CN111198761A - Resource scheduling and allocating device, method and computer readable storage medium - Google Patents
Resource scheduling and allocating device, method and computer readable storage medium Download PDFInfo
- Publication number
- CN111198761A CN111198761A CN201911090547.4A CN201911090547A CN111198761A CN 111198761 A CN111198761 A CN 111198761A CN 201911090547 A CN201911090547 A CN 201911090547A CN 111198761 A CN111198761 A CN 111198761A
- Authority
- CN
- China
- Prior art keywords
- resource
- target application
- application program
- scene
- dql
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Stored Programmes (AREA)
Abstract
The application relates to the technical field of power consumption parameter setting, and provides a resource scheduling and allocating device, a resource scheduling and allocating method and a computer readable storage medium, wherein the resource scheduling and allocating method comprises the following steps: the method comprises the steps of obtaining the current resource use state of a target application program, judging the estimated resource use configuration of the target application program in a target time period according to the current resource use state by adopting a preset deep reinforcement learning DQL model, and carrying out resource scheduling and distribution actions on the target application program according to the estimated resource use configuration when the target time period is entered. By means of the method, the resource use condition of the application program can be modeled and analyzed by the deep reinforcement learning model, and pre-estimation is carried out according to the analysis result, so that the resource is distributed in advance, excessive resource occupation between the application programs is avoided, the resource distribution problem between the application programs and the system is coordinated, and optimal allocation of the resource is realized.
Description
Technical Field
The present application relates to the field of performance parameter setting technologies, and in particular, to a resource scheduling and allocating method, a resource scheduling and allocating apparatus using the resource scheduling and allocating method, and a computer-readable storage medium.
Background
With the use of resource scheduling and allocating devices such as smart phones and tablet computers becoming more and more frequent, the problem of power consumption becomes more and more prominent. The power consumption of the terminal is reduced, the service life of the battery is prolonged, and the method has very important significance for improving user experience.
At present, the number of application programs (apps) on a resource scheduling and allocating device is huge, the quality is uneven, and most of the power consumption of the resource scheduling and allocating device is consumed on the apps under normal conditions. The power consumption of the App is divided into a foreground part and a background part: the foreground is responsible for direct interaction with users, and the background runs various services and the like. Partial App background activities are frequent, and in many cases, the activities are not necessary at all, but a large amount of power consumption is wasted, and the battery service time is shortened. Meanwhile, many apps are updated frequently, and the power consumption of part of versions is obviously increased when the front desk is used.
However, most of the power consumption parameters of the current mobile phone are in the mobile phone, the power consumption parameters of a certain scene are fixed, or the power consumption parameters aiming at the certain scene are lacked, so that it is not difficult to see that the parameter condition is far from realizing intellectualization, the accuracy is not high, especially many parameters are only tested in a laboratory, and are directly written into the mobile phone of a user to execute operation, the requirements of the user can obviously not be met, and the optimal operation state is not brought to the user.
It is easy to see that the prior art has many defects, for example, if the resources are allocated in advance, the specific allocation is more difficult to determine; if real-time resource allocation is adopted, timeliness cannot be grasped, and some games may need low latency; furthermore, the resource utilization is not efficient, because other applications also need to occupy resources at some time.
In view of the various defects in the prior art, the inventors of the present application have made extensive studies to provide a resource scheduling and allocating apparatus, a resource scheduling and allocating method, and a computer-readable storage medium.
Disclosure of Invention
The present application aims to provide a resource scheduling and allocating apparatus, a resource scheduling and allocating method, and a computer-readable storage medium, which can use a deep reinforcement learning model to perform modeling analysis on the resource usage of an application program, and perform estimation according to the analysis result, thereby allocating resources in advance, avoiding mutual occupation of too many resources between the application programs, coordinating the resource allocation problem between the application programs and a system, and achieving optimal allocation of resources.
In order to solve the above technical problem, the present application provides a resource scheduling and allocating method, as one embodiment, the resource scheduling and allocating method includes:
s100, acquiring the current resource use state of a target application program;
s200, judging the estimated resource use configuration of the target application program in a target time period according to the current resource use state by adopting a preset deep reinforcement learning DQL model;
s300, when the target time slot is entered, resource scheduling and distributing actions are carried out on the target application program according to the pre-estimated resource use configuration.
