Computing power resource determination method and device, equipment, medium and product

Abstract

The invention discloses a computing power resource determination method, apparatus and device, a medium and a product. The method comprises the steps of obtaining real-time computing power resource data of a natural language demand and a resource library; performing intention analysis on the natural language demand based on the historical performance database, and determining a computing power task instruction; and on the basis of the real-time computing power resource data, performing dual-algorithm collaborative matching on the computing power task instruction, and determining a target resource combination meeting requirements from a resource library. According to the method, intention analysis is carried out on natural language requirements in combination with a historical performance database, fuzziness of intention analysis is eliminated from the source, cooperative matching is carried out through a bi-pass algorithm so as to reduce matching deviation of computing power resources, accurate analysis of the requirements is achieved, and then accurate matching of the computing power resources is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer technology, and in particular to a computing power resource determination method, device, equipment, medium and product. BACKGROUND

[0002] With the rapid development of digital economy, computing power resources have become the core production factors of the digital transformation of governments and enterprises. Intelligent trading of computing power resources, as a key link connecting supply and demand, directly affects resource allocation efficiency and business landing effectiveness. The intention-driven computing power resource trading mode gradually becomes an important technical direction for improving trading efficiency by analyzing user natural language requirements and automatically completing resource matching and transaction execution.

[0003] In the prior art, computing power resource trading relies on manual communication or simple automated systems: after the user proposes computing power requirements in the form of text and voice, etc., the technical parameters need to be manually sorted and converted; the system filters resources based on preset rules or a basic database to generate a preliminary quotation.

[0004] However, user requirements are mostly non-standardized expressions, and manual conversion is prone to understanding bias, resulting in low accuracy of subsequent resource matching and demand-resource mismatching problems. In the entire process, the demand understanding, resource matching and price generation links are relatively independent, and lack of dynamic coordination mechanism. SUMMARY

[0005] The present application provides a computing power resource determination method, device, equipment, medium and product to realize accurate understanding of demand and accurate matching of computing power resources.

[0006] According to a first aspect of the present application, a computing power resource determination method is provided, comprising:

[0007] acquiring real-time computing power resource data of natural language requirements and a resource library;

[0008] performing intention analysis on the natural language requirements based on a historical performance database to determine computing power task instructions;

[0009] based on the real-time computing power resource data, performing double-algorithm collaborative matching on the computing power task instructions to determine a target resource combination that meets the requirements from the resource library.

[0010] According to a second aspect of the present application, a computing power resource determination device is provided, comprising:

[0011] a demand acquisition module configured to acquire real-time computing power resource data of natural language requirements and a resource library;

[0012] an intention analysis module configured to perform intention analysis on the natural language requirements based on a historical performance database to determine computing power task instructions;

[0013] The resource determination module is used to perform dual-algorithm collaborative matching on the computing power task instructions based on the real-time computing power resource data, and determine the target resource combination that meets the requirements from the resource library.

[0014] According to a third aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0015] At least one processor; and

[0016] A memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the computing resource determination method according to any embodiment of the present invention.

[0018] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the computing resource determination method according to any embodiment of the present invention.

[0019] According to a fifth aspect of the present invention, embodiments of the present invention also provide a computer program product, the computer program product including a computer program, which, when executed by a processor, implements the computing resource determination method of any embodiment of the present invention.

[0020] The technical solution of this invention involves acquiring real-time computing power resource data from natural language requirements and a resource library; performing intent parsing on natural language requirements based on a historical fulfillment database to determine computing power task instructions; and using a dual-algorithm collaborative matching process on the computing power task instructions based on real-time computing power resource data to determine a target resource combination that meets the requirements from the resource library. By combining the historical fulfillment database with intent parsing of natural language requirements, the ambiguity of intent analysis is eliminated at the source. The dual-algorithm collaborative matching reduces the matching deviation of computing power resources, achieving accurate parsing of requirements and thus accurate matching of computing power resources.

