Method, device, storage medium and electronic device for adjusting computing resources

Abstract

The present disclosure relates to a method, device, storage medium and electronic device for adjusting computing power resources. The method comprises: obtaining target configuration information corresponding to multiple delivery channels in a specified time period, the target configuration information comprising a target configuration corresponding to each delivery channel, the target configuration comprising a target queue length of an elastic queue, a target model size of an elastic model and a target link complexity of an elastic link, the target queue length of an elastic queue is used to indicate the amount of promotional content to be input into the elastic link, the elastic model is set in the elastic link, the elastic link is used to determine a revenue indicator according to the promotional content, obtain the actual computing power resource ratio when the promotional content is delivered according to the target configuration information, adjust the target configuration information according to the actual computing power resource ratio and the preset computing power resource ratio, and deliver the promotional content according to the adjusted target configuration information to optimize the actual computing power resource ratio.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular, to a method, device, storage medium and electronic device for adjusting computing resources. Background Art

[0002] With the continuous development of advertising delivery services, the overall traffic of advertising delivery services is increasing, and the computing resources consumed are also increasing. In order to better utilize computing resources and obtain higher business benefits, elastic computing power allocation is mainly achieved through PID (English: Proportion Integration Differentiation, Chinese: Proportional Integral Differentiation) controller, using elastic queue length control. However, using this method, the utilization of computing resources may still be difficult to achieve expectations, which will affect business benefits. Summary of the invention

[0003] In order to solve the problems existing in the related art, the present disclosure provides a method, device, storage medium and electronic device for adjusting computing resources.

[0004] In order to achieve the above objective, according to a first aspect of an embodiment of the present disclosure, a method for adjusting computing resources is provided, the method comprising:

[0005] Obtain target configuration information corresponding to multiple delivery channels in a specified time period; the target configuration information includes a target configuration corresponding to each delivery channel, the target configuration includes a target queue length of an elastic queue, a target model size of an elastic model, and a target link complexity of an elastic link; the target queue length of the elastic queue is used to indicate the amount of promotional content to be input into the elastic link, the elastic model is set in the elastic link, and the elastic link is used to determine a revenue indicator according to the promotional content;

[0006] Obtaining the actual computing power resource ratio when delivering the promotion content according to the target configuration information;

[0007] The target configuration information is adjusted according to the actual computing power resource ratio and the preset computing power resource ratio, and the promotion content is delivered according to the adjusted target configuration information to optimize the actual computing power resource ratio.

[0008] Optionally, the obtaining of target configuration information corresponding to multiple delivery channels in a specified time period includes:

[0009] Acquire multiple types of candidate configuration information; each type of the candidate configuration information includes a candidate configuration corresponding to each of the delivery channels, and the candidate configuration includes a candidate queue length of an elastic queue, a candidate model size of an elastic model, and a candidate link complexity of an elastic link; and each two types of the candidate configuration information are different;

[0010] Determine the benefit indicator corresponding to each of the candidate configuration information;

[0011] According to the benefit indicator, the target configuration information is determined from a plurality of the candidate configuration information.

[0012] Optionally, the determining a benefit indicator corresponding to each of the to-be-selected configuration information includes:

[0013] For each type of the configuration information to be selected, a simulation of placing the promotional content is performed according to the configuration information to be selected to obtain a revenue indicator corresponding to the configuration information to be selected.

[0014] Optionally, determining the target configuration information from a plurality of the candidate configuration information according to the benefit indicator includes:

[0015] The candidate configuration information whose corresponding revenue index and corresponding computing power resource ratio meet the preset constraints is used as the target configuration information.

[0016] Optionally, obtaining the actual computing resource ratio when delivering the promotion content according to the target configuration information includes:

[0017] For each of the delivery channels, the promotion content is delivered according to the target configuration corresponding to the delivery channel indicated by the target configuration information, and a channel computing power resource ratio of the delivery channel is obtained;

[0018] The sum of the channel computing power resource proportions of all delivery channels is taken as the actual computing power resource proportion.