As an embodiment, before the step of obtaining the current resource usage state of the target application, the method further includes:
acquiring scene use resource data of the target application program;
and establishing a DQL model corresponding to the target application program according to the scene usage resource data.
As an implementation manner, the step of acquiring the scene usage resource data of the target application specifically includes:
extracting scene resource characteristics of the target application program through a convolutional neural network;
and acquiring scene use resource data according to the scene resource characteristics.
The target application program comprises game software, multimedia playing software, friend-making chat software, webpage surfing software and/or network live broadcast software; the step of extracting the scene resource features of the target application program through the convolutional neural network specifically includes:
decomposing the scene page of the target application program into frame pictures;
and extracting the scene resource characteristics of the target application program from the frame picture through a convolutional neural network.
As an embodiment, the step of establishing a DQL model corresponding to the target application using the resource data according to the scenario further includes:
and performing Q function fitting on the DQL model to train and obtain an optimized DQL model.
As an embodiment, the step of performing Q function fitting on the DQL model specifically includes:
acquiring and extracting scene resource comprehensive characteristics of the target application program, the similar application program and the self resource use condition of the equipment through a convolutional neural network;
and performing Q function fitting training on the DQL model by using the scene resource comprehensive characteristics.
As one embodiment, the step S200 further includes:
acquiring historical use data of the target application program;
and judging the pre-estimated resource use configuration of the target application program in a target time period according to the historical use data and the current resource use state.
As one embodiment, the historical usage data includes:
at least one of a usage period, a usage frequency, and an operating frequency.
In order to solve the above technical problem, the present application further provides a resource scheduling allocation apparatus, which is configured with a processor, as one embodiment, the processor is configured to execute a resource scheduling allocation program to implement the resource scheduling allocation method including the foregoing.
To solve the above technical problem, the present application further provides a computer-readable storage medium as one embodiment, which is used for storing a resource scheduling assignment program, and when the resource scheduling assignment program is executed by a processor, the resource scheduling assignment program implements a resource scheduling assignment method including the above.
The resource scheduling and allocating device, method and computer readable storage medium provided by the present application, the resource scheduling and allocating method includes the steps of: the method comprises the steps of obtaining the current resource use state of a target application program, judging the estimated resource use configuration of the target application program in a target time period according to the current resource use state by adopting a preset deep reinforcement learning DQL model, and carrying out resource scheduling and distribution actions on the target application program according to the estimated resource use configuration when the target time period is entered. By means of the method, the resource use condition of the application program can be modeled and analyzed by the deep reinforcement learning model, and pre-estimation is carried out according to the analysis result, so that the resource is distributed in advance, excessive resource occupation between the application programs is avoided, the resource distribution problem between the application programs and the system is coordinated, and optimal allocation of the resource is realized.
The foregoing description is only an overview of the technical solutions of the present application, and in order to make the technical means of the present application more clearly understood, the present application may be implemented in accordance with the content of the description, and in order to make the above and other objects, features, and advantages of the present application more clearly understood, the following preferred embodiments are described in detail with reference to the accompanying drawings.
Drawings
Fig. 1A is a flowchart illustrating a resource scheduling allocation method according to an embodiment of the present application.
Fig. 1B is a flowchart illustrating an embodiment of step S100 shown in fig. 1A.
Fig. 1C is a schematic flowchart of an embodiment of step S200 shown in fig. 1A.
Fig. 2 is a functional block diagram of an embodiment of a resource scheduling and allocating apparatus according to the present application.
Detailed Description
To further clarify the technical measures and effects taken by the present application to achieve the intended purpose, the present application will be described in detail below with reference to the accompanying drawings and preferred embodiments.
While the present application has been described in terms of specific embodiments and examples for achieving the desired objects and objectives, it is to be understood that the invention is not limited to the disclosed embodiments, but is to be accorded the widest scope consistent with the principles and novel features as defined by the appended claims.
Referring to fig. 1A, fig. 1A is a schematic flowchart illustrating a resource scheduling allocation method according to an embodiment of the present application.
It should be noted that the resource scheduling and allocating method according to the present embodiment may include, but is not limited to, the following steps.
