A vehicle-mounted edge server scheduling method, system and storage medium to minimize the number of vehicle-mounted edge servers

Abstract

The present invention relates to a vehicle-mounted edge server scheduling method, system and storage medium for minimizing the number of vehicle-mounted edge servers, and relates to the field of mobile edge computing. The method comprises: obtaining a set of regions and warehouses, establishing a network model; determining the total energy consumption of the vehicle-mounted edge servers in the warehouse in the driving path, establishing a problem of minimizing the number of vehicle-mounted edge servers; optimizing and solving the problem of minimizing the number, determining a scheduling scheme for the vehicle-mounted edge servers, and scheduling the vehicle-mounted edge servers in the warehouse according to the scheduling scheme. The method can realize dynamic scheduling of vehicle-mounted edge servers as well as scheduling flexibility and balance, reduce the number of vehicle-mounted edge servers, save the scheduling cost of vehicle-mounted edge servers, and solve the problem of resource waste and shortage caused by statically deploying edge servers in mobile edge computing scenarios.

Description

Technical Field

[0001] The present invention relates to the field of mobile edge computing, and in particular to a vehicle-mounted edge server scheduling method, system and storage medium that minimizes the number of vehicle-mounted edge servers. Background Art

[0002] Mobile edge computing can offload computationally intensive tasks running on terminal devices to nearby mobile edge computing servers for computing, so that users' computing tasks do not need to be uploaded to remote cloud servers, thereby significantly reducing the latency of applications. Mobile edge computing has attracted widespread attention from academia and industry. Some industry associations and standardization organizations, including the 5G Automotive Association and the European Telecommunications Standards Institute, have identified many use cases for mobile edge computing, such as augmented reality technology, autonomous driving, etc.

[0003] The main research contents of mobile edge computing are computing task offloading, edge service caching and edge system deployment. At present, a large number of works are based on statically deployed edge servers, that is, research based on existing fixed edge servers on the ground. However, in practical applications, due to user mobility, dynamic resource demand and regional differences, the load of fixed edge servers is highly dynamic and unbalanced. In statically deployed edge servers, once the deployment plan is executed, it cannot be changed. When the task request volume of a large number of users changes in time and space, the edge server cannot move according to the changes in actual demand, nor can it expand resources according to the changes in actual demand, resulting in the inability to meet the needs of some users in hot spots, and the resources of edge servers in remote areas are wasted. Summary of the invention

[0004] The main purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, and to provide a vehicle-mounted edge server scheduling method, system and storage medium that minimize the number of vehicle-mounted edge servers, which can realize dynamic scheduling of vehicle-mounted edge servers as well as scheduling flexibility and balance, reduce the number of vehicle-mounted edge servers, save the scheduling cost of vehicle-mounted edge servers, and solve the resource waste and shortage problems caused by static deployment of edge servers in mobile edge computing scenarios.

[0005] According to one aspect of the present invention, the present invention provides a method for scheduling an on-vehicle edge server to minimize the number of on-vehicle edge servers, the method comprising:

[0006] Obtain a set of regions and warehouses and establish a network model, wherein the vehicle-mounted edge servers in the warehouse provide computing services for multiple regions, the vehicle-mounted edge servers in the same warehouse are of the same type, and the vehicle-mounted edge servers in different warehouses are of different types;

[0007] Determine the total energy consumption of the on-board edge servers in the warehouse during the driving path and establish the problem of minimizing the number of on-board edge servers;

[0008] The quantity minimization problem is optimized and solved, a scheduling scheme for the vehicle-mounted edge servers is determined, and the vehicle-mounted edge servers in the warehouse are scheduled according to the scheduling scheme.

[0009] Preferably, determining the total energy consumption of the vehicle-mounted edge server in the warehouse during the driving path includes:

[0010] Warehouses j The vehicle-mounted edge server m is on the driving path l j，m Total energy consumption in Calculated by the following formula:

[0011]

[0012]

[0013] Among them, ρ i Represents computing task a i The number of CPU cycles required, Represents the energy consumption of a CPU cycle, Indicates the capacitance switch related to the vehicle chip; is the driving energy consumption per unit driving distance of the vehicle-mounted edge server, e ij Indicates that the vehicle-mounted edge computing server is from the warehouse s j Arrival Area v i The operator “||” indicates the length of the edge; es j (v i ) is warehouse s j The vehicle-mounted edge server in the execution area v i The computing energy consumption, w(e ij ) is the vehicle-mounted edge computing server from the warehouse s j Arrival Area v i Driving energy consumption; For warehouses j The vehicle-mounted edge server m is on the driving path l j，m The driving energy consumption under For warehouses j In-vehicle edge server m execution The computational energy consumption of all regions in the collection.

[0014] Preferably, the problem of establishing the minimization of the number of vehicle-mounted edge servers includes:

[0015] The problem of minimizing the number of onboard edge servers is expressed as:

[0016]

[0017]

[0018]

[0019]

[0020]

[0021] Among them, M j Represents each warehouse s j The set of scheduled vehicle-mounted edge servers; B j Represents warehouses j The energy of the onboard edge server in the vehicle; Represents the set of warehouses whose service area set is not empty; V j For warehouses j The set of regions where services are provided; i，m Represents region v i Is it served by the vehicle-mounted edge server m? If it is served by the vehicle-mounted edge server m, then g i，m =1, otherwise g i，m =0.

