Intelligent computing power scheduling method and system based on data analysis

Abstract

The invention belongs to the technical field of data analysis, and particularly relates to an intelligent computing power scheduling method and system based on data analysis, and the method comprises the steps: determining a reasonable computing power demand through dynamic analysis of a plurality of computing power requests and parameters thereof, and selecting an optimal computing power resource for scheduling. The method comprises the following specific steps: receiving a plurality of computing power requests and parameters, and analyzing and determining a preliminary computing power demand; selecting a first candidate computing power resource according to the preliminary demand; adjusting candidate resources to obtain a second candidate computing power resource set in combination with the second computing power request and a parameter updating demand; calculating an optimal allocation scheme by using the set, and generating a scheduling instruction; and executing the scheduling instruction and monitoring the effect, and optimizing the subsequent request processing logic according to feedback. According to the method, the computing power resource utilization rate is improved, and the response speed and service quality of the system are enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of data analysis, and specifically relates to an intelligent computing power scheduling method and system based on data analysis. Background Art

[0002] With the development of cloud computing and big data technologies, data centers are facing an increasing demand for computing power. Traditional computing power scheduling methods are often based on static resource allocation strategies and lack an effective response mechanism to dynamically changing computing power demands. This fixed resource allocation method may lead to low resource utilization and failure to respond to users' computing power requests in a timely manner, thus affecting service quality.

[0003] In the prior art, users' computing power requests are usually processed by using predetermined rules or simple queuing mechanisms. Although this method is simple and easy to implement, it is inadequate when faced with complex and ever-changing application scenarios. For example, when multiple users submit computing power requests at the same time, how to reasonably allocate limited computing power resources becomes an urgent problem to be solved.

[0004] Existing computing power scheduling methods are difficult to adapt to rapidly changing computing power requirements, especially in the case of concurrent requests from multiple users, which may lead to uneven resource allocation or even waste. In addition, due to the lack of an effective evaluation mechanism for scheduling results, it is difficult to adjust the scheduling strategy according to actual conditions, thus failing to ensure efficient resource utilization and service quality. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent computing power scheduling method and system based on data analysis, which not only improves the utilization rate of computing power resources, but also continuously optimizes the subsequent request processing logic according to the scheduling result feedback, thereby significantly improving the system's response speed and service quality, so as to solve the problems raised in the above background technology.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent computing power scheduling method based on data analysis, comprising:

[0007] Step 1: receiving multiple computing power requests and corresponding parameters, and analyzing and determining the preliminary computing power requirements based on the received first computing power request and parameters;

[0008] Step 2: Select at least a portion of computing power resources from a preset computing power pool as first candidate computing power resources according to the preliminary computing power demand;

[0009] Step 3: Based on the first candidate computing power resources, the second computing power request and the parameter update computing power requirement are combined to form a second computing power requirement, and the first candidate computing power resources are adjusted based on the updated computing power requirement to obtain a second candidate computing power resource set;

[0010] Step 4: Using the second candidate computing power resource set, calculate the optimal computing power allocation plan that matches the third computing power request and parameters, allocate the second candidate computing power resource set according to the optimal computing power allocation plan, and generate an actual scheduling instruction;

[0011] Step 5: Execute the actual scheduling instruction, monitor the execution effect to evaluate the scheduling result, and optimize the subsequent computing power request processing logic based on the scheduling result feedback.

[0012] Preferably, receiving multiple computing power requests and corresponding parameters includes:

[0013] For multiple computing power requests and corresponding parameters received, the formula is used:

[0014] F(x)=w1·x1+w2·x2+...+w n ·x n Calculate the weight value for each request, where x i Indicates the parameters of the i-th request, w i is the weight coefficient of the corresponding parameter; according to the calculated weight value F(x), the computing power requests are sorted to form a priority list;

[0015] The computing power requests are processed in sequence according to the priority list, and the remaining available computing power resources are updated after each request is processed.

[0016] Preferably, the analysis and determination of the preliminary computing power requirements include:

[0017] For the first computing power request and its parameters received, apply the formula To quantify the computing power requirements, where p i Represents the priority of the i-th parameter, c i is the corresponding computing power consumption value;

[0018] Based on the quantified computing power requirement D, identify a set of computing power resources that meet the conditions, and according to the computing power requirement D, select the computing power resources that meet or are closest to the requirements from the identified computing power resource set as the initial computing power configuration.

[0019] Preferably, selecting at least a portion of computing power resources from a preset computing power pool as the first candidate computing power resources includes:

[0020] According to the preliminary computing power demand D, a threshold T is set to filter the resources in the preset computing power pool, where T is the maximum acceptable computing power gap;

[0021] Calculate the computing resources R in the preset computing pool i The gap Δ between the initial computing power requirement D i =|R i -D|; Select all Δi ≤T computing resources R i , forming the first candidate computing power resource set.

