A multi-user computing offloading method based on 5G private network splitter

By building the objective function of joint unloading total delay and resource allocation balance in the 5G private network shunt, the weighted resource optimization algorithm is used to solve the problems of multi-user calculation unloading delay and load balancing, and efficient resource allocation and Load balancing improves network availability and economic benefits.

CN114173379BActive Publication Date: 2024-11-12CHONGQING UNIV OF POSTS & TELECOMM
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202111519999.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-13
Publication Date
2024-11-12
Estimated Expiration
2041-12-13

AI Technical Summary

Technical Problem

The prior art is difficult to effectively coordinate the different resource needs of multiple users in the case of multiple mobile users, especially how to solve the problems of total delay minimization and load balancing at the same time when multi-user computing and unloading, and lack practical solutions.

Method used

Using a multi-user computing and unloading method based on 5G private network shunt, the objective function of joint unloading total delay and resource allocation balance is constructed, and the objective function is solved using the weighted resource optimization algorithm to determine the optimal unloading decision and realize the effective allocation of resources. and load balancing.

Benefits of technology

Effectively reduce the average offload delay and transmission delay of multiple users, balance the workload of each mobile edge terminal device or edge server, improve the flexibility and availability of the network, and have good economic benefits.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114173379B_ABST
    Figure CN114173379B_ABST
Patent Text Reader

Abstract

The present invention belongs to the field of wireless communications, and specifically relates to a multi-user computing offloading method based on a 5G private network splitter; the method comprises: constructing a multi-user computing offloading framework based on a 5G private network splitter; constructing an objective function of the total delay of joint offloading and the balance of resource allocation according to the multi-user computing offloading framework; solving the objective function of the total delay of joint offloading and the balance of resource allocation to obtain an optimal solution of the objective function; the user determines an offloading decision according to the optimal solution of the objective function; the user performs task offloading according to the determined offloading decision; the present invention has good scalability and security, effectively utilizes the bandwidth and throughput of each UE device, enhances the network data processing capability, and improves the flexibility and availability of the network; can effectively reduce the average offloading delay and the average transmission delay of multiple users, and at the same time balance the workload of each mobile edge terminal device or edge server, and has good economic benefits.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of wireless communications, and in particular relates to a multi-user computing offloading method based on a 5G private network splitter. Background Art

[0002] In recent years, mobile user devices (mobile phones, tablets, etc.) are becoming important tools for learning, entertainment, social networking and business development, and mobile data traffic has grown rapidly, and this growth is expected to continue in the future. Due to the rapid growth of mobile data traffic, the amount of computing and data exchange between mobile user devices and the cloud is also increasing rapidly. In response to this situation, the industry has introduced Mobile Edge Computing (MEC) technology to supplement and enrich cloud computing. Since mobile edge computing technology directly deploys mobile edge computing servers (MECS) on base stations, mobile user devices that need to offload computing do not necessarily have to choose remote cloud servers, but can adopt a proximity strategy, that is, offload the computing load to the edge server nearby, or directly offload to the local terminal node with other idle resource capabilities, so that mobile users can receive short-distance cloud computing services within the coverage of the wireless access network, allowing users to enjoy a high-quality network experience, that is, mobile edge computing extends computing resources and storage resources from the core wide area network to the edge network.

[0003] The current problem encountered by mobile user devices when offloading local computing tasks to MECS is that the data transmission during the offloading process causes a non-negligible delay between the mobile device and the MECS. When multiple mobile users need to offload computing, it is still difficult to coordinate and handle the different resource requirements of multiple users. In the existing related technologies, most methods use linear regression to predict the processing time of each task on each candidate MECS, and at the same time, according to the previously observed state of the MECS, offload the computing tasks of mobile users to the appropriate MECS. However, when the number of mobile users who need to offload computing is large and there are multiple MECSs that can accept offloading services, there is still a lack of practical solutions to simultaneously solve the problem of minimizing the total delay of multi-user offloading and balancing the load of multiple MECSs. Summary of the invention

[0004] In view of this, the present invention proposes a multi-user computing offloading method based on a 5G private network splitter, the method comprising:

[0005] S1: Build a multi-user computing offloading framework based on 5G private network splitter; the 5G private network splitter includes the user plane function device UPF and the small mobile edge computing platform miniMEP;

[0006] S2: Construct the objective function of the total delay of joint offloading and the balance of resource allocation according to the multi-user computing offloading framework;

[0007] S3: Solve the objective function of the total delay of joint offloading and the balance of resource allocation to obtain the optimal solution of the objective function; the user determines the offloading decision based on the optimal solution of the objective function;

[0008] S4: The user offloads the task according to the determined offloading decision.

