Cooperative service deployment and computing unloading method in mobile edge computing network

Abstract

The invention discloses a cooperative service deployment and computation unloading method in a mobile edge computing network. The method comprises the following steps: S1, establishing a three-layer system model for supporting service deployment in a D2D (Device-to-Device) MEC (Mobile Edge Computing) network; s2, service deployment and unloading task calculation are carried out in the MEC network through a three-layer system model; and S3, completing the minimization of the total cost of the request service under the constraint of the cache resources and the computing resources of the system through the joint optimization of the cache and the computing resources. According to the method, services are cooperatively deployed on a mobile device and an edge server, a service deployment and task allocation model is established, a request service total cost minimization problem is formalized, a local search algorithm is proposed to obtain a service deployment and request mechanism, the problem is solved within polynomial time, and the service cost of an MEC network supporting D2D is reduced. Compared with other benchmark tests, the DMSP service deployment mechanism provided by the invention can effectively reduce the system service cost.

Description

Technical Field

[0001] The present invention belongs to the technical field of service deployment and resource allocation in a mobile edge computing network, and specifically relates to a collaborative service deployment and computing offloading method in a mobile edge computing network. Background Art

[0002] Edge caching caches programs and / or data in MEC networks, which can reduce the frequency of repeated data transmission and task calculations. It is considered to be a cost-effective way to reduce computing latency, energy consumption, and bandwidth costs. In MEC networks, cached items are not limited to data content, but also include data related to the deployed services, such as execution programs, data sets, or libraries. Service deployment deploys service-related data at the edge to execute corresponding tasks on specific edge resources. Therefore, service deployment can be regarded as another form of edge caching.

[0003] Service deployment deploys relevant data on edge nodes so that they can perform corresponding tasks requested by users. To process computing tasks offloaded by users, edge servers need to provide deployed compatible programs and necessary data in advance. For example: in VR / AR services, target data sets and pre-trained models need to be cached in MEC nodes to run target classification and target recognition tasks. When the MEC server does not have data related to the business to be processed, it can request and download the data required for the business from a remote cloud data center, or send the service to the cloud server and complete the computing task there. Both methods will generate delays, the former is the delay of service transmission, and the latter is the delay of task forwarding.

[0004] With the development of communication and manufacturing technologies, the computing and caching capabilities of smart mobile devices are constantly increasing. Mobile devices with free cache resources and computing resources can share their resources with other devices through direct D2D communication. In this way, mobile devices can provide collaborative task processing services, reduce the processing burden of edge servers, and obtain reasonable returns.

[0005] However, the limited cache capacity of mobile devices or edge servers is not enough to cache the data required by all types of services. Only a few services can be deployed on mobile devices, and the number of users who can access the services is also limited. Therefore, it is necessary to optimize deployment decisions and select the services to be deployed and the nodes that provide the services to maximize system performance. Summary of the invention

[0006] In view of the above-mentioned deficiencies in the prior art, the present invention provides a collaborative service deployment and computing offloading method in a mobile edge computing network, which can optimize the location of service deployment and the node problem of processing services, minimize the total cost of services under the constraints of cache and computing resources, and solve the problem that the limited cache capacity of current mobile devices or edge servers is insufficient to cache the data required for all types of services.

[0007] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a collaborative service deployment and computing offloading method in a mobile edge computing network, comprising the following steps:

[0008] S1. Establish a three-layer system model for service deployment in MEC network supporting D2D;

[0009] S2, deploy services and offload computing tasks in the MEC network through a three-layer system model;

[0010] S3. By jointly optimizing cache and computing resources, the cost of requesting services is minimized, and the total cost of requesting services is minimized under the constraints of system cache resources and computing resources.

[0011] Further: in said S1, the three-layer system model includes a remote cloud server, an edge server and a plurality of mobile devices;

[0012] When a mobile device requests a service, it obtains the service from nearby mobile devices through a direct D2D communication link between devices, or obtains the service from an edge server through a wireless channel, or obtains the service from a remote cloud server through a backhaul link.

[0013] Further: In S2, the total cost C of deploying all services on the mobile device deploy The specific expression of (X) is:

[0014]

[0015] In the formula, is the deployment decision of service i on mobile device m, N is the set of mobile devices, is the cost of deploying service i on mobile device m, I is the service set, and X is defined as the service deployment decision.