[0022] Preferably, forming the second computing power requirement in combination with the second computing power request and the parameter update computing power requirement includes:

[0023] Obtain the parameter set P2 of the second computing power request, and apply the formula D′2=f(P2) to evaluate the computing power requirement D′2 corresponding to the request, where f is the evaluation function;

[0024] The total computing power C of the first candidate computing power resource set total Compare with the initial computing power requirement D and the newly evaluated requirement D′2, and update the computing power requirement to D2=D+D′2, ensuring that D2≤C total ;

[0025] If D2>C total , then adjust D2 to C proportionally total , forming the final second computing power requirement.

[0026] Preferably, adjusting the first candidate computing power resources based on the updated computing power requirement to obtain a second candidate computing power resource set includes:

[0027] Based on the updated computing power requirement D2, re-evaluate the applicability of each resource in the first candidate computing power resource set;

[0028] For each resource R in the first candidate computing resource set i , calculate its matching degree with D2 Where R i ≥D2; filter out all matching degree M i Resources close to 1 form the second candidate computing power resource set.

[0029] Preferably, calculating an optimal computing power allocation scheme that matches the third computing power request and parameters includes:

[0030] Obtain the parameter set P3 of the third computing power request, and calculate the computing power requirement D3 corresponding to the request;

[0031] For each resource R in the second candidate computing resource set i , apply formula A i =min(D3, R i ) to determine the maximum computing power A that can be allocated to the third computing power request i ;

[0032] Compare each A i value, choose so that A i The computing resources closest to D3 are used as part of the optimal allocation plan.

[0033] Preferably, allocating the second candidate computing resource set to generate an actual scheduling instruction includes:

[0034] Determine the resource R in the optimal computing power allocation plan i And its corresponding distribution amount A i , where A i By R i The provided computing power is balanced with the demand of the third computing power request D3;

[0035] Construct scheduling instruction L i , which contains the resource ID i and the distribution amount A i , that is, L i =(ID i , A i ); According to the scheduling instruction L i , allocate the resource R i The computing power A i Request to the third computing power;

[0036] Record the allocated computing power A i , update the remaining computing resources R′ i =R i -A i , to reflect resource usage;

[0037] Repeat the above steps until all resources in the second candidate computing resource set are properly allocated, generating a complete scheduling instruction set I = {I1, I2, ..., I n}, used for actual computing power scheduling.

[0038] Preferably, executing the actual scheduling instruction, monitoring the execution effect to evaluate the scheduling result, and optimizing the subsequent computing power request processing logic according to the scheduling result feedback include:

[0039] Execute the scheduling instruction set I, implement computing resource allocation, and monitor the performance index P during the execution process;

[0040] According to the performance index P, the effectiveness E of the scheduling result is evaluated: E = g(P), where g is the evaluation function;

[0041] According to the effectiveness E of the scheduling result, adjust the parameters in the computing power request processing logic. After updating the parameters, recalculate the processing logic of subsequent computing power requests to form a new computing power demand assessment and resource allocation plan, and apply the optimized processing logic to new computing power requests.

[0042] On the other hand, the present invention proposes an intelligent computing power scheduling system based on data analysis, comprising:

[0043] A computing power request receiving and preliminary demand analysis module, used to receive multiple computing power requests and corresponding parameters, and analyze and determine preliminary computing power requirements based on the received first computing power request and parameters;

[0044] A preliminary computing power resource configuration module, configured to select at least a portion of computing power resources from a preset computing power pool as first candidate computing power resources according to the preliminary computing power demand;

[0045] A computing power requirement update and resource adjustment module, configured to form a second computing power requirement based on the first candidate computing power resource, in combination with the second computing power request and the parameter update computing power requirement, and adjust the first candidate computing power resource based on the updated computing power requirement to obtain a second candidate computing power resource set;

[0046] an optimal computing power allocation and scheduling module, configured to use the second candidate computing power resource set to calculate an optimal computing power allocation scheme that matches the third computing power request and parameters, allocate the second candidate computing power resource set according to the optimal computing power allocation scheme, and generate an actual scheduling instruction;

[0047] The scheduling execution and feedback optimization module is used to execute the actual scheduling instructions, monitor the execution effect to evaluate the scheduling results, and optimize the subsequent computing power request processing logic based on the scheduling result feedback.

