Method and system for solving time slot allocation of dynamic frame structure in multi-user MEC network
By dividing multiple frames in a multi-user mobile edge computing network and using the Lagrangian dual method, the optimization problem of minimizing end-to-end average error probability is solved, and the reliability improvement of dynamic frame structure design and time slot allocation method is solved. The problem of this achieves higher network reliability, especially when the number of sensors is small.
Patent Information
- Application Number
- CN202411227108.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-03
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-09-03
AI Technical Summary
The prior art has failed to effectively solve the dynamic frame structure design and time slot allocation method in multi-user mobile edge computing network, resulting in insufficient improvement of network reliability, especially lack of solutions in end-to-end average reliability modeling.
By dividing the delay sensitivity task collected by the sensor into multiple frames, and constructing end-to-end average error probability based on the decoding error probability and delay violation probability, minimizing optimization problems, using Lagrangian dual method to convert from time domain to The channel domain is decomposed into a sub-problem to solve the time slot allocation under the dynamic frame structure.
The reliability of multi-user edge computing network is significantly improved, especially when the number of sensors is small, the effect is more obvious. The time slot allocation method of dynamic frame structure is better than the traditional fixed frame structure.
Smart Images

Figure CN118741609B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of edge computing, and in particular relates to a method and system for solving time slot allocation of a dynamic frame structure of a multi-user MEC network. Background Art
[0002] The introduction and application of mobile edge computing has greatly improved the reliability of time-sensitive applications in wireless communication transmission, especially in the field of industrial Internet of Things, such as multi-machine collaborative operation, and achieved reliable and efficient data transmission and calculation. However, the current mobile edge computing network still does not fully consider the multi-user scenario, especially the dynamic frame structure design and time slot allocation solution in the multi-user scenario. Therefore, the research on dynamic frame structure design and time slot allocation to improve the reliability of multi-user mobile edge computing networks is still a key issue.
[0003] Thanks to the research on the finite code length theorem by Polyanskiy et al., the decoding error probability in the wireless communication transmission process can be well modeled. However, Polyanskiy et al. only analyzed the decoding error probability in the point-to-point wireless communication transmission process, and the end-to-end reliability research is still lacking. With the continuous improvement of hardware performance, the probability of calculation error can often be ignored, but the delay violation probability that the current task cannot be completed on time due to insufficient time still exists. Bennis et al. proposed to use the extreme value theorem to analyze the tail distribution of the queue to make up for this deficiency and completed the modeling of the queue delay violation probability. Although the above work has completed the reliability modeling of point-to-point wireless communication and server computing process, the modeling of the end-to-end average reliability of the edge computing network and the dynamic frame structure time slot allocation method are still lacking, especially a dynamic frame structure design and time slot allocation solution method for improving the reliability of multi-user mobile edge computing networks. Summary of the invention
[0004] Facing the future high-reliability and low-latency edge computing scenarios, the current lack of end-to-end average reliability modeling, especially the time slot allocation method under the dynamic frame structure. Therefore, the patent of this invention provides a dynamic frame structure design and time slot allocation solution method for improving the reliability of multi-user mobile edge computing networks, aiming to improve the reliability of multi-user edge computing networks.
[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:
[0006] The method for solving time slot allocation of a dynamic frame structure of a multi-user MEC network includes the following steps:
[0007] Step 1: Based on the industrial IoT scenario, the entire process of sending the delay-sensitive tasks collected by sensors to the edge computing server is divided intoN Frames, model the decoding error probability of each frame communication transmission process, and characterize the delay violation probability;
[0008] Step 2: Based on the decoding error probability and delay violation probability of the communication transmission process in step 1, the end-to-end error probability within a single frame is constructed, and based on this, the end-to-end average error probability under the multi-frame task is obtained; an optimization problem for minimizing the end-to-end average error probability with time length as the optimization variable is constructed;
[0009] Step 3: Based on the constructed optimization problem of minimizing the end-to-end average error probability, convert the optimization problem from the time domain to the channel domain to obtain its equivalent optimization problem; and use the Lagrange dual method to obtain the dual problem of the optimization problem in the channel domain, and decompose the dual problem into sub-problems with the same number of sub-problems as the number of channels;
[0010] Step 4: By solving the first-order derivative of the objective function of the dual problem sub-problem with respect to the allocation time length, the optimal value of the Lagrange multiplier of each sub-problem is derived, and the time slot allocation solution under the dynamic frame structure is obtained based on the optimal value of the Lagrange multiplier of each sub-problem.