[0016] The decision to offload computing tasks is as follows:

[0017]

[0018] In the formula, The proportion of service i requested by mobile device n being assigned to mobile device m for processing, and the set of D2D devices that provide services to mobile device n is defined as For service i, the proportion of computing tasks offloaded to the edge server and the remote cloud is defined as and in The proportion of service requests that can be processed satisfies the following formula:

[0019]

[0020] Further: in said S2, the calculation of the offloading cost includes the task transmission cost and the task calculation cost;

[0021] When mobile device n requests a service, the task transmission cost C trans The specific expression of (Y) is:

[0022]

[0023] In the formula, The transmission delay of the offloaded computing task required to serve the requested service, is the transmission energy consumption of the input data, α is the delay weighting coefficient of mobile device n sending service i, β is the energy consumption weighting coefficient of mobile device n sending service i, and the set Y is defined as the total task allocation decision.

[0024] The computational cost of the task offloaded to the device, edge server or remote cloud server is C com The specific expression of (Y) is:

[0025]

[0026] In the formula, The computational latency incurred for executing all computational tasks on mobile device n, C EC (Y) is the task computation cost when the computing task generated by the mobile device requesting service is offloaded to the edge server or remote cloud server, and its specific expression is:

[0027]

[0028] In the formula, The proportion of computing tasks offloaded to the edge server for service i requested by mobile device n, b i is the input data size of the requested service i, The proportion of computing tasks offloaded to the remote cloud by service i requested by mobile device n, and c0 is the computing cost per unit data using the edge cloud or cloud.

[0029] Further: In S3, the expression of the total cost C(X,Y) is specifically:

[0030] C(X,Y)=C deploy (X)+C offload (Y)

[0031] In the formula, C offload (Y) is the offloading cost required to calculate the service request, and its specific expression is:

[0032] C offload (Y) = C trans (Y)+C com (Y).

[0033] Further: In S3, the total cost minimization of requesting services under the constraints of system cache resources and computing resources is modeled as the P0 problem, and its specific expression is:

[0034] P0:

[0035] st

[0036]

[0037] Further: the local search algorithm DMSP algorithm is used to optimize the P0 problem, so as to minimize the total cost of requesting services under the constraints of system cache resources and computing resources. The method of using the local search algorithm DMSP algorithm to optimize the P0 problem is as follows:

[0038] Initialize the service deployment strategy X to an empty set. According to the model described in P0, fix the service deployment decision, simplify the service deployment and calculate the offloading cost to C * (X), update the service deployment cost according to three different ways of adding, replacing and deleting services, and then calculate the total cost, and output the current optimal service deployment and computing offloading strategy X,Y;

[0039] The method of updating the service deployment cost according to the three different ways of adding, replacing and deleting services is as follows:

[0040] S31, determine whether the cost of adding the service deployment decision satisfies the following formula, if yes, proceed to S32, if not, proceed to S33;

[0041]

[0042] In the formula, C * (X) is a fixed service deployment decision, simplified service deployment and computational offloading cost, p(N,ε)=N 2 / ε, The cost of adding a service deployment decision, Q is an empty set;

[0043] S32. Add a service deployment item, and update the service deployment cost to the original cost and sum;

[0044] S33, judging whether the cost of the replacement service deployment decision satisfies the following formula, if yes, proceeding to S34, if no, proceeding to S35;

[0045]

[0046] In the formula, The cost of making deployment decisions for replacement services,

[0047] S34, replace the service deployment decision, and update the service deployment cost to the original cost and C replace The sum of (i,m);

[0048] S35, judging whether the cost of deleting the service deployment decision satisfies the following formula, if yes, proceeding to S36, if no, proceeding to S37;

[0049]

[0050] In the formula, C remove (i,m) is the cost of deleting the service deployment decision,

[0051] S36, delete the service deployment decision, and update the service deployment cost to the original cost and C remove The sum of (i,m).

[0052] Further: In S31, the cost of adding service deployment decision The specific expression is:

[0053]

[0054] In the formula, The cost of deploying service i on mobile device m;

[0055] In S33, the cost of replacing the service deployment decision The specific expression is:

[0056]

[0057] In the formula, The deployment cost of adding a new service deployment decision, Deploy decision to remove service and reassign Y r i The cost of all requests in Y is reduced; r iis the set of offloading decisions for service i on the reallocated mobile device r;

[0058] In S35, the cost C of deleting the service deployment decision remove The specific expression of (i,m) is:

[0059]

[0060] In the formula, Add an upper limit to the cost of service i on mobile device r0, The deployment cost of adding a new service i on mobile device r0, The part of mobile device m that processes the request for service i from mobile device n. The computing resources allocated to service i for mobile device m, The unused computing resources of mobile device r are reallocated. is the unused computing resources of mobile device m,