[0048] Technical effects and advantages of the present invention: Compared with the prior art, the intelligent computing power scheduling method and system based on data analysis proposed by the present invention has the following advantages:

[0049] The present invention dynamically analyzes multiple received computing power requests and their parameters to determine reasonable computing power requirements, and on this basis selects the optimal computing power resources for scheduling. This method not only improves the utilization rate of computing power resources, but also continuously optimizes the subsequent request processing logic based on the scheduling result feedback, thereby significantly improving the system's response speed and service quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a flow chart of the intelligent computing power scheduling method based on data analysis of the present invention;

[0051] Figure 2 This is a block diagram of the intelligent computing power scheduling system based on data analysis of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. The specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0053] The present invention provides an intelligent computing power scheduling method based on data analysis, which aims to determine reasonable computing power requirements by dynamically analyzing multiple computing power requests and their parameters, and select the optimal computing power resources for scheduling. The specific steps include: receiving multiple computing power requests and parameters, analyzing and determining the preliminary computing power requirements; selecting a first candidate computing power resource based on the preliminary requirements; adjusting the candidate resources to obtain a second candidate computing power resource set in combination with the second computing power request and parameter update requirements; using the set to calculate the optimal allocation plan and generate scheduling instructions; executing the scheduling instructions and monitoring the effects, and optimizing the subsequent request processing logic based on feedback. This method improves the utilization rate of computing power resources and enhances the response speed and service quality of the system. Specifically as follows:

[0054] like Figure 1 As shown, in this embodiment, a smart computing power scheduling method based on data analysis includes the following steps:

[0055] Step 1: receiving multiple computing power requests and corresponding parameters, and analyzing and determining the preliminary computing power requirements based on the received first computing power request and parameters;

[0056] Further, receiving multiple computing power requests and corresponding parameters includes: the system first receives multiple computing power requests from different users, each request is accompanied by a corresponding parameter set P i , these parameters include processing power, storage requirements, network bandwidth, etc.

[0057] For multiple computing power requests and corresponding parameters received, the formula is used:

[0058] F(x)=w1·x1+w2·x2+...+w n ·x n Calculate the weight value for each request, where x i Indicates the parameters of the i-th request, w i is the weight coefficient of the corresponding parameter; according to the calculated weight value F(x), the computing power requests are sorted to form a priority list; the weight coefficient can be pre-set according to the actual application scenario, and all computing power requests are sorted from high to low according to the weight value to form a priority list.

[0059] According to the priority list, computing power requests are processed in sequence, and the remaining available computing power resources are updated after each request is processed. By introducing weight value calculation and priority sorting, this embodiment can more intelligently determine the processing order when receiving multiple computing power requests. This not only helps to improve the fairness and efficiency of resource allocation, but also ensures that high-priority requests are responded to in a timely manner. Specifically:

[0060] Improved resource utilization: By giving priority to requests with higher weights, limited computing resources can be used more effectively and idle resources can be reduced.

[0061] Faster response time: High-priority requests are processed first, allowing important tasks to be responded to more quickly and improving the overall system response speed.

[0062] Improved service quality: Dynamically adjusting the processing order based on the importance and urgency of requests can better meet the diverse needs of users and thus improve service quality.

[0063] Furthermore, the initial computing power requirements are analyzed and determined to include:

[0064] For the first computing power request and its parameters received, apply the formula To quantify the computing power requirements, where p i Represents the priority of the i-th parameter, c i is the corresponding computing power consumption value; for example, if the processing power has a higher priority, then p i The weight of i Indicates the computing power consumption per unit of processing power.

[0065] Based on the quantified computing power requirement D, identify a set of computing power resources that meet the conditions, and according to the computing power requirement D, select the computing power resources that meet or are closest to the requirements from the identified computing power resource set as the initial computing power configuration.

[0066] Assuming p1=0.5, p2=0.3, p3=0.2, and c1=10, c2=5, c3=2, the computing power requirement D can be calculated as: D1=0.5·10+0.3·5+0.2·2=5+1.5+0.4=6.9.

[0067] Identify a set of computing resources that meet the conditions: Based on the quantified computing power demand D, the system identifies a set of computing power resources from the preset computing power pool that can meet or approach D. Assume that there are resources RA, RB, and RC in the computing power pool, and their computing powers are 6, 7, and 5 respectively.

[0068] Filter the initial computing power configuration: Filter the computing power resources closest to the requirement D from the above computing power resource set as the initial computing power configuration. In this example, RB's computing power is 7, which is closest to D, so it is used as the initial computing power configuration.

[0069] Step 2: Select at least a portion of computing power resources from a preset computing power pool as first candidate computing power resources according to the preliminary computing power demand;

[0070] Further, selecting at least a portion of computing power resources from the preset computing power pool as the first candidate computing power resources includes:

[0071] According to the preliminary computing power demand D, a threshold T is set to filter the resources in the preset computing power pool, where T is the maximum acceptable computing power gap; for example, assuming T=1, it means that the system only selects those resources whose computing power gap does not exceed 1.

[0072] Calculate the computing resources R in the preset computing pool i The gap Δ between the initial computing power requirement D i =|R i -D|; Select all Δ i ≤T computing resources R i , forming the first candidate computing power resource set.