[0019] Optionally, adjusting the target configuration information according to the actual computing resource ratio and the preset computing resource ratio includes:

[0020] The difference between the preset computing power resource ratio and the actual computing power resource ratio is used as feedback, and the target configuration information is adjusted through a PID controller.

[0021] According to a second aspect of an embodiment of the present disclosure, a computing resource adjustment device is provided, the device comprising:

[0022] an acquisition module configured to acquire target configuration information corresponding to a plurality of delivery channels in a specified time period; the target configuration information includes a target configuration corresponding to each of the delivery channels, the target configuration includes a target queue length of an elastic queue, a target model size of an elastic model, and a target link complexity of an elastic link; the target queue length of the elastic queue is used to indicate the amount of promotional content to be input into the elastic link, the elastic model is set in the elastic link, and the elastic link is used to determine a revenue indicator according to the promotional content;

[0023] The acquisition module is further configured to acquire the actual computing resource ratio when the promotion content is delivered according to the target configuration information;

[0024] The adjustment module is configured to adjust the target configuration information according to the actual computing power resource ratio and the preset computing power resource ratio, and to deliver the promotion content according to the adjusted target configuration information to optimize the actual computing power resource ratio.

[0025] Optionally, the acquisition module includes:

[0026] The first acquisition submodule is configured to acquire a plurality of candidate configuration information; each of the candidate configuration information includes a candidate configuration corresponding to each of the delivery channels, and the candidate configuration includes a candidate queue length of an elastic queue, a candidate model size of an elastic model, and a candidate link complexity of an elastic link; and each two of the candidate configuration information are different;

[0027] A determination submodule is configured to determine a revenue indicator corresponding to each of the to-be-selected configuration information;

[0028] The determination submodule is further configured to determine the target configuration information from a plurality of the candidate configuration information according to the benefit indicator.

[0029] Optionally, the determination submodule is configured to simulate, for each type of the candidate configuration information, the delivery of the promotional content according to the candidate configuration information, and obtain a revenue indicator corresponding to the candidate configuration information.

[0030] Optionally, the determination submodule is configured to use the candidate configuration information whose corresponding revenue indicator and corresponding computing power resource ratio satisfy preset constraints as the target configuration information.

[0031] Optionally, the acquisition module includes:

[0032] The second acquisition submodule is configured to deliver the promotion content to each delivery channel according to the target configuration corresponding to the delivery channel indicated by the target configuration information, and acquire the channel computing power resource ratio of the delivery channel;

[0033] The processing submodule is configured to take the sum of the channel computing power resource proportions of all delivery channels as the actual computing power resource proportion.

[0034] Optionally, the adjustment module is configured to use the difference between the preset computing power resource ratio and the actual computing power resource ratio as a feedback amount, and adjust the target configuration information through a PID controller.

[0035] According to a third aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the steps of the method described in the first aspect above are implemented.

[0036] According to a fourth aspect of an embodiment of the present disclosure, there is provided an electronic device, including:

[0037] a memory having a computer program stored thereon;

[0038] A processor is used to execute the computer program in the memory to implement the steps of the method described in the first aspect above.

[0039] Through the above technical solution, the present disclosure first obtains the target configuration information corresponding to multiple delivery channels in a specified time period, wherein the target configuration information includes the target configuration corresponding to each delivery channel, the target configuration includes the target queue length of the elastic queue, the target model size of the elastic model and the target link complexity of the elastic link, the target queue length of the elastic queue is used to indicate the number of promotional contents to be input into the elastic link, the elastic model is set in the elastic link, and the elastic link is used to determine the revenue index according to the promotional content, and then obtains the actual computing power resource ratio when the promotional content is delivered according to the target configuration information, and then adjusts the target configuration information according to the actual computing power resource ratio and the preset computing power resource ratio, and delivers the promotional content according to the adjusted target configuration information to optimize the actual computing power resource ratio. The present disclosure improves the utilization of computing power resources by obtaining the target configuration information that can meet the maximization of computing power resources under the condition of maximum revenue, and delivers the promotional content according to the target configuration information, and at the same time, the target configuration information can be adjusted according to the actual computing power resource ratio and the preset computing power resource ratio, and the promotional content can be delivered according to the adjusted target configuration information, so that the actual computing power resource ratio is closer to the preset computing power resource ratio, thereby optimizing the utilization of computing power resources and improving business revenue.