[0021] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of a method for determining computing resources according to Embodiment 1 of the present invention;

[0024] Figure 2 This is a flowchart of a method for determining computing resources according to Embodiment 2 of the present invention;

[0025] Figure 3 This is a schematic diagram of a computing resource determination device according to Embodiment 3 of the present invention;

[0026] Figure 4 This is a schematic diagram of the structure of an electronic device that implements an embodiment of the present invention. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] Example 1

[0030] Figure 1This is a flowchart illustrating a method for determining computing resources according to Embodiment 1 of the present invention. This embodiment is applicable to situations involving the matching of computing resources. The method can be executed by a computing resource determination device, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method includes:

[0031] S110. Obtain real-time computing power resource data for natural language requirements and resource libraries.

[0032] In this embodiment, natural language requirements can be understood as computing resource demand information expressed in natural language form. The resource library can be understood as a library used to determine all available computing resources. Real-time computing resource data can be understood as attributes representing each current computing resource, such as resource type, performance parameters, real-time price, supplier information, and resource status data.

[0033] Specifically, the processor can obtain the user's natural language input requirements, as well as real-time computing resource data from the associated resource library.

[0034] S120. Based on the historical performance database, perform intent parsing on natural language requirements to determine computing power task instructions.

[0035] In this embodiment, the historical fulfillment database can be understood as a database consisting of historically resolved demands. The computing power task instructions can be understood as structured task instructions.

[0036] Specifically, the processor can perform word segmentation and entity recognition on the natural language requirements input by the user, extract the core elements of computing power type, duration, performance requirements, and industry scenario; then, through the actual performance vector of resources in the historical performance database, it can calibrate the fuzzy expressions in the natural language requirements to obtain computing power task instructions, so as to ensure the accuracy of the quantification parameters in the output structured task instructions through calibration.

[0037] S130. Based on real-time computing power resource data, dual-algorithm collaborative matching is performed on computing power task instructions to determine the target resource combination that meets the requirements from the resource library.

[0038] In this embodiment, the target resource combination can be understood as the optimal combination of computing resources that satisfies the needs of natural language processing. The needs refer to the computing power requirements corresponding to the computing task instructions.

[0039] Specifically, the processor can use the first algorithm in the dual algorithm to combine real-time computing power resource data to first select multiple candidate computing power resource combinations that can meet the computing power task instructions from the resource library. Then, the second algorithm in the dual algorithm can further select the candidate computing power resource combinations to determine the best target resource combination that meets the requirements.

[0040] The technical solution of this invention involves acquiring real-time computing power resource data from natural language requirements and a resource library; performing intent parsing on natural language requirements based on a historical fulfillment database to determine computing power task instructions; and using a dual-algorithm collaborative matching process on the computing power task instructions based on real-time computing power resource data to determine a target resource combination that meets the requirements from the resource library. By combining the historical fulfillment database with intent parsing of natural language requirements, the ambiguity of intent analysis is eliminated at the source. The dual-algorithm collaborative matching reduces the matching deviation of computing power resources, achieving accurate parsing of requirements and thus accurate matching of computing power resources.

[0041] As a first optional embodiment in this example, based on the above embodiment, it further includes:

[0042] When the data update conditions are met, the explicit performance data and implicit status data of each computing resource in the resource library are obtained to obtain real-time computing resource data.

[0043] In this embodiment, data update conditions can be understood as conditions used to trigger updates to real-time computing resource data. These conditions may include timed updates with a base frequency of a preset time interval / time, and immediate updates when computing resource price fluctuations exceed a preset threshold, fault conditions are abnormal, or load fluctuations exceed a threshold. Explicit performance data are performance indicators used to characterize the basic capabilities of operator resources, and may include hardware indicators such as CPU clock speed and GPU computing power. Implicit status data can be understood as status indicators reflecting resource stability, reliability, and network transmission quality, and may include dynamic information such as the past 30 days of operation logs and real-time load rates.