[0073] Assume that there are three resources RA, RB and RC in the computing power pool, and their computing power is 6, 7 and 5 respectively. The initial computing power requirement D has been calculated to be 6.9.

[0074] For RA, the gap ΔA = |6-6.9| = 0.9

[0075] For RB, the gap ΔB = |7-6.9| = 0.1

[0076] For RC, the gap ΔC = |5-6.9| = 1.9.

[0077] Assuming T = 1, RA and RB are selected because their differences are less than or equal to 1. Therefore, the first candidate computing resource set is {RA, RB}.

[0078] Setting the threshold T allows the system to adjust the screening criteria according to actual conditions, thereby enhancing the flexibility of the scheduling strategy. For example, when the computing power demand fluctuates greatly, the size of T can be adjusted to adapt to different demand conditions. By quickly screening out qualified resources, the system can generate a set of candidate resources more quickly, thereby accelerating the scheduling process and improving the system response speed. Through precise screening and efficient scheduling, the system can better meet user needs, improve service quality, and thus enhance user experience.

[0079] Step 3: Based on the first candidate computing power resources, the second computing power request and the parameter update computing power requirement are combined to form a second computing power requirement, and the first candidate computing power resources are adjusted based on the updated computing power requirement to obtain a second candidate computing power resource set;

[0080] Further, the second computing power requirement is formed by combining the second computing power request and the parameter update computing power requirement, including:

[0081] Obtaining the parameter set P2 of the second computing power request is also achieved by receiving the computing power request information of the user. This process involves receiving the user's request and extracting necessary parameters therefrom.

[0082] The specific steps are as follows:

[0083] The user submits a computing power request, which contains the parameter information required to perform a specific task.

[0084] The system receives this computing power request and parses the required parameters from the request, which constitute parameter set P2.

[0085] For example, if the user's request is to perform a data processing task, then P2 includes the following parameters: the number of CPU cores or processing power required, the necessary memory size, storage space requirements, network bandwidth requirements, task priority, and estimated running time.

[0086] When the system receives this computing power request, it will parse the relevant information in the request and organize this information into parameter set P2 for subsequent computing power demand assessment and resource allocation. This means that P2 is directly extracted from the user's computing power request as the basis for further processing by the system.

[0087] After parameter set P2, the formula D′2=f(P2) is applied to evaluate the computing power requirement D′2 corresponding to the request, where f is the evaluation function; the total computing power C of the first candidate computing power resource set is total Compare with the initial computing power requirement D and the newly evaluated requirement D′2, and update the computing power requirement to D2=D+D′2, ensuring that D2≤C total ; If D2>C total , then adjust D2 to C proportionally total , forming the final second computing power requirement.

[0088] For example:

[0089] Obtaining the parameter set of the second computing power request: The system receives the second computing power request and its parameter set P2. Assume that P2 includes parameters such as processing capacity p1, storage requirement p2, and network bandwidth p3.

[0090] Apply the formula to evaluate the computing power demand: Apply the evaluation function f to evaluate the computing power demand D′2 corresponding to the second computing power request. Assume that the evaluation function is:

[0091] f(P2)=p1·c1+p2·c2+p3·c3, where c1, c2, and c3 are the corresponding computing power consumption values ​​respectively.

[0092] For example, assuming p1 = 0.6, p2 = 0.2, p3 = 0.2, and c1 = 8, c2 = 4, c3 = 3, the computing power requirement D′2 can be calculated as:

[0093] D'2=0.6·8+0.2·4+0.2·3=4.8+0.8+0.6=6.2.

[0094] Compare and update computing power requirements: Set the total computing power C of the first candidate computing power resource set total Compare with the initial computing power requirement D and the newly evaluated requirement D′2. Assume that the total computing power of the first candidate computing power resource set is C total is 15, the initial computing power requirement D is 6.9, and the newly assessed demand D′2 is 6.2, then the total demand D2 is D+D′2=6.9+6.2=13.1.

[0095] Update computing power requirements: Ensure that the updated computing power requirement D2 does not exceed the total computing power C of the first candidate computing power resource set total In this case, D2 = 13.1, which is less than C total =15, so no adjustment is needed, and D2 is directly used as the second computing power requirement.

[0096] Adjust proportionally: If D2>C total , then adjust D2 to C proportionally total .

[0097] For example, if D2 = 18, which is greater than C total =15, then adjust D2 proportionally to D2=(C total / D2)·D2=15 / 18, forming the final second computing power requirement.