[0040] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present disclosure but do not constitute a limitation of the present disclosure. In the accompanying drawings:

[0042] Figure 1 is a flow chart of a method for adjusting computing resources according to an exemplary embodiment;

[0043] Figure 2 is based on Figure 1 The illustrated embodiment shows a flow chart of step 101;

[0044] Figure 3 is a block diagram of a computing resource adjustment device according to an exemplary embodiment;

[0045] Figure 4 is based on Figure 3 A block diagram of an acquisition module shown in the illustrated embodiment;

[0046] Figure 5 is based on Figure 3 A block diagram of another acquisition module shown in the illustrated embodiment;

[0047] Figure 6 It is a block diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0048] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0049] Figure 1 FIG. 1 is a flow chart of a method for adjusting computing resources according to an exemplary embodiment. Figure 1 As shown, the method may include the following steps:

[0050] Step 101: Obtain target configuration information corresponding to multiple delivery channels in a specified time period.

[0051] Among them, the target configuration information includes the target configuration corresponding to each delivery channel. The target configuration includes the target queue length of the elastic queue, the target model size of the elastic model and the target link complexity of the elastic link. The target queue length of the elastic queue is used to indicate the number of promotional content to be input into the elastic link. The elastic model is set in the elastic link. The elastic link is used to determine the revenue indicator based on the promotional content.

[0052] For example, when using various delivery channels to deliver promotional content, it is necessary to implement elastic computing power allocation through elastic queues, elastic models, and elastic links, so as to better utilize computing power resources and obtain higher business benefits. Among them, the delivery channel can be a platform that supports the delivery of promotional content, for example, it can include multiple channels such as APP (English: Application, Chinese: Application), mini-programs, web pages, etc., and the promotional content can be advertising, news, and product information. In order to further improve the utilization of computing power resources, dynamic planning and combination can be performed in advance in an offline manner to determine the optimal combination of elastic queues, elastic models, and elastic links corresponding to all delivery channels in each time period of the day (for example, every 15 minutes can be used as a time period), and then use the optimal combination to deliver promotional content through all delivery channels, and in the delivery process, according to the actual computing power resource ratio and the preset computing power resource ratio, the PID control strategy is used to further adjust the optimal combination, so as to ensure the utilization of computing power resources and improve business benefits.

[0053] Specifically, first, the target configuration corresponding to each delivery channel in the specified time period (for example, the time period from 12:00 to 12:15) can be obtained by offline dynamic planning combination, and then the target configuration information corresponding to all delivery channels in the specified time period can be obtained. Among them, the target configuration corresponding to each delivery channel can be understood as the optimal combination of elastic queues, elastic models and elastic links of the delivery channel in the specified time period, which can specifically include the target queue length of the elastic queue, the target model size of the elastic model and the target link complexity of the elastic link. The target queue length of the elastic queue is used to indicate the number of promotional contents to be input into the elastic link. The queue lengths of the elastic queues allocated to delivery channels with different value flows are different (that is, the number of promotional contents allocated is different). The elastic model is set in the elastic link. The elastic model is used to estimate the revenue indicators of the promotional content (for example, the number of orders, clicks and promotional content revenue, etc.) according to the promotional content. Different elastic models can be assigned to different value flows. The large model has a better estimation effect than the small model, but it consumes more computing resources. Elastic links are used to determine revenue indicators based on promotional content. Different value flows can be assigned elastic links of different complexity. Elastic links with higher complexity consume more computing resources. The complexity of an elastic link is expressed as the number of processing steps included in the link. For example, a complete elastic link can include a series of processing steps such as recall, rough sorting, fine sorting, mechanism, and floating. The complexity of a complete elastic link is the highest. If the rough sorting processing step is removed from the complete elastic link, the complexity of the elastic link will also be reduced accordingly.

[0054] Step 102, obtaining the actual computing power resource ratio when the promotion content is delivered according to the target configuration information.