[0009] Preferably, constructing a multi-user computing offloading framework based on a 5G private network splitter includes constructing a computing offloading total delay framework and constructing a computing resource allocation balance framework.

[0010] Furthermore, constructing a total latency framework for computing offload includes:

[0011] MiniMEP is used as a mobile edge computing collaborative server to record the resource availability status of each idle service node; j ,c j ) represents the available status of computing resources and storage resources of service node j, where p j represents the computing resources of the service node, c j represents the storage resources of the service node; the intensive computing task of mobile user k is defined as T k , using joint resources to (z k ,e k ) represents the resource demand of mobile user k, where the service node is task T k The computing resources provided are z k , the service node is task T k The storage resources provided are e k ;

[0012] Every t time, mobile user k sends its joint resource demand to (z k ,e k ) is sent to miniMEP;

[0013] MiniMEP calculates the remaining computing resources, remaining storage resources of each service node and the joint resource requirements of mobile users. k ,e k ), calculate the total unloading delay of each mobile user.

[0014] Furthermore, the formula for calculating the total unloading delay is:

[0015]

[0016]

[0017]

[0018] Where f1(x,y) represents the objective function of the total offloading delay, K represents the total number of users, M represents the total number of service nodes, and d kj represents the computing resource margin of the mobile edge computing collaborative server, z i Indicates the computing resources requested by the user, e i represents the storage resources requested by the user, J represents the service node set, x kj Indicates uninstallation judgment parameters.

[0019] Furthermore, the framework for calculating the balance of resource allocation includes:

[0020] Define the concentration parameter u that represents the offloading decision occupying each service node resource, and define the vector A that represents the cumulative load of each Wufu node resource = (a1, a2, a3...a M ), where a j represents the normalized cumulative workload of service node j;

[0021] The resource allocation balance is calculated based on the concentration parameter and the vector of the cumulative load of each service node's resources.

[0022] Furthermore, the cumulative resource load of each service node is j Defined as:

[0023]

[0024] Among them, a j represents the normalized cumulative workload of service node j, k1 represents the emphasis on computing resource usage, k2 represents the emphasis on storage resource usage, K represents the total number of users, and x kj Indicates uninstallation judgment parameters.

[0025] Furthermore, the formula for calculating the balance of resource allocation is:

[0026]

[0027]

[0028]

[0029] Among them, f2(x,y) represents the objective function of resource allocation balance, M represents the total number of service nodes, and y j Indicates busy determination parameter, a j The elements in the vector representing the cumulative load of each service node resource, represents the average value of the cumulative load of resources of each service node, and v represents the mean square error of the cumulative load of resources of each service node.

[0030] Preferably, the objective function of the combined unloading total delay and resource allocation balance is:

[0031] min[f(x,y)]=min[w1f1(x,y)+w2f2(x,y)]

[0032] C1: 0≤w1≤1, 0≤w2≤1;

[0033] C2: w1+w2=1

[0034] C3:

[0035] C4:

[0036] Where f1(x, y) represents the objective function of the total offloading delay, f2(x, y) represents the objective function of the resource allocation balance, w1 represents the weight value assigned to the total offloading delay, w2 represents the weight value of resource allocation using the balance, K represents the total number of users, M represents the total number of service nodes, and z k Indicates that the service node is task T k The computing resources provided, x kj Indicates uninstallation judgment parameter, p j represents the computing resources of the service node, y j Indicates the busy judgment parameter, e k Indicates that the service node is task T k The storage resources provided, c j Represents the storage resources of a service node.

[0037] Preferably, the process of solving the objective function of the total delay of joint unloading and the balance degree of resource allocation includes: using a weighted resource optimization algorithm to solve the objective function, and the solving steps include:

[0038] S1: Determine the weight w1 of the objective function f1(x,y) and the weight w2 of the function f2(x,y);

[0039] S2: Preprocess the objective function and inequality constraints according to weights w1 and w2 to obtain the coefficient matrix x of each variable in the objective function;

[0040] S3: traverse the elements in the coefficient matrix x, and use the elements in the coefficient matrix as the values ​​of intcon in turn;

[0041] S4: Initialize the coefficient matrix A of the inequality constraint and the constraint vector b of the inequality constraint; initialize the coefficient matrix Aeq of the equality constraint and the constraint vector beq of the equality constraint; the optimization interval lb of the variable x takes the all-zero vector, and ub takes the all-one vector;

[0042] S5: Call the intlinprog() function to implement linear programming for the objective function f(x,y) according to the initialized inequality constraint coefficient matrix A, the initialized inequality constraint constraint vector b, the initialized equality constraint coefficient matrix Aeq, the initialized equality constraint constraint vector beq and the value of intcon. The final obtained x is the optimal solution xopt.