[0098] By comparing the total computing power of the candidate computing resource set with the computing power requirements and adjusting according to the actual situation, it is ensured that resource allocation meets the requirements without exceeding resource limits, thereby improving the flexibility and rationality of resource utilization. By adjusting the computing power requirements in proportion, the system can allocate resources more reasonably and avoid insufficient resources due to excessive demand, thereby optimizing resource allocation. By dynamically updating the computing power requirements, the system can respond more quickly to ensure efficient use of resources and service quality. By accurately evaluating and reasonably adjusting the computing power requirements, the system can better meet user needs and improve service quality and user experience.

[0099] Furthermore, the first candidate computing resources are adjusted based on the updated computing power requirement to obtain a second candidate computing power resource set, including:

[0100] Based on the updated computing power requirement D2, re-evaluate the applicability of each resource in the first candidate computing power resource set; for each resource R in the first candidate computing power resource set i , calculate its matching degree with D2 Where R i ≥D2; filter out all matching degree M i Resources close to 1 form the second candidate computing power resource set.

[0101] For example:

[0102] Assume that after the processing of step 2, the updated computing power requirement D2 is 13.1 (if D2>Ctotal, it is adjusted proportionally to C total ).

[0103] Based on the updated computing power requirement D2, the applicability of each resource in the first candidate computing power resource set is re-evaluated. Assume that the first candidate computing power resource set includes resources RA, RB, and RC, whose computing powers are 6, 7, and 5 respectively.

[0104] For each resource in the first candidate computing resource set, calculate its matching degree M with the updated computing power requirement D2 i . Matching degree M i Defined as resource R i The ratio of computing power to computing power requirement D2, i.e. Mi = D2 / R i (Note that the expression here is slightly adjusted for consistency). For each resource R i , matching degree M i Should satisfy R i ≥D2.

[0105] For RA, the matching degree MA = 13.1 / 6 ≈ 2.18 (does not meet the conditions)

[0106] For RB, the matching degree MB = 13.1 / 7≈1.87 (does not meet the conditions)

[0107] For RC, the matching degree MC = 13.1 / 5≈2.62 (does not meet the conditions)

[0108] Since the computing power of RA, RB, and RC are all less than the updated computing power requirement D2, suppose we need to find a resource with a computing power greater than or equal to D2. Suppose there are other resources in the computing power pool, such as RD and RE, whose computing power is 15 and 14 respectively.

[0109] For RD, the matching degree MD = 13.1 / 15≈0.87 (does not meet the conditions)

[0110] For RE, the matching degree ME = 13.1 / 14 ≈ 0.94 (close to 1)

[0111] All resources with a matching degree close to 1 are screened out to form a second set of candidate computing resources. In this example, the matching degree of RE is close to 1, so it is included as part of the second set of candidate computing resources.

[0112] By calculating the matching degree, the system can more accurately screen out resources that are suitable for the updated computing power requirements and ensure efficient use of resources. By selecting resources with a matching degree close to 1, the system can better meet the updated computing power requirements and avoid over-allocation or under-allocation of resources, thereby improving resource utilization. By re-evaluating the applicability and calculating the matching degree, the system can flexibly adjust resource selection according to the updated computing power requirements to ensure flexibility and adaptability of resource allocation. By quickly screening out resources with a high matching degree, the system can generate a second set of candidate computing power resources more quickly, thereby accelerating the scheduling process and improving system response speed.

[0113] Step 4: Using the second candidate computing power resource set, calculate the optimal computing power allocation plan that matches the third computing power request and parameters, allocate the second candidate computing power resource set according to the optimal computing power allocation plan, and generate an actual scheduling instruction;

[0114] Further, calculating the optimal computing power allocation scheme that matches the third computing power request and parameters includes:

[0115] Obtaining the parameter set P3 of the third computing power request, obtaining the parameter set P3 of the third computing power request is achieved by receiving the request information of the user. The process here involves receiving the computing power request submitted by the user and extracting necessary parameters therefrom.

[0116] The specific steps are as follows:

[0117] The user submits a computing power request, which contains the parameter information required to perform a specific task.

[0118] The system receives this request and parses the required parameters from the request, which constitute parameter set P3.

[0119] For example, if the user's request is to execute a data processing task, then P3 includes the following parameters: processing speed requirements (such as the amount of data processed per second), the required storage space size, the priority of the task, and the estimated running time.

[0120] When the system receives this request, it will parse the relevant information in the request and organize the information into parameter set P3 for subsequent computing power demand assessment and resource allocation. This means that P3 is directly extracted from the user's computing power request as the basis for further processing by the system.

[0121] After obtaining P3, calculate the computing power requirement D3 corresponding to the request; for each resource R in the second candidate computing power resource set i , apply formula A i =min(D3,R i ) to determine the maximum computing power A that can be allocated to the third computing power request i ; Compare each A i value, choose so that A i The computing resources closest to D3 are used as part of the optimal allocation plan.

