CN110908804A - Computing resource allocation method and device and training method and device of model thereof - Google Patents
Computing resource allocation method and device and training method and device of model thereof Download PDFInfo
- Publication number
- CN110908804A CN110908804A CN201911179666.7A CN201911179666A CN110908804A CN 110908804 A CN110908804 A CN 110908804A CN 201911179666 A CN201911179666 A CN 201911179666A CN 110908804 A CN110908804 A CN 110908804A
- Authority
- CN
- China
- Prior art keywords
- computing resource
- sequence
- resource allocation
- computing
- edge cloud
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5072—Grid computing
Abstract
The application provides a computing resource allocation method, a computing resource allocation device and a training method and a training device of a model thereof, which relate to the technical field of resource allocation and comprise the following steps: receiving computing resource supply quantity sent by a plurality of edge cloud servers and computing resource demand quantity sent by a plurality of user sides; arranging a plurality of computing resource supply quantities in an ascending order to obtain a first sequence; performing descending arrangement on the computing resource demand quantities to obtain a second sequence; determining an intersection point with the same amount of computing resources between the first sequence and the second sequence; and determining a computing resource distribution result between the edge cloud server and the user side according to the intersection point by applying a pre-trained computing resource distribution model, so that the technical problem of unreasonable computing resource distribution result is solved.
Description
Technical Field
The present application relates to the field of resource allocation technologies, and in particular, to a method and an apparatus for allocating computing resources and a method and an apparatus for training a model thereof.
Background
In the edge cloud computing environment, a computing task is uploaded to an edge cloud server closest to a user (physical distance) for computing, and then a computing result is returned to the user. However, unlike cloud servers which have nearly unlimited computing power, the computing resources and computing power of edge cloud servers are limited, and how to reasonably allocate these limited resources becomes a difficult problem when there are too many users around the edge cloud servers.
At present, in the existing computing resource allocation, computing resources are allocated to a user side according to the amount of the computing resources that can be provided by an edge cloud server, and the method easily causes unreasonable results of computing resource allocation.
Disclosure of Invention
The invention aims to provide a computing resource allocation method and device and a training method and device of a model thereof, so as to solve the technical problem that a computing resource allocation result is unreasonable.
In a first aspect, an embodiment of the present application provides a computing resource allocation method, where the method includes:
receiving computing resource supply quantity sent by a plurality of edge cloud servers and computing resource demand quantity sent by a plurality of user sides;
arranging a plurality of computing resource supply quantities in an ascending order to obtain a first sequence; performing descending arrangement on the computing resource demand quantities to obtain a second sequence;
determining an intersection point with the same amount of computing resources between the first sequence and the second sequence;
and determining a computing resource distribution result between the edge cloud server and the user side according to the intersection point by applying a pre-trained computing resource distribution model.
In one possible implementation, the pre-trained computational resource allocation model is an Empirical Weighting Algorithm (EWA) model.
In a second aspect, a method for training a computational resource allocation model is provided, the method comprising:
receiving computing resource supply quantity sent by a plurality of edge cloud servers and computing resource demand quantity sent by a plurality of user sides;
arranging a plurality of computing resource supply quantities in an ascending order to obtain a first sequence; performing descending arrangement on the computing resource demand quantities to obtain a second sequence;
if the maximum value of the first sequence is larger than or equal to the minimum value of the second sequence, determining a calculation resource distribution result between the edge cloud server and the user side according to an intersection point with the same calculation resource amount between the first sequence and the second sequence until the maximum value of the first sequence is smaller than the minimum value of the second sequence;
and training the initial EWA model according to the calculation resource distribution result to obtain a pre-trained calculation resource distribution model.
In one possible implementation, the method further includes:
judging whether the calculation resource allocation result of the current resource allocation period is the same as the calculation resource allocation result of the previous resource allocation period;
if so, determining that the calculation resource distribution result of the next resource distribution cycle is the calculation resource distribution result of the current resource distribution cycle.