[0043] Furthermore, the formula for solving the objective function using the weighted resource optimization algorithm is:

[0044] [x,f(x,y)]=intlinprog(f,intcon,A,b,Aeq,beq,lb,ub)

[0045] A×x≤b

[0046] Aeq×x=beq

[0047] lb≤x≤ub

[0048] Among them, x represents the coefficient matrix of the variables in the objective function, f(x,y) is the objective function, intlinprog() is the function for linear programming of the objective function, intcon represents the position of the integer decision variable in x, A represents the coefficient matrix of the inequality constraint, b represents the constraint vector of the inequality constraint, Aeq represents the coefficient matrix of the equality constraint, beq represents the constraint vector of the equality constraint, lb represents the lower limit of the constraint interval of the variable x, and ub represents the upper limit of the constraint interval of the variable x.

[0049] The beneficial effects of the present invention are as follows: the present invention proposes a multi-user computing offloading method based on 5G private network splitters such as UPF devices and mini MEP, which has good scalability and security. It not only ensures the high availability of industrial Internet equipment through a cluster composed of multiple edge terminal devices with idle resources, but also effectively utilizes the bandwidth and throughput of each UE device through load balancing, enhances the network data processing capability, and improves the flexibility and availability of the network; it can effectively reduce the average unloading delay and average transmission delay of multiple users, and balance the workload of each mobile edge terminal device or edge server, and has good economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a structural diagram of the system framework in the present invention;

[0051] Figure 2 It is a flow chart of a multi-user computing offloading method combining UPF and miniMEP in the present invention;

[0052] Figure 3 To construct a linear optimization objective function flow chart in the present invention;

[0053] Figure 4 This is a flow chart of the weighted resource optimization algorithm in the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0055] The present invention proposes a multi-user computing offloading method based on a 5G private network splitter. The structure of the system framework of the present invention is as follows: Figure 1 As shown in the figure, it mainly includes the overall architecture and the user plane function device UPF (User Plane Function) required in the applicable 5G scenario, the small mobile edge computing platform miniMEP (mini Mobile Edge Platform) and the location of the local terminal device; the base station bound to the miniUPF is used to obtain the data information of the industrial Internet device transmitted by the CPE and the task information that the local device wants to offload, and transmit it to the mini UPF for data screening and processing; the miniUPF device is used to receive and obtain the data information transmitted by the base station and screen and determine, receive the data packets that meet the rules to the factory edge cloud, and the mini MEP collects the load of each CPE and performs corresponding processing, and forwards the non-compliant data packets to the core network, thereby reducing the delay that must be passed through the core network to access the local network. Mini MEP is used as an edge collaborative server for collaborative mobile edge terminals and servers. It will summarize the offload tasks and related information sent from the current user terminal device, determine which edge terminal devices and servers in the local network have sufficient computing and storage resources, and after calculating the objective function, determine which edge terminal device and service to offload it to is the best, thereby solving the problem that the current terminal device does not have sufficient computing task capabilities.

[0056] The overall process of implementing user unloading according to UPF and miniMEP in the present invention is as follows: Figure 2 As shown, specifically, Figure 3As shown, a multi-user computing offloading method based on a 5G private network splitter in the present invention includes:

[0057] S1: Build a multi-user computing offloading framework based on 5G private network splitter; the 5G private network splitter includes UPF equipment and small mobile edge computing platform miniMEP;

[0058] S2: Construct the objective function of the total delay of joint offloading and the balance of resource allocation according to the multi-user computing offloading framework; the total delay of offloading includes the offloading delay and the transmission delay;

[0059] S3: Solve the objective function of the total delay of joint offloading and the balance of resource allocation to obtain the optimal solution of the objective function; the user determines the offloading decision based on the optimal solution of the objective function;

[0060] S4: The user offloads the task according to the determined offloading decision.