[0122] For example:

[0123] The system receives a third computing power request and its parameter set P3. Assume that P3 includes parameters such as processing capacity p1, storage requirement p2, and network bandwidth p3.

[0124] Calculate computing power requirements: Apply the evaluation function f to evaluate the computing power requirement D3 corresponding to the third computing power request.

[0125] Assume the evaluation function is:

[0126] f(P3)=p1·c1+p2·c2+p3·c3, where c1, c2, c3 are the corresponding computing power consumption values. For example, assuming p1=0.5, p2=0.3, p3=0.2, and c1=10, c2=5, c3=2, the computing power requirement D3 can be calculated as:

[0127] D3=0.5·10+0.3·5+0.2·2=5+1.5+0.4=6.9.

[0128] For each resource R in the second candidate computing resource set i , apply formula A i =min(D3,R i ) to determine the maximum computing power A that can be allocated to the third computing power request i Assume that the second candidate computing power resource set includes resources RA, RB, and RC, whose computing powers are 15, 14, and 13 respectively.

[0129] For RA, the maximum hashrate is AA = 6.9

[0130] For RB, the maximum hashrate AB==6.9

[0131] For RC, the maximum hashrate AC==6.9

[0132] Compare each A i value, choose so that A i The computing resources closest to D3 are used as part of the optimal allocation solution. In this example, all A i The values ​​are both 6.9, so any one of the resources can be selected as part of the optimal allocation solution.

[0133] By calculating the maximum computing power that each resource can allocate to the third computing power request, the system can allocate resources more accurately and ensure efficient use of resources. i The resources closest to D3, the system can meet the needs of the third computing power request to the greatest extent and avoid waste of resources.

[0134] Further, the second candidate computing power resource set is allocated to generate an actual scheduling instruction, including: determining the resource R in the optimal computing power allocation solution i And its corresponding distribution amount A i , where A i By R i The provided computing power is balanced with the demand of the third computing power request D3; construct the scheduling instruction L i , which contains the resource ID i and distribution amount A i , that is, L i =(ID i , A i );According to the scheduling instruction L i , allocate the resource R i The computing power A i Request to the third computing power; record the allocated computing power A i , update the remaining computing resources R′ i =R i -A i , to reflect resource usage;

[0135] Repeat the above steps until all resources in the second candidate computing resource set are properly allocated, generating a complete scheduling instruction set I = {I1, I2, ..., I n}, used for actual computing power scheduling.

[0136] By constructing scheduling instructions that include resource identifiers and allocation amounts, the system can accurately control resource allocation and ensure that each computing power request is supported by appropriate resources. By recording the allocated computing power and updating the remaining computing power resources, the system can track resource usage in real time, avoid over-allocation or idle resources, and improve resource utilization. By generating a complete set of scheduling instructions, the system can automatically perform scheduling tasks, reduce manual intervention, and improve the degree of automation of the scheduling process. By quickly generating and executing scheduling instructions, the system can respond to computing power requests more quickly, improving the system's response speed and service quality. By dynamically updating resource status, the system can better maintain the stability and reliability of resource use and avoid service interruptions caused by improper resource allocation.

[0137] Step 5: Execute the actual scheduling instruction, monitor the execution effect to evaluate the scheduling result, and optimize the subsequent computing power request processing logic based on the scheduling result feedback. The details are as follows:

[0138] Execute the actual scheduling instruction, monitor the execution effect to evaluate the scheduling result, and optimize the subsequent computing power request processing logic according to the scheduling result feedback, including: execute the scheduling instruction set I, implement computing power resource allocation, and monitor the performance index P during the execution process; according to the performance index P, evaluate the effectiveness E of the scheduling result: E=g(P), where g is the evaluation function;

[0139] According to the effectiveness E of the scheduling result, adjust the parameters in the computing power request processing logic. After updating the parameters, recalculate the processing logic of subsequent computing power requests to form a new computing power demand assessment and resource allocation plan, and apply the optimized processing logic to new computing power requests.

[0140] For example:

[0141] According to the generated scheduling instruction set I = {IA, IB, IC}, computing power resource allocation is performed. For example, IA = (IDA, 6.9) I, IB = (IDB, 6.9), IC = (IDC, 6.9) is executed to allocate the specified computing power to the corresponding computing power request.

[0142] After computing resources are allocated, the performance indicators P during the execution process are monitored. These indicators may include task response time, resource utilization, task completion time, etc. Assume that the monitored performance indicators are P = {p1, p2, p3}, where p1 represents task response time, p2 represents resource utilization, and p3 represents task completion time.

[0143] The effectiveness of the scheduling result is evaluated using the evaluation function E=g(P) according to the performance index P. Assuming that the evaluation function g is a comprehensive evaluation function, the effectiveness of the scheduling result is measured according to the weighted average of various performance indicators.