In a possible implementation, after the step of determining whether the calculation resource allocation result of the current resource allocation period is the same as the calculation resource allocation result of the previous resource allocation period, the method further includes:
and if not, re-receiving the next-period computing resource supply quantity sent by the edge cloud servers and the next-period computing resource demand quantity sent by the user sides in the next resource allocation period.
In one possible implementation, the next cycle computing resource supply and the next cycle computing resource demand are policy adjusted computing resource amounts.
In a third aspect, an apparatus for allocating computing resources is provided, the apparatus comprising:
the receiving module is used for receiving the computing resource supply quantity sent by the edge cloud servers and the computing resource demand quantity sent by the user sides;
the arrangement module is used for carrying out ascending arrangement on the plurality of computing resource supply quantities to obtain a first sequence; performing descending arrangement on the computing resource demand quantities to obtain a second sequence;
a first determining module, configured to determine an intersection point where the amount of computing resources is the same between the first sequence and the second sequence;
and the second determining module is used for applying a pre-trained computing resource distribution model and determining a computing resource distribution result between the edge cloud server and the user side according to the intersection point.
In a fourth aspect, there is provided a training apparatus for a computational resource allocation model, the apparatus comprising:
the receiving unit is used for receiving the computing resource supply quantity sent by the edge cloud servers and the computing resource demand quantity sent by the user sides;
the arrangement unit is used for carrying out ascending arrangement on the plurality of computing resource supply quantities to obtain a first sequence and carrying out descending arrangement on the plurality of computing resource demand quantities to obtain a second sequence;
a determining unit, configured to determine, if the maximum value of the first sequence is greater than or equal to the minimum value of the second sequence, a result of computing resource allocation between the edge cloud server and the user side according to an intersection point where the amount of computing resources between the first sequence and the second sequence is the same until the maximum value of the first sequence is smaller than the minimum value of the second sequence;
and the training unit is used for training the initial EWA model according to the calculation resource distribution result to obtain a pre-trained calculation resource distribution model.
In a fifth aspect, this application provides an electronic device, which includes a memory and a processor, where the memory stores a computer program that is executable on the processor, and the processor executes the computer program to implement the method of the first aspect or the second aspect.
In a sixth aspect, this embodiment of the present application further provides a computer-readable storage medium storing machine executable instructions, which, when invoked and executed by a processor, cause the processor to perform the method of the first or second aspect.
The embodiment of the application brings the following beneficial effects:
the embodiment of the application provides a computing resource allocation method, a computing resource allocation device and a training method and a training device of a model thereof, which can receive computing resource supply sent by a plurality of edge cloud servers and computing resource demand sent by a plurality of clients, then, perform ascending arrangement on the computing resource supply to obtain a first sequence, perform descending arrangement on the computing resource demand to obtain a second sequence, then determine an intersection point between the first sequence and the second sequence, wherein the computing resource supply is the same as the computing resource demand, preferably, apply a pre-trained computing resource allocation model, determine a computing resource allocation result between the edge cloud servers and the clients according to the intersection point, model a resource allocation mode in an edge cloud environment by using a bidirectional resource allocation method between the edge cloud servers and the clients, construct a bidirectional resource allocation mechanism, so that the computing resource allocation result can better accord with the situation of the clients and the edge cloud servers, the reasonability of the calculation resource distribution result is improved, and the technical problem of unreasonable calculation resource distribution result in the prior art is solved.
In order to make the aforementioned objects, features and advantages of the present application more comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the detailed description of the present application or the technical solutions in the prior art, the drawings needed to be used in the detailed description of the present application or the prior art description will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic flowchart of a computing resource allocation method according to an embodiment of the present disclosure;
FIG. 2 is a flowchart of a method for training a computational resource allocation model according to an embodiment of the present disclosure;
FIG. 3 is a schematic structural diagram of a computing resource allocation apparatus according to an embodiment of the present application;
FIG. 4 is a schematic structural diagram of a training apparatus for a computational resource allocation model according to an embodiment of the present disclosure;
fig. 5 is a schematic structural diagram illustrating an electronic device provided in an embodiment of the present application.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the present application will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The terms "comprising" and "having," and any variations thereof, as referred to in the embodiments of the present application, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements but may alternatively include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
In the description of the present application, the meaning of "at least one" means one or more than one unless otherwise stated.