[0022] Preferably, the optimizing and solving the quantity minimization problem to determine the scheduling scheme of the vehicle-mounted edge server includes:

[0023] (4.1) Input undirected graph G = (V ∪ S, E), where V = {v 1 , v 2 , ..., v n} represents a set of n regions, S = {s 1 ,s 2 , ..., s h} represents the set of h warehouses, E represents the set of paths between warehouses and regions and between regions;

[0024] (4.2) Initialize the driving path set of the vehicle edge server The set of warehouses whose service area set is not an empty set

[0025] (4.3) Traverse all warehouses, if V j is not an empty set, then

[0026] (4.4) Traversal For all warehouses in , execute steps (4.5) to (4.13) until all warehouses are traversed, and execute step (4.14);

[0027] (4.5) For each warehouse s j , construct warehouses j An undirected graph G′ with only edge weights j , the weight of each edge is as follows:

[0028] G′ j Middle side ii′ The weight w′(e ii′ )for:

[0029]

[0030] G′ j Middle side ij The weight w′(e ij )for:

[0031]

[0032] (4.6) Initialization Among them, the operator Indicates rounding up;

[0033] (4.7) Warehouse s j Service area collection V j The area in is divided into z+1 categories, traversing the loop variable x from 0 to z, warehouse s j The x-th region set V j x Initialize to an empty set;

[0034] (4.8) Traverse V j Each region v in i , execute steps (4.9), (4.10), and (4.11) until all areas have been traversed, and then execute step (4.12);

[0035] (4.9) If V j 0 =V j 0 ∪{v i};

[0036] (4.10) If V j z =V j z ∪{v i};

[0037] (4.11) Traverse the loop variable x from 1 to z-1, if Then update V j x =Vj x ∪{v i};

[0038] (4.12) Traverse the loop variable x from 0 to z and construct the warehouse s j The graph G″ after partitioning j =(V j x , E(V j x )), E(V j x ) represents the set V j x The edge set composed of warehouse s j The set of paths in the partitioned graph unrooted DVRP stands for unrooted vehicle routing algorithm with distance constraints;

[0039] (4.13) Traversal All paths in renew

[0040] (4.14) Output the driving path set of the vehicle-mounted edge server.

[0041] Preferably, the unrooted DVRP comprises the following steps:

[0042] (5.1) Input warehouse s j The graph G″ after partitioning j =(V j x , E(V j x )), let γ = 2 x Delta-1;

[0043] (5.2) Initialize variable k to 1;

[0044] (5.3) If k≤|V j x |, execute step (5.4), otherwise execute step (5.10);

[0045] (5.4) Generate a minimum weight forest F with nk edges based on graph G. First, sort the edges in ascending order of weight, and then add the edges one by one. If the newly added edge does not cause a loop in the forest, add the new edge to the forest until a minimum weight forest with nk edges is generated. At the same time, generate the connected component C from F. 1 , C 2 , ..., C k ;Path collection

[0046] (5.5) Initialize variable i to 1;

[0047] (5.6) If i≤k, then initialize the number of vehicle-mounted edge servers N(i, k) = 1 under the current i and k values, according to C i The generation loop TOUR i , the TOUR i Delete an edge to generate path P i , then execute step (5.7); otherwise, execute step (5.9);

[0048] (5.7) If P i If the path length is greater than γ, then from path P i Find the path P′ whose path length is closest to γ ​​but less than γ i , update path P i =P i \P′ i , SQL k =SQL k ∪P′ i , N(i, k) = N(i, k) + 1; otherwise, execute step (5.8); where “\” indicates finding the difference of sets;

[0049] (5.8)SQL k =SQL k ∪P i , i=i+1, then execute step (5.6);

[0050] (5.9) k = k + 1, execute step (5.3);

[0051] (5.10) Output SQL k′ ;in express The value of k when the value is minimum.

[0052] According to another aspect of the present invention, the present invention provides a vehicle-mounted edge server scheduling system for minimizing the number of vehicle-mounted edge servers, the system comprising:

[0053] An acquisition module is used to acquire a set of regions and warehouses and establish a network model, wherein the vehicle-mounted edge servers in the warehouse provide computing services for multiple regions, the vehicle-mounted edge servers in the same warehouse are of the same type, and the vehicle-mounted edge servers in different warehouses are of different types;

[0054] Establish a module to determine the total energy consumption of the vehicle-mounted edge servers in the warehouse during the driving path and establish the problem of minimizing the number of vehicle-mounted edge servers;

[0055] The scheduling module is used to optimize and solve the quantity minimization problem, determine the scheduling plan of the vehicle-mounted edge server, and schedule the vehicle-mounted edge server in the warehouse according to the scheduling plan.

[0056] Preferably, the establishing module determines the total energy consumption of the vehicle-mounted edge server in the warehouse in the driving path, including:

[0057] Warehouses j The vehicle-mounted edge server m is on the driving path l j，m Total energy consumption in Calculated by the following formula:

[0058]

[0059]

[0060] Among them, ρ i Represents computing task a i The number of CPU cycles required, Represents the energy consumption of a CPU cycle, Indicates the capacitance switch related to the vehicle chip; is the driving energy consumption per unit driving distance of the vehicle-mounted edge server, e ij Indicates that the vehicle-mounted edge computing server is from the warehouse s j Arrival Area v i The operator “||” indicates the length of the edge; ex j (v i ) is warehouse s j The vehicle-mounted edge server in the execution area v i The computing energy consumption, w(e ij ) is the vehicle-mounted edge computing server from the warehouse s j Arrival Area v i Driving energy consumption; For warehouses j The vehicle-mounted edge server m is on the driving path l j，m The driving energy consumption under For warehouses j In-vehicle edge server m execution The computational energy consumption of all regions in the collection.