[0144] For example, g(P)=w1·p1+w2·p2+w3·p3, where w1, w2, w3 are weight coefficients of various performance indicators.

[0145] If E is higher, it means the scheduling result is valid.

[0146] If E is low, it means that the scheduling results need to be optimized.

[0147] Adjust the parameters in the computing power request processing logic according to the effectiveness of the scheduling result E. For example, if E is low, you may need to adjust the weight coefficients w1, w2, w3, or reset the threshold T.

[0148] Recalculate the processing logic of subsequent computing power requests: After updating the parameters, recalculate the processing logic of subsequent computing power requests to form a new computing power demand evaluation and resource allocation plan. For example, recalculate the new computing power demand D3′ and the matching degree M i ′, generate a new scheduling instruction set. Apply the optimized processing logic to new computing power requests to ensure more reasonable resource allocation and improve the performance of the overall system.

[0149] By monitoring the performance indicators during the execution process in real time and evaluating the effectiveness of the scheduling results based on these indicators, the system can promptly discover and solve problems and ensure the quality of the scheduling results. Based on the feedback of the scheduling results, the system can dynamically adjust the parameters in the computing power request processing logic, thereby continuously optimizing the scheduling strategy and improving the rationality and efficiency of resource allocation. Through continuous optimization of the scheduling results, the system can better utilize computing power resources, avoid resource waste, and improve the overall utilization of resources. Through continuous feedback and optimization, the system can better adapt to the ever-changing computing power requirements and improve the flexibility and adaptability of the system.

[0150] On the other hand, the present invention proposes an intelligent computing power scheduling system based on data analysis, such as Figure 2 As shown, it includes: a computing power request receiving and preliminary demand analysis module, a preliminary computing power resource configuration module, a computing power demand update and resource adjustment module, an optimal computing power allocation and scheduling module, and a scheduling execution and feedback optimization module.

[0151] A computing power request receiving and preliminary demand analysis module, used to receive multiple computing power requests and corresponding parameters, and analyze and determine preliminary computing power requirements based on the received first computing power request and parameters;

[0152] A preliminary computing power resource configuration module, configured to select at least a portion of computing power resources from a preset computing power pool as first candidate computing power resources according to the preliminary computing power demand;

[0153] A computing power requirement update and resource adjustment module, configured to form a second computing power requirement based on the first candidate computing power resource, in combination with the second computing power request and the parameter update computing power requirement, and adjust the first candidate computing power resource based on the updated computing power requirement to obtain a second candidate computing power resource set;

[0154] an optimal computing power allocation and scheduling module, configured to use the second candidate computing power resource set to calculate an optimal computing power allocation scheme that matches the third computing power request and parameters, allocate the second candidate computing power resource set according to the optimal computing power allocation scheme, and generate an actual scheduling instruction;

[0155] The scheduling execution and feedback optimization module is used to execute the actual scheduling instructions, monitor the execution effect to evaluate the scheduling results, and optimize the subsequent computing power request processing logic based on the scheduling result feedback.

[0156] In addition, the above-mentioned computing power request receiving and preliminary demand analysis module, preliminary computing power resource configuration module, computing power demand update and resource adjustment module, optimal computing power allocation and scheduling module, and scheduling execution and feedback optimization module are also used to implement the other steps of the above-mentioned intelligent computing power scheduling method based on data analysis during execution, which will not be elaborated here one by one.

[0157] In summary, the present invention dynamically analyzes multiple received computing power requests and their parameters to determine reasonable computing power requirements, and selects the optimal computing power resources for scheduling on this basis. This method not only improves the utilization rate of computing power resources, but also continuously optimizes the subsequent request processing logic based on the scheduling result feedback, thereby significantly improving the system's response speed and service quality.

[0158] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent computing power scheduling method based on data analysis, characterized in that: include: Step 1: receiving multiple computing power requests and corresponding parameters, and analyzing and determining the preliminary computing power requirements based on the received first computing power request and parameters; Step 2: Select at least a portion of computing power resources from a preset computing power pool as first candidate computing power resources according to the preliminary computing power demand; Step 3: Based on the first candidate computing power resources, the second computing power request and the parameter update computing power requirement are combined to form a second computing power requirement, and the first candidate computing power resources are adjusted based on the updated computing power requirement to obtain a second candidate computing power resource set; Step 4: Using the second candidate computing power resource set, calculate the optimal computing power allocation plan that matches the third computing power request and parameters, allocate the second candidate computing power resource set according to the optimal computing power allocation plan, and generate an actual scheduling instruction; Step 5: Execute the actual scheduling instruction, monitor the execution effect to evaluate the scheduling result, and optimize the subsequent computing power request processing logic based on the scheduling result feedback.