Step S100, acquiring the current resource use state of the target application program.
And step S200, judging the estimated resource use configuration of the target application program in a target time period according to the current resource use state by adopting a preset deep reinforcement learning DQL model.
And step S300, performing resource scheduling and allocation actions on the target application program according to the pre-estimated resource use configuration when the target time period is entered.
It should be noted that, in this embodiment, the target time period may refer to a certain time period that is extended, or may be a certain time node, or a time period when the target application enters a certain running state.
In addition, in this embodiment, the estimated resource usage configuration of the target application program in the target time period may be periodically, irregularly, or in real time, and of course, the steps of the method may be executed by performing intelligent triggering according to the running state of the application program.
In this embodiment, before the step of obtaining the current resource usage state of the target application, the method further includes: acquiring scene use resource data of the target application program; and establishing a DQL model corresponding to the target application program according to the scene usage resource data.
Further, referring to fig. 1B, the step of acquiring the scene usage resource data of the target application program according to the present embodiment may specifically include: s110, extracting scene resource characteristics of the target application program through a convolutional neural network; and S120, acquiring scene use resource data according to the scene resource characteristics.
In addition, as shown in fig. 1C, the step S200 according to this embodiment may further include:
s210, acquiring historical use data of the target application program;
s220, judging the estimated resource use configuration of the target application program in the target time period according to the historical use data and the current resource use state.
It should be understood that, except for the special case of the first installation use, the general application program may be used occasionally and frequently by the user in the device, and for these usage records, the present embodiment may perform statistics and estimate the usage trend, and obtain the result of estimated resource usage configuration with little difference from the actual situation.
For example, the historical usage data in this embodiment may include: at least one of a usage period, a usage frequency, and an operating frequency.
For example, the target application is "oral english practice" with usage periods of 7-8 am and 19-20 pm, the error is basically not more than ± 5 minutes, and the usage is adhered to in this period every day, so as to determine the resource usage configuration required to ensure "oral english practice" in this period, and to appropriately shut down other apparently conflicting resource configurations of the application to ensure normal usage of "oral english practice".
For example, if the target application is "current news" with a frequency of use of three refreshes per hour substantially during daytime hours, wherein each refresh may be viewed for 5 minutes, then one of the historical usage data may be used as a resource usage configuration to determine the need to ensure this frequency of use of "current news" and may be used to appropriately shut down other apparently conflicting resource configurations of the application at the expected usage period (or frequency) to ensure normal use of "current news".
For example, if the target application is "go" and the operation frequency is once every 3 minutes after the application is started, wherein each refresh may be browsing for 5 minutes, then the historical usage data as one of them may be used as the resource usage configuration for judging that this operation frequency of "go" needs to be ensured, and other apparently conflicting resource configurations of the application may be appropriately turned off when the expected operation frequency comes, so as to ensure the normal use of "go".
Further, any combination of three or more of the above historical usage data as understood by those skilled in the art may be used, such as 20-22 pm for the target application "go", every 3 minutes after the application is started, every monday, three, or five three days for the frequency of operation, and then the other apparently conflicting resource configurations of the application may be turned off for the combined historical usage data to ensure the resource usage configuration and normal usage of this frequency of operation of "go".
Specifically, the step of extracting the scene resource feature of the target application program through the convolutional neural network according to the embodiment may specifically include: decomposing the scene page of the target application program into frame pictures; and extracting the scene resource characteristics of the target application program from the frame picture through a convolutional neural network.
It should be added that the frame picture may refer to a picture that is decomposed frame by frame, and a target application, such as a game interface, may be decomposed frame by frame.
It should be noted that, in this embodiment, the step of establishing the DQL model corresponding to the target application according to the scene usage resource data may further include: and performing Q function fitting on the DQL model to train and obtain an optimized DQL model.
Similarly, the step of performing Q function fitting on the DQL model in this embodiment specifically includes: acquiring and extracting scene resource comprehensive characteristics of the target application program, the similar application program and the self resource use condition of the equipment through a convolutional neural network; and performing Q function fitting training on the DQL model by using the scene resource comprehensive characteristics.