[0055] In this step, for each delivery channel, promotional content can be delivered according to the target configuration corresponding to the delivery channel indicated by the target configuration information (that is, each delivery channel delivers promotional content according to the optimal combination of its corresponding elastic queue, elastic model and elastic link), and the channel computing power resource ratio of the delivery channel is obtained. Then, the sum of the channel computing power resource ratios of all delivery channels can be used as the actual computing power resource ratio. The actual computing power resource ratio is used to indicate the ratio of computing power resources used by multiple delivery channels when delivering promotional content according to the target configuration information to the total computing power resources.

[0056] Step 103, according to the actual computing power resource ratio and the preset computing power resource ratio, the target configuration information is adjusted, and the promotion content is delivered according to the adjusted target configuration information to optimize the actual computing power resource ratio.

[0057] For example, a preset computing power resource ratio (for example, 90%) can be set in advance, and the preset resource ratio is used to represent the ratio of computing power resources to total computing power resources required by multiple delivery channels when delivering promotional content as expected by the user. Ideally, the actual computing power resource ratio of multiple delivery channels when delivering promotional content according to the target configuration information should be the same as the preset computing power resource ratio. However, in actual situations, each channel may not be able to use up the allocated computing power resources, which often makes the actual computing power resource ratio lower than the preset resource ratio. In order to ensure the utilization of computing power resources, a PID control strategy can be adopted to adjust the target configuration information according to the actual computing power resource ratio and the preset computing power resource ratio according to a preset period (for example, once every 10 seconds) (that is, adjust the optimal combination of elastic queues, elastic models and elastic links corresponding to each delivery channel within a specified time period). For example, the queue length of the elastic queue of the delivery channel can be adjusted, and the model size of the elastic model of the delivery channel can also be adjusted. After the target configuration information is adjusted, promotional content can be delivered according to the adjusted target configuration information to optimize the actual computing power resource ratio so that the actual computing power resource ratio is close to the preset computing power resource ratio.

[0058] In summary, the present disclosure first obtains target configuration information corresponding to multiple delivery channels in a specified time period, wherein the target configuration information includes the target configuration corresponding to each delivery channel, the target configuration includes the target queue length of the elastic queue, the target model size of the elastic model and the target link complexity of the elastic link, the target queue length of the elastic queue is used to indicate the number of promotional contents to be input into the elastic link, the elastic model is set in the elastic link, and the elastic link is used to determine the revenue index according to the promotional content, and then obtains the actual computing power resource ratio when the promotional content is delivered according to the target configuration information, and then adjusts the target configuration information according to the actual computing power resource ratio and the preset computing power resource ratio, and delivers the promotional content according to the adjusted target configuration information to optimize the actual computing power resource ratio. The present disclosure improves the utilization of computing power resources by obtaining target configuration information that can meet the maximization of computing power resources under the condition of maximum revenue, and delivers the promotional content according to the target configuration information, and at the same time, the target configuration information can be adjusted according to the actual computing power resource ratio and the preset computing power resource ratio, and the promotional content can be delivered according to the adjusted target configuration information, so that the actual computing power resource ratio is closer to the preset computing power resource ratio, thereby optimizing the utilization of computing power resources and improving business revenue.

[0059] Figure 2 is based on Figure 1 The embodiment shown is a flow chart of step 101. Figure 2 As shown, step 101 may include the following steps:

[0060] Step 1011, obtain multiple types of candidate configuration information, each type of candidate configuration information includes a candidate configuration corresponding to each delivery channel, the candidate configuration includes a candidate queue length of the elastic queue, a candidate model size of the elastic model, and a candidate link complexity of the elastic link, and each two types of candidate configuration information are different.

[0061] For example, first, for a specified time period, elastic queues, elastic models, elastic links, and multiple delivery channels can be combined to obtain multiple types of configuration information to be selected. Each type of configuration information to be selected can include a configuration to be selected corresponding to each delivery channel. Further, elastic queues, elastic models, elastic links, and multiple delivery channels can be combined according to the following table, where the content corresponding to each cell in the table is the traffic proportion of the delivery channel corresponding to the cell determined based on historical data to the access traffic proportion of all delivery channels.

[0062] Table 1

[0063]

[0064] Step 1012: determine the revenue index corresponding to each type of configuration information to be selected.