2. According to claim 1, a method for intelligent computing power scheduling based on data analysis is characterized in that: Receive multiple hashrate requests and corresponding parameters including: For multiple computing power requests and corresponding parameters received, the formula is used: F(x)=w1·x1+w2·x2+...+w n ·x n Calculate the weight value for each request, where x i Indicates the parameters of the i-th request, w i is the weight coefficient of the corresponding parameter; according to the calculated weight value F(x), the computing power requests are sorted to form a priority list; The computing power requests are processed in sequence according to the priority list, and the remaining available computing power resources are updated after each request is processed.

3. According to claim 2, a method for intelligent computing power scheduling based on data analysis is characterized in that: The initial computing power requirements determined by analysis include: For the first computing power request and its parameters received, apply the formula To quantify the computing power requirements, where p i Represents the priority of the i-th parameter, c i is the corresponding computing power consumption value; Based on the quantified computing power requirement D, a set of computing power resources that meet the conditions is identified, and according to the computing power requirement D, the computing power resources that meet or are closest to the requirements are selected from the identified computing power resource set as the initial computing power configuration.

4. The intelligent computing power scheduling method based on data analysis according to claim 3 is characterized in that: Selecting at least a portion of computing power resources from the preset computing power pool as first candidate computing power resources includes: According to the preliminary computing power demand D, a threshold T is set to filter the resources in the preset computing power pool, where T is the maximum acceptable computing power gap; Calculate the computing resources R in the preset computing pool i The gap Δ between the initial computing power requirement D i =|R i -D|; Select all Δ i ≤T computing resources R i , forming the first candidate computing power resource set.

5. The intelligent computing power scheduling method based on data analysis according to claim 4 is characterized in that: The second computing power requirement is formed by combining the second computing power request and the parameter update computing power requirement, including: Obtain the parameter set P2 of the second computing power request, and apply the formula D′2=f(P2) to evaluate the computing power requirement D′2 corresponding to the request, where f is the evaluation function; The total computing power C of the first candidate computing power resource set total Compare with the initial computing power requirement D and the newly evaluated requirement D′2, and update the computing power requirement to D2=D+D′2, ensuring that D2≤C total ; If D2>C total , then adjust D2 to C proportionally total , forming the final second computing power requirement.

6. The intelligent computing power scheduling method based on data analysis according to claim 5 is characterized in that: The first candidate computing resources are adjusted based on the updated computing power requirement to obtain a second candidate computing power resource set, including: Based on the updated computing power requirement D2, re-evaluate the applicability of each resource in the first candidate computing power resource set; For each resource R in the first candidate computing resource set i , calculate its matching degree with D2 Where R i ≥D2; filter out all matching degree M i Resources close to 1 form the second candidate computing power resource set.

7. The intelligent computing power scheduling method based on data analysis according to claim 6 is characterized in that: Calculate the optimal computing power allocation solution that matches the third computing power request and parameters, including: Obtain the parameter set P3 of the third computing power request, and calculate the computing power requirement D3 corresponding to the request; For each resource R in the second candidate computing resource set i , apply formula A i =min(D3, R i ) to determine the maximum computing power A that can be allocated to the third computing power request i ; Compare each A i value, choose so that A i The computing resources closest to D3 are used as part of the optimal allocation plan.

8. The intelligent computing power scheduling method based on data analysis according to claim 7 is characterized in that: Allocating the second candidate computing resource set to generate an actual scheduling instruction includes: Determine the resource R in the optimal computing power allocation plan i And its corresponding distribution amount A i , where A i By R i The provided computing power is balanced with the demand of the third computing power request D3; Construct scheduling instruction L i , which contains the resource ID i and the distribution amount A i , that is, L i =(ID i , A i ); According to the scheduling instruction L i , allocate the resource R i The computing power A i Request to the third computing power; Record the allocated computing power A i , update the remaining computing resources R′ i =R i -A i , to reflect resource usage; Repeat the above steps until all resources in the second candidate computing resource set are properly allocated, generating a complete scheduling instruction set I = {I1, I2, ..., I n }, used for actual computing power scheduling.

9. The intelligent computing power scheduling method based on data analysis according to claim 8 is characterized in that: Execute the actual scheduling instructions, monitor the execution effect to evaluate the scheduling results, and optimize the subsequent computing power request processing logic based on the scheduling result feedback, including: Execute the scheduling instruction set I, implement computing resource allocation, and monitor the performance index P during the execution process; According to the performance index P, the effectiveness E of the scheduling result is evaluated: E = g(P), where g is the evaluation function; According to the effectiveness E of the scheduling result, adjust the parameters in the computing power request processing logic. After updating the parameters, recalculate the processing logic of subsequent computing power requests to form a new computing power demand assessment and resource allocation plan, and apply the optimized processing logic to new computing power requests.

10. An intelligent computing power scheduling system based on data analysis, characterized in that: include: A computