Dynamic vehicle-mounted cloud enabled undirected weighted graph task hybrid scheduling method

Abstract

The embodiment of the invention provides a dynamic vehicle-mounted cloud enabled undirected weighted graph task hybrid scheduling method. The method comprises the following steps: regarding a compute-intensive task and a vehicle-mounted cloud as an undirected weighted graph, performing task graph model modeling on the compute-intensive task to obtain a task graph model, and performing vehicle-mounted cloud model modeling on the vehicle-mounted cloud to obtain a vehicle-mounted cloud model; carrying out task completion time modeling, data interaction cost modeling and uncertainty modeling on the basis to obtain task completion time, data interaction cost and uncertainty; on the basis, the scheduling problem of the compute-intensive tasks is converted into an optimization problem, and the aim of the optimization problem is to find an optimal task mapping template so as to minimize a target function of task scheduling; and if the solving result is not available, solving the optimization problem by adopting an offline task scheduling mode, and if the solving result is not available, solving the optimization problem by adopting an online task scheduling mode. In this way, compute-intensive tasks can be efficiently and reliably scheduled.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle networking, and particularly to an undirected weighted graph task hybrid scheduling method empowered by dynamic vehicle-mounted cloud. Background Art

[0002] With the rapid development of vehicle networking technology, vehicle-mounted cloud, as an emerging computing platform, is playing an increasingly important role in intelligent transportation systems. However, existing task scheduling methods often have problems such as low scheduling efficiency, low resource utilization rate, and long task completion time when facing a dynamically changing vehicle-mounted cloud environment. Especially for computationally intensive tasks with an internal processing topology structure, such as large data analysis, image processing, etc., the scheduling difficulty is greater. Therefore, how to design an efficient and reliable task scheduling method to make full use of the computing resources of the vehicle-mounted cloud has become an urgent technical problem to be solved. Summary of the Invention

[0003] In a first aspect, an embodiment of the present invention provides an undirected weighted graph task hybrid scheduling method empowered by dynamic vehicle-mounted cloud, the method comprising:

[0004] Regarding computationally intensive tasks and vehicle-mounted cloud as an undirected weighted graph, performing task graph model modeling on the computationally intensive tasks to obtain a task graph model, and performing vehicle-mounted cloud graph model modeling on the vehicle-mounted cloud to obtain a vehicle-mounted cloud graph model;

[0005] Based on the task graph model and the vehicle-mounted cloud graph model, performing task completion time modeling, data interaction cost modeling, and uncertainty modeling to obtain task completion time, data interaction cost, and uncertainty;

[0006] Based on the task graph model, the vehicle-mounted cloud graph model, the task completion time, the data interaction cost, and the uncertainty, converting the scheduling problem of computationally intensive tasks into an optimization problem, the goal of which is to find an optimal task mapping template to minimize the objective function of task scheduling;

[0007] Using an offline task scheduling mode to solve the optimization problem, and if the solution result is unavailable, using an online task scheduling mode to solve the optimization problem.

[0008] In some feasible implementations of the first aspect, the task graph model is represented as:

[0009] G task ={V task ,E task ,W task};

[0010] Wherein, V task is a vertex set, representing a set of task components, and each task component has attribute data; E taskis the edge set, representing the set of data exchange requirements between task components. Each edge has a weight, represented as the weight set W task , describing the time or bandwidth consumption required for data exchange;

[0011] The vehicle-mounted cloud map model is represented as:

[0012] G serv ={V serv , E serv , W serv};

[0013] Among them, V serv is the vertex set, representing the set of service providers in the vehicle-mounted cloud. Each service provider has attribute data; E serv is the edge set, representing the set of connection relationships between service providers. Each edge has a weight, represented as the weight set W serv , describing the connection duration and connection quality.

[0014] In some realizable ways of the first aspect, the task completion time is represented as:

[0015]

[0016] Among them, is the task completion time; α n,m is a binary variable. When α n,m = 1, it means that the task component v n is scheduled to be executed on the service provider s m . Otherwise, α n,m = 0; A is the set of all α n,m ; is the final completion time of the task component v n .