Features and exemplary embodiments of various aspects of the present application will be described in detail below. In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the present application by illustrating examples thereof. The present application is in no way limited to any specific configuration and algorithm set forth below, but rather covers any modification, replacement or improvement of elements, components or algorithms without departing from the spirit of the present application. In the drawings and the following description, well-known structures and techniques are not shown in order to avoid unnecessarily obscuring the present application.
At present, a service area of one edge cloud server has many users to obtain services, and for one user, the service area may exist within the service range of a plurality of edge cloud servers.
The traditional resource allocation method is a mathematical analysis method, and the resource allocation scheme can be solved by constructing a utility convex function and a utility concave function and utilizing the analysis method in an iterative manner. However, the traditional mathematical analysis method has a defect in the edge cloud computing scene, and the utility function constructed in the scene is not necessarily a simple concave-convex function, so that the computation is complex. In addition, the duration of the resource allocation result of each round of the resource allocation environment is short, the resource allocation scheme needs to be changed continuously to deal with the number of participants in the resource allocation environment and the dynamic property of the user's own requirements, and the traditional analytic method cannot give the allocation result continuously. Moreover, the current computing resource allocation is to allocate computing resources from the edge cloud server to the user side in a unidirectional manner, which is likely to cause unreasonable allocation results.
Based on this, the method and the device for allocating computing resources and the method and the device for training the model thereof provided by the embodiments of the present application can solve the technical problem that the result of allocating computing resources is unreasonable in the prior art.
For the convenience of understanding of the present embodiment, a method and an apparatus for allocating computing resources and a method and an apparatus for training a model thereof disclosed in the embodiments of the present application will be described in detail first.
Fig. 1 is a flowchart illustrating a computing resource allocation method according to an embodiment of the present application. As shown in fig. 1, the method includes:
and S110, receiving the computing resource supply quantity sent by the edge cloud servers and the computing resource demand quantity sent by the user side.
For example, a bidirectional resource allocation server may be constructed in an edge cloud service area, where the service range covers a plurality of Edge Cloud Servers (MECs) in the area and a plurality of clients that need to use the edge cloud services, and the allocated resources are limited computing resources possessed by each edge cloud server. In the resource allocation process, an intermediate server Broker of a third party can be introduced to complete resource allocation, the computing resource supply amount of each edge cloud server and the computing resource demand amount of a user side are sent to the Broker, and the Broker finally determines which edge cloud server and which user side achieve resource allocation. Since the total amount of computing resources is much less than the demand of users, there is a need to enable the allocation of computing resources to the most demanding users.
S120, performing ascending arrangement on the plurality of computing resource supply quantities to obtain a first sequence; and performing descending order arrangement on the plurality of computing resource demand quantities to obtain a second sequence.
Illustratively, the Broker sorts the received computing resource demand sequence from the client in descending order and the received computing resource supply sequence from the edge cloud server in ascending order.
S130, determining the intersection points with the same calculation resource quantity between the first sequence and the second sequence.
The Broker finds the intersection point of the first sequence and the second sequence, namely the point where the demand of the computing resources is the same as the supply of the computing resources.
And S140, determining a computing resource distribution result between the edge cloud server and the user side according to the intersection point by applying the pre-trained computing resource distribution model.
The computing resource allocation method provided by the embodiment of the application can be used as a distributed edge computing resource allocation method, a bidirectional resource allocation method between an edge cloud server and a user side is used, a resource allocation mode in an edge cloud environment is modeled, and a bidirectional resource allocation mechanism is constructed.
The above steps are described in detail below.
In some embodiments, the pre-trained computing resource allocation model is an EWA model.