[0061] Preferably, the establishment module establishes the problem of minimizing the number of vehicle-mounted edge servers, including:

[0062] The problem of minimizing the number of onboard edge servers is expressed as:

[0063]

[0064]

[0065]

[0066]

[0067]

[0068] Among them, M j Represents each warehouse s j The set of scheduled vehicle-mounted edge servers; B j Represents warehouses j The energy of the onboard edge server in the vehicle; Represents the set of warehouses whose service area set is not empty; V j For warehouses j The set of regions where services are provided; i，m Represents region v i Is it served by the vehicle-mounted edge server m? If it is served by the vehicle-mounted edge server m, then g i，m =1, otherwise g i，m =0.

[0069] According to another aspect of the present invention, the present invention provides a vehicle-mounted edge server scheduling system for minimizing the number of vehicle-mounted edge servers, the system comprising: a processor, a memory, the memory storing computer-executable instructions, and the computer-executable instructions implement the above method steps when executed by the processor.

[0070] According to another aspect of the present invention, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions implement the above method steps when executed by a processor.

[0071] Beneficial effect: The present invention takes into account the resource waste and shortage problems caused by the static deployment of edge servers, and determines the scheduling scheme for the warehouse to schedule the on-board edge servers to provide computing services for the region based on the regional set of warehouse services. It can realize the dynamic scheduling of on-board edge servers as well as the flexibility and balance of scheduling, reduce the number of on-board edge servers, save the scheduling cost of on-board edge servers, and solve the resource waste and shortage problems caused by the static deployment of edge servers in mobile edge computing scenarios.

[0072] The features and advantages of the present invention will become clear through reference to the following drawings and detailed description of specific embodiments of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 Flowchart of a method for scheduling vehicle-mounted edge servers to minimize the number of vehicle-mounted edge servers;

[0074] Figure 2 It is a schematic diagram of the network model;

[0075] Figure 3 This is the flow chart of the scheduling algorithm for the vehicle-mounted edge server;

[0076] Figure 4 This is the unrooted DVRP algorithm flow chart;

[0077] Figure 5 Schematic diagram of the on-vehicle edge server scheduling system to minimize the number of on-vehicle edge servers. DETAILED DESCRIPTION

[0078] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings 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.

[0079] Example 1

[0080] Figure 1 Flowchart of the vehicle-mounted edge server scheduling method for minimizing the number of vehicle-mounted edge servers. Figure 1 As shown, this embodiment provides a vehicle-mounted edge server scheduling method and system that minimizes the number of vehicle-mounted edge servers, which can realize dynamic scheduling of vehicle-mounted edge servers as well as scheduling flexibility and balance, reduce the number of vehicle-mounted edge servers, save the scheduling cost of vehicle-mounted edge servers, and solve the resource waste and shortage problems caused by static deployment of edge servers in mobile edge computing scenarios.

[0081] According to one aspect of the present invention, the present invention provides a method for scheduling an on-vehicle edge server to minimize the number of on-vehicle edge servers, the method comprising:

[0082] S1: Obtain a set of regions and warehouses and establish a network model, in which the vehicle-mounted edge servers in the warehouse provide computing services for multiple regions, the vehicle-mounted edge servers in the same warehouse are of the same type, and the vehicle-mounted edge servers in different warehouses are of different types.

[0083] Specifically, refer to Figure 2 , on the two-dimensional plane Ω, let V = {v 1 , v 2 , ..., v n} represents a set of n regions, let S = {s 1 ,s 2 , ..., s h} represents a set of h warehouses, and defines a graph G = (V∪S, E), where E represents the set of paths between warehouses and regions and between regions.

[0084] Each area v i ∈V generates computing task a i , let ρ i Represents computing task a i The number of CPU cycles required;

[0085] Each warehouse has at least n vehicle-mounted edge servers. The warehouse dispatches vehicle-mounted edge servers to provide computing services for multiple regions. The number of vehicle-mounted edge servers that can be dispatched by each warehouse is different. Let M j Represents each warehouse s j ∈S is a set of vehicle-mounted edge servers scheduled by the system. The vehicle-mounted edge servers start from the warehouse, arrive at the regional location to perform computing tasks, and finally return to the warehouse;

[0086] Assume that the vehicle-mounted edge servers in the same warehouse are of the same type, and the CPU frequency, energy, and driving speed of the vehicle-mounted edge servers in the same warehouse are the same; the vehicle-mounted edge servers in different warehouses are of different types, and the CPU frequency, energy, and driving speed of the vehicle-mounted edge servers in different warehouses are different; let F j , B j and R j Respectively represent warehouses j The CPU frequency, energy and driving speed of the vehicle-mounted edge server in the figure; assuming that the energy of any vehicle-mounted edge server is sufficient to reach the area farthest from the warehouse;

[0087] Each warehouse j Each of them has a set of regions V that they serve j ; A set of warehouses that represent a non-empty set of service regions;

[0088] L j Represents warehouses j The set of driving paths of all vehicle-mounted edge servers in , l j，m Represents warehouses j In-vehicle edge server m∈M j The driving route starts from the warehouse, serves multiple areas, and then returns to the warehouse; Indicates that the vehicle-mounted edge server m is on the driving path l j，m The set of regions served on Indicates from warehouse s j The set of edges from which to begin traveling the route.