[0017] In some realizable ways of the first aspect, the data interaction cost is represented as:

[0018]

[0019] Among them, is the data interaction cost; β m,m′ is a binary variable. β m , m′ = 1 means that when two associated task components are assigned to different service providers, data interaction cost will be generated during the processing of the task components. Otherwise, β m,m′ = 0; B is the set of all β m,m′ ; is for two service providers s m and s m′The data interaction cost.

[0020] In some realizable ways of the first aspect, the uncertainty is represented as the connection time between service providers The data interaction cost between service providers The computing power f of the service provider m And the data transmission rate r from the task vehicle to the service vehicle m The corresponding random numbers respectively, and the random numbers follow a known probability distribution.

[0021] In some realizable ways of the first aspect, in the offline task scheduling mode, the objective function corresponding to the optimization problem is expressed as the weighted sum of the mathematical expectations of the task completion time and the data interaction cost; the constraint conditions corresponding to the optimization problem include that the service provider for scheduling the task component is unique; the probability that the task component completion time exceeds the time limit is less than a threshold; if two associated task components are calculated on two service providers respectively, then these two service providers maintain a certain connection with a certain probability;

[0022] Using the offline task scheduling mode to solve the optimization problem, including:

[0023] Using an algorithm based on region exploration and heterogeneous subgraph search to solve the optimization problem.

[0024] In some realizable ways of the first aspect, in the online task scheduling mode, the objective function corresponding to the optimization problem is expressed as the weighted sum of the actual task completion time and the actual data interaction cost; the constraint conditions corresponding to the optimization problem include that the service provider for scheduling the task component is unique; the task component must be completed within the specified time; the service providers for processing two associated task components must maintain a certain communication time to meet the interaction of intermediate data;

[0025] Using the online task scheduling mode to solve the optimization problem, including:

[0026] Using an instantaneous heterogeneous subgraph search algorithm to solve the optimization problem.

[0027] In the second aspect, an embodiment of the present invention provides a dynamic vehicle-mounted cloud-enabled undirected weighted graph task hybrid scheduling device, and the device includes:

[0028] A modeling module, configured to regard the compute-intensive task and the vehicle-mounted cloud as an undirected weighted graph, perform task graph model modeling on the compute-intensive task to obtain a task graph model, and perform vehicle-mounted cloud graph model modeling on the vehicle-mounted cloud to obtain a vehicle-mounted cloud graph model;

[0029] The modeling module is also used to perform task completion time modeling, data interaction cost modeling, and uncertainty modeling based on the task graph model and the vehicle-mounted cloud graph model, so as to obtain the task completion time, data interaction cost, and uncertainty;

[0030] The conversion module is used to convert the scheduling problem of compute-intensive tasks into an optimization problem based on the task graph model, the vehicle-mounted cloud graph model, the task completion time, the data interaction cost, and the uncertainty. The goal is to find the optimal task mapping template to minimize the objective function of task scheduling;

[0031] The solving module is used to solve the optimization problem in an offline task scheduling mode. If the solution result is unavailable, it will solve the optimization problem in an online task scheduling mode.

[0032] In a third aspect, an embodiment of the present invention provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as described above.

[0033] In a fourth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium storing computer instructions, and the computer instructions are used to cause a computer to execute the method as described above.

[0034] In the embodiments of the present invention, by constructing a graph model of compute-intensive tasks and vehicle-mounted clouds, a scheduling strategy combining offline and online is designed to achieve efficient and reliable scheduling of compute-intensive graph tasks in a dynamically changing vehicle-mounted cloud environment.