In practical applications, an Artificial Intelligence (AI) method can be used to solve the problem of computing resource allocation, and the problem of computing resource allocation can be solved based on the characteristics of an enhanced learning method that integrates an empirical weight charm value algorithm (EWA) for enhanced learning. The resource allocation strategies of each time can be learned through model training, and the strategies of each user side and each edge cloud server and the change of the edge cloud service environment can be observed. In addition, the method has no limitation on the construction of utility functions by the traditional method, so that the universality and the generalization of the method provided by the application are enhanced.
Fig. 2 is a schematic flowchart of a training method of a computational resource allocation model according to an embodiment of the present disclosure. As shown in fig. 2, the method includes:
s210, receiving the computing resource supply quantity sent by the edge cloud servers and the computing resource demand quantity sent by the user sides.
For example, a proxy server capable of running the EWA mechanism may be allocated to each resource allocation participant, i.e., the edge cloud server and the user side, and the proxy server may be capable of sending the computing resource supply amount of the edge cloud server and the computing resource demand amount of the user to the Broker server.
S220, performing ascending arrangement on the plurality of computing resource supply quantities to obtain a first sequence; and performing descending order arrangement on the plurality of computing resource demand quantities to obtain a second sequence.
For example, the Broker server Broker may sort the sequence of the computing resource demands received from the user side in a descending order and sort the sequence of the computing resource supplies received from the edge cloud server in an ascending order.
And S230, if the maximum value of the first sequence is greater than or equal to the minimum value of the second sequence, determining a calculation resource distribution result between the edge cloud server and the user side according to the intersection point with the same calculation resource amount between the first sequence and the second sequence until the maximum value of the first sequence is smaller than the minimum value of the second sequence.
The middle server Broker may find an intersection point of the two sequences, and if the remaining computing resource amount of the corresponding edge cloud server is sufficient for the user side, that is, the maximum value of the first sequence is greater than or equal to the minimum value of the second sequence, the resource allocation process of the corresponding edge cloud server and the user side is completed, and this step is repeated until the maximum value of the ascending sequence is less than the minimum value of the descending sequence.
S240, training the initial EWA model according to the calculation resource distribution result to obtain a pre-trained calculation resource distribution model.
It should be noted that reinforcement learning, which is a unique learning method in the artificial intelligence method, can form a strategy through interaction with the environment.
In the embodiment of the application, the problem of bidirectional resource allocation is solved by using an artificial intelligence method of the EWA, participants in each bidirectional resource allocation environment can operate based on reinforcement learning, and finally can mutually interact to complete transactions and learn a resource allocation strategy suitable for the participants, so that the edge cloud server can meet the service requirement of a user side as much as possible.
The above steps are described in detail below.
In some embodiments, the method may further comprise the steps of:
step a, judging whether the calculation resource distribution result of the current resource distribution cycle is the same as the calculation resource distribution result of the previous resource distribution cycle; if yes, the step b is carried out.
And b, determining the calculation resource distribution result of the next resource distribution cycle as the calculation resource distribution result of the current resource distribution cycle.
Since the respective conditions of the participants in the resource allocation environment are different, an allocation scheme exists only for a short period, and the allocation scheme needs to be re-modified to adapt to the respective conditions of the participants in the resource allocation environment every period. For example, the profit (value) of the computing resource that the edge cloud computing service can bring to a certain user terminal changes more with time, and for the edge cloud server, the profit compensation cost (cost) needs to be obtained by selling the computing resource, and the cost also changes with time. In addition, the number of edge cloud servers and clients is also dynamically changing.
The method provided by the embodiment of the application can quickly determine the resource allocation mode of the next period in a short period, and meanwhile, the method can adapt to the time dynamic change of the resource allocation environment. In practical application, if the computing resource demand and the computing resource supply in the current resource allocation cycle are the same as those in the previous resource allocation cycle, it is proved that the best strategy is found by each client and the edge cloud server, the global benefit of the resource allocation environment is the highest under the strategy sets, and the allocation scheme is issued to each edge cloud server for execution so as to adapt to the dynamic change of the resource allocation environment along with time.
In some embodiments, after the step a, the method may further include the steps of:
and c, if the calculation resource distribution result of the current resource distribution cycle is different from the calculation resource distribution result of the previous resource distribution cycle, re-receiving the calculation resource supply of the next cycle sent by the edge cloud servers and the calculation resource demand of the next cycle sent by the user sides in the next resource distribution cycle.
In practical application, if the calculation resource demand and the calculation resource supply in the current resource allocation cycle are different from the calculation resource allocation result in the previous resource allocation cycle, the initial step S210 is returned to in the next resource allocation cycle, and the above steps are repeated until the policies of the calculation resource demand and the calculation resource supply converge, that is, the calculation resource allocation result in the current resource allocation cycle is the same as the calculation resource allocation result in the previous resource allocation cycle.
In the embodiment of the application, under the condition that the calculation resource allocation result of the current resource allocation period is different from the calculation resource allocation result of the previous resource allocation period, the next resource allocation period can be ensured to receive the calculation resource supply amount and the calculation resource demand amount again, so that the adjustment of the resource amount and the learning of the allocation method are continuously carried out in the next resource allocation period, and the learning process of artificial intelligence is strengthened.
In some embodiments, the next cycle computing resource supply and the next cycle computing resource demand are policy adjusted computing resource amounts.
The client calculates utility (net return) based on computing resource demand and return, and the edge cloud server calculates utility (net return) based on cost and payment from the client. Based on this return, the proxy server based on the EWA algorithm updates the computation resource demand amount or the computation resource supply amount policy. The artificial intelligence method based on reinforcement learning can dynamically adapt to the change of external conditions so as to adjust own strategy and ensure the maximum benefit.
In the embodiment of the application, a reinforcement learning method is used for simulating human interaction so as to solve the problem of bidirectional resource allocation. Each participant becomes an agent and operates based on an EWA reinforcement learning algorithm that assists the participant in making decisions in each round to give the price of the transaction. After each round, each participant can adjust the calculation resource demand strategy according to the profit condition of the round. And then new resource allocation is carried out until the final computing resource demand and asking price strategy converge. And when the balance point is reached, the profit of each participant is proved to reach the maximum value, so that the global optimization is reached. After the calculation of the resource demand mode converges, the allocation of the resources further converges, and this allocation scheme is used as the allocation scheme in the next short time slice, and after the next short time slice starts, the allocation scheme in the next time slice is calculated based on the new resource allocation environment condition.
Therefore, when the profit that can be brought by using the edge cloud computing resources increases, the computing resource demand amount of the user can be increased to obtain the resources. When the amount of the resource required by the system is increased, the amount of the resource required tends to be calculated, so that the resource allocation result is balanced as much as possible.
Fig. 3 provides a schematic structural diagram of a computing resource allocation apparatus. As shown in fig. 3, the computing resource allocation apparatus 700 includes:
a receiving module 701, configured to receive a computing resource supply amount sent by a plurality of edge cloud servers and a computing resource demand amount sent by a plurality of clients;
an arranging module 702, configured to arrange the plurality of computing resource provision amounts in an ascending order to obtain a first sequence; performing descending order arrangement on the plurality of computing resource demand quantities to obtain a second sequence;
a first determining module 703, configured to determine an intersection point where the amount of computing resources is the same between the first sequence and the second sequence;
a second determining module 704, configured to apply the pre-trained computing resource allocation model, and determine a computing resource allocation result between the edge cloud server and the user side according to the intersection point.
In some embodiments, the pre-trained computing resource allocation model is an EWA model.
The computing resource allocation device provided in the embodiment of the present application has the same technical features as the computing resource allocation method provided in the above embodiment, so that the same technical problems can be solved, and the same technical effects can be achieved.