[0044] Specifically, when the data update conditions are met, the processor can collect explicit performance data and implicit status data of computing resources from the corresponding computing power supplier interface in the resource library, the real-time monitoring system, and the industry database. The processor verifies the accuracy of the collected data by comparing the deviation between the supplier's nominal performance and the actual test performance. When the deviation exceeds the preset verification threshold, the processor marks the data status and triggers the review mechanism to ensure the reliability of the implicit status data. The verified data is then standardized to obtain real-time computing power resource data and synchronized to the resource library for storage, providing real-time and complete data support for resource matching and price calculation.

[0045] For example, the processor can connect to the supplier's system via an HTTPS interface to collect explicit data such as GPU model and memory capacity, and obtain implicit data such as CPU load and network packet loss rate in real time via the SNMP protocol. The data update conditions can be set to update every 10 minutes for ordinary resources and every 2 minutes for popular resources such as GPUs. When a price fluctuation of more than 10% or a failure lasting more than 5 minutes is detected, an update is triggered immediately and synchronized to the Redis global cache to ensure data consistency. Stress tests are performed on newly connected resources. If the nominal computing power deviates from the actual test value by more than 15%, it is marked as "pending verification" and an email notification is sent to the supplier. Resources that have not been verified will not participate in the matching.

[0046] The first optional embodiment of the present invention ensures the reliability of explicit / implicit resource data through a triple mechanism of data collection, dynamic updating of data update conditions, and quality verification (deviation review), providing high-quality input for algorithm matching.

[0047] Example 2

[0048] Figure 2 This is a flowchart of a method for determining computing resources according to Embodiment 2 of the present invention. This embodiment is a further refinement of the above embodiment. Figure 2 As shown, the method includes:

[0049] S201. Obtain real-time computing power resource data for natural language requirements and resource libraries.

[0050] S202. Extract core elements related to computing power from natural language requirements.

[0051] In this embodiment, the core elements related to computing power can be understood as those used to characterize the core requirements for computing power.

[0052] Specifically, the processor can segment and recognize entities from natural language requirements through natural language processing, and extract core elements such as computing power type, duration, performance requirements and industry scenarios from the natural language requirements as core elements related to computing power.

[0053] For example, a user's input natural language requirement could be "Provide computing power support for AI training in our research institute, for about a week, with better performance." The processor can use semantic understanding technology based on a large language model (or a pre-trained model based on BERT) to decompose the user's non-standardized expression (i.e., the above natural language requirement) into multiple levels, extracting the original requirement elements such as "subject = research institute", "scenario = AI training", "duration = one week" and "core requirement = high performance" as core elements related to computing power.

[0054] S203. Vectorize the fuzzy expressions in natural language requirements to obtain fuzzy expression vectors.

[0055] In this embodiment, a fuzzy expression can be understood as an expression whose intent cannot be accurately identified, such as "better performance" in the example above. A fuzzy expression vector can be understood as a fuzzy expression represented by vectorization.

[0056] Specifically, the processor can vectorize the fuzzy representations in natural language requirements to obtain fuzzy representation vectors.

[0057] S204. Match the fuzzy representation vectors based on the historical performance database to obtain the calibrated representation.

[0058] In this embodiment, the calibration statement can be understood as the precise requirement statement after calibrating the fuzzy statement, that is, translating the fuzzy statement into a requirement statement that the processor can understand.

[0059] Specifically, the processor can match historical similar demand cases stored in the historical performance database with the fuzzy expression vector (for example, cosine similarity matching can be used) to determine the expression content corresponding to the fuzzy expression vector as the calibration expression.

[0060] For example, a precise translation mechanism is initiated for ambiguous statements (such as "performance should be better"). By combining historical similar demand cases stored in the historical performance database (such as "high performance in AI training" corresponding to "single-precision floating-point computing power ≥100PFlops, video memory ≥512GB"), the expression deviation of the ambiguous statement is calibrated by cosine similarity matching (similarity ≥0.9), and the calibrated statement is obtained as peak computing power ≥100PFlops.

[0061] S205. Based on the calibration description and core elements related to computing power, determine the computing power task instructions.