[0035] It should be understood that the content described in the summary of the invention is not intended to limit the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] With reference to the accompanying drawings and the following detailed description, the above and other features, advantages, and aspects of the embodiments of the present invention will become more apparent. The drawings are used to better understand the present invention and do not constitute a limitation to the present invention. In the drawings, the same or similar reference numerals represent the same or similar elements, where:

[0037] Figure 1 is a flowchart of a method for hybrid scheduling of undirected weighted graph tasks empowered by a dynamic vehicle-mounted cloud provided by an embodiment of the present invention;

[0038] Figure 2 is a schematic diagram of a graph task scheduling scenario for a vehicle-mounted cloud provided by an embodiment of the present invention;

[0039] Figure 3 Schematic diagram of the offline task scheduling mode and the online task scheduling mode provided by the embodiments of the present invention;

[0040] Figure 4 Schematic diagram of the graph-structured task tested in the simulation provided by the embodiments of the present invention;

[0041] Figure 5 Schematic diagram of the performance analysis of the objective function provided by the embodiments of the present invention;

[0042] Figure 6 Schematic diagram of the performance analysis of the time cost of policy generation provided by the embodiments of the present invention;

[0043] Figure 7 Schematic diagram of the performance analysis of the graph task completion time provided by the embodiments of the present invention;

[0044] Figure 8 Structure diagram of a dynamic vehicle-mounted cloud-enabled undirected weighted graph task hybrid scheduling device provided by the embodiments of the present invention;

[0045] Figure 9 Structure diagram of an exemplary electronic device capable of implementing the embodiments of the present invention. Detailed implementation manners

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0047] In addition, the term "and / or" in the present invention is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in the present invention generally represents an "or" relationship between the front and rear associated objects.

[0048] To solve the technical problems in the background art, the embodiments of the present invention provide a dynamic vehicle-mounted cloud-enabled undirected weighted graph task hybrid scheduling method, device, equipment, and storage medium. The following will, with reference to the accompanying drawings, elaborate on a dynamic vehicle-mounted cloud-enabled undirected weighted graph task hybrid scheduling method, device, equipment, and storage medium provided by the embodiments of the present invention through specific embodiments.

[0049] Figure 1This is a flowchart of a dynamic vehicle-mounted cloud-enabled undirected weighted graph task hybrid scheduling method provided by an embodiment of the present invention. As Figure 1 shown, the undirected weighted graph task hybrid scheduling method 100 may include:

[0050] S110. Regard the compute-intensive tasks and the vehicle-mounted cloud as an undirected weighted graph, perform task graph model modeling on the compute-intensive tasks to obtain a task graph model, and perform vehicle-mounted cloud graph model modeling on the vehicle-mounted cloud to obtain a vehicle-mounted cloud graph model.

[0051] S120. Based on the task graph model and the vehicle-mounted cloud graph model, perform task completion time modeling, data interaction cost modeling, and uncertainty modeling to obtain the task completion time, data interaction cost, and uncertainty.

[0052] S130. Based on the task graph model, the vehicle-mounted cloud graph model, the task completion time, the data interaction cost, and the uncertainty, convert the scheduling problem of the compute-intensive tasks into an optimization problem, the goal of which is to find the optimal task mapping template to minimize the objective function of task scheduling.

[0053] S140. Solve the optimization problem using an offline task scheduling mode. If the solution result is unavailable, then solve the optimization problem using an online task scheduling mode.

[0054] In the embodiment of the present invention, by constructing a graph model of compute-intensive tasks and the vehicle-mounted cloud, a scheduling strategy combining offline and online is designed to achieve efficient and reliable scheduling of compute-intensive graph tasks in a dynamically changing vehicle-mounted cloud environment.

[0055] For the convenience of further understanding, the above content will be described in detail below in combination with Figure 2 the scenario shown:

[0056] (1) Modeling

[0057] (1.1) Task graph model modeling

[0058] The task graph model is expressed as:

[0059] G task ={V task , E task , W task};

[0060] Among them, V task is the vertex set, representing the task component set. Each task component has attribute data, for example, specific computing requirements (such as CPU, memory, storage, etc.), data size, and tolerable completion time, etc.; E task is the edge set, representing the data exchange requirement set between task components. Each edge has a weight, expressed as the weight set Wtask , which describes the time or bandwidth consumption required for data exchange.