[0061] The construction of a multi-user computing offloading framework based on a 5G private network splitter includes constructing a computing offloading total delay framework and constructing a computing resource allocation balance framework; the computing offloading framework based on a 5G private network splitter can allow user data to enter the local enterprise network directly through the UPF splitter without passing through the core network, reducing transmission delay and improving the security of the overall offloading process; constructing a computing offloading total delay framework includes: using miniMEP as a mobile edge computing collaborative server to record the resource availability status of each idle service node; using joint resource pair (p j ,c j ) represents the available status of computing resources and storage resources of service node j, where p j represents the computing resources of the service node, c j represents the storage resources of the service node; the intensive computing task of mobile user k is defined as T k , using joint resources to (z k ,e k ) represents the resource demand of mobile user k, where the service node is task T k The computing resources provided are z k , the service node is task T k The storage resources provided are e k ; The service nodes include mobile edge terminals and edge servers; Every t time, at the beginning of each t time period, mobile user k sends its joint resource demand to (z k ,e k ) is sent to miniMEP; miniMEP calculates (z k ,e k ), and calculate the total offloading delay for each mobile user.

[0062] Computing resource margin of mobile edge computing collaborative server d kj for:

[0063]

[0064] Among them, z i Indicates the computing resources requested by the user, e i Represents the storage resources requested by the user, and J represents the service node set.

[0065] In order to express the service and service relationship between mobile users, mobile edge terminals and edge servers, we introduce the matrix X = X i1×M Represents the overall unloading decision, the matrix has k rows and j columns, and the elements x of the matrix are kj It is defined as the offloading decision parameter, indicating whether mobile user k offloads the computing task to service node j, where k∈K, j∈M. K is the total number of mobile users, M is the total number of service nodes, and the offloading decision parameter x kj The assignment is as follows:

[0066]

[0067] Assume that a mobile terminal can only offload computing tasks to one edge terminal node or edge server at most, so in each row of the matrix X, at most one element has a value of 1, and the rest of the elements in this row are 0. Further, introduce the vector Y=y i1×M To identify whether service node j provides offloading service for mobile users, its element y i (j∈M) is defined as a busy determination parameter, indicating whether service node j provides offload service for mobile users. The busy determination parameter is assigned as:

[0068]

[0069] The framework for building a balanced computing resource allocation includes:

[0070] In order to evaluate the load balancing of the offloading decision, two parameters are defined: one is the concentration parameter u that represents the service node occupied by the offloading scheme, and the other is the vector A = (a1, a2, a3...a M ). The element a of A j (j∈M) and parameter u are both normalized parameters with values ​​in the range of [0,1]. The definition of the concentration parameter u is:

[0071]

[0072] The vector of the accumulated load of each service node resource A=(a1,a2,a3...a M ) element a in j Defined as:

[0073]

[0074] Among them, a j (j∈M) represents the normalized cumulative workload of service node j (taking into account the usage load of computing resources and storage resources), and the normalized cumulative resource loads of all M service nodes a1, a2, a3...a M The cumulative resource load vector A = (a1, a2, a3...a M ); k1 represents the importance attached to the usage of computing resources, k2 represents the importance attached to the usage of storage resources, and the weights k1 and k2 satisfy 0≤k1, k2≤1, and k1+k2=1; preferably, the usage of computing resources and storage resources of edge terminal nodes or edge servers are equally important, that is, k1=k2=0.5.

[0075] The objective functions for constructing the total latency of joint offloading and the balance of resource allocation include:

[0076] First, the total latency of computing offloading for mobile users should be minimized to improve the mobile user experience. Second, the resource usage load on the service nodes should be balanced to avoid the unbalanced computing offloading scheme from exhausting the resources on the service nodes too early. Therefore, the optimization of computing offloading decision is the optimization of the two objectives of total offloading latency and resource allocation balance. It is converted into an integer-based linear optimization problem, and the objective function of the combined total offloading latency and resource allocation balance is constructed:

[0077] min[f(x,y)]=min[w1f1(x,y)+w2f2(x,y)]

[0078] C1: 0≤w1≤1, 0≤w2≤1;

[0079] C2: w1+w2=1

[0080] C3:

[0081] C4:

[0082]

[0083]

[0084]

[0085] Among them, f1(x,y) represents the objective function of the total unloading delay, f2(x,y) represents the objective function of the resource allocation balance, w1 represents the weight value assigned to the total unloading delay, and w2 represents the weight value assigned to the resource utilization balance. represents the average value of the cumulative load of resources of each service node, and v represents the mean square error of the cumulative load of resources of each service node.