[0089] S2: Determine the total energy consumption of the on-board edge servers in the warehouse during the driving path and establish the problem of minimizing the number of on-board edge servers;

[0090] S3: Optimize and solve the quantity minimization problem, determine a scheduling plan for the vehicle-mounted edge servers, and schedule the vehicle-mounted edge servers in the warehouse according to the scheduling plan.

[0091] This method can realize dynamic scheduling of vehicle-mounted edge servers as well as scheduling flexibility and balance, reduce the number of vehicle-mounted edge servers, save the scheduling cost of vehicle-mounted edge servers, and solve the resource waste and shortage problems caused by static deployment of edge servers in mobile edge computing scenarios.

[0092] Preferably, determining the total energy consumption of the vehicle-mounted edge server in the warehouse during the driving path includes:

[0093] Warehouses j The vehicle-mounted edge server m is on the driving path l j，m Total energy consumption in Calculated by the following formula:

[0094]

[0095]

[0096] Among them, ρ i Represents computing task a i The number of CPU cycles required, Represents the energy consumption of a CPU cycle, Indicates the capacitance switch related to the vehicle chip; is the driving energy consumption per unit driving distance of the vehicle-mounted edge server, e ij Indicates that the vehicle-mounted edge computing server is from the warehouse s j Arrival Area v i The operator “||” indicates the length of the edge; es j (v i ) is warehouse s j The vehicle-mounted edge server in the execution area v i The computing energy consumption, w(e ij ) is the vehicle-mounted edge computing server from the warehouse s j Arrival Area v i Driving energy consumption; For warehouses j The vehicle-mounted edge server m is on the driving path l j，m The driving energy consumption under For warehouses j In-vehicle edge server m execution The computational energy consumption of all regions in the collection.

[0097] Specifically, warehouses jThe vehicle-mounted edge server in the execution area v i The computing energy consumption j (v i )for:

[0098]

[0099] in, Represents the energy consumption of a CPU cycle, Indicates the capacitance switch related to the vehicle chip,

[0100] Warehouses j In-vehicle edge server m execution The computational energy consumption of all regions in the collection can be calculated as follows:

[0101]

[0102] The vehicle-mounted edge server will generate driving energy consumption when driving. The driving energy consumption per unit driving distance of the vehicle-mounted edge server is defined as Then the vehicle-mounted edge computing server is from area v i Arrival Area v i′ The driving energy consumption w(e ij )for:

[0103]

[0104] Among them, e ii′ Indicates that the vehicle-mounted edge computing server is from area v i Arrival Area v i′ The operator “||” means to find the length of the edge;

[0105] On-board edge computing servers from warehouses j Arrival Area v i The driving energy consumption is

[0106]

[0107] Among them, e ij Indicates that the vehicle-mounted edge computing server is from the warehouse s j Arrival Area v i The edge of

[0108] Warehouses j The vehicle-mounted edge server m is on the driving path l j，m The corresponding driving energy consumption can be calculated by the following formula:

[0109]

[0110] Among them, e represents The edge elements of

[0111] Warehouses j The vehicle-mounted edge server m is on the driving path l j，m The total energy consumption can be calculated as follows:

[0112]

[0113] Preferably, the problem of establishing the minimization of the number of vehicle-mounted edge servers includes:

[0114] The problem of minimizing the number of onboard edge servers is expressed as:

[0115]

[0116]

[0117]

[0118]

[0119]

[0120] Among them, M j Represents each warehouse s j The set of scheduled vehicle-mounted edge servers; B j Represents warehouses j The energy of the onboard edge server in the vehicle; Represents the set of warehouses whose service area set is not empty; V j For warehouses j The set of regions where services are provided; i，m Represents region v i Is it served by the vehicle-mounted edge server m? If it is served by the vehicle-mounted edge server m, then g i，m =1, otherwise g i，m =0.

[0121] Specifically, combined with the vehicle edge server scheduling model, the problem of minimizing the number of vehicle edge servers is formalized, namely:

[0122]

[0123]

[0124]

[0125]

[0126]

[0127] Among them, constraint (7-3) ensures that the energy consumption of the vehicle edge server does not exceed the energy of the vehicle edge server; constraint (7-2) means that any warehouse s j The vehicle-mounted edge server must provide services for all regions in the region set of the warehouse to which it belongs. Constraint (7-3) indicates that all regions will be served once and only once by the vehicle-mounted edge server. i，m Represents region v i Whether it is served by the vehicle edge server m, if it is served by the vehicle edge server m, g i，m =1, otherwise g i，m =0.