[0065] For example, for each type of configuration information to be selected, a simulation of the promotional content can be performed according to the configuration information to be selected to obtain the revenue indicator corresponding to the configuration information to be selected. For example, when the specified time period is 12:00-12:15 and the delivery channel is APP, the selected configuration corresponding to the APP can be that the length of the selected queue of the elastic queue is 80, the size of the selected model of the elastic model is 100G, and the complexity of the selected link of the elastic link is the highest complexity (that is, the elastic link is a complete link). In this way, simulation can be performed offline to obtain revenue indicators such as exposure, clicks, orders, and content delivery revenue. It should be noted that in this way, when performing dynamic planning, traffic simulation simulation is no longer required, and the simulation results can be directly extracted, thereby speeding up the speed of dynamic planning and adjustment using PID control strategy.

[0066] Step 1013, determining target configuration information from a plurality of candidate configuration information according to the benefit index.

[0067] In this step, according to the profit index, the target configuration information whose corresponding profit index and corresponding computing power resource ratio meet the preset constraint conditions can be determined from a variety of candidate configuration information. The preset constraint condition is to select the candidate configuration information that can maximize the computing power resources while ensuring the maximum profit as the target configuration information.

[0068] Optionally, step 1013 may be implemented in the following manner:

[0069] The candidate configuration information whose corresponding profit index and corresponding computing power resource ratio meet the preset constraints is used as the target configuration information.

[0070] For example, the process of determining the target configuration information can actually be converted into the simulation results of simulating the promotion content based on each candidate configuration information to find the optimal path. This path can be expressed as a problem of maximizing benefits under computing power constraints. The expression of this path can be expressed as:

[0071] Max∑∑X(ij)*F revenue (TYPEi, COMBj) (1)

[0072] St.M<∑∑X(ij)*F revenue / CU (2)

[0073] Wherein, expression (1) is used to express the maximum benefit that can be achieved by delivering promotional content according to the selected configuration information under ideal conditions, and expression (2) is used to express the minimum benefit to be achieved by delivering promotional content according to the selected configuration information. TYPEi is the type of delivery channel, COMBj is the type of elastic queue, elastic model and elastic link combination, X(ij) is the corresponding traffic proportion under TYPEi and COMBj, and F revenue (TYPEi, COMBj) is used to represent the corresponding profit indicators under TYPEi and COMBj, CU is the computing power resource, and M is the minimum profit threshold to be achieved.

[0074] Assuming that the higher the computing power resources invested, the greater the benefits, then F revenue / CU is a monotonically decreasing function, and through Lagrange duality we can get:

[0075] Min∑∑X(ij)*F revenue (TYPEi, COMBj)+alpha*∑∑X(ij)*F revenue / CU (3)

[0076] Extracting X(ij) from expression (3) shows that maximizing revenue is only related to computing resources, and the position of maximizing computing resources while ensuring maximum revenue is: CU*M=(1+alpha)*revenue. Then, through this position, the candidate configuration information that can meet the maximization of computing resources while ensuring maximum revenue can be selected as the target configuration information.

[0077] Optionally, step 103 may be implemented in the following manner:

[0078] The difference between the preset computing power resource ratio and the actual computing power resource ratio is used as feedback, and the target configuration information is adjusted through the PID controller.

[0079] For example, after obtaining the actual computing power resource ratio, the difference between the preset computing power resource ratio and the actual computing power resource ratio can be used as the feedback quantity through the PID controller, and then the target configuration information can be used as the control quantity, and the PID controller parameters can be determined by performing PID control based on the parameter optimization method of the Pareto parameter optimality. Then, the target configuration information can be adjusted according to the determined parameters of the PID controller. After the target configuration information is adjusted, promotional content can be delivered according to the adjusted target configuration information to optimize the actual computing power resource ratio so that the actual computing power resource ratio is close to the preset computing power resource ratio.