[0061] The beneficial effects of the present invention are as follows: the present invention proposes a service deployment optimization scheme for joint caching and computing resource allocation in a mobile edge computing network, collaboratively deploys services on mobile devices and edge servers, establishes a service deployment and task allocation model, formalizes the problem of minimizing the total cost of services, proposes a local search algorithm to obtain service deployment and request mechanisms, solves the problem in polynomial time, and reduces the service cost of MEC networks supporting D2D. Compared with other benchmark tests, the DMSP service deployment mechanism proposed by the present invention can effectively reduce the system cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 Flowchart of a service deployment optimization scheme for joint caching and computing resource allocation in a mobile edge computing network.

[0063] Figure 2 The figure is a comparison of the total cost of smart mobile devices with different computing capabilities under the DMSP mechanism and the traditional mechanism;

[0064] Figure 3 The figure is a comparison of the total cost of smart mobile devices with different storage capacities under the DMSP mechanism and the traditional mechanism;

[0065] Figure 4 This is a comparison diagram of the total cost with different numbers of service requests under the DMSP mechanism and the traditional mechanism. DETAILED DESCRIPTION

[0066] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0067] like Figure 1 As shown, in one embodiment of the present invention, a collaborative service deployment and computing offloading method in a mobile edge computing network includes the following steps:

[0068] S1. Establish a three-layer system model for service deployment in MEC network supporting D2D;

[0069] S2, through the three-layer system model, the service deployment and offload task calculation are performed in the MEC network;

[0070] S3. By jointly optimizing cache and computing resources, the cost of requesting services for offloading computing tasks is minimized, and the total cost of requesting services is minimized under the constraints of system cache resources and computing resources.

[0071] In S1, the three-layer system model includes a remote cloud server, an edge server, and several local mobile devices;

[0072] When a mobile device requests a service, it obtains the service from nearby mobile devices through a direct D2D communication link between devices, or obtains the service from an edge server through a wireless channel, or obtains the service from a remote cloud server through a backhaul link.

[0073] In this embodiment, due to the limited cache and computing resources of mobile devices and edge servers, the present invention proposes the location of service deployment and the node optimization problem of processing services, and minimizes the system service cost by jointly optimizing cache and computing resources.

[0074] In S2, the deployment cost C of deploying all services on the mobile device deploy The specific expression of (X) is:

[0075]

[0076] In the formula, is the deployment decision of service i on mobile device m, N is the set of mobile devices, is the cost of deploying service i on mobile device m, I is the service set, and X is defined as the service deployment decision.

[0077] The decision to offload computing tasks is as follows:

[0078]

[0079] In the formula, The proportion of service i requested by mobile device n being assigned to mobile device m for processing, and the set of D2D devices that provide services to mobile device n is defined as For service i, the proportion of computing tasks offloaded to the edge server and the remote cloud is defined as and in The proportion of service requests that can be processed satisfies the following formula:

[0080]

[0081] In said S2, the calculation of offloading cost includes task transmission cost and task calculation cost;

[0082] When mobile device n requests a service, the task transmission cost C trans The specific expression of (Y) is:

[0083]

[0084] In the formula, The transmission delay of the offloaded computing task required to serve the requested service, is the transmission energy consumption of the input data, α is the delay weighting coefficient of mobile device n sending service i, β is the energy consumption weighting coefficient of mobile device n sending service i, and the set Y is defined as the total task allocation decision.

[0085]

[0086] The computational cost of the task offloaded to the device, edge server or remote cloud server is C com The specific expression of (Y) is:

[0087]

[0088] In the formula, The computational latency incurred for executing all computational tasks on mobile device n, C EC (Y) is the task computation cost when the computing task of the mobile device is offloaded to the edge server or remote cloud server, and its specific expression is:

[0089]

[0090] In the formula, The proportion of computing tasks offloaded to the edge server for service i requested by mobile device n, b iis the input data size of the requested service i, The proportion of computing tasks offloaded to the remote cloud by service i requested by mobile device n, and c0 is the computing cost per unit data using the edge cloud or cloud.

[0091] In this embodiment, the system needs to consider both the service deployment strategy X and the computing offloading strategy Y, that is, the total cost in the system includes the service deployment cost and the computing task offloading cost, wherein the computing task offloading cost includes the task transmission cost and the computing cost.