[0128] Preferably, the optimizing and solving the quantity minimization problem to determine the scheduling scheme of the vehicle-mounted edge server includes:

[0129] (4.1) Input undirected graph G = (V ∪ S, E), where V = {v 1 , v 2 , ..., v n} represents a set of n regions, S = {s 1 ,s 2 , ..., s h} represents the set of h warehouses, E represents the set of paths between warehouses and regions and between regions;

[0130] (4.2) Initialize the driving path set of the vehicle edge server The set of warehouses whose service area set is not an empty set

[0131] (4.3) Traverse all warehouses, if V j is not an empty set, then

[0132] (4.4) Traversal For all warehouses in , execute steps (4.5) to (4.13) until all warehouses are traversed, and execute step (4.14);

[0133] (4.5) For each warehouse s j , construct warehouses j An undirected graph G′ with only edge weights j , the weight of each edge is as follows:

[0134] G′ j Middle side ii′ The weight w′(e ii′ )for:

[0135]

[0136] G′ jMiddle side ij The weight w′(e ij )for:

[0137]

[0138] (4.6) Initialization Among them, the operator Indicates rounding up;

[0139] (4.7) Warehouse s j Service area collection V j The area in is divided into z+1 categories, traversing the loop variable x from 0 to z, warehouse s j The x-th region set V j x Initialize to an empty set;

[0140] (4.8) Traverse V j Each region v in i , execute steps (4.9), (4.10), and (4.11) until all areas have been traversed, and then execute step (4.12);

[0141] (4.9) If V j 0 =V j 0 ∪{v i};

[0142] (4.10) If V j z =V j z ∪{v i};

[0143] (4.11) Traverse the loop variable x from 1 to z-1, if Then update V j x =V j x ∪{v i};

[0144] (4.12) Traverse the loop variable x from 0 to z and construct the warehouse s j The graph G″ after partitioning j =(V j x , E(V j x )), E(V j x ) represents the set V jx The edge set composed of warehouse s j The set of paths in the partitioned graph unrooted DVRP stands for unrooted vehicle routing algorithm with distance constraints;

[0145] (4.13) Traversal All paths in renew

[0146] (4.14) Output the driving path set of the vehicle-mounted edge server.

[0147] Specifically, the implementation process of this step can be found in Figure 3 The process shown.

[0148] Preferably, the unrooted DVRP comprises the following steps:

[0149] (5.1) Input warehouse s j The graph G″ after partitioning j =(V j x , E(V j x )), let γ = 2 x Delta-1;

[0150] (5.2) Initialize variable k to 1;

[0151] (5.3) If k≤|V j x |, execute step (5.4), otherwise execute step (5.10);

[0152] (5.4) Generate a minimum weight forest F with nk edges based on graph G. First, sort the edges in ascending order of weight, and then add the edges one by one. If the newly added edge does not cause a loop in the forest, add the new edge to the forest until a minimum weight forest with nk edges is generated. At the same time, generate the connected component C from F. 1 , C 2 , ..., C k ;Path collection

[0153] (5.5) Initialize variable i to 1;

[0154] (5.6) If i≤k, then initialize the number of vehicle-mounted edge servers N(i, k) = 1 under the current i and k values, according to C i The generation loop TOUR i , the TOUR i Delete an edge to generate path P i, then execute step (5.7); otherwise, execute step (5.9);

[0155] (5.7) If P i If the path length is greater than γ, then from path P i Find the path P′ whose path length is closest to γ ​​but less than γ i , update path P i =P i \P′ i , SQL k =SQL k ∪P′ i , N(i, k) = N(i, k) + 1; otherwise, execute step (5.8); where “\” indicates finding the difference of sets;

[0156] (5.8)SQL k =SQL k ∪P i , i=i+1, then execute step (5.6);

[0157] (5.9) k = k + 1, execute step (5.3);

[0158] (5.10) Output SQL k′ ;in express The value of k when the value is minimum.

[0159] Specifically, the implementation process of this step can be found in Figure 3 The process shown.

[0160] This embodiment takes into account the resource waste and shortage problems caused by the static deployment of edge servers. According to the regional set of warehouse services, a scheduling scheme is determined for the warehouse to schedule the on-board edge servers to provide computing services for the region. This can realize dynamic scheduling of on-board edge servers as well as scheduling flexibility and balance, reduce the number of on-board edge servers, save the scheduling cost of on-board edge servers, and solve the resource waste and shortage problems caused by the static deployment of edge servers in mobile edge computing scenarios.

[0161] In order to more clearly understand the implementation process of the present invention, another specific example is given below.

[0162] This embodiment takes a certain mobile edge computing scenario as an example, and the vehicle-mounted edge server scheduling method includes the following steps:

[0163] On the two-dimensional plane Ω, V={v 1 , v 2 , v 3 , v 4 , v 5} represents the region set, S = {s 1 ,s 2} represents a warehouse set. The parameters of the warehouse are shown in Table 1.

[0164] Table 1 Parameters of the vehicle-mounted edge server

[0165]

[0166] The service request parameters of the region are shown in Table 2:

[0167] Table 2: Service request parameters for a region

[0168] area CPU Requirements <![CDATA[v 1 ]]> 795 <![CDATA[v 2 ]]> 868 <![CDATA[v 3 ]]> 901 <![CDATA[v 4 ]]> 609 <![CDATA[v 5 ]]> 689

[0169] The distance parameters between warehouses and regions are shown in Table 3:

[0170] Table 3: Distance parameters between warehouses and regions

[0171]

[0172] (4.1) Input undirected graph G = (V∪S, E), the set of energy and service areas of the onboard edge servers of all warehouses;

[0173] (4.2) Initialize the driving path set of the vehicle edge server The set of warehouses whose service area set is not an empty set