[0061] (1.2) Modeling of the in-vehicle cloud map model

[0062] The in-vehicle cloud map model is expressed as:

[0063] G serv = {V serv , E serv , W serv};

[0064] Among them, V serv is the set of vertices, representing the set of service providers (abbreviated as SP) in the in-vehicle cloud. Each service provider has attribute data, for example, specific computing capabilities (such as the number of CPU cores, memory size, storage capacity, etc.), geographical location, connection speed, etc.; E serv is the set of edges, representing the set of connection relationships between service providers. Each edge has a weight, expressed as the weight set W serv , which describes the connection duration (i.e., the time period during which service providers can maintain a connection) and connection quality (such as bandwidth, latency, etc.).

[0065] (1.3) Modeling of the task completion time

[0066] The computing time of the task component v n on the service provider s m can be expressed as: The data transmission delay can be expressed as: Among them, q n represents the computing resources required by this task component, f m represents the computing ability of this service provider, d n represents the data volume of this task component, r m represents the data transmission rate of this task component's data from the task vehicle to the service provider. Therefore, the final completion time of the task component v n can be expressed as: On this basis, the task completion time can be expressed as:

[0067]

[0068] Among them, is the task completion time; α n,m is a binary variable. When α n,m = 1, it means that the task component v n is scheduled to be executed on the service provider s m , otherwise α n,m = 0; A is the set of all α n,m ; For the final completion time of task component v n

[0069] (1.4) Data interaction cost modeling

[0070] Here, the data interaction cost between service providers in the vehicle cloud is introduced. Use to represent the data interaction cost between two service providers s m and s m′ (if they process two connected task components, they need to exchange intermediate data during the data processing, thus generating a certain time or energy consumption cost). In addition, to indicate whether this cost is generated, a binary variable β m,m′ is introduced:

[0071]

[0072] where β m,m′ = 1 indicates that two associated task components are assigned to different service providers, and data interaction cost will be generated during the processing of task components. Conversely, β m,m′ = 0.

[0073] Furthermore, use B to represent the set of all β m,m′ , and represent the total data interaction cost of processing the entire task as:

[0074]

[0075] where is the data interaction cost; β m,m′ , B are as described above; is the data interaction cost between two service providers s m and s m′ .

[0076] (1.5) Uncertainty modeling

[0077] To be closer to the actual vehicle networking scenario, uncertainty is introduced here, that is, the connection time between service providers the data interaction cost between service providers the computing power f of service providers m and the data transmission rate r between the task vehicle and the service vehicle m all change with time. Therefore, they are all modeled as random numbers here, and each random number follows a known probability distribution (these distributions can be obtained from the historical data of the real dataset).

[0078] (2) Optimization problem

[0079] ​Based on the above modeling, the task scheduling problem of compute-intensive tasks is transformed into an optimization problem here, and its goal is to find an optimal task mapping template to minimize the objective function of task scheduling. This objective function comprehensively considers multiple factors such as task completion time, data exchange cost, and resource utilization efficiency.

[0080] However, because there are uncertain random numbers in the optimization, in the offline task scheduling mode, the objective function corresponding to the optimization problem is expressed as the weighted sum of the mathematical expectations of the task completion time and the data interaction cost; the constraint conditions corresponding to the optimization problem include that the service provider for scheduling task components is unique (C1); the probability that the task component completion time times out is less than a threshold (C2); if two associated task components are calculated on two service providers respectively, then these two service providers maintain a certain connection with a certain probability (C3);

[0081] To sum up, in the offline task scheduling mode, the optimization problem can be expressed as:

[0082]

[0083] s.t.,

[0084]

[0085] Among them, the constraint condition (C1) means that each task component can be scheduled to at most one service provider; the constraint condition (C2) means that if the task component v n is calculated on the service provider s m then the probability that the completion time times out must be less than a threshold ξ; (C3) means that if two associated task components v n and v n′ are calculated on the service providers s m and s m′ respectively, then these two service providers need to maintain a certain connection with a certain probability. A * represents the offline scheduling decision obtained by solving the above optimization problem. The symbol represents the expectation.