In a specific example of this embodiment, in the step of fitting a Q function to the DQL model, the Q function includes:
wherein, in formula 1, s is the current state of the resource allocation amount, a is the resource scheduling allocation action of increasing or decreasing the resource, Π is the current dynamic resource allocation strategy,
is a function of the return, and r is abnormal when the operation at time t is abnormaltIs a negative number, otherwise 0, gamma is a constant,
is a probability function of the state transition to s' when the resource scheduling assignment action a is taken in state s, vπ(s’)＝∑Aπ(a|s’)Qπ(s ', a) represents the sum of the state value functions of the resource scheduling assignment action a in the state s'.
Further, the performing Q function fitting on the DQL model according to the present embodiment includes:
training an objective function
Wherein Q is an action value function obtained by definition,
is a fitting value of the action value function learned by the neural network model, and θ is a model parameter.
In this embodiment, in the step of determining, by using a preset deep reinforcement learning DQL model, a resource scheduling and allocating action of the target application in a target time period according to the current resource usage state, the pre-estimation function of the resource scheduling and allocating action a includes:
a ═ argmax (Q (s, a')) -formula 2
In equation 2, Q is an action value function, S is a current state space, a' is an action space, and a is an optimal action selected by the function.
It is easily understood that the present embodiment can adjust the resource usage of, for example, GPU, CPU, RAM, etc. by predicting the resource usage configuration.
Specifically, the embodiment can set the APP-related running data on the resource scheduling and allocating device, which can be a network state, a bluetooth state, a GPS state, location information, a CPU/DDR/GPU frequency point, foreground power consumption, background power consumption, foreground running resources, background running resources, foreground running time, background running time, and the like.
According to the method and the device, the deep reinforcement learning model can be adopted to carry out modeling analysis on the resource use condition of the application program, and estimation is carried out according to the analysis result, so that the resource is distributed in advance, the problem that too many resources are occupied by the application programs, the resource distribution problem between the application programs and the system is coordinated, and the optimal configuration of the resource is realized.
Referring to fig. 2, the present application further provides a resource scheduling and allocating apparatus, as an embodiment, the apparatus is configured with a processor 21, and the processor 21 is configured to execute a resource scheduling and allocating procedure to implement the resource scheduling and allocating method as described above.
In this embodiment, the resource scheduling and allocating device includes one of a mobile phone, a tablet computer, a wearable device, and a notebook computer.
In a specific working process, the target application program comprises game software, multimedia playing software, friend-making chat software, webpage surfing software and/or network live broadcast software.
Specifically, the processor 21 according to this embodiment is configured to obtain a current resource usage state of the target application.
The processor 21 is configured to determine, by using a preset deep reinforcement learning DQL model, an estimated resource usage configuration of the target application program in a target time period according to the current resource usage state.
The processor 21 is configured to perform a resource scheduling and allocating action on the target application according to the pre-estimated resource usage configuration when entering the target time period.
It should be noted that, in this embodiment, the target time period may refer to a certain time period that is extended, or may be a certain time node, or a time period when the target application enters a certain running state.
In addition, the processor 21 of this embodiment is configured to obtain historical usage data of the target application; the processor 21 is configured to determine an estimated resource usage configuration of the target application program in a target time period according to the historical usage data and the current resource usage state.
It should be understood that, except for the special case of the first installation use, the general application program may be used occasionally and frequently by the user in the device, and for these usage records, the present embodiment may perform statistics and estimate the usage trend, and obtain the result of estimated resource usage configuration with little difference from the actual situation.
For example, the historical usage data in this embodiment may include: at least one of a usage period, a usage frequency, and an operating frequency.
For example, the target application is "oral english practice" with usage periods of 7-8 am and 19-20 pm, the error is basically not more than ± 5 minutes, and the usage is adhered to in this period every day, so as to determine the resource usage configuration required to ensure "oral english practice" in this period, and to appropriately shut down other apparently conflicting resource configurations of the application to ensure normal usage of "oral english practice".
For example, if the target application is "current news" with a frequency of use of three refreshes per hour substantially during daytime hours, wherein each refresh may be viewed for 5 minutes, then one of the historical usage data may be used as a resource usage configuration to determine the need to ensure this frequency of use of "current news" and may be used to appropriately shut down other apparently conflicting resource configurations of the application at the expected usage period (or frequency) to ensure normal use of "current news".