[0174] (4.3) Traverse all warehouses, if V j is not an empty set, then

[0175] at this time,

[0176] (4.4) Traversal For all warehouses in , execute steps (4.5) to (4.13) until all warehouses are traversed, and execute step (4.14);

[0177] (4.5) For each warehouse s j , construct warehouses j An undirected graph G′ with only edge weights j , the weight value of each edge is as follows:

[0178] G′ j Middle side ii The weight w′(e ii′ )for

[0179]

[0180] G′j Middle side ij The weight w′(e ij )for

[0181]

[0182] (4.6) Initialization Among them, the operator Indicates rounding up;

[0183] At this time, Δ = 0.9, δ = 24.5, z = 0;

[0184] (4.7) Warehouse s j Service area collection V j The area in is divided into z+1 categories, traversing the loop variable x from 0 to z, warehouse s j The x-th region set V j x Initialize to an empty set;

[0185] (4.8) Traverse V j Each region v in i , execute steps (4.9), (4.10), and (4.11) until all areas have been traversed, and then execute step (4.12);

[0186] (4.9) If V j 0 =V j 0 ∪{v i};

[0187] (4.10) If V j z =V j z ∪{v i};

[0188] (4.11) Traverse the loop variable x from 1 to z-1, if Then update V j x =V j x ∪{v i};

[0189] (4.12) Traverse the loop variable x from 0 to z and construct the warehouse s j The graph G″ after partitioning j =(V j x , E(V j x)), E(V j x ) represents the set V j x The edge set composed of warehouse s j The set of paths in the partitioned graph unrooted DVRP stands for unrooted vehicle routing algorithm with distance constraints;

[0190] (4.13) Traversal All paths in renew

[0191] (4.14) Output the driving path set of the vehicle-mounted edge server.

[0192] at this time, There is only one path in Therefore only one on-board edge server is required.

[0193] In order to better verify the technical effects of the method of the present invention, two algorithms are designed to conduct actual comparative tests with the vehicle-mounted edge server scheduling algorithm of the present invention to verify the real effects of the method of the present invention, specifically including:

[0194] (1) Path segmentation algorithm: The algorithm first calls the Christofides algorithm to generate a path that includes all areas, and then performs path segmentation based on the energy consumption constraints of the vehicle-mounted edge server.

[0195] (2) Greedy algorithm: First, each area is assigned to the nearest warehouse. j Each vehicle-mounted edge server finds the region set V j The area with the lowest energy consumption for driving and executing regional computing is selected. If the energy of the vehicle-mounted edge server is sufficient to reach the area and complete the computing tasks in the area, and the remaining energy is sufficient to return to the warehouse, the area is added to the path, and then the V j The selected area is removed from the list. Otherwise, the vehicle-mounted edge server directly returns to the warehouse until all areas are served.

[0196] In this embodiment, the test range is set to 10km*10km, and the range is randomly distributed in 30 areas. Assume that the number of CPU cycles required for tasks in each area is different, and they are uniformly distributed in the interval [500, 1500]Mcycles. There are 5 warehouses, and the warehouse operation and maintenance costs are uniformly distributed in the interval [50, 150]. Each warehouse can provide multiple vehicle-mounted edge servers. The CPU frequency of the vehicle-mounted edge server is uniformly distributed in the interval [150, 400]Mcycles / s. The driving energy consumption per unit time of the vehicle-mounted edge server is 0.12kWh / k, and the driving speed of the vehicle-mounted edge server is uniformly distributed in the interval [30, 50]km / h.

[0197] Considering the impact of the driving energy consumption of the vehicle edge servers on the number of vehicle edge servers, the vehicle edge server scheduling algorithm is always better than other algorithms. The number of vehicle edge servers is reduced by 64.1% and 60.4% compared with the greedy algorithm and the path segmentation algorithm, respectively.

[0198] Considering the impact of the number of CPU cycles required for the task on the number of vehicle edge servers, the vehicle edge server scheduling algorithm is always better than other algorithms. The number of vehicle edge servers is reduced by 58.3% and 62.4% compared with the greedy algorithm and path segmentation algorithm, respectively.

[0199] Considering the impact of the number of regions on the number of vehicle edge servers, the vehicle edge server scheduling algorithm is always better than other algorithms. The number of vehicle edge servers is reduced by 41.6% and 50% compared with the greedy algorithm and path segmentation algorithm, respectively.

[0200] Example 2

[0201] Figure 5 Schematic diagram of the on-board edge server scheduling system to minimize the number of on-board edge servers. Figure 5 As shown, this embodiment provides a vehicle-mounted edge server scheduling system that minimizes the number of vehicle-mounted edge servers, and the system includes:

[0202] An acquisition module 501 is used to acquire a set of regions and warehouses and establish a network model, wherein the vehicle-mounted edge servers in the warehouse provide computing services for multiple regions, the vehicle-mounted edge servers in the same warehouse are of the same type, and the vehicle-mounted edge servers in different warehouses are of different types;

[0203] Establishing module 502, for determining the total energy consumption of the vehicle-mounted edge servers in the warehouse in the driving path, and establishing a problem of minimizing the number of vehicle-mounted edge servers;

[0204] The scheduling module 503 is used to optimize and solve the quantity minimization problem, determine the scheduling plan of the vehicle-mounted edge server, and schedule the vehicle-mounted edge server in the warehouse according to the scheduling plan.