[0086] In the online task scheduling mode, the objective function corresponding to the optimization problem is expressed as the weighted sum of the actual task completion time and the actual data interaction cost; the constraint conditions corresponding to the optimization problem include that the service provider for scheduling task components is unique (C1); the task component must be completed within the specified time (C4); the service providers for two processing-associated task components must maintain a certain communication time to meet the interaction of intermediate data (C5);

[0087] To sum up, in the online task scheduling mode, the optimization problem can be expressed as:

[0088]

[0089] s.t.,

[0090]

[0091] Among them, the constraint condition (C1) is as described above; the constraint condition (C4) means that the component must be completed within the specified time, and the constraint condition (C5) means that the service providers of two processing-related task components must maintain a certain communication time to meet the interaction of intermediate data; A ** represents the scheduling decision obtained by solving this online optimization objective.

[0092] (3) Algorithm Design

[0093] The offline task scheduling mode and the online task scheduling mode can be as Figure 3 shown.

[0094] (3.1) Offline Task Scheduling Mode

[0095] An algorithm based on region exploration and heterogeneous subgraph search (RA-PilotISS) is used to solve the optimization problem.

[0096] (3.1.1) Region Exploration:

[0097] Within the given time limit, by traversing the vertices (service providers) in the vehicle-mounted cloud graph model, a set of candidate service providers with similar attributes to the vertices (task components) in the task graph model is found. The purpose of this step is to narrow the search scope and improve the search efficiency. Optionally, the similar attributes can include computing power, geographical location, connection speed, etc.

[0098] (3.1.2) Heterogeneous Subgraph Search:

[0099] In the set of candidate service providers, an optimal task mapping template that meets the constraint conditions is found through the heterogeneous subgraph search algorithm. The core of this step is to design an efficient search strategy to find the optimal solution within a limited time. Optionally, the heterogeneous subgraph search algorithm can adopt optimization algorithms such as heuristic search algorithms, genetic algorithms, simulated annealing, etc.

[0100] (3.1.3) Template Optimization:

[0101] The found optimal task mapping template is further optimized to further improve the efficiency and quality of task scheduling.

[0102] The optimization methods can include adjusting weight coefficients, optimizing resource allocation strategies, etc.

[0103] (3.2) Online Task Scheduling Mode

[0104] The transient heterogeneous subgraph search algorithm (TE-InstaISS) is used to solve the optimization problem.

[0105] (3.2.1) Transient heterogeneous subgraph search algorithm:

[0106] Under the current network conditions, quickly find an available task mapping scheme.

[0107] The algorithm first constructs a real-time on-vehicle cloud map model based on the connection status and resource usage of the current service provider.

[0108] Then, using heuristic search or local search strategies, quickly find a task mapping scheme that meets the constraints in the real-time on-vehicle cloud map model.

[0109] (3.2.2) Dynamic adjustment strategy:

[0110] During the task execution process, dynamically adjust the task mapping scheme according to the real-time changes in the connection status and resource usage of the service provider.

[0111] The adjustment strategy can include reallocating task components, migrating task components, etc.

[0112] (4) Simulation experiments

[0113] To verify the effectiveness of the proposed method, simulation experiments are conducted here. The experimental results show that under problems of different scales, the method of the present invention is superior to the existing methods in terms of task completion time, data exchange cost, resource utilization efficiency, etc. Specifically, by comparing the performance of the method of the present invention (P-HTS) with several benchmark methods (such as InstaISS, InstaSISO, TPTS, DPTS, RTS, etc.) in terms of task completion time (TCT), data exchange cost (DEC), and cost function (CF), it can be found that the method of the present invention performs excellently in all aspects. The simulation experiments test three types of graph tasks as Figure 4 shown; the performance analysis of the objective function can be as Figure 5 shown; the performance analysis of the strategy generation time cost can be as Figure 6 shown; the performance analysis of the graph task completion time can be as Figure 7 shown.

[0114] It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0115] The above is the introduction of the method embodiments. The following will further illustrate the solution of the present invention through device embodiments.

