where the integer part represents the number of the MEC side that executes this task, and the size of the decimal part represents the priority of the subtask. The smaller the decimal part number, the higher the priority, which indicates the priority execution of this subtask, that is, it means that it will be executed preferentially by the MEC side with the strongest task processing ability.

[0023] S12, Resource Allocation: Based on the offloading decision, assign a mobile edge server to each offloading task or divided subtask.

[0024] Resource allocation is performed through a three-layer game, that is, the user side, the cluster head MEC side, and the lower MEC side play a resource game to distribute computation offload tasks, thereby minimizing the overall expenditure, minimizing the weighted sum of time delay and energy consumption, and completing the entire computation offload process efficiently and orderly.

[0025] S121: According to different tasks to be offloaded, the user side, the cluster head MEC side and the subordinate MEC side play a non-cooperative game to solve the Nash equilibrium state of the offloading decision.

[0026] S122, the MEC performs calculations. After the Nash equilibrium state is obtained, the resource allocation is performed according to the offload decision in the Nash equilibrium state, and the corresponding MEC performs the corresponding offload calculations and returns the processing results to the user side.

[0027] S2. Based on the above model, computation offloading and resource allocation are performed by a multi-layer game algorithm, and optimal computation offloading and resource allocation plans are obtained under the condition of minimizing the weighted sum of time delay and energy consumption.

[0028] S21, the game resource allocation is performed according to the requirement that the participants in each game, namely the user side, the cluster head MEC side and the lower MEC side, reach the Nash equilibrium state. The three parties need to satisfy the requirement of the three-layer Stackelberg game, in which the cluster head MEC side is the first leader of the game, the lower MEC side is the second leader of the game, and the user side is the follower of the game. The requirement is expressed by the following formula.

number

[0029] Wherein, N1 represents the cluster head MEC side, which first generates its own optimal pure strategy S1; N2 represents the lower MEC side, which generates the corresponding optimal response strategy S2 based on S1; N3 represents the user side, which generates the corresponding optimal response strategy S3 based on S1 and S2; the optimal response strategy set S=(S1, S2, S3) is called a Stackelberg equilibrium solution of the game; the self-determination variable of N1 is U1; N2 generates the self-determination variable U2 based on S1; N3 generates the self-determination variable U3 based on S1 and S2;

number

number

number

number

[0030] In the process of solving the game using the mixed particle swarm optimization (PSO) algorithm, multiple strategies are generated at each update of the particle swarm, and after each iteration, each game participant wants to save the strategy with the highest utility function. However, since each participant is connected and influenced by each other, the optimal strategy of the leader among the participants may bring bad results to the followers among the participants, that is, it will have different results for various decisions of each participant. Nash equilibrium is achieved by reaching an acceptable stable optimal solution of all decision makers, and if any decision maker does not change his / her decision, it will cause a decrease in utility as long as other decision makers change their decisions. In the algorithm, a game is generated for each participant at each swarm update, and each particle represents a pure strategy of the game. Each particle searches for the optimal strategy in the strategy space and applies the optimal strategy to the game. In the iterative game, each participant adjusts their strategy according to the characteristics of Nash equilibrium to achieve a strategy that satisfies each participant. In each iteration of the algorithm, each particle in the swarm learns to its Nash equilibrium solution partner, and the particles in different participants adjust their strategies through learning so that the particles approach the final Nash equilibrium.

[0031] If there are M edge servers (MEC side) in the system, the set M = {1, 2, 3, , M} represents the edge servers, and if there are N mobile devices (user side) in a static environment, each mobile device has a computation task that needs to be executed, which is denoted as N = {1, 2, 3, , N}. These tasks are represented as a task set T = {T1, T2, T3, , Tn}, and each task Ti is executed in the following three ways.

[0032] Local Execution: When a user performs processing locally, the time delay for local computation is given by

number

[0033] Among them, β i represents the amount of computational data used in task processing, and C lrepresents the number of clocks required for each bit data amount, and f l represents the local computational power, The local computational energy consumption is given by: E local =T local *P local Among them, P. local represents the power consumption of the local device.