[0205] Preferably, the establishment module 502 determines the total energy consumption of the vehicle-mounted edge server in the warehouse during the driving path, including:

[0206] Warehouses j The vehicle-mounted edge server m is on the driving path l j，m Total energy consumption in Calculated by the following formula:

[0207]

[0208]

[0209] Among them, ρ i Represents computing task a i The number of CPU cycles required, Represents the energy consumption of a CPU cycle, Indicates the capacitance switch related to the vehicle chip; is the driving energy consumption per unit driving distance of the vehicle-mounted edge server, e ij Indicates that the vehicle-mounted edge computing server is from the warehouse s j Arrival Area v i The operator “||” indicates the length of the edge; es j (v i ) is warehouse s j The vehicle-mounted edge server in the execution area v i The computing energy consumption, w(e ij ) is the vehicle-mounted edge computing server from the warehouse s j Arrival Area v i Driving energy consumption; For warehouses j The vehicle-mounted edge server m is on the driving path l j，m The driving energy consumption under For warehouses j In-vehicle edge server m execution The computational energy consumption of all regions in the collection.

[0210] Preferably, the establishment module 502 establishes the problem of minimizing the number of vehicle-mounted edge servers, including:

[0211] The problem of minimizing the number of onboard edge servers is expressed as:

[0212]

[0213]

[0214]

[0215]

[0216]

[0217] Among them, M j Represents each warehouse s j The set of scheduled vehicle-mounted edge servers; B j Represents warehouses j The energy of the onboard edge server in the vehicle; Represents the set of warehouses whose service area set is not empty; V j For warehouses j The set of regions where services are provided; i，m Represents region v i Is it served by the vehicle-mounted edge server m? If it is served by the vehicle-mounted edge server m, then g i，m =1, otherwise g i，m =0.

[0218] The specific implementation process of the functions realized by each module in this embodiment 2 is the same as the implementation process in embodiment 1, and will not be repeated here.

[0219] Example 3

[0220] This embodiment provides a vehicle-mounted edge server scheduling system that minimizes the number of vehicle-mounted edge servers. The system includes: a processor and a memory. The memory stores computer-executable instructions. When the computer-executable instructions are executed by the processor, the method steps in Example 1 are implemented. The specific implementation process can refer to the implementation process of the method steps in Example 1, which will not be repeated here.

[0221] Example 4

[0222] This embodiment provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are executed by a processor, the method steps in Embodiment 1 are implemented. The specific implementation process can refer to the implementation process of the method steps in Embodiment 1, which will not be repeated here.

[0223] It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0224] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0225] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0226] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A vehicle-mounted edge server scheduling method for minimizing the number of vehicle-mounted edge servers, characterized in that: The method comprises: Obtain a set of regions and warehouses and establish a network model, wherein the vehicle-mounted edge servers in the warehouse provide computing services for multiple regions, the vehicle-mounted edge servers in the same warehouse are of the same type, and the vehicle-mounted edge servers in different warehouses are of different types; Determine the total energy consumption of the on-board edge servers in the warehouse during the driving path and establish the problem of minimizing the number of on-board edge servers; Optimizing and solving the quantity minimization problem, determining a scheduling scheme for the vehicle-mounted edge servers, and scheduling the vehicle-mounted edge servers in the warehouse according to the scheduling scheme; Determining the total energy consumption of the vehicle-mounted edge server in the warehouse during the driving path includes: Warehouses j The vehicle-mounted edge server m is on the driving path l j,m Total energy consumption in Calculated by the following formula: Among them, ρ i Represents computing task a i The number of CPU cycles required, Represents the energy consumption of a CPU cycle, Indicates the capacitance switch related to the vehicle chip; is the driving energy consumption per unit driving distance of the vehicle-mounted edge server, e ij Indicates that the vehicle-mounted edge computing server is from the warehouse s j Arrival Area v i The operator "||" indicates the length of the edge; ex j (v i ) is warehouse s j The vehicle-mounted edge server in the execution area v i The computing energy consumption, w(e ij ) is the vehicle-mounted edge computing server from the warehouse s j Arrival Area v i Driving energy consumption; For warehouses j The vehicle-mounted edge server m is on the driving path l j,m The driving energy consumption under For warehouses j In-vehicle edge server m execution The computational energy consumption of all regions in the collection; For path l j,m The set of edges on ; The problem of minimizing the number of vehicle-mounted edge servers includes: The problem of minimizing the number of onboard edge servers is expressed as: Among them, M j Represents each warehouse s j The set of scheduled vehicle-mounted edge servers; B j Represents warehouses j The energy of the onboard edge server in the vehicle; Represents the set of warehouses whose service area set is not empty; V j For warehouses j The set of regions where services are provided; i,m Represents region v i Is it served by the vehicle-mounted edge server m? If it is served by the vehicle-mounted edge server m, then g i,m =1, otherwise g i,m =0; The optimizing and solving the quantity minimization problem to determine the scheduling scheme of the vehicle-mounted edge server includes: (4.1) Input undirected graph G = (V ∪ S, E), where V = {v1, v2, …, v n } represents a set of n regions, S = {s1, s2, …, s h } represents the set of h warehouses, E represents the set of paths between warehouses and regions and between regions; (4.2) Initialize the driving path set of the vehicle edge server The set of warehouses whose service area set is not an empty set (4.3) Traverse all warehouses, if V j is not an empty set, then (4.4) Traversal For all warehouses in , execute steps (4.5) to (4.13) until all warehouses are traversed, and execute step (4.14); (4.5) For each warehouse s j , construct warehouses j An undirected graph G with only edge weights j ', the weight value of each edge is as follows: G j 'Middle side ii' The weight w'(e ii' )for: G j 'Middle side ij The weight w'(e ij )for: (4.6) Initialization Among them, the operator Indicates rounding up to an integer; (4.7) Warehouse s j Service area collection V j The area in is divided into z+1 categories, traversing the loop variable x from 0 to z, warehouse s j The x-th region set Initialize to an empty set; (4.8) Traverse V j Each region v in i , execute steps (4.9), (4.10), and (4.11) until all areas have been traversed, and then execute step (4.12); (4.9) If (4.10) If (4.11) Traverse the loop variable x from 1 to z-1, if Update (4.12) Traverse the loop variable x from 0 to z and construct the warehouse s j The divided graph Representing a collection The edge set composed of warehouse s j The set of paths in the partitioned graph unrooted DVRP stands for unrooted vehicle routing algorithm with distance constraints; (4.13) Traversal All paths in renew (4.14) Output the driving path set of the vehicle-mounted edge server.