For example, if the target application is "go" and the operation frequency is once every 3 minutes after the application is started, wherein each refresh may be browsing for 5 minutes, then the historical usage data as one of them may be used as the resource usage configuration for judging that this operation frequency of "go" needs to be ensured, and other apparently conflicting resource configurations of the application may be appropriately turned off when the expected operation frequency comes, so as to ensure the normal use of "go".
Further, any combination of three or more of the above historical usage data as understood by those skilled in the art may be used, such as 20-22 pm for the target application "go", every 3 minutes after the application is started, every monday, three, or five three days for the frequency of operation, and then the other apparently conflicting resource configurations of the application may be turned off for the combined historical usage data to ensure the resource usage configuration and normal usage of this frequency of operation of "go".
In this embodiment, the processor 21 is configured to obtain scene usage resource data of the target application; and establishing a DQL model corresponding to the target application program according to the scene usage resource data.
Further, the processor 21 in this embodiment is configured to extract scene resource features of the target application through a convolutional neural network; the processor 21 is configured to obtain scene usage resource data according to the scene resource features.
Specifically, the processor 21 according to this embodiment is configured to extract the scene resource feature of the target application through a convolutional neural network, and specifically may include: the processor 21 is configured to decompose a scene page of the target application into frame pictures; the processor 21 is configured to extract scene resource features of the target application from the frame picture through a convolutional neural network.
It should be added that the frame picture may refer to a picture that is decomposed frame by frame, and a target application, such as a game interface, may be decomposed frame by frame.
It should be noted that, in this embodiment, the processor 21 is configured to establish a DQL model corresponding to the target application according to the scene usage resource data, and may further include: the processor 21 is configured to perform Q function fitting on the DQL model to train to obtain an optimized DQL model.
Similarly, the processor 21 in this embodiment is configured to perform Q function fitting on the DQL model, and specifically may include: the processor 21 is configured to collect and extract scene resource comprehensive characteristics of the target application program, the similar application program, and the device resource usage status through a convolutional neural network; and performing Q function fitting training on the DQL model by using the scene resource comprehensive characteristics.
In a specific embodiment of this embodiment, the processor 21 is configured to perform Q function fitting on the DQL model, where the Q function includes:
wherein, in formula 3, s is the current state of the resource allocation amount, a is the resource scheduling allocation action of increasing or decreasing the resource, Π is the current dynamic resource allocation strategy,
is a function of the return, and r is abnormal when the operation at time t is abnormaltIs a negative number, otherwise 0, gamma is a constant,
is a probability function of the state transition to s' when the resource scheduling assignment action a is taken in state s, vπ(s’)＝∑Aπ(a|s’)Qπ(s ', a) represents the sum of the state value functions of the resource scheduling assignment action a in the state s'.
Further, the performing Q function fitting on the DQL model according to the present embodiment includes:
training an objective function
Wherein Q is an action value function obtained by definition,
is a fitting value of the action value function learned by the neural network model, and θ is a model parameter.
In this embodiment, the processor 21 adopts a preset deep reinforcement learning DQL model to determine, according to the current resource usage state, a resource scheduling and allocating action of the target application program in a target time period, where an estimation function of the resource scheduling and allocating action a includes:
a ═ argmax (Q (s, a')) -formula 4
In equation 4, Q is an action value function, S is a current state space, a' is an action space, and a is an optimal action selected by the function.
It is easily understood that the present embodiment can adjust the resource usage of, for example, GPU, CPU, RAM, etc. by predicting the resource usage configuration.
Specifically, the embodiment can set the APP-related running data on the resource scheduling and allocating device, which can be a network state, a bluetooth state, a GPS state, location information, a CPU/DDR/GPU frequency point, foreground power consumption, background power consumption, foreground running resources, background running resources, foreground running time, background running time, and the like.
According to the method and the device, the deep reinforcement learning model can be adopted to carry out modeling analysis on the resource use condition of the application program, and estimation is carried out according to the analysis result, so that the resource is distributed in advance, the problem that too many resources are occupied by the application programs, the resource distribution problem between the application programs and the system is coordinated, and the optimal configuration of the resource is realized.