[0116] Figure 8 The following is a structural diagram of a dynamic vehicle-mounted cloud-enabled undirected weighted graph task hybrid scheduling device provided by an embodiment of the present invention. As Figure 8 shown, the undirected weighted graph task hybrid scheduling device 800 may include:

[0117] A modeling module 810, configured to regard a compute-intensive task and a vehicle-mounted cloud as an undirected weighted graph, perform task graph model modeling on the compute-intensive task to obtain a task graph model, and perform vehicle-mounted cloud graph model modeling on the vehicle-mounted cloud to obtain a vehicle-mounted cloud graph model.

[0118] The modeling module 810 is further configured to perform task completion time modeling, data interaction cost modeling, and uncertainty modeling based on the task graph model and the vehicle-mounted cloud graph model to obtain the task completion time, the data interaction cost, and the uncertainty.

[0119] A conversion module 820, configured to convert the scheduling problem of the compute-intensive task into an optimization problem based on the task graph model, the vehicle-mounted cloud graph model, the task completion time, the data interaction cost, and the uncertainty. The goal is to find an optimal task mapping template to minimize the objective function of task scheduling;

[0120] A solving module 830, configured to solve the optimization problem in an offline task scheduling mode. If the solution result is unavailable, the optimization problem is solved in an online task scheduling mode.

[0121] It can be understood that Figure 8 each module / unit in the undirected weighted graph task hybrid scheduling device 800 shown has the function of implementing Figure 1 each step in the undirected weighted graph task hybrid scheduling method 100 shown and can achieve its corresponding technical effects. For the sake of brevity, it will not be described in detail here.

[0122] Figure 9It is a structural diagram of an exemplary electronic device capable of implementing the embodiments of the present invention. The electronic device 900 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device 900 can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown in the present invention, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed in the present invention.

[0123] As Figure 9 shown, the electronic device 900 may include a computing unit 901, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the electronic device 900 can also be stored. The computing unit 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0124] A plurality of components in the electronic device 900 are connected to the I / O interface 905, including: an input unit 906, such as a keyboard, a mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a magnetic disk, an optical disk, etc.; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 909 allows the electronic device 900 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0125] The computing unit 901 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 executes the various methods and processes described above, such as method 100. For example, in some embodiments, method 100 can be implemented as a computer program product, including a computer program tangibly embodied in a computer-readable medium, such as the storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 900 via the ROM 902 and / or the communication unit 909. When the computer program is loaded into the RAM 903 and executed by the computing unit 901, one or more steps of the method 100 described above can be executed. Alternatively, in other embodiments, the computing unit 901 can be configured to execute method 100 in any other suitable manner (e.g., by means of firmware).

[0126] The various embodiments described above in the present invention can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on a chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special or general programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0127] The program code for implementing the methods of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to the processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the program codes are executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program codes can be executed entirely on the machine, partially on the machine, as an independent software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0128] In the context of the present invention, a computer-readable medium may be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of the computer-readable storage medium would include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0129] It should be noted that the present invention also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute method 100 and achieve the corresponding technical effects achieved by the method of the embodiments of the present invention. For the sake of concise description, it will not be elaborated herein.

[0130] In addition, the present invention also provides a computer program product, which includes a computer program that implements method 100 when executed by a processor.

[0131] It should be understood that the various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present invention can be achieved. The present invention places no restrictions herein.

[0132] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A dynamic vehicle-mounted cloud-enabled undirected weighted graph task hybrid scheduling method, characterized in that: The method comprises: The computing-intensive task and the vehicle-mounted cloud are regarded as an undirected weighted graph, a task graph model is modeled for the computing-intensive task to obtain a task graph model, and a vehicle-mounted cloud graph model is modeled for the vehicle-mounted cloud to obtain a vehicle-mounted cloud graph model; Based on the task graph model and the vehicle cloud graph model, the task completion time, data interaction cost and uncertainty are modeled to obtain the task completion time, data interaction cost and uncertainty. Based on the task graph model, vehicle cloud graph model, task completion time, data interaction cost, and uncertainty, the scheduling problem of computationally intensive tasks is converted into an optimization problem. Its goal is to find the optimal task mapping template to minimize the objective function of task scheduling. The optimization problem is solved using the offline task scheduling mode. If the solution is not available, the optimization problem is solved using the online task scheduling mode.