FIG. 4 provides a schematic diagram of a training apparatus for a computational resource allocation model. As shown in fig. 4, the training apparatus 800 for computing resource allocation model includes:
a receiving unit 801, configured to receive computing resource supply amounts sent by multiple edge cloud servers and computing resource demand amounts sent by multiple user sides;
an arranging unit 802, configured to perform ascending arrangement on the multiple computing resource supply amounts to obtain a first sequence, and perform descending arrangement on the multiple computing resource demand amounts to obtain a second sequence;
a first determining unit 803, configured to determine, if the maximum value of the first sequence is greater than or equal to the minimum value of the second sequence, a result of computing resource allocation between the edge cloud server and the user side according to an intersection point where the amount of computing resources between the first sequence and the second sequence is the same until the maximum value of the first sequence is smaller than the minimum value of the second sequence;
and the training unit 804 is configured to train the initial EWA model according to the calculation resource allocation result to obtain a pre-trained calculation resource allocation model.
In some embodiments, the apparatus further comprises:
the judging unit is used for judging whether the calculation resource distribution result of the current resource distribution cycle is the same as the calculation resource distribution result of the previous resource distribution cycle;
and if so, determining that the calculation resource allocation result of the next resource allocation cycle is the calculation resource allocation result of the current resource allocation cycle.
In some embodiments, the receiving unit 801 is further configured to re-receive, in the next resource allocation cycle, the next cycle computing resource supply amount sent by the plurality of edge cloud servers and the next cycle computing resource demand amount sent by the plurality of user terminals.
In some embodiments, the next cycle computing resource supply and the next cycle computing resource demand are policy adjusted computing resource amounts.
The training device of the computational resource allocation model provided by the embodiment of the application has the same technical characteristics as the training method of the computational resource allocation model provided by the embodiment, so that the same technical problems can be solved, and the same technical effects can be achieved.
As shown in fig. 5, an electronic device 900 includes a memory 901 and a processor 902, where the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the steps of the method provided in the foregoing embodiment.
Referring to fig. 5, the electronic device 900 further includes: a bus 903 and a communication interface 904, the processor 902, the communication interface 904, and the memory 901 are connected by the bus 903; the processor 902 is used to execute executable modules, such as computer programs, stored in the memory 901.
The Memory 901 may include a Random Access Memory (RAM), and may further include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 904 (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network, and the like can be used.
Bus 903 may be an ISA bus, PCI bus, EISA bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one double-headed arrow is shown in FIG. 5, but this does not indicate only one bus or one type of bus.
The memory 901 is used for storing a program, and the processor 902 executes the program after receiving an execution instruction, and the method performed by the apparatus defined by the process disclosed in any of the foregoing embodiments of the present application may be applied to the processor 902, or implemented by the processor 902.
The processor 902 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 902. The Processor 902 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in the memory 901, and the processor 902 reads the information in the memory 901, and completes the steps of the above method in combination with the hardware thereof.
An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to perform the steps of the above method.
The computing resource allocation apparatus and the training apparatus of the computing resource allocation model provided in the embodiments of the present application may be specific hardware on a device, or software or firmware installed on a device, or the like. The device provided by the embodiment of the present application has the same implementation principle and technical effect as the foregoing method embodiments, and for the sake of brief description, reference may be made to the corresponding contents in the foregoing method embodiments where no part of the device embodiments is mentioned. It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the foregoing systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the mobile control method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus once an item is defined in one figure, it need not be further defined and explained in subsequent figures, and moreover, the terms "first", "second", "third", etc. are used merely to distinguish one description from another and are not to be construed as indicating or implying relative importance.
Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present application, and are used for illustrating the technical solutions of the present application, but not limiting the same, and the scope of the present application is not limited thereto, and although the present application is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope disclosed in the present application; such modifications, changes or substitutions do not depart from the scope of the embodiments of the present application. Are intended to be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. A method of computing resource allocation, the method comprising:
receiving computing resource supply quantity sent by a plurality of edge cloud servers and computing resource demand quantity sent by a plurality of user sides;
arranging a plurality of computing resource supply quantities in an ascending order to obtain a first sequence; performing descending arrangement on the computing resource demand quantities to obtain a second sequence;
determining an intersection point with the same amount of computing resources between the first sequence and the second sequence;
and determining a computing resource distribution result between the edge cloud server and the user side according to the intersection point by applying a pre-trained computing resource distribution model.