[0062] Specifically, the processor can perform a true analysis of the calibration description and core elements related to computing power to determine the computing power task instructions.

[0063] For example, the final computing power task instruction obtained in the above example is: "Computing power type = AI training dedicated GPU cluster, peak computing power ≥ 100 PFlops, duration = 168 hours, stability requirement = failure rate ≤ 0.1%", ensuring that ambiguity is eliminated from the source and providing "zero deviation" input for subsequent resource matching. Another example is the natural language requirement of "financial industry data computing power, approximately 3 days". This is processed using a BERT pre-trained model, and after word segmentation, keywords such as "financial industry", "computing power", and "3 days" are extracted. Entity recognition locates "industry scenario = finance" and "duration = 72 hours", using these as core elements related to computing power. The vague expression "approximately needed" is corrected by retrieving historical performance data of similar requirements from the historical performance database (e.g., "financial industry computing power" corresponds to "encrypted computing power support, compliance certification") and using vector similarity matching (similarity ≥ 0.85) to calibrate "approximately needed" to "peak computing power ≥ 40 PFlops", eliminating the ambiguity of the expression.

[0064] S206. By combining real-time computing power resource data with the first algorithm, candidate resource combinations are determined from the resource library.

[0065] In this embodiment, the first algorithm can be a globally optimal algorithm. The candidate resource combination can be understood as a combination of computing resources that satisfies the optimal objective of the first algorithm.

[0066] Specifically, the processor can use the first algorithm to match the globally optimal candidate resource combination from real-time computing power resource data, and generate reference transaction prices and reference price details based on the real-time cost elements and market supply and demand factors of the candidate resource combination.

[0067] For example, the processor can first select resource pools that meet the performance parameters from the resource library based on real-time computing power resource data, and then use the first algorithm to perform multi-dimensional calculations on the performance, cost and utilization of the resources in the resource pool, and output 3-5 sets of candidate resource combinations with the highest cost performance; when generating the reference price, it combines the real-time utilization rate of resources (e.g., utilization rate > 90% triggers premium) and the supplier's historical quotation deviation rate (deviation rate < 5% grants price discount) to form a detailed list including basic costs and premium ratios, so that users can quickly understand the cost range.

[0068] For example, the first algorithm can be an optimization algorithm for matching the globally optimal resource combination, and its core objective function is:

[0069]

[0070] Where F represents the cost-effectiveness score of the resource combination, with a higher value indicating higher cost-effectiveness; P represents the resource performance compliance, ranging from 0 to 1, calculated based on the matching degree between the performance parameters in the structured task instructions and the actual performance of the resources; U represents the real-time resource utilization rate, ranging from 0 to 1, provided by the real-time resource module, reflecting the current load status of the resources; C represents the total cost of the resource combination (unit: yuan), which is the sum of the basic costs of each resource in the combination; the first algorithm maximizes the objective function F through iterative calculation and outputs the globally optimal candidate resource combination.

[0071] S207. Based on the second algorithm, high-risk resource combinations are eliminated from the candidate resource combinations to obtain the target resource combination.

[0072] In this embodiment, the second algorithm can be a reinforcement learning-based implicit constraint verification algorithm. High-risk resource combinations can be understood as resource combinations among candidate resource combinations that do not satisfy the implicit constraints.

[0073] Specifically, the processor can use a second algorithm to eliminate high-risk resource combinations from the candidate resource combinations to obtain the target resource combination.

[0074] For example, the second algorithm is a latent constraint verification algorithm based on reinforcement learning, and its core reward function is:

[0075]

[0076] Where R represents the real-time reward value for resource fulfillment quality, ranging from 0 to 100, with a higher value indicating more reliable quality; F t R represents the failure rate of the resource within a preset period, with a value range of 0-1, calculated from historical failure data in the real-time resource module; L represents the CPU load fluctuation coefficient, with a value range of 0-1, calculated based on the load standard deviation within a preset time period, reflecting resource stability; D represents the network packet loss rate, with a value range of 0-1, provided by real-time monitoring data, reflecting network transmission quality; when R < preset threshold, the second algorithm triggers the resource elimination mechanism, replacing it with resources that meet implicit constraints.