Furthermore, the present application may also provide a computer-readable storage medium for storing a resource scheduling assignment program, which when executed by a processor, implements a resource scheduling assignment method including the resource scheduling assignment method as described in fig. 1A and the embodiments thereof.
It should be noted that the method embodiments described herein may be machine or computer implemented, at least in part. Some examples may include a computer-readable storage medium or machine-readable storage medium encoded with instructions that may be used to configure an electronic device to perform the methods described in the above embodiments. Implementations of these methods may include code, such as microcode, assembly language code or higher level language code, and the like. Such code may include computer readable instructions for performing various methods. The code may form part of a resource scheduling assignment program product. The code can be tangibly stored on one or more volatile, non-transitory, or non-volatile tangible computer-readable media, e.g., during execution or at other times. Examples of such tangible computer-readable storage media may include, but are not limited to, hard disks, removable magnetic disks, removable optical disks (e.g., compact disks and digital video disks), magnetic cassettes, memory cards or sticks, Random Access Memories (RAMs), Read Only Memories (ROMs), and the like.
Although the present application has been described with reference to a preferred embodiment, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the application, and all changes, substitutions and alterations that fall within the spirit and scope of the application are to be understood as being included within the following description of the preferred embodiment.
Claims (10)
1. A resource scheduling and allocating method is characterized by comprising the following steps:
s100, acquiring the current resource use state of a target application program;
s200, judging the estimated resource use configuration of the target application program in a target time period according to the current resource use state by adopting a preset deep reinforcement learning DQL model;
s300, when the target time slot is entered, resource scheduling and distributing actions are carried out on the target application program according to the pre-estimated resource use configuration.
2. The method of claim 1, wherein the step of obtaining the current resource usage state of the target application is preceded by the step of:
acquiring scene use resource data of the target application program;
and establishing a DQL model corresponding to the target application program according to the scene usage resource data.
3. The method according to claim 2, wherein the step of S100 specifically includes:
s110, extracting scene resource characteristics of the target application program through a convolutional neural network;
and S120, acquiring scene use resource data according to the scene resource characteristics.
4. The method of claim 3, wherein the target application comprises game software, multimedia playing software, friend-making chat software, web surfing software and/or web-casting software;
the step of extracting the scene resource features of the target application program through the convolutional neural network specifically includes:
decomposing the scene page of the target application program into frame pictures;
and extracting the scene resource characteristics of the target application program from the frame picture through a convolutional neural network.
5. The method according to claim 2, wherein the step of using the resource data to build the DQL model corresponding to the target application according to the scenario further comprises:
and performing Q function fitting on the DQL model to train and obtain an optimized DQL model.
6. The method according to claim 5, wherein the step of performing Q-function fitting on the DQL model specifically comprises:
acquiring and extracting scene resource comprehensive characteristics of the target application program, the similar application program and the self resource use condition of the equipment through a convolutional neural network;
and performing Q function fitting training on the DQL model by using the scene resource comprehensive characteristics.
7. The method according to any one of claims 1 to 6, wherein the step S200 further comprises:
acquiring historical use data of the target application program;
and judging the pre-estimated resource use configuration of the target application program in a target time period according to the historical use data and the current resource use state.
8. The method of claim 7, wherein the historical usage data comprises:
at least one of a usage period, a usage frequency, and an operating frequency.
9. A resource scheduling assignment arrangement, characterized in that it is configured with a processor for executing a resource scheduling assignment procedure to implement a method comprising the resource scheduling assignment according to any of claims 1-8.