[0086] The method for solving the objective function of the total delay of joint offloading and the balance of resource allocation is to use a weighted resource optimization algorithm to solve the objective function, including:

[0087] For a fully distributed computing offloading solution, when all edge terminals or edge servers have exactly the same resource usage load ratio, f2(x, y) achieves the minimum possible value of 0; as the difference in resource usage load ratios of different edge terminals or edge servers becomes larger, and the distribution of edge terminals or edge servers occupied by offloaded computing tends to be concentrated, the value of f2(x, y) also increases, asymptotically reaching 1, at which time the degree of resource usage imbalance is the highest. The constraint on the resource load of each edge terminal or edge server is very simple, that is, the sum of the resources allocated to each mobile user must not exceed its total resource upper limit. The functions f1(x, y) and f2(x, y) are aggregated together using the linear weighted method to obtain the objective function f(x, y) that represents the entire optimization problem. When f(x, y) achieves the minimum value, it indicates that the total offloading delay and the balance between service node resource allocation of this computing offloading solution are generally optimal; such as Figure 4 As shown, the solution steps include:

[0088] S1: Determine the weight w1 of the objective function f1(x,y) and the weight w2 of the function f2(x,y);

[0089] S2: Preprocess the objective function and inequality constraints according to weights w1 and w2 to obtain the coefficient matrix x of each variable in the objective function;

[0090] S3: traverse the elements in the coefficient matrix x, and use the elements in the coefficient matrix as the values ​​of intcon in turn;

[0091] S4: Initialize the coefficient matrix A of the inequality constraint and the constraint vector b of the inequality constraint; initialize the coefficient matrix Aeq of the equality constraint and the constraint vector beq of the equality constraint; the optimization interval lb of the variable x takes the all-zero vector, and ub takes the all-one vector;

[0092] S5: Call the intlinprog() function to implement linear programming for the objective function f(x,y) according to the initialized inequality constraint coefficient matrix A, the initialized inequality constraint constraint vector b, the initialized equality constraint coefficient matrix Aeq, the initialized equality constraint constraint vector beq and the value of intcon. The final obtained x is the optimal solution xopt.

[0093] The formula for solving the objective function using the weighted resource optimization algorithm is:

[0094] [x,f(x,y)]=intlinprog(f,intcon,A,b,Aeq,beq,lb,ub)

[0095] A×x≤b

[0096] Aeq×x=beq

[0097] lb≤x≤ub

[0098] Among them, x represents the coefficient matrix of the variables in the objective function, f(x,y) is the objective function, intlinprog() is the function for linear programming of the objective function, intcon represents the position of the integer decision variable in x, A represents the coefficient matrix of the inequality constraint, b represents the constraint vector of the inequality constraint, Aeq represents the coefficient matrix of the equality constraint, beq represents the constraint vector of the equality constraint, lb represents the lower limit of the constraint interval of the variable x, and ub represents the upper limit of the constraint interval of the variable x.

[0099] The optimal solution xopt is the optimal solution to the objective function of the combined offloading total delay and resource allocation balance. The system then guides each mobile terminal to implement offloading according to the offloading plan corresponding to xopt.

[0100] The present invention proposes a multi-user computing offloading method based on 5G private network splitters such as UPF devices and mini MEP, which has good scalability and security. It not only ensures the high availability of industrial Internet equipment through a cluster composed of multiple edge terminal devices with idle resources, but also effectively utilizes the bandwidth and throughput of each UE device through load balancing, enhances the network data processing capability, and improves the flexibility and availability of the network; it can effectively reduce the average unloading delay and average transmission delay of multiple users, and balance the workload of each mobile edge terminal device or edge server, and has good economic benefits.