[0077] Furthermore, after performing dual-algorithm collaborative matching on computing task instructions based on real-time computing resource data to determine the target resource combination that meets the requirements from the resource library, it also includes:

[0078] The matching difference between the candidate resource combination and the target resource combination is used as error data; based on the error data, the algorithm parameters and data update conditions of the first algorithm and the second algorithm are updated.

[0079] In this embodiment, the matching difference can be understood as the difference between operators in the resource combination. Error data is used for subsequent parameter updates.

[0080] Specifically, the processor can compare the matching differences between candidate resource combinations and target resource combinations as error data. Based on this error data, the processor can adjust the objective function parameters of the first algorithm to improve its matching accuracy, reduce the optimization intervention frequency of the second algorithm, and optimize the above data update conditions according to the error data type, thereby improving the accuracy and efficiency of price calculation.

[0081] Based on the above embodiments, the steps for updating the algorithm parameters and data update conditions of the first algorithm and the second algorithm based on error data can be refined as follows:

[0082] Based on the error data, determine the resource replacement rate and error factors; adjust the algorithm parameters of the first algorithm and the second algorithm based on the resource replacement rate and the preset threshold range; update the data update conditions based on the error factors.

[0083] In this embodiment, the resource replacement rate can be understood as the result of dividing the number of replaced resources by the number of candidate resources. Error factors can be understood as the reasons for errors leading to resource differences. The preset threshold range can be understood as a judgment threshold set for adjusting different parameters of the algorithm.

[0084] Specifically, the processor can determine the resource replacement rate by dividing the number of replaced resources (i.e., the number of resources replaced by the target resource combination in the candidate resource combination) in the error data by the total number of candidate resources (i.e., the total number of resources in the candidate resource combination). The processor can also identify error factors through the replaced resources in the error data. Furthermore, the processor can compare the resource replacement rate with a preset threshold range to determine the applicable threshold range and the corresponding algorithm parameters to be adjusted, thereby adjusting the algorithm parameters of both the first and second algorithms. Finally, the processor can update the data update conditions based on the error factors.

[0085] For example, when the resource replacement rate exceeds 20%, the weight of the implicit parameter of P value (stability and network quality) in the first algorithm can be increased (from 30% to 40%); when the resource replacement rate is ≤5% for 3 consecutive times, the second algorithm is changed from full verification to 20% sampling verification; if the error is caused by the delay of price data, the update frequency of popular resources in the data update conditions is shortened from 2 minutes to 1 minute, and system loss is reduced through dynamic optimization.

[0086] Furthermore, the second algorithm simultaneously generates a risk decision report when the resource removal mechanism is triggered. This report includes: a chain of implicit state data evidence for the removed resource, a health comparison matrix of the replacement resource, and a cost change analysis resulting from risk avoidance. The risk decision report includes a chain of evidence with the removed resource's failure timestamps for the past 7 days, a load fluctuation curve, and a screenshot of packet loss rate monitoring. The health comparison matrix horizontally compares five indicators, including failure rate, load fluctuation, and packet loss rate, indicating the percentage by which the replacement resource outperforms the removed resource in each indicator (e.g., a 40% reduction in failure rate). The cost change analysis quantifies the increase in basic costs after replacement (e.g., 0.3 million yuan) and the expected reduction in failure losses (e.g., 2 million yuan), clearly defining the net benefit of risk avoidance and providing users with a basis for decision-making.