[0092] In S3, the total cost C(X,Y) is specifically expressed as:

[0093] C(X,Y)=C deploy (X)+C offload (Y)

[0094] In the formula, C offload (Y) is the offloading cost required to calculate the offloading service request, and its specific expression is:

[0095] C offload (Y) = C trans (Y)+C com (Y).

[0096] The computing task offloading is complementary and competitive with edge caching, that is, a certain amount of service-related data is pre-deployed on mobile devices and edge servers. When serving, intelligent mobile devices or edge networks are used to process requests near users at a lower cost. By jointly optimizing the location of service deployment and the problem of processing service nodes, the total cost of the service is minimized under the constraints of cache and computing resources.

[0097] In S3, the total cost minimization of requesting services under the constraints of system cache resources and computing resources is modeled as the P0 problem, and its specific expression is:

[0098] P0:

[0099] st

[0100]

[0101] In this embodiment, the constraints of the P0 problem are as follows: the request task can be executed only when the service is deployed on mobile device m, the sum of the processing ratios of the services at different locations is 1, the computing resource allocation of the mobile device cannot exceed its total computing capacity, and the deployment decision of service i on mobile device m is The value is a binary variable 0 or 1, which indicates the proportion of service i assigned to mobile device m for processing. The value range is [0,1].

[0102] P0 is a mixed integer linear programming problem, which is an NP-hard problem. Considering the NP-hard characteristics of problem P0, its optimal solution usually requires exponential computing time. In order to solve this problem, it is usually considered to design a low-complexity algorithm to obtain a suboptimal solution in polynomial time. To this end, the present invention adopts three local search algorithms and proposes a DMSP algorithm for joint service deployment and task allocation. This mechanism repeatedly executes one of the local search algorithms, thereby reducing the total cost of the current solution mechanism. The local search algorithm DMSP algorithm is used to optimize the P0 problem, and the total cost of requesting services under the constraints of system cache resources and computing resources is minimized. The method of optimizing the P0 problem using the local search algorithm DMSP algorithm is specifically as follows:

[0103] Initialize the service deployment strategy X to an empty set. According to the model described in P0, fix the service deployment decision, simplify the service deployment and calculate the offloading cost to C * (X), update the service deployment cost according to three different ways of adding, replacing and deleting services, and then calculate the total cost, and output the current optimal service deployment and computing offloading strategy X,Y;

[0104] The method of updating the service deployment cost according to the three different ways of adding, replacing and deleting services is as follows:

[0105] S31, determine whether the cost of adding the service deployment decision satisfies the following formula, if yes, proceed to S32, if not, proceed to S33;

[0106]

[0107] In the formula, C * (X) is a fixed service deployment decision, simplified service deployment and computational offloading cost, p(N,ε)=N 2 / ε, The cost of adding a service deployment decision, Q is an empty set;

[0108] S32. Add a service deployment item, and update the service deployment cost to the original cost and sum;

[0109] S33, judging whether the cost of the replacement service deployment decision satisfies the following formula, if yes, proceeding to S34, if no, proceeding to S35;

[0110]

[0111] In the formula, The cost of making deployment decisions for replacement services,

[0112] S34, replace the service deployment decision, and update the service deployment cost to the original cost and C replace The sum of (i,m);

[0113] S35, judging whether the cost of deleting the service deployment decision satisfies the following formula, if yes, proceeding to S36, if no, proceeding to S37;

[0114]

[0115] In the formula, C remove (i,m) is the cost of deleting the service deployment decision,

[0116] S36, delete the service deployment decision, and update the service deployment cost to the original cost and C remove The sum of (i,m).

[0117] In S31, the cost of adding service deployment decision The specific expression is:

[0118]

[0119] In the formula, The cost of deploying service i on mobile device m;

[0120] In this embodiment, a service deployment decision is added, that is, a service deployment item is added to the existing service deployment decision, and based on the new service deployment decision Get an optimized task allocation mechanism.

[0121] In S33, the cost of replacing the service deployment decision The specific expression is:

[0122]

[0123] In the formula, The deployment cost of adding a new service deployment decision, Deploy decision to remove service and reassign Y r i The cost of all requests in Y is reduced; r i is the set of offloading decisions for service i on the reallocated mobile device r;

[0124] In this embodiment, the service deployment decision is replaced, that is, a service item is added and other service deployment items are deleted.