10. A computer-readable storage medium for storing a resource scheduling assignment program which, when executed by a processor, implements a method comprising the resource scheduling assignment of any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911090547.4A CN111198761A (en) | 2019-11-08 | 2019-11-08 | Resource scheduling and allocating device, method and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911090547.4A CN111198761A (en) | 2019-11-08 | 2019-11-08 | Resource scheduling and allocating device, method and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111198761A true CN111198761A (en) | 2020-05-26 |
Family
ID=70746286
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911090547.4A Pending CN111198761A (en) | 2019-11-08 | 2019-11-08 | Resource scheduling and allocating device, method and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111198761A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112162861A (en) * | 2020-09-29 | 2021-01-01 | 广州虎牙科技有限公司 | Thread allocation method and device, computer equipment and storage medium |
| CN112380012A (en) * | 2020-11-16 | 2021-02-19 | Oppo（重庆）智能科技有限公司 | Resource adjusting method and device, terminal equipment and storage medium |
| CN112600906A (en) * | 2020-12-09 | 2021-04-02 | 中国科学院深圳先进技术研究院 | Resource allocation method and device for online scene and electronic equipment |
| CN113225830A (en) * | 2021-06-07 | 2021-08-06 | 维沃移动通信有限公司 | Data network uplink scheduling method and device and electronic equipment |
| CN116777730A (en) * | 2023-08-25 | 2023-09-19 | 湖南马栏山视频先进技术研究院有限公司 | GPU efficiency improvement method based on resource scheduling |
-
2019
- 2019-11-08 CN CN201911090547.4A patent/CN111198761A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112162861A (en) * | 2020-09-29 | 2021-01-01 | 广州虎牙科技有限公司 | Thread allocation method and device, computer equipment and storage medium |
| CN112162861B (en) * | 2020-09-29 | 2024-04-19 | 广州虎牙科技有限公司 | Thread allocation method, thread allocation device, computer equipment and storage medium |
| CN112380012A (en) * | 2020-11-16 | 2021-02-19 | Oppo（重庆）智能科技有限公司 | Resource adjusting method and device, terminal equipment and storage medium |
| CN112600906A (en) * | 2020-12-09 | 2021-04-02 | 中国科学院深圳先进技术研究院 | Resource allocation method and device for online scene and electronic equipment |
| CN113225830A (en) * | 2021-06-07 | 2021-08-06 | 维沃移动通信有限公司 | Data network uplink scheduling method and device and electronic equipment |
| CN116777730A (en) * | 2023-08-25 | 2023-09-19 | 湖南马栏山视频先进技术研究院有限公司 | GPU efficiency improvement method based on resource scheduling |
| CN116777730B (en) * | 2023-08-25 | 2023-10-31 | 湖南马栏山视频先进技术研究院有限公司 | GPU efficiency improvement method based on resource scheduling |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111198761A (en) | Resource scheduling and allocating device, method and computer readable storage medium | |
| CN108874539B (en) | Resource allocation method, device, terminal and storage medium | |
| CN107547746B (en) | Resource allocation method and related product | |
| US20170192819A1 (en) | Method and electronic device for resource allocation | |
| WO2019042169A1 (en) | Resource allocation method and related products | |
| CN107547745B (en) | Resource allocation method and related product | |
| EP3748470A1 (en) | Resource scheduling method and terminal device | |
| US20120064893A1 (en) | Prediction of mobile bandwidth and usage requirements | |
| US10031572B2 (en) | Method and device for processing core of processor, and terminal | |
| WO2021093365A1 (en) | Gpu video memory management control method and related device | |
| CN105677378A (en) | Application management method and mobile terminal | |
| US20130035107A1 (en) | System and method for adaptive traffic prioritization and bandwidth allocation on mobile data networks | |
| KR102219015B1 (en) | Use of network support protocols to improve network usage | |
| US20170177255A1 (en) | Method and electronic device for allocating dynamic memory resources | |
| CN111414070A (en) | Case power consumption management method and system, electronic device and storage medium | |
| WO2023077813A1 (en) | Method and apparatus for determining fake traffic in live broadcast room | |
| CN111045925A (en) | Application program performance parameter optimization method and device and computer readable storage medium | |
| CN112997543B (en) | Application grouping adjustment method and device, storage medium and electronic equipment | |
| EP3352048A1 (en) | Mobile terminal and working method thereof | |
| CN105607901B (en) | Method and device for optimizing system memory and terminal equipment | |
| CN113439435A (en) | Video playing method and device, storage medium and electronic equipment | |
| CN113204425B (en) | Method, device, electronic equipment and storage medium for process management internal thread | |
| CN113032142B (en) | Resource adjusting method and device, electronic equipment and computer readable storage medium | |
| CN114035940A (en) | Resource allocation method and device | |
| CN113918002A (en) | Frequency modulation method, frequency modulation device, storage medium and electronic equipment |
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 |