[0087] The technical solution of this invention extracts the core elements of natural language requirements and calibrates fuzzy expressions by combining resource performance vectors from a historical performance database, generating zero-bias structured computing power task instructions to solve the problem of misinterpretation of requirements. A first algorithm performs global optimization to achieve explicit constraints, determining candidate resource combinations. A second algorithm verifies implicit constraints, eliminating high-risk resources to obtain the target resource combination, achieving dual verification of explicit and implicit constraints. First, a reference price is generated based on the candidate combinations from the first algorithm to meet user expectations. Then, the second algorithm optimizes the price and combines it with an auction mechanism to generate the exact price, simultaneously outputting details to balance efficiency and transparency. After determining the target resource combination, the parameters of the first algorithm are adjusted based on the matching error, reducing the frequency of intervention by the second algorithm and optimizing the resource data update frequency, enabling the system to self-iterate and improve long-term matching accuracy and efficiency.

[0088] Example 3

[0089] Figure 3 This is a schematic diagram of a computing resource determination device provided in Embodiment 3 of the present invention. Figure 3 As shown, the device includes:

[0090] The demand acquisition module 31 is used to acquire real-time computing power resource data of natural language demand and resource library;

[0091] The intent parsing module 32 is used to parse the natural language requirements based on the historical performance database to determine the computing power task instructions;

[0092] The resource determination module 33 is used to perform dual-algorithm collaborative matching on the computing power task instructions based on the real-time computing power resource data, and determine the target resource combination that meets the requirements from the resource library.

[0093] The technical solution of this invention involves acquiring real-time computing power resource data from natural language requirements and a resource library; performing intent parsing on natural language requirements based on a historical fulfillment database to determine computing power task instructions; and using a dual-algorithm collaborative matching process on the computing power task instructions based on real-time computing power resource data to determine a target resource combination that meets the requirements from the resource library. By combining the historical fulfillment database with intent parsing of natural language requirements, the ambiguity of intent analysis is eliminated at the source. The dual-algorithm collaborative matching reduces the matching deviation of computing power resources, achieving accurate parsing of requirements and thus accurate matching of computing power resources.

[0094] Furthermore, the intent parsing module 32 is specifically used for:

[0095] Extract core computing power-related elements from the natural language requirements;

[0096] The fuzzy expressions in the natural language requirements are vectorized to obtain fuzzy expression vectors;

[0097] The fuzzy representation vector is matched based on the historical performance database to obtain a calibrated representation;

[0098] Based on the calibration description and the core elements related to computing power, the computing power task instructions are determined.

[0099] Furthermore, the resource determination module 33 is specifically used for:

[0100] By combining the real-time computing power resource data with the first algorithm, candidate resource combinations are determined from the resource library;

[0101] The second algorithm is used to eliminate high-risk resource combinations from the candidate resource combinations to obtain the target resource combination.

[0102] Furthermore, the device also includes:

[0103] The error determination module is used to determine the matching difference between the candidate resource combination and the target resource combination as error data after the dual-algorithm collaborative matching of the computing power task instruction based on the real-time computing power resource data and the resource library.

[0104] The error update module is used to update the algorithm parameters and data update conditions of the first algorithm and the second algorithm based on the error data.

[0105] Specifically, the error update module is used for:

[0106] Based on the error data, the resource replacement rate and error factors are determined;

[0107] Based on the resource replacement rate and the preset threshold range, the algorithm parameters of the first algorithm and the second algorithm are adjusted.

[0108] The data update conditions are updated based on the aforementioned error factors.

[0109] Optionally, the device further includes:

[0110] The data update module is used to obtain the explicit performance data and implicit status data of each computing resource in the resource library when the data update conditions are met, so as to obtain real-time computing resource data.

[0111] The computing resource determination device provided in this embodiment of the invention can execute the computing resource determination method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0112] Example 4

[0113] Figure 4 A schematic diagram of an electronic device 40 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0114] like Figure 4 As shown, the electronic device 40 includes at least one processor 41 and a memory, such as a read-only memory (ROM) 42 or a random access memory (RAM) 43, communicatively connected to the at least one processor 41. The memory stores computer programs executable by the at least one processor. The processor 41 can perform various appropriate actions and processes based on the computer program stored in the ROM 42 or loaded from storage unit 48 into the RAM 43. The RAM 43 may also store various programs and data required for the operation of the electronic device 40. The processor 41, ROM 42, and RAM 43 are interconnected via a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.