[0034] Execution on the cluster head MEC side: When one of the cluster head MEC sides receives a task, the transfer rate of the task is expressed as:

number

[0035] In the formula, d1 represents the transmission distance, ρ1 represents the noise density on the cluster head MEC side, ρ2 represents the noise density on the follower MEC side, ρ0 represents the noise density on the task side, and B e1 represents the broadband resources allocated to the cluster head MEC, and P represents the transmission capacity of the cluster head MEC.

[0036] The time delay of the tasks is given by:

number

[0037] In the formula, α i Task T i represents the amount of input data, and γ i represents the amount of data to be fed back as the calculation results, and P u and P d represent the task upload and task download efforts, respectively, and f e represents the computational power of the cluster head MEC, and C e represents the number of clocks required for each bit data amount.

[0038] The current calculated energy consumption on the MEC side is expressed by the following formula:

number

[0039] In the formula, P mec represents the current calculation effort on the MEC side.

[0040] Follower MEC Execution: When the current cluster head MEC offloads a task to the nearest follower MEC with sufficient resources, the task transfer rate is expressed as follows:

number

[0041] The time delay on the follower MEC side is expressed by the following equation.

number

[0042] The computational energy consumption of a task is expressed as follows:

number

[0043] The total time delay through the system is given by:

number

[0044] The total energy consumption of the entire system is expressed by the following formula:

number

[0045] The total energy consumption function for the entire model is expressed as follows: F sum =w*Tsum +(1-w)*E sum

[0046] where W is an influence factor and the range value of W is from 0 to 1.

[0047] The entire system satisfies the following constraint requirements:

number

[0048] In the formula, α, β, and γ are energy consumption constraint coefficients.

[0049] S22: The edge node (MEC side) performs calculations. It allocates resources according to the game results, and transmits the allocation results to the corresponding calculation node (MEC side) for calculation.

[0050] S23, return the result: according to the obtained calculation result, the task is returned to the user side, and the calculation offload task is completed.

[0051] Experiments are conducted in real task scenarios, and tests are performed according to different scales of mobile edge servers and task quantities. The comparison algorithms of the present invention adopt a weighted particle swarm optimization algorithm (wPSO), an algorithm combining particle swarm optimization algorithm and genetic algorithm (GAPSO), and a game theory algorithm (GT). The above three comparison algorithms are compared with the algorithm combining game theory algorithm and particle swarm optimization algorithm (GTPSO) of the present invention.

[0052] When the number of tasks is 50, the comparative effect of the four algorithms when gradually increasing the server scale is shown in Figure 4. When the number of servers is 20, the comparative effect of the four algorithms when gradually increasing the number of tasks is shown in Figure 5. As shown in Figure 2, taking the task offloading in three clusters as an example, each cluster head will lead the lower MEC side group closest to it. The overall relationship of the game is shown in Figure 3.

[0053] Finally, the following should be noted: The above embodiments are merely for describing the technical means of the present invention, and are not intended to limit the same, and the present invention has been described in detail with reference to the above embodiments. However, the technical means described in the above embodiments can be modified or equivalently replaced in part or all of the technical features, and it is obvious to those skilled in the art that such modifications or replacements do not depart from the essence of the corresponding technical means of the embodiments of the present invention.

[0054] (Additional Note) (Appendix 1) S1. Establish a mobile edge computing computation offloading and resource allocation model in multi-user and multi-task scenarios; S2. According to the model, a multi-layer game algorithm is used to perform computation offloading and resource allocation, and an optimal computation offloading scheme and resource allocation scheme are obtained under the condition of minimizing the weighted sum of time delay and energy consumption; In the S1, the computation offloading includes: determining whether the current task needs to be offloaded according to the computation resource size required by the task and the computation capability of the user side; determining whether the task needs to be divided according to the task attributes; and determining whether to perform complete computation offloading or partial computation offloading as an offloading decision; In the S1, the resource allocation includes allocating a mobile edge server to each offload task or divided subtask based on the offload decision; The present invention provides a method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games, comprising:

[0055] (Appendix 2) In the above S1, When performing complete computation offloading, the task granularity is 1, and there is no need to prioritize tasks. When performing partial computation offloading, the granularity of the task is greater than 0 and less than 1, and the method includes a step of dividing the task according to the granularity of the task and prioritizing the divided subtasks; The method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games as described in Appendix 1, characterized in that:

[0056] (Appendix 3) In S2, performing computation offloading and resource allocation using the multi-layer game algorithm includes: The multi-layer game algorithm uses a three-layer game resource allocation model including three layers: the user side, the cluster head MEC side, and the lower MEC side. The user side, the cluster head MEC side, and the lower MEC side play resource games to allocate computation offload tasks and minimize the weighted sum of time delay and energy consumption. Calculation is performed on the lower MEC side, Allocate resources according to the game results, and communicate the results to the appropriate MEC for calculation. and returning the task processing result to the user side according to the obtained calculation result to complete the calculation of the offloaded task. The method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games as described in Appendix 2.

[0057] (Appendix 4) The three-level game resource allocation method includes: According to different tasks to be offloaded, the user side, the cluster head MEC side, and the subordinate MEC side play a non-cooperative game to solve a Nash equilibrium state of offloading decision; After the Nash equilibrium state is obtained, perform resource allocation according to the offloading decision in the Nash equilibrium state, perform corresponding offloading calculation on the corresponding MEC side, and return the result to the user side. The method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games as described in Supplementary Note 3.

[0058] (Appendix 5) The step of solving the Nash equilibrium state includes: The user side, the cluster head MEC side, and the lower MEC side satisfy the requirements of a three-level Stackelberg game, in which the cluster head MEC side is the first leader of the game, the lower MEC side is the second leader of the game, and the user side is the follower of the game, and the requirements of the three-level Stackelberg game are as follows:

number

number

number

number

number

[0059] (Appendix 6) The corresponding MEC performs the corresponding offloading calculation and returns the result to the user side, specifically, Local execution: When a user does the work locally, the time delay of the local computation is given by

number

number

number

number

number

number

number

number

number

[0060] (Appendix 7) The entire system satisfies the following constraints:

number

Claims

1. S1. Establish a mobile edge computing computation offloading and resource allocation model in multi-user and multi-task scenarios; S2. According to the model, a multi-layer game algorithm is used to perform computation offloading and resource allocation, and an optimal computation offloading plan and resource allocation plan are obtained under the condition of minimizing the weighted sum of time delay and energy consumption; In the S1, the computation offloading includes: determining whether the current task needs to be computationally offloaded according to the computation resource size required for the task and the computation capability of the user side; determining whether the task needs to be divided according to the task attributes; and determining whether to perform complete computation offloading or partial computation offloading as offloading decision; In the S1, the resource allocation includes allocating a mobile edge server to each offload task or divided subtask based on the offload decision; The present invention provides a method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games, comprising:

2. In the above S1, When performing complete computation offloading, the task granularity is 1, and there is no need to prioritize tasks. When performing partial computation offloading, the granularity of a task is greater than 0 and less than 1, and the method includes a step of dividing a task according to the granularity of the task and prioritizing the divided subtasks; The method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games as claimed in claim 1.

3. In S2, performing computation offloading and resource allocation using the multi-layer game algorithm includes: The multi-layer game algorithm uses a three-layer game resource allocation model including three layers: the user side, the cluster head MEC side, and the lower MEC side, and the user side, the cluster head MEC side, and the lower MEC side play resource games to allocate computation offload tasks and minimize the weighted sum of time delay and energy consumption; Calculation is performed on the lower MEC side; Allocate resources according to the game results, and transmit the allocation results to the appropriate MEC for calculation. and returning the task processing result to the user side according to the obtained calculation result to complete the calculation of the offloaded task. The method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games as claimed in claim 2.

4. The three-level game resource allocation method includes: According to different tasks to be offloaded, the user side, the cluster head MEC side, and the lower MEC side play a non-cooperative game to solve a Nash equilibrium state of offloading decision; After the Nash equilibrium state is obtained, perform resource allocation according to the offloading decision in the Nash equilibrium state, perform corresponding offloading calculation on the corresponding MEC side, and return the result to the user side. The method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games as claimed in claim 3.

5. The step of solving the Nash equilibrium state includes: The user side, the cluster head MEC side, and the lower MEC side satisfy requirements of a three-level Stackelberg game in which the cluster head MEC side is the first leader of the game, the lower MEC side is the second leader of the game, and the user side is the follower of the game, and the requirements of the three-level Stackelberg game are as follows: [0010] Among them, N １ represents the cluster head MEC side, and first selects its optimal pure strategy S １ Created N ２ represents the lower MEC side, S １ Based on the corresponding optimal response strategy S ２ Created N ３ represents the user side, and S １ and S ２ Based on the corresponding optimal response strategy S ３ Create an optimal set of response strategies S = (S １ , S ２ , S ３ ) is called a Stackelberg equilibrium solution of the game, and N １ The self-determination variable of U １ and N ２ Is S １ Based on the self-determination variable U ２ Created N ３ Is S １ and S ２ Based on the self-determination variable U ３ Created [0025] and Among them, [0030] But, N １ is the utility function of [0045] But, N ２ is the utility function of [0050] But, N ３ is the utility function of The method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games as claimed in claim 4.

6. The corresponding MEC performs the corresponding offloading calculation and returns the result to the user, specifically, Local execution: When a user does the work locally, the time delay of the local computation is expressed as: [006] Among them, β ｉ represents the amount of calculation data used in task processing, and C ｌ represents the number of clocks required for each bit data amount, and f ｌ represents the local computational power, The local computational energy consumption is given by: E ｌｏｃａｌ =T ｌｏｃａｌ *P ｌｏｃａｌ Among them, P ｌｏｃａｌ represents the power consumption of the local device; Execution on the cluster head MEC side: When one of the cluster head MEC sides receives a task, the transfer rate of the task is expressed by the following formula: [0070] In the formula, d １ represents the transmission distance, and ρ １ represents the noise density on the cluster head MEC side, and ρ ２ represents the noise density on the follower MEC side, and ρ ０ represents the noise density on the task side, and B ｅ１ represents the broadband resource allocated to the cluster head MEC, and P represents the transmission capacity of the cluster head MEC; The time delay of the task is expressed as follows: [0080] In the formula, α ｉ Task T ｉ represents the amount of input data, and γ ｉ represents the amount of data to be fed back as a result of the calculation, and P ｕ and P ｄ f represents the upload and download time of a task, respectively. ｅ represents the computational power of the cluster head MEC, and C ｅ represents the number of clocks required for each bit data amount, The current calculated energy consumption on the MEC side is expressed by the following formula: [0097] In the formula, P ｍｅｃ represents the current computational effort on the MEC side, Follower MEC execution: When the current cluster head MEC offloads a task to the nearest follower MEC with sufficient resources, the transfer rate of the task is expressed as follows: [0010] In the formula, B ｅ２ represents the broadband resource allocated to the follower MEC; The time delay on the follower MEC side is expressed by the following equation: ##EQU00011## The computational energy consumption of a task is given by the following equation: ##EQU00012## The total time delay through the system is given by: ##EQU00013## The total energy consumption of the entire system is expressed by the following formula: ##EQU00014## The total energy consumption function for the entire model is expressed as follows: F ｓｕｍ =w*T ｓｕｍ +(1-w)*E ｓｕｍ where W is an influence factor and the range value of W is from 0 to 1; The method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games as claimed in claim 4.

7. The entire system satisfies the following constraints: ##EQU00015## where α, β, and γ are energy consumption constraint coefficients. The method for optimizing the combination of computation offloading and resource allocation in a mobile edge computing network based on hierarchical games as claimed in claim 6.

Images