[0080] In summary, the present disclosure first obtains target configuration information corresponding to multiple delivery channels in a specified time period, wherein the target configuration information includes the target configuration corresponding to each delivery channel, the target configuration includes the target queue length of the elastic queue, the target model size of the elastic model and the target link complexity of the elastic link, the target queue length of the elastic queue is used to indicate the number of promotional contents to be input into the elastic link, the elastic model is set in the elastic link, and the elastic link is used to determine the revenue index according to the promotional content, and then obtains the actual computing power resource ratio when the promotional content is delivered according to the target configuration information, and then adjusts the target configuration information according to the actual computing power resource ratio and the preset computing power resource ratio, and delivers the promotional content according to the adjusted target configuration information to optimize the actual computing power resource ratio. The present disclosure improves the utilization of computing power resources by obtaining target configuration information that can meet the maximization of computing power resources under the condition of maximum revenue, and delivers the promotional content according to the target configuration information, and at the same time, the target configuration information can be adjusted according to the actual computing power resource ratio and the preset computing power resource ratio, and the promotional content can be delivered according to the adjusted target configuration information, so that the actual computing power resource ratio is closer to the preset computing power resource ratio, thereby optimizing the utilization of computing power resources and improving business revenue.

[0081] Figure 3 FIG. 1 is a block diagram of a device for adjusting computing resources according to an exemplary embodiment. Figure 3 As shown, the device 200 includes:

[0082] The acquisition module 201 is configured to acquire target configuration information corresponding to multiple delivery channels in a specified time period. The target configuration information includes the target configuration corresponding to each delivery channel, and the target configuration includes the target queue length of the elastic queue, the target model size of the elastic model, and the target link complexity of the elastic link. The target queue length of the elastic queue is used to indicate the number of promotional contents to be input into the elastic link. The elastic model is set in the elastic link, and the elastic link is used to determine the revenue indicator according to the promotional content.

[0083] The acquisition module 201 is further configured to acquire the actual computing power resource ratio when the promotion content is delivered according to the target configuration information.

[0084] The adjustment module 202 is configured to adjust the target configuration information according to the actual computing power resource ratio and the preset computing power resource ratio, and to deliver promotional content according to the adjusted target configuration information to optimize the actual computing power resource ratio.

[0085] Figure 4 is based on Figure 3 The embodiment shown is a block diagram of an acquisition module. Figure 4 As shown, the acquisition module 201 includes:

[0086] The first acquisition submodule 2011 is configured to acquire multiple candidate configuration information. Each candidate configuration information includes a candidate configuration corresponding to each delivery channel, and the candidate configuration includes a candidate queue length of an elastic queue, a candidate model size of an elastic model, and a candidate link complexity of an elastic link, and each two candidate configuration information are different.

[0087] The determination submodule 2012 is configured to determine the revenue indicator corresponding to each type of configuration information to be selected.

[0088] The determination submodule 2012 is further configured to determine target configuration information from a plurality of candidate configuration information according to the benefit index.

[0089] Optionally, the determination submodule 2012 is configured to simulate the placement of promotional content for each type of configuration information to be selected according to the type of configuration information to be selected, and obtain a revenue indicator corresponding to the type of configuration information to be selected.

[0090] Optionally, the determination submodule 2012 is configured to use the candidate configuration information whose corresponding revenue index and corresponding computing power resource ratio satisfy preset constraints as the target configuration information.

[0091] Figure 5 is based on Figure 3 The embodiment shown is a block diagram of another acquisition module. Figure 5 As shown, the acquisition module 201 includes:

[0092] The second acquisition submodule 2013 is configured to deliver promotional content to each delivery channel according to the target configuration corresponding to the delivery channel indicated by the target configuration information, and to obtain the channel computing power resource ratio of the delivery channel.

[0093] The processing submodule 2014 is configured to take the sum of the channel computing power resource proportions of all delivery channels as the actual computing power resource proportion.

[0094] Optionally, the adjustment module 202 is configured to use the difference between the preset computing power resource ratio and the actual computing power resource ratio as a feedback amount, and adjust the target configuration information through a PID controller.