[0115] Multiple components in electronic device 40 are connected to I / O interface 45, including: input unit 46, such as keyboard, mouse, etc.; output unit 47, such as various types of monitors, speakers, etc.; storage unit 48, such as disk, optical disk, etc.; and communication unit 49, such as network card, modem, wireless transceiver, etc. Communication unit 49 allows electronic device 40 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0116] Processor 41 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 41 performs the various methods and processes described above, such as methods for determining computing resources.

[0117] In some embodiments, the computing resource determination method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 40 via ROM 42 and / or communication unit 49. When the computer program is loaded into RAM 43 and executed by processor 41, one or more steps of the computing resource determination method described above may be performed. Alternatively, in other embodiments, processor 41 may be configured to perform the computing resource determination method by any other suitable means (e.g., by means of firmware).

[0118] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0119] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0120] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0121] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0122] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0123] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0124] In one embodiment, the present invention further includes a computer program product, which includes a computer program that, when executed by a processor, implements the computing resource determination method of any embodiment of the present invention.

[0125] In implementing the computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0126] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0127] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A computing power resource determination method, characterized in that, The method comprises the following steps: acquiring natural language demand and real-time computing resource data of a resource library; performing intention analysis on the natural language demand based on a historical performance database to determine a computing task instruction; based on the real-time computing resource data, performing double-algorithm collaborative matching on the computing task instruction to determine a target resource combination from the resource library that meets the demand.

2. The method of claim 1, wherein, The intention analysis on the natural language demand based on the historical performance database to determine the computing task instruction comprises the following steps: extracting computing-related core elements from the natural language demand; vectorizing fuzzy expressions in the natural language demand to obtain a fuzzy expression vector; matching the fuzzy expression vector based on the historical performance database to obtain a calibrated expression; determining a computing task instruction according to the calibrated expression and the computing-related core elements.

3. The method of claim 1, wherein, The double-algorithm collaborative matching on the computing task instruction based on the real-time computing resource data to determine a target resource combination from the resource library that meets the demand comprises the following steps: determining a candidate resource combination from the resource library by combining the real-time computing resource data with a first algorithm; based on a second algorithm, eliminating a high-risk resource combination from the candidate resource combination to obtain a target resource combination.

4. The method of claim 3, wherein, After the double-algorithm collaborative matching on the computing task instruction based on the real-time computing resource data to determine a target resource combination from the resource library that meets the demand, the method further comprises the following steps: taking a matching difference between the candidate resource combination and the target resource combination as error data; based on the error data, updating algorithm parameters and data update conditions of the first algorithm and the second algorithm.

5. The method of claim 4, wherein, The updating of the algorithm parameters and the data update conditions of the first algorithm and the second algorithm based on the error data comprises the following steps: determining a resource replacement rate and an error factor according to the error data; adjusting algorithm parameters of the first algorithm and the second algorithm according to the resource replacement rate and a preset threshold range; updating data update conditions according to the error factor.

6. The method of claim 1, wherein, The method further comprises the following steps: when the data update condition is met, acquiring explicit performance data and implicit state data of each computing resource in the resource library to obtain real-time computing resource data.

7. A computing resource determination apparatus, characterized by comprising: The method comprises the following steps: a demand acquisition module configured to acquire natural language demand and real-time computing resource data of a resource library; an intention analysis module configured to perform intention analysis on the natural language demand based on a historical performance database to determine a computing task instruction; a resource determination module configured to, based on the real-time computing resource data, perform double-algorithm collaborative matching on the computing task instruction to determine a target resource combination from the resource library that meets the demand.

8. An electronic device, comprising: The electronic device comprises: at least one processor; and a memory connected in communication with the at least one processor; wherein the memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the computing resource determination method of any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for causing the processor to implement the computing resource determination method in any one of claims 1-6 when executed.

10. A computer program product, characterised in that, The computer program product comprises a computer program which, when executed by the processor, implements the computing resource determination method according to any one of claims 1-6.

Images