2. The method of claim 1, wherein the pre-trained computing resource allocation model is an EWA model.
3. A method of training a computational resource allocation model, the method comprising:
receiving computing resource supply quantity sent by a plurality of edge cloud servers and computing resource demand quantity sent by a plurality of user sides;
arranging a plurality of computing resource supply quantities in an ascending order to obtain a first sequence; performing descending arrangement on the computing resource demand quantities to obtain a second sequence;
if the maximum value of the first sequence is larger than or equal to the minimum value of the second sequence, determining a calculation resource distribution result between the edge cloud server and the user side according to an intersection point with the same calculation resource amount between the first sequence and the second sequence until the maximum value of the first sequence is smaller than the minimum value of the second sequence;
and training the initial EWA model according to the calculation resource distribution result to obtain a pre-trained calculation resource distribution model.
4. The training method of claim 3, further comprising:
judging whether the calculation resource allocation result of the current resource allocation period is the same as the calculation resource allocation result of the previous resource allocation period;
if so, determining that the calculation resource distribution result of the next resource distribution cycle is the calculation resource distribution result of the current resource distribution cycle.
5. The training method according to claim 4, wherein after the step of determining whether the calculation resource allocation result of the current resource allocation period is the same as the calculation resource allocation result of the previous resource allocation period, the method further comprises:
and if not, re-receiving the next-period computing resource supply quantity sent by the edge cloud servers and the next-period computing resource demand quantity sent by the user sides in the next resource allocation period.
6. The training method of claim 5, wherein the next cycle computing resource supply and the next cycle computing resource demand are policy adjusted computing resource quantities.
7. An apparatus for allocating computing resources, the apparatus comprising:
the receiving module is used for receiving the computing resource supply quantity sent by the edge cloud servers and the computing resource demand quantity sent by the user sides;
the arrangement module is used for carrying out ascending arrangement on the plurality of computing resource supply quantities to obtain a first sequence; performing descending arrangement on the computing resource demand quantities to obtain a second sequence;
a first determining module, configured to determine an intersection point where the amount of computing resources is the same between the first sequence and the second sequence;
and the second determining module is used for applying a pre-trained computing resource distribution model and determining a computing resource distribution result between the edge cloud server and the user side according to the intersection point.
8. An apparatus for training a computational resource allocation model, the apparatus comprising:
the receiving unit is used for receiving the computing resource supply quantity sent by the edge cloud servers and the computing resource demand quantity sent by the user sides;
the arrangement unit is used for carrying out ascending arrangement on the plurality of computing resource supply quantities to obtain a first sequence and carrying out descending arrangement on the plurality of computing resource demand quantities to obtain a second sequence;
a determining unit, configured to determine, if the maximum value of the first sequence is greater than or equal to the minimum value of the second sequence, a result of computing resource allocation between the edge cloud server and the user side according to an intersection point where the amount of computing resources between the first sequence and the second sequence is the same until the maximum value of the first sequence is smaller than the minimum value of the second sequence;
and the training unit is used for training the initial EWA model according to the calculation resource distribution result to obtain a pre-trained calculation resource distribution model.
9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program operable on the processor, and wherein the processor implements the steps of the method of any of claims 1 to 6 when executing the computer program.