[0101] The above embodiments further illustrate the purpose, technical solutions and advantages of the present invention in detail. It should be understood that the above embodiments are only preferred implementation modes of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made to the present invention within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-user computing offloading method based on a 5G private network splitter, characterized in that: include: S1: Build a multi-user computing offloading framework based on 5G private network splitter, including building a computing offloading total latency framework and building a computing resource allocation balance framework; The 5G private network splitter includes the user plane function device UPF and the small mobile edge computing platform miniMEP; The total latency framework for computing offload includes: MiniMEP is used as a mobile edge computing collaborative server to record the resource availability status of each idle service node; j , c j ) represents the available status of computing resources and storage resources of service node j, where p j represents the computing resources of the service node, c j represents the storage resources of the service node; the intensive computing task of mobile user k is defined as T k , using joint resources to (z k , e k ) represents the resource demand of mobile user k, where the service node is task T k The computing resources provided are z k , the service node is task T k The storage resources provided are e k ; Every t time, mobile user k sends its joint resource demand to (z k , e k ) is sent to miniMEP; MiniMEP calculates the remaining computing resources, remaining storage resources of each service node and the joint resource requirements of mobile users. k , e k ) Calculate the total unloading delay of each mobile user; the formula for calculating the total unloading delay is: Where f1(x, y) represents the objective function of the total offloading delay, K represents the total number of users, M represents the total number of service nodes, and d kj represents the computing resource margin of the mobile edge computing collaborative server, z i Indicates the computing resources requested by the user, e i represents the storage resources requested by the user, J represents the service node set, x kj Indicates uninstallation determination parameters; The framework for building a balanced computing resource allocation includes: Define a parameter u that represents the concentration of resources occupied by each service node in the offloading decision, and define a vector A that represents the cumulative load of resources of each service node = (a1, a2, a3...a M ), where a j represents the normalized cumulative workload of service node j; the cumulative resource load a of each service node j The expression is: Among them, a j represents the normalized cumulative workload of service node j, k1 represents the emphasis on computing resource usage, and k2 represents the emphasis on storage resource usage; The resource allocation balance is calculated based on the concentration parameter and the vector of the cumulative load of each service node resource. The resource allocation balance formula is: Among them, f2(x,y) represents the objective function of resource allocation balance, y j Indicates busy determination parameters. represents the average value of the cumulative load of resources of each service node, and v represents the mean square error of the cumulative load of resources of each service node; S2: According to the multi-user computing offloading framework, the objective function of the total delay of joint offloading and the balance of resource allocation is constructed; the objective function of the total delay of joint offloading and the balance of resource allocation is: min[f(x,y)]=min[w1f1(x,y)+w2f2(x,y)] C1: 0≤w1≤1, 0≤w2≤1; C2: w1+w2=1 Wherein, w1 represents the weight value assigned to the total unloading delay, and w2 represents the weight value of resource allocation using the balance degree; S3: Solve the objective function of the total delay of joint offloading and the balance of resource allocation to obtain the optimal solution of the objective function; the user determines the offloading decision based on the optimal solution of the objective function; S4: The user offloads the task according to the determined offloading decision.

2. According to claim 1, a multi-user computing offloading method based on a 5G private network splitter is characterized in that: The process of solving the objective function of the total delay of joint offloading and the balance degree of resource allocation includes: using a weighted resource optimization algorithm to solve the objective function, and the steps of solving include: S1: Determine the weight w1 of the objective function f1(x,y) and the weight w2 of the function f2(x,y); S2: Preprocess the objective function and inequality constraints according to weights w1 and w2 to obtain the coefficient matrix x of each variable in the objective function; S3: traverse the elements in the coefficient matrix x, and use the elements in the coefficient matrix as the values ​​of intcon in turn; S4: Initialize the coefficient matrix A of the inequality constraint and the constraint vector b of the inequality constraint; initialize the coefficient matrix Aeq of the equality constraint and the constraint vector beq of the equality constraint; the optimization interval lb of the variable x takes the all-zero vector, and ub takes the all-one vector; S5: Call the intlinprog() function to implement linear programming for the objective function f(x,y) according to the initialized inequality constraint coefficient matrix A, the initialized inequality constraint constraint vector b, the initialized equality constraint coefficient matrix Aeq, the initialized equality constraint constraint vector beq and the value of intcon, and the final obtained x is the optimal solution xopt; The formula for solving the objective function using the weighted resource optimization algorithm is: [x,f(x,y)]=intlinprog(f,intcon,A,b,Aeq,beq,lb,ub) A×x≤b Aeq×x=beq lb≤x≤ub Among them, x represents the coefficient matrix of the variables in the objective function, f(x,y) is the objective function, intlinprog() is the function for linear programming of the objective function, intcon represents the position of the integer decision variable in x, A represents the coefficient matrix of the inequality constraint, b represents the constraint vector of the inequality constraint, Aeq represents the coefficient matrix of the equality constraint, beq represents the constraint vector of the equality constraint, lb represents the lower limit of the constraint interval of the variable x, and ub represents the upper limit of the constraint interval of the variable x.

Citation Information

Patent Citations

  • MEC-assisted data shunting method in 5G network

    CN108174421A

  • Task unloading and resource allocation joint optimization method based on edge cooperation

    CN111586762A

  • Load balancing method based on fog node cooperation in fog computing network

    CN111641973A

  • Apparatus for constructing energy-efficient communication and computation resource allocation framework for multiple communication service and method thereof

    US20210160888A1