2. The method according to claim 1, characterized in that: The task graph model is expressed as: G task ={V task ,E task ,W task }; Among them, V task is a vertex set, representing a set of task components, each of which has attribute data; task is a set of edges, representing the set of data exchange requirements between task components. Each edge has a weight, represented as the weight set W task , describes the time or bandwidth consumption required for data exchange; The vehicle-mounted cloud image model is expressed as: G serv ={V serv ,E serv ,W serv }; Among them, V serv is a vertex set, representing the set of service providers in the vehicle-mounted cloud, each service provider has attribute data; serv is a set of edges, representing the set of connections between service providers. Each edge has a weight, represented as the weight set W serv , describing the connection duration and connection quality.

3. The method according to claim 2, characterized in that The task completion time is expressed as: in, is the task completion time; α n,m is a binary variable, α n,m =1, indicating that the task component v n Dispatched to service providers m Execute on, otherwise α n,m =0; A is all α n,m A collection of; For the task component v n The final completion time.

4. The method according to claim 3, characterized in that The data interaction cost is expressed as: in, is the data interaction cost; β m,m′ is a binary variable, β m,m′ =1, indicating that two related task components are assigned to different service providers, which will generate data interaction costs during the task component processing. m,m′ =0; B is all β m,m′ A collection of; For two service providers m and m′ The data interaction cost.

5. The method according to claim 4, characterized in that The uncertainty is expressed as the connection time between service providers Data exchange cost between service providers The computing power of the service provider f m and the data transmission rate r between the mission vehicle and the service vehicle m The corresponding random numbers obey the known probability distribution.

6. The method according to claim 5, characterized in that In the offline task scheduling mode, the objective function corresponding to the optimization problem is expressed as the weighted sum of the mathematical expectation of the task completion time and the data interaction cost; the constraints corresponding to the optimization problem include that the service provider of the scheduling task component is unique; the probability of the task component completion time exceeding a threshold is less than a threshold; two associated task components are calculated on two service providers respectively, and the two service providers maintain a certain connection with a certain probability; The offline task scheduling mode is used to solve the optimization problem, including: An algorithm based on region exploration and heterogeneous subgraph search is used to solve the optimization problem.

7. The method according to claim 6, characterized in that In the online task scheduling mode, the objective function corresponding to the optimization problem is expressed as the weighted sum of the actual task completion time and the actual data interaction cost; the constraints corresponding to the optimization problem include that the service provider of the scheduling task component is unique; the task component must be completed within the specified time; The service providers of two processing-related task components must maintain a certain communication time to satisfy the interaction of intermediate data; The online task scheduling mode is used to solve the optimization problem, including: The instantaneous heterogeneous subgraph search algorithm is used to solve the optimization problem.

8. A dynamic vehicle-mounted cloud-enabled undirected weighted graph task hybrid scheduling device, characterized in that: The device comprises: A modeling module is used to regard the computing-intensive task and the vehicle-mounted cloud as an undirected weighted graph, perform task graph modeling on the computing-intensive task to obtain a task graph model, and perform vehicle-mounted cloud graph modeling on the vehicle-mounted cloud to obtain a vehicle-mounted cloud graph model; The modeling module is also used to perform task completion time modeling, data interaction cost modeling and uncertainty modeling based on the task graph model and the vehicle-mounted cloud graph model to obtain the task completion time, data interaction cost and uncertainty; The conversion module is used to convert the scheduling problem of computationally intensive tasks into an optimization problem based on the task graph model, the vehicle cloud graph model, the task completion time, the data interaction cost, and the uncertainty. Its goal is to find the optimal task mapping template to minimize the objective function of task scheduling. The solution module is used to solve the optimization problem in an offline task scheduling mode. If the solution result is not available, the optimization problem is solved in an online task scheduling mode.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 7.

Images