10. A computer readable storage medium having stored thereon machine executable instructions which, when invoked and executed by a processor, cause the processor to execute the method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911179666.7A CN110908804A (en) | 2019-11-26 | 2019-11-26 | Computing resource allocation method and device and training method and device of model thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911179666.7A CN110908804A (en) | 2019-11-26 | 2019-11-26 | Computing resource allocation method and device and training method and device of model thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110908804A true CN110908804A (en) | 2020-03-24 |
Family
ID=69819975
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911179666.7A Pending CN110908804A (en) | 2019-11-26 | 2019-11-26 | Computing resource allocation method and device and training method and device of model thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110908804A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112700116A (en) * | 2020-12-29 | 2021-04-23 | 北京嘀嘀无限科技发展有限公司 | Method for determining service achievement amount, resource allocation method and device |
| CN113300982A (en) * | 2020-06-08 | 2021-08-24 | 阿里巴巴集团控股有限公司 | Resource allocation method, device, system and storage medium |
| WO2021212347A1 (en) * | 2020-04-21 | 2021-10-28 | Oppo广东移动通信有限公司 | Communication method and related device |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110266770A (en) * | 2019-05-30 | 2019-09-20 | 湖南大学 | Idle cloud resource dispatching method and device based on game theory |
-
2019
- 2019-11-26 CN CN201911179666.7A patent/CN110908804A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110266770A (en) * | 2019-05-30 | 2019-09-20 | 湖南大学 | Idle cloud resource dispatching method and device based on game theory |
Non-Patent Citations (1)
| Title |
|---|
| 姚海鹏 等: "Reinforcement-Learning- and Belief-Learning-Based Double Auction Mechanism for Edge Computing Resource Allocation", 《IEEE INTERNET OF THINGS JOURNAL》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021212347A1 (en) * | 2020-04-21 | 2021-10-28 | Oppo广东移动通信有限公司 | Communication method and related device |
| CN113300982A (en) * | 2020-06-08 | 2021-08-24 | 阿里巴巴集团控股有限公司 | Resource allocation method, device, system and storage medium |
| CN113300982B (en) * | 2020-06-08 | 2022-08-23 | 阿里巴巴集团控股有限公司 | Resource allocation method, device, system and storage medium |
| CN112700116A (en) * | 2020-12-29 | 2021-04-23 | 北京嘀嘀无限科技发展有限公司 | Method for determining service achievement amount, resource allocation method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110908804A (en) | Computing resource allocation method and device and training method and device of model thereof | |
| CN108287666B (en) | Data storage method and device for cloud storage environment | |
| CN109714797B (en) | Mobile edge network resource allocation method based on auction theory | |
| Ghosh et al. | Pricing for profit in internet of things | |
| Shojaiemehr et al. | A three-phase process for SLA negotiation of composite cloud services | |
| CN111340244A (en) | Prediction method, training method, device, server and medium | |
| CN114816738A (en) | Method, device and equipment for determining calculation force node and computer readable storage medium | |
| CN110928956B (en) | Method, device and equipment for descending cloud service in edge-cloud computing environment | |
| CN108667877B (en) | Method and device for determining recommendation information, computer equipment and storage medium | |
| CN110427160A (en) | Put in order a group equalization methods and device for distribution | |
| CN111491382B (en) | Satellite capacity allocation method and device and electronic equipment | |
| CN115964182B (en) | Resource scheduling method and system | |
| Zhou et al. | Maximizing the profits of cloud service providers via dynamic virtual resource renting approach | |
| CN109379413B (en) | Method and device for acquiring user contribution value and storage medium | |
| CN116777518A (en) | Transaction management method, device, storage medium and equipment | |
| CN108093083B (en) | Cloud manufacturing task scheduling method and device and terminal | |
| CN116775304A (en) | Task allocation method, system, device and nonvolatile storage medium | |
| CN115640454A (en) | Product recommendation method, system, electronic device and storage medium | |
| CN115794421A (en) | Resource allocation method and device and electronic equipment | |
| CN115421930A (en) | Task processing method, system, device, equipment and computer readable storage medium | |
| KR102590112B1 (en) | Coded and Incentive-based Mechanism for Distributed Training of Machine Learning in IoT | |
| Liu et al. | An online combinatorial auction based resource allocation and pricing mechanism for network slicing in 5G | |
| CN108958931B (en) | Calculation migration method and device | |
| CN110111135B (en) | Power generation side member dynamic quotation decision method, device and equipment | |
| CN112836971A (en) | Quota resource determination method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200324 |