[0125] In S35, the cost C of deleting the service deployment decision remove The specific expression of (i,m) is:

[0126]

[0127] In the formula, Add an upper limit to the cost of service i on mobile device r0, The deployment cost of adding a new service i on mobile device r0, The part of mobile device m that processes the request for service i from mobile device n. The computing resources allocated to service i for mobile device m, The unused computing resources of mobile device r are reallocated. is the unused computing resources of mobile device m,

[0128] In this embodiment, the service deployment decision is deleted, that is, the service i is removed from one mobile device and deployed to other mobile devices.

[0129] The invention also provides a simulation experiment to verify the effect of the invention.

[0130] Depend on Figure 2 It can be seen that when the computing capacity of the mobile device increases, the Service Placement On the Device (SPOD) mechanism has a greater impact on the total cost than other mechanisms, so it decreases faster. The Edge and Device Service Placement (EDSP) mechanism allows services to be deployed on mobile devices, but only caches the currently requested services, and there is no collaboration between mobile devices. The Service Placement At the Edge (SPAE) mechanism does not deploy services on devices, and the increase in computing power has no effect on its cost. The DMSP mechanism proposed in the present invention takes into account the collaborative deployment of services between devices and reduces the wireless channel transmission overhead. Therefore, its total cost is minimal.

[0131] Depend on Figure 3 It can be seen that as the cache capacity of the mobile device increases, the total cost of the SPOD, EDSP and DMSP mechanisms decreases. When the cache capacity increases from 100MB to 200MB, the total cost decreases by 11.22%, 12.94% and 13.41% respectively. This is because with the increase in the cache capacity of the mobile device, although the cost of wireless channel transmission is reduced, the cost of deploying services is also increased, and the reduction in cost will be offset.

[0132] Depend on Figure 4It can be seen that when the amount of tasks increases, the DMSP mechanism has the lowest cost. When the number of services is 15, the cost of the DMSP mechanism is 4.69% lower than the EDSP mechanism, 10.94% lower than the SPAE mechanism, and 31.25% lower than the SPOD mechanism. This is because the DMSP mechanism, compared with the EDSP mechanism, collaboratively deploys services at the mobile device, the SPAE mechanism does not utilize the resources at the mobile device, and the SPOD mechanism does not occupy the resources of the edge server.

[0133] The beneficial effects of the present invention are as follows: the present invention proposes a service deployment optimization scheme for joint caching and computing resource allocation in a mobile edge computing network, collaboratively deploys services on mobile devices and edge servers, establishes a service deployment and task allocation model, formalizes the problem of minimizing the total cost of requesting services, proposes a local search algorithm to obtain service deployment and request mechanisms, solves the problem in polynomial time, and reduces the service cost of MEC networks supporting D2D. Compared with other benchmark tests, the DMSP service deployment mechanism proposed by the present invention can effectively reduce system costs.

[0134] In the description of the present invention, it is necessary to understand that the orientation or positional relationship indicated by the terms "center", "thickness", "upper", "lower", "horizontal", "top", "bottom", "inner", "outer", "radial", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", and "third" are used only for descriptive purposes, and cannot be understood as indicating or implying the relative importance or the number of implicitly specified technical features. Therefore, the features defined by "first", "second", and "third" may explicitly or implicitly include one or more of the features.

Claims

1. A collaborative service deployment and computation offloading method in a mobile edge computing network, characterized in that: The following steps are involved: S1. Establish a three-layer system model for service deployment in MEC network supporting D2D; S2, through the three-layer system model, the service deployment and offload task calculation are performed in the MEC network; S3: By jointly optimizing cache and computing resources, the total cost of request service is minimized under the constraints of system cache resources and computing resources.

2. The method for collaborative service deployment and computation offloading in a mobile edge computing network according to claim 1, characterized in that: In S1, the three-layer system model includes a remote cloud server, an edge server, and several local mobile devices; When a mobile device requests a service, it obtains the service from nearby mobile devices through a direct D2D communication link between devices, or obtains the service from an edge server through a wireless channel, or obtains the service from a remote cloud server through a backhaul link.

3. The method for collaborative service deployment and computation offloading in a mobile edge computing network according to claim 2, characterized in that: In S2, the deployment cost C of deploying all services on the mobile device deploy The specific expression of (X) is: In the formula, is the deployment decision of service i on mobile device m, N is the set of mobile devices, is the cost of deploying service i on mobile device m, I is the service set, and X is defined as the service deployment decision. The decision to offload computing tasks is as follows: In the formula, The proportion of service i requested by mobile device n being assigned to mobile device m for processing, and the set of D2D devices that provide services to mobile device n is defined as For service i, the proportion of computing tasks offloaded to the edge server and the remote cloud is defined as and in The proportion of service requests that can be processed satisfies the following formula:

4. The method for collaborative service deployment and computation offloading in a mobile edge computing network according to claim 3, characterized in that: In said S2, the calculation of offloading cost includes task transmission cost and task calculation cost; When mobile device n requests a service, the task transmission cost C trans The specific expression of (Y) is: In the formula, The transmission delay of the offloaded computing task required to serve the requested service, is the transmission energy consumption of the input data, α is the delay weighting coefficient of mobile device n sending service i, β is the energy consumption weighting coefficient of mobile device n sending service i, and the set Y is defined as the total task allocation decision. The computational cost of the task offloaded to the device, edge server or remote cloud server is C com The specific expression of (Y) is: In the formula, The computational latency incurred for executing all computational tasks on mobile device n, C EC (Y) is the task computation cost when the computing task of the mobile device is offloaded to the edge server or remote cloud server, and its specific expression is: In the formula, The proportion of computing tasks offloaded to the edge server for service i requested by mobile device n, b i is the input data size of the requested service i, The proportion of computing tasks offloaded to the remote cloud by service i requested by mobile device n, and c0 is the computing cost per unit data using the edge cloud or cloud.

5. The method for collaborative service deployment and computation offloading in a mobile edge computing network according to claim 4, characterized in that: In S3, the total cost C(X,Y) is specifically expressed as: C(X,Y)=C deploy (X)+C offload (Y) In the formula, C offload (Y) is the offloading cost required to calculate the service request, and its specific expression is: C offload (Y)=C trans (Y)+C com (Y)。 6. The method for collaborative service deployment and computation offloading in a mobile edge computing network according to claim 5, characterized in that: In S3, the total cost minimization of requesting services under the constraints of system cache resources and computing resources is modeled as the P0 problem, and its specific expression is: P0: st 7. The method for collaborative service deployment and computation offloading in a mobile edge computing network according to claim 6, characterized in that: The local search algorithm DMSP algorithm is used to optimize the P0 problem, and the total cost of requesting services is minimized under the constraints of system cache resources and computing resources. The method of using the local search algorithm DMSP algorithm to optimize the P0 problem is as follows: Initialize the service deployment strategy X to an empty set. According to the model described in P0, fix the service deployment decision, simplify the service deployment and calculate the offloading cost to C * (X), update the service deployment cost according to three different ways of adding, replacing and deleting services, and then calculate the total cost, and output the current optimal service deployment and computing offloading strategy X,Y; The method of updating the service deployment cost according to the three different ways of adding, replacing and deleting services is as follows: S31, determine whether the cost of adding the service deployment decision satisfies the following formula, if yes, proceed to S32, if not, proceed to S33; In the formula, C * (X) is a fixed service deployment decision, simplified service deployment and computational offloading cost, p(N,ε)=N 2 / ε, The cost of adding a service deployment decision, Q is an empty set; S32. Add a service deployment item, and update the service deployment cost to the original cost and sum; S33, judging whether the cost of the replacement service deployment decision satisfies the following formula, if yes, proceeding to S34, if no, proceeding to S35; In the formula, The cost of making deployment decisions for replacement services, S34, replace the service deployment decision, and update the service deployment cost to the original cost and C replace The sum of (i,m); S35, judging whether the cost of deleting the service deployment decision satisfies the following formula, if yes, proceeding to S36, if no, proceeding to S37; In the formula, C remove (i,m) is the cost of deleting the service deployment decision, S36, delete the service deployment decision, and update the service deployment cost to the original cost and C remove The sum of (i,m).

8. The method for collaborative service deployment and computation offloading in a mobile edge computing network according to claim 7, characterized in that: In S31, the cost of adding service deployment decision The specific expression is: In the formula, The cost of deploying service i on mobile device m; In S33, the cost of replacing the service deployment decision The specific expression is: In the formula, The deployment cost of adding a new service deployment decision, Deploy decision to remove service and reassign Y r i The cost of all requests in Y is reduced; r i is the set of offloading decisions for service i on the reallocated mobile device r; In S35, the cost C of deleting the service deployment decision remove The specific expression of (i,m) is: In the formula, Add an upper limit to the cost of service i on mobile device r0, The deployment cost of adding a new service i on mobile device r0, The part of mobile device m that processes the request for service i from mobile device n. The computing resources allocated to service i for mobile device m, The unused computing resources of mobile device r are reallocated. is the unused computing resources of mobile device m,

Images