[0011] Furthermore, the industrial Internet of Things scenario in step 1 is:
[0012] The radius of the IIoT area is R , Randomly fix the location of edge computing servers, K Sensors are placed around industrial equipment to collect data on its operation;
[0013] In the communication transmission part, the sensor sends data to the edge computing server using frequency division multiple access.
[0014] Furthermore, the decoding error probability of each frame communication transmission process in step 1 is:
[0015]
[0016] in, For the n The first k The decoding error probability from sensor to server, Indicates n Intraframe k The signal-to-noise ratio from the sensor to the server, For the n Intraframe k The encoding rate from sensor to server, For the n Intraframe k The channel transmission capacity between each sensor and the server; Indicates the n The code length allocated to the communication time within the frame, represents the Q function, Indicates n Intraframe k The channel dispersion between the sensor and the server is is the intermediate variable, For about The probability function of
[0017] For the server computing process, n Frame queue delay violation probability It is expressed as:
[0018]
[0019] in, For the time threshold d The cumulative distribution function of To complete the n The time required for the frame task, is the range parameter, is the shape parameter, G is the GPD distribution, For the n The frame server calculates the duration of the phase, and max() is the maximum value function.
[0020] Furthermore, the end-to-end error probability within a single frame in step 2 is expressed as:
[0021]
[0022] in, For the n The end-to-end error probability within the frame structure, For the n Communication error probability of a frame;
[0023] The end-to-end average error probability under multi-frame tasks is expressed as:
[0024]
[0025] in, is the average end-to-end error probability under multi-frame tasks, K Indicates the maximum number of sensors; n Indicates n frame.
[0026] Furthermore, the optimization problem of minimizing the end-to-end average error probability with time length as the optimization variable in step 2 is for:
[0027]
[0028]
[0029]
[0030]
[0031]
[0032] in, As constraints, For the n The average of the total time spent in the frame communication transmission phase and the server calculation phase, For the n The length of the frame communication transmission phase; For the n The maximum time length given by a frame, represents the average time given, represents a given maximum end-to-end error probability, is the time set allocated to the communication phase for the entire task in the time domain, It is the time set allocated to the calculation stage of the entire task in the time domain. A collection of frames.
[0033] Furthermore, in step 3, the equivalent optimization problem is obtained by converting from the time domain to the channel domain: for:
[0034]
[0035]
[0036]
[0037]
[0038]
[0039] in, is the expected value of the end-to-end error probability during the entire service cycle; For the The average time actually consumed in the communication phase and the calculation phase under each channel; For the The time allocated to the communication phase under each channel; For the The time allocated to the calculation phase under each channel; is the kth sensor in the Decoding error probability under channels; For the The probability of error in the calculation under the channel is For the The maximum time length given under a channel, is the time set allocated to the communication phase for the entire task in the channel domain, is the time set allocated to the calculation stage for the entire task in the channel domain, A collection of channels.
[0040] Furthermore, in step 3, the dual equation obtained by using the Lagrange dual method is:
[0041]
[0042] in, It is a constraint The Lagrange multiplier of , the corresponding pair problem is:
[0043]
[0044] in, Expressed as The corresponding pair of questions;
[0045] , the optimization problem P1 is decomposed into L The sub-problem is expressed as:
[0046]
[0047]
[0048]
[0049]
[0050]
[0051] in, , and the feasible domain of the optimization variable is expressed as:
[0052]
[0053] in, For the The dual equation under channels; For the channel, For the Channel value under each channel; For the End-to-end error probability under channels; is the feasible domain of communication transmission time; The feasible domain of computing time for the server; for K The maximum data in the data packet transmitted by a sensor; The duration of each symbol; is the SNR threshold; is the set of channel gains, Threshold for end-to-end error probability for each frame service; For the The actual time consumed in each channel; The threshold given in the calculation stage, is the given channel gain threshold.