2. The method according to claim 1, characterized in that The unrooted DVRP includes the following steps: (5.1) Input warehouse s j The divided graph Let γ = 2 x Delta-1; (5.2) Initialize variable k to 1; (5.3) If Execute step (5.4), otherwise execute step (5.10); (5.4) Generate a minimum weight forest F with nk edges based on graph G. First, sort the edges in ascending order of weight, and then add the edges one by one. If the newly added edge does not cause a loop in the forest, the new edge is added to the forest until a minimum weight forest with nk edges is generated. At the same time, generate connected components C1, C2, …, C k ;Path collection (5.5) Initialize variable i to 1; (5.6) If i≤k, then initialize the number of vehicle-mounted edge servers N(i,k)=1 under the current i and k values, according to C i The generation loop TOUR i , the TOUR i Delete an edge to generate path P i , then execute step (5.7); otherwise, execute step (5.9); (5.7) If P i If the path length is greater than γ, then from path P i Find the path P whose path length is closest to γ ​​but less than γ i ', update path P i =P i \P i ', SQL k =SQL k ∪P i ', N(i,k)=N(i,k)+1; otherwise, execute step (5.8); where "\" means to find the difference of sets; (5.8)SQL k =SQL k ∪P i , i=i+1, then execute step (5.6); (5.9) k = k + 1, execute step (5.3); (5.10) Output SQL k' ;in express The value of k when the value is minimum.

3. A system using the vehicle edge server scheduling method for minimizing the number of vehicle edge servers according to any one of claims 1 to 2, characterized in that: The system comprises: An acquisition module is used to acquire a set of regions and warehouses and establish a network model, wherein the vehicle-mounted edge servers in the warehouse provide computing services for multiple regions, the vehicle-mounted edge servers in the same warehouse are of the same type, and the vehicle-mounted edge servers in different warehouses are of different types; Establish a module to determine the total energy consumption of the vehicle-mounted edge servers in the warehouse during the driving path and establish the problem of minimizing the number of vehicle-mounted edge servers; The scheduling module is used to optimize and solve the quantity minimization problem, determine the scheduling plan of the vehicle-mounted edge server, and schedule the vehicle-mounted edge server in the warehouse according to the scheduling plan.

4. The system according to claim 3, characterized in that The establishment module determines the total energy consumption of the vehicle-mounted edge server in the warehouse during the driving path, including: Warehouses j The vehicle-mounted edge server m is on the driving path l j,m Total energy consumption in Calculated by the following formula: Among them, ρ i Represents computing task a i The number of CPU cycles required, Represents the energy consumption of a CPU cycle, Indicates the capacitance switch related to the vehicle chip; is the driving energy consumption per unit driving distance of the vehicle-mounted edge server, e ij Indicates that the vehicle-mounted edge computing server is from the warehouse s j Arrival Area v i The operator "||" indicates the length of the edge; ex j (v i ) is warehouse s j The vehicle-mounted edge server in the execution area v i The computing energy consumption, w(e ij ) is the vehicle-mounted edge computing server from the warehouse s j Arrival Area v i Driving energy consumption; For warehouses j The vehicle-mounted edge server m is on the driving path l j,m The driving energy consumption under For warehouses j In-vehicle edge server m execution The computational energy consumption of all regions in the collection.

5. The system according to claim 4, characterized in that The establishment module establishes the problem of minimizing the number of vehicle-mounted edge servers, including: The problem of minimizing the number of onboard edge servers is expressed as: Among them, M j Represents each warehouse s j The set of scheduled vehicle-mounted edge servers; B j Represents warehouses j The energy of the onboard edge server in the vehicle; Represents the set of warehouses whose service area set is not empty; V j For warehouses j The set of regions where services are provided; i,m Represents region v i Is it served by the vehicle-mounted edge server m? If it is served by the vehicle-mounted edge server m, then g i,m =1, otherwise g i,m =0.

6. A vehicle-mounted edge server scheduling system for minimizing the number of vehicle-mounted edge servers, characterized in that: The system comprises: a processor and a memory, wherein the memory stores computer executable instructions, and when the computer executable instructions are executed by the processor, the method according to any one of claims 1 to 2 is implemented.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, the method according to any one of claims 1 to 2 is implemented.

Images