[0095] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0096] In summary, the present disclosure first obtains target configuration information corresponding to multiple delivery channels in a specified time period, wherein the target configuration information includes the target configuration corresponding to each delivery channel, the target configuration includes the target queue length of the elastic queue, the target model size of the elastic model and the target link complexity of the elastic link, the target queue length of the elastic queue is used to indicate the number of promotional contents to be input into the elastic link, the elastic model is set in the elastic link, and the elastic link is used to determine the revenue index according to the promotional content, and then obtains the actual computing power resource ratio when the promotional content is delivered according to the target configuration information, and then adjusts the target configuration information according to the actual computing power resource ratio and the preset computing power resource ratio, and delivers the promotional content according to the adjusted target configuration information to optimize the actual computing power resource ratio. The present disclosure improves the utilization of computing power resources by obtaining target configuration information that can meet the maximization of computing power resources under the condition of maximum revenue, and delivers the promotional content according to the target configuration information, and at the same time, the target configuration information can be adjusted according to the actual computing power resource ratio and the preset computing power resource ratio, and the promotional content can be delivered according to the adjusted target configuration information, so that the actual computing power resource ratio is closer to the preset computing power resource ratio, thereby optimizing the utilization of computing power resources and improving business revenue.

[0097] Figure 6 is a block diagram of an electronic device 300 according to an exemplary embodiment. For example, the electronic device 300 may be provided as a server. Figure 6 The electronic device 300 includes a processor 322, which may be one or more, and a memory 332 for storing a computer program executable by the processor 322. The computer program stored in the memory 332 may include one or more modules, each corresponding to a set of instructions. In addition, the processor 322 may be configured to execute the computer program to perform the above-mentioned method for adjusting computing resources.

[0098] In addition, the electronic device 300 may further include a power supply component 326 and a communication component 350. The power supply component 326 may be configured to perform power management of the electronic device 300, and the communication component 350 may be configured to implement communication of the electronic device 300, for example, wired or wireless communication. In addition, the electronic device 300 may further include an input / output (I / O) interface 358. The electronic device 300 may operate based on an operating system stored in the memory 332, such as Windows Server 2008. TM , Mac OSX TM , Unix TM , Linux TM etc.

[0099] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by the processor, the steps of the above-mentioned method for adjusting computing power resources are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 332 including program instructions, and the above-mentioned program instructions can be executed by the processor 322 of the electronic device 300 to complete the above-mentioned method for adjusting computing power resources.

[0100] In another exemplary embodiment, a computer program product is also provided, which includes a computer program that can be executed by a programmable device, and the computer program has a code portion for executing the above-mentioned method for adjusting computing power resources when executed by the programmable device.

[0101] The preferred embodiments of the present disclosure are described in detail above in conjunction with the accompanying drawings; however, the present disclosure is not limited to the specific details in the above embodiments. Within the technical concept of the present disclosure, a variety of simple modifications can be made to the technical solution of the present disclosure, and these simple modifications all fall within the protection scope of the present disclosure.

[0102] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present disclosure will not further describe various possible combinations.

[0103] In addition, various embodiments of the present disclosure may be arbitrarily combined, and as long as they do not violate the concept of the present disclosure, they should also be regarded as the contents disclosed by the present disclosure.

Claims

1. A method for adjusting computing resources, characterized in that: The method comprises: Obtain target configuration information corresponding to multiple delivery channels in a specified time period; the target configuration information includes a target configuration corresponding to each delivery channel, the target configuration includes a target queue length of an elastic queue, a target model size of an elastic model, and a target link complexity of an elastic link; the target queue length of the elastic queue is used to indicate the amount of promotional content to be input into the elastic link, the elastic model is set in the elastic link, and the elastic link is used to determine a revenue indicator according to the promotional content; Obtaining the actual computing power resource ratio when delivering the promotion content according to the target configuration information; According to the actual computing power resource ratio and the preset computing power resource ratio, the target configuration information is adjusted, and the promotion content is delivered according to the adjusted target configuration information to optimize the actual computing power resource ratio; The preset computing power resource ratio is used to indicate the ratio of computing power resources to total computing power resources required by multiple delivery channels expected by the user when delivering promotional content; the actual computing power resource ratio is used to indicate the ratio of computing power resources used by multiple delivery channels when delivering promotional content according to target configuration information to total computing power resources; The obtaining of the actual computing resource ratio when delivering the promotion content according to the target configuration information includes: For each of the delivery channels, the promotion content is delivered according to the target configuration corresponding to the delivery channel indicated by the target configuration information, and a channel computing power resource ratio of the delivery channel is obtained; The sum of the channel computing power resource ratios of all delivery channels is taken as the actual computing power resource ratio; The step of adjusting the target configuration information according to the actual computing resource ratio and the preset computing resource ratio includes: The difference between the preset computing power resource ratio and the actual computing power resource ratio is used as feedback, and the target configuration information is adjusted through a PID controller.