[0054] Furthermore, in step 4, for any given channel, when is determined by solving about and The first-order derivative of obtains the solution of the optimization problem P2, including:
[0055] The first part is to determine the optimal :First, randomly given ; Determine the current The communication time length and calculation time length under the value; update until convergence, and get the optimal ;
[0056] The second part is to determine the most efficient time allocation for each frame under the dynamic frame structure: Get the optimal communication time length and calculation time length;
[0057] According to the optimal solution obtained, substitute about and The end-to-end average error probability is obtained by taking the first-order derivative of .
[0058] On the other hand, the present invention provides a time slot allocation solution system for a multi-user MEC network dynamic frame structure, comprising the following steps:
[0059] Module 1: It is used to send the delay-sensitive tasks collected by sensors to the edge computing server based on the industrial Internet of Things scenario. The whole process is divided into N frames, model the decoding error probability of the communication transmission process, and characterize the delay violation probability;
[0060] Module 2: It is used to construct the end-to-end error probability within a single frame based on the decoding error probability and delay violation probability of the communication transmission process, and based on this, obtain the end-to-end average error probability under the multi-frame task; construct an optimization problem for minimizing the end-to-end average error probability with time length as the optimization variable;
[0061] Module 3: It is used to transform the optimization problem from the time domain to the channel domain based on the constructed optimization problem of minimizing the end-to-end average error probability, and obtain its equivalent optimization problem; and use the Lagrange dual method to obtain the dual problem of the optimization problem in the channel domain, and decompose the dual problem into sub-problems with the same number of channels;
[0062] Module 4: It is used to derive the optimal value of the Lagrange multiplier of each sub-problem by solving the first-order derivative of the objective function of the dual problem sub-problem with respect to the allocation time length, and obtain the time slot allocation solution under the dynamic frame structure based on the optimal value of the Lagrange multiplier of each sub-problem.
[0063] Compared with the prior art, the present invention has the following beneficial effects:
[0064] Compared with the traditional fixed frame structure, the design of the time slot allocation method of the dynamic frame structure in the multi-user MEC network can greatly improve the reliability of the network, and the fewer the number of sensors, the more obvious the increase in reliability. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0066] Figure 1 An edge computing network model including multiple sensors and a single edge computing server proposed by the present invention;
[0067] Figure 2 A schematic diagram of the frame structure of a multi-user edge computing network proposed in the present invention;
[0068] Figure 3 A flowchart of a dynamic frame structure design and time slot allocation solution method for improving the reliability of a multi-user mobile edge computing network proposed by the present invention;
[0069] Figure 4 This is a performance demonstration of a dynamic frame structure design and time slot allocation solution method for improving the reliability of a multi-user mobile edge computing network proposed in the present invention. DETAILED DESCRIPTION
[0070] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0071] Example 1
[0072] This embodiment provides a method for solving time slot allocation of a dynamic frame structure of a multi-user MEC network, including the following steps:
[0073] Step 1: Based on the industrial IoT scenario, the entire process of sending the delay-sensitive tasks collected by sensors to the edge computing server is divided into N Frames, model the decoding error probability of each frame communication transmission process, and characterize the delay violation probability;
[0074] Step 2: Based on the decoding error probability and delay violation probability of the communication transmission process in step 1, the end-to-end error probability within a single frame is constructed, and based on this, the end-to-end average error probability under the multi-frame task is obtained; an optimization problem for minimizing the end-to-end average error probability with time length as the optimization variable is constructed;
[0075] Step 3: Based on the constructed optimization problem of minimizing the end-to-end average error probability, convert the optimization problem from the time domain to the channel domain to obtain its equivalent optimization problem; and use the Lagrange dual method to obtain the dual problem of the optimization problem in the channel domain, and decompose the dual problem into sub-problems with the same number of sub-problems as the number of channels;
[0076] Step 4: By solving the first-order derivative of the objective function of the dual problem sub-problem with respect to the allocation time length, the optimal value of the Lagrange multiplier of each sub-problem is derived, and the time slot allocation solution under the dynamic frame structure is obtained based on the optimal value of the Lagrange multiplier of each sub-problem.
[0077] Figure 1 The industrial Internet of Things scenario is proposed for this invention: including K The edge computing network model of a single edge computing server with multiple sensors. In this embodiment, multiple users refer to multiple sensors. The communication between each sensor and a single edge computing server constitutes an independent communication link. The radius of the proposed multi-user mobile edge computing scenario is R = 150 m, the transmission power of each user is 20 dBm, the bandwidth is 5 MHz, and the path loss model is , the duration of each symbol is set to ms. The maximum error probability is given by In order to model and optimize the effective insecure transmission throughput, the frame structure is as follows: Figure 2 The present invention proposes a dynamic frame structure time slot allocation solution method for improving the reliability of a multi-user mobile edge computing network. The algorithm flow is as follows: Figure 3 As shown;
[0078] Step 1 K A sensor is defined as a combination, expressed as follows:
[0079]
[0080] in, K Indicates the maximum number of sensors in this group. K The sensors are used to monitor industrial equipment and send the collected delay-sensitive tasks to the edge computing server. The total time of the entire task is expressed as The whole process is divided into N frames, and the n The frame length is , n Represented as the index of the frame, this N A frame can be defined as a set, expressed as:
[0081]
[0082] Each frame structure consists of two parts: the communication transmission phase and the server calculation phase. For the nth frame, the corresponding time length is expressed as and , satisfying the following relationship:
[0083]
[0084] The maximum time length under the nth frame;
[0085] In this industrial Internet of Things scenario, the downlink transmission data of the edge computing network is often a logical signal, so it is not included in this downlink transmission process.
[0086] For the communication transmission process, the decoding error probability can be expressed as:
[0087]
[0088] in, For the n The first k The decoding error probability from sensor to server, Indicates n Intraframe kThe signal-to-noise ratio from the sensor to the server, For the n Intraframe k The encoding rate from sensor to server, For the n Intraframe k The channel transmission capacity between each sensor and the server; Indicates the n The code length allocated to the communication time within the frame, represents the Q function, Indicates n Intraframe k The channel dispersion between the sensor and the server is is the intermediate variable, For about The probability function of
[0089] For the server computing process, n Frame queue delay violation probability It is expressed as:
[0090]
[0091] in, For the time threshold d The cumulative distribution function of To complete the n The time required for the frame task, is the range parameter, is the shape parameter, G is the GPD distribution (generalized Pareto distribution), For the n The frame server calculates the duration of the phase, and max() is the maximum value function.
[0092] In step 2, the task execution in the current frame is considered reliable only when the data decoding of all links is successful and the server completes the calculation on time. Therefore, the end-to-end error probability in a single frame structure can be expressed as:
[0093]
[0094] in, For the n The end-to-end error probability within the frame structure, For the n Communication error probability of a frame;
[0095] Since this patent considers the average reliability analysis under a multi-frame structure, the end-to-end average error probability can be expressed as:
[0096]
[0097] in, is the average end-to-end error probability under multi-frame tasks, K Indicates the maximum number of sensors; n Indicates n frame.
[0098] In order to minimize the end-to-end average error probability, that is, to maximize the end-to-end average reliability, the corresponding optimization problem can be expressed as:
[0099]
[0100]
[0101]
[0102]
[0103]
[0104] in, For the n The average of the total time spent in the frame communication transmission phase and the server calculation phase, For the n The length of the frame communication transmission phase; For the n The maximum time length given by a frame, represents the average time given, represents a given maximum end-to-end error probability, is the time set allocated to the communication phase for the entire task in the time domain, It is the time set allocated to the calculation stage of the entire task in the time domain.
[0105] Indicates that the average of the total time spent should not exceed the average of the given total time, This means that the total duration should not exceed the given threshold. The decoding error probability and delay violation error probability are limited to not exceed the threshold.
[0106] In step 3, a set of channel state combinations is first introduced, which is expressed as:
[0107]
[0108] in, represents the joint probability density function of the channel, Indicates this LThe resolution of the combination satisfies the following relationship:
[0109]
[0110] And the above relationship is accurate when L approaches infinity. Therefore, the optimization problem OP established in the time domain can be transformed into an equivalent problem in the channel domain, and the equivalent problem is expressed as:
[0111]
[0112]
[0113]
[0114]
[0115]
[0116] in, is the expected value of the end-to-end error probability during the entire service cycle; For the The average time actually consumed in the communication phase and the calculation phase under each channel; For the The time allocated to the communication phase under each channel; For the The time allocated to the calculation phase under each channel; is the kth sensor in the Decoding error probability under channels; For the The probability of error in the calculation under the channel is For the The maximum time length given under a channel, is the time set allocated to the communication phase for the entire task in the channel domain, is the time set allocated to the calculation phase of the entire task in the channel domain, A collection of channels.
[0117] At this point, the optimization problem P1 is an obvious dynamic programming problem. The Lagrange dual method is used to obtain its dual equation:
[0118]
[0119] in, is the constrained Lagrange multiplier, and its corresponding pair problem is:
[0120]
[0121] because , the dual problem The strong duality of is still guaranteed, then the optimization problem P1 can be decomposed into L Subproblem, denoted as P2:
[0122]
[0123]
[0124]
[0125]
[0126]
[0127] in, , and the feasible domain of the optimization variable can be expressed as:
[0128]
[0129] At this point, the solution to the optimization problem OP can be transformed into the solution of L sub-problems P2.
[0130] For any given channel as described in step 4, when is determined, then the solution to the optimization problem P2 can be solved by about and The first-order derivative of is obtained, and its first-order derivative is as follows:
[0131]
[0132] Other , we can get the following expression:
[0133]
[0134] Obviously, when After determination, the time length of the communication phase can be directly obtained by comparing expressions, but due to the complexity of the Q function, the time length of the communication phase cannot be directly determined and needs to be obtained through iteration. Therefore, this embodiment proposes the following iterative algorithm, which mainly includes two parts.
[0135] The first part is to determine the optimal :First, randomly given ; According to the above expression, determine the current The communication time length and calculation time length under the value; update until convergence. Therefore, the optimal .
[0136] The second part is to determine the optimal time allocation for each frame under the dynamic frame structure. First, the channel state under the current frame and the optimal Substituting into the above expressions, we can get the optimal communication time length and calculation time length.
[0137] According to the optimal solution obtained, the end-to-end average error probability can be obtained by substituting it back into the end-to-end average error probability formula. At this point, the time slot allocation result under the dynamic frame structure of multi-user mobile edge computing can be obtained.
[0138] like Figure 4 As shown in the figure, under different sensor numbers and different transmission distances, the reliability of the time slot allocation method with dynamic frame structure is always better than that with static frame structure, and when the number of sensors is small, the improvement of system reliability by dynamic frame is more obvious.
[0139] Example 2
[0140] This embodiment provides a time slot allocation solution system for a multi-user MEC network dynamic frame structure, including the following steps:
[0141] Module 1: It is used to send the delay-sensitive tasks collected by sensors to the edge computing server based on the industrial Internet of Things scenario. The whole process is divided into N frames, model the decoding error probability of the communication transmission process, and characterize the delay violation probability;
[0142] Module 2: It is used to construct the end-to-end error probability within a single frame based on the decoding error probability and delay violation probability of the communication transmission process, and based on this, obtain the end-to-end average error probability under the multi-frame task; construct an optimization problem for minimizing the end-to-end average error probability with time length as the optimization variable;
[0143] Module 3: It is used to transform the optimization problem from the time domain to the channel domain based on the constructed optimization problem of minimizing the end-to-end average error probability, and obtain its equivalent optimization problem; and use the Lagrange dual method to obtain the dual problem of the optimization problem in the channel domain, and decompose the dual problem into sub-problems with the same number of channels;
[0144] Module 4: It is used to derive the optimal value of the Lagrange multiplier of each sub-problem by solving the first-order derivative of the objective function of the dual problem sub-problem with respect to the allocation time length, and obtain the time slot allocation solution under the dynamic frame structure based on the optimal value of the Lagrange multiplier of each sub-problem.
[0145] It should be understood that parts not elaborated in detail in this specification belong to the prior art.
[0146] It should be understood that the above description of the preferred embodiment is relatively detailed, and it cannot be considered as limiting the scope of protection of the patent of the present invention. It is not necessary and impossible to list all the embodiments here. Under the enlightenment of the present invention, ordinary technicians in this field can also make substitutions or modifications without departing from the scope of protection of the claims of the present invention, which all fall within the scope of protection of the present invention. The scope of protection of the present invention shall be based on the attached claims.
Claims
1. A method for solving time slot allocation of dynamic frame structure in multi-user MEC network, characterized in that: The following steps are involved: Step 1: Based on the industrial IoT scenario, the entire process of sending the delay-sensitive tasks collected by sensors to the edge computing server is divided into N Frames, model the decoding error probability of each frame communication transmission process, and characterize the delay violation probability; Step 2: Based on the decoding error probability and delay violation probability of the communication transmission process in step 1, the end-to-end error probability within a single frame is constructed, and based on this, the end-to-end average error probability under the multi-frame task is obtained; an optimization problem for minimizing the end-to-end average error probability with time length as the optimization variable is constructed; Step 3: Based on the constructed optimization problem of minimizing the end-to-end average error probability, convert the optimization problem from the time domain to the channel domain to obtain its equivalent optimization problem; and use the Lagrange dual method to obtain the dual problem of the optimization problem in the channel domain, and decompose the dual problem into sub-problems with the same number of sub-problems as the number of channels; Step 4: By solving the first-order derivative of the objective function of the dual problem sub-problem with respect to the allocation time length, the optimal value of the Lagrange multiplier of each sub-problem is derived, and the time slot allocation solution under the dynamic frame structure is obtained based on the optimal value of the Lagrange multiplier of each sub-problem.
2. The method for solving time slot allocation of a dynamic frame structure of a multi-user MEC network according to claim 1 is characterized in that: The industrial Internet of Things scenario in step 1 is: The radius of the IIoT area is R , Randomly fix the location of edge computing servers, K Sensors are placed around industrial equipment to collect data on its operation; In the communication transmission part, the sensor sends data to the edge computing server using frequency division multiple access.
3. The method for solving time slot allocation of a dynamic frame structure of a multi-user MEC network according to claim 1, characterized in that: The decoding error probability of each frame communication transmission process in step 1 is: in, For the n The first k The decoding error probability from sensor to server, Indicates n Intraframe k The signal-to-noise ratio from the sensor to the server, For the n Intraframe k The encoding rate from sensor to server, For the n Intraframe k The channel transmission capacity between each sensor and the server; Indicates the n The code length allocated to the communication time within the frame, represents the Q function, Indicates n Intraframe k The channel dispersion between the sensor and the server is is the intermediate variable, About The probability function of For the server computing process, n Frame queue delay violation probability It is expressed as: in, For the time threshold d The cumulative distribution function of To complete the n The time required for the frame task, is the range parameter, is the shape parameter, G is the GPD distribution, For the n The frame server calculates the duration of the phase, and max() is the maximum value function.
4. The method for solving time slot allocation of a dynamic frame structure of a multi-user MEC network according to claim 3 is characterized in that: The end-to-end error probability within a single frame in step 2 is expressed as: in, For the n The end-to-end error probability within the frame structure, For the n Communication error probability of a frame; The end-to-end average error probability under multi-frame tasks is expressed as: in, is the average end-to-end error probability under multi-frame tasks, K Indicates the maximum number of sensors; n Indicates n frame.
5. The method for solving time slot allocation of a dynamic frame structure of a multi-user MEC network according to claim 4 is characterized in that: The optimization problem of minimizing the end-to-end average error probability with time length as the optimization variable in step 2 is for: in, As constraints, For the n The average of the total time spent in the frame communication transmission phase and the server calculation phase, For the n Length of frame communication transmission phase; For the n The maximum time length given by a frame, represents the average time given, represents a given maximum end-to-end error probability, is the time set allocated to the communication phase for the entire task in the time domain, It is the time set allocated to the calculation stage of the entire task in the time domain. A collection of frames.
6. The method for solving time slot allocation of a dynamic frame structure of a multi-user MEC network according to claim 5, characterized in that: In step 3, the equivalent optimization problem is obtained by converting from the time domain to the channel domain. for: in, is the expected value of the end-to-end error probability during the entire service cycle; For the The average time actually consumed in the communication phase and the calculation phase under each channel; For the The time allocated to the communication phase under each channel; For the The time allocated to the calculation phase under each channel; is the kth sensor in the Decoding error probability under channels; For the The probability of error in the calculation under channels is For the The maximum time length given under a channel, is the time set allocated to the communication phase for the entire task in the channel domain, is the time set allocated to the calculation phase of the entire task in the channel domain, A collection of channels.
7. The method for solving time slot allocation of a dynamic frame structure of a multi-user MEC network according to claim 6, characterized in that: In step 3, the dual equation obtained by using the Lagrange dual method is: in, It is a constraint The Lagrange multiplier of , the corresponding pair problem is: in, Expressed as The corresponding pair of questions; , the optimization problem P1 is decomposed into L The sub-problem is expressed as: in, , and the feasible domain of the optimization variable is expressed as: in, For the The dual equation under channels; For the channel, For the Channel value under each channel; For the End-to-end error probability under channels; is the feasible domain of communication transmission time; The feasible domain of computing time for the server; for K The maximum data in the data packet transmitted by a sensor; The duration of each symbol; is the SNR threshold; is the set of channel gains, Threshold for end-to-end error probability for each frame service; For the The actual time consumed in each channel; The threshold given in the calculation stage, is the given channel gain threshold.
8. The method for solving time slot allocation of a dynamic frame structure of a multi-user MEC network according to claim 7, characterized in that: In step 4, for any given channel, when is determined by solving about and The first-order derivative of obtains the solution of the optimization problem P2, including: The first part is to determine the optimal :First, randomly given ; Determine the current The communication time length and calculation time length under the value; update until convergence, and get the optimal ; The second part is to determine the most efficient time allocation for each frame under the dynamic frame structure: Get the optimal communication time length and calculation time length; According to the optimal solution obtained, substitute about and The end-to-end average error probability is obtained by taking the first-order derivative of .
9. A time slot allocation solution system for a dynamic frame structure of a multi-user MEC network, characterized in that: The following steps are involved: Module 1: It is used to send the delay-sensitive tasks collected by sensors to the edge computing server based on the industrial Internet of Things scenario. The whole process is divided into N frames, model the decoding error probability of the communication transmission process, and characterize the delay violation probability; Module 2: It is used to construct the end-to-end error probability within a single frame based on the decoding error probability and delay violation probability of the communication transmission process, and based on this, obtain the end-to-end average error probability under the multi-frame task; construct an optimization problem for minimizing the end-to-end average error probability with time length as the optimization variable; Module 3: It is used to transform the optimization problem from the time domain to the channel domain based on the constructed optimization problem of minimizing the end-to-end average error probability, and obtain its equivalent optimization problem; and use the Lagrange dual method to obtain the dual problem of the optimization problem in the channel domain, and decompose the dual problem into sub-problems with the same number of channels; Module 4: It is used to derive the optimal value of the Lagrange multiplier of each sub-problem by solving the first-order derivative of the objective function of the dual problem sub-problem with respect to the allocation time length, and obtain the time slot allocation solution under the dynamic frame structure according to the optimal value of the Lagrange multiplier of each sub-problem; The time slot allocation solution system for the multi-user MEC network dynamic frame structure is used to execute the steps in the time slot allocation solution method for the multi-user MEC network dynamic frame structure described in any one of claims 1-8.