2. The method according to claim 1, characterized in that The obtaining of target configuration information corresponding to multiple delivery channels in a specified time period includes: Acquire multiple types of candidate configuration information; each type of the candidate configuration information includes a candidate configuration corresponding to each of the delivery channels, and the candidate configuration includes a candidate queue length of an elastic queue, a candidate model size of an elastic model, and a candidate link complexity of an elastic link; and each two types of the candidate configuration information are different; Determine the benefit indicator corresponding to each of the candidate configuration information; According to the benefit indicator, the target configuration information is determined from a plurality of the candidate configuration information.

3. The method according to claim 2, characterized in that The determining of the benefit indicator corresponding to each of the candidate configuration information includes: For each type of the configuration information to be selected, a simulation of placing the promotional content is performed according to the configuration information to be selected to obtain a revenue indicator corresponding to the configuration information to be selected.

4. The method according to claim 2, characterized in that: The step of determining the target configuration information from a plurality of the candidate configuration information according to the benefit indicator includes: The candidate configuration information whose corresponding revenue index and corresponding computing power resource ratio meet the preset constraints is used as the target configuration information.

5. A computing resource adjustment device, characterized in that: The device comprises: an acquisition module configured to acquire target configuration information corresponding to a plurality of delivery channels in a specified time period; the target configuration information includes a target configuration corresponding to each of the delivery channels, the target configuration includes a target queue length of an elastic queue, a target model size of an elastic model, and a target link complexity of an elastic link; the target queue length of the elastic queue is used to indicate the amount of promotional content to be input into the elastic link, the elastic model is set in the elastic link, and the elastic link is used to determine a revenue indicator according to the promotional content; The acquisition module is further configured to acquire the actual computing resource ratio when the promotion content is delivered according to the target configuration information; an adjustment module, configured to adjust the target configuration information according to the actual computing resource ratio and the preset computing resource ratio, and deliver the promotion content according to the adjusted target configuration information to optimize the actual computing resource ratio; The preset computing power resource ratio is used to indicate the ratio of computing power resources to total computing power resources required by multiple delivery channels expected by the user when delivering promotional content; the actual computing power resource ratio is used to indicate the ratio of computing power resources used by multiple delivery channels when delivering promotional content according to target configuration information to total computing power resources; The obtaining of the actual computing resource ratio when delivering the promotion content according to the target configuration information includes: For each of the delivery channels, the promotion content is delivered according to the target configuration corresponding to the delivery channel indicated by the target configuration information, and a channel computing power resource ratio of the delivery channel is obtained; The sum of the channel computing power resource ratios of all delivery channels is taken as the actual computing power resource ratio; The step of adjusting the target configuration information according to the actual computing resource ratio and the preset computing resource ratio includes: The difference between the preset computing power resource ratio and the actual computing power resource ratio is used as feedback, and the target configuration information is adjusted through a PID controller.

6. The device according to claim 5, characterized in that The acquisition module comprises: The first acquisition submodule is configured to acquire a plurality of candidate configuration information; each of the candidate configuration information includes a candidate configuration corresponding to each of the delivery channels, and the candidate configuration includes a candidate queue length of an elastic queue, a candidate model size of an elastic model, and a candidate link complexity of an elastic link; and each two of the candidate configuration information are different; A determination submodule is configured to determine a revenue indicator corresponding to each of the to-be-selected configuration information; The determination submodule is further configured to determine the target configuration information from a plurality of the candidate configuration information according to the benefit indicator.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 4 are implemented.

8. An electronic device, characterized in that: include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1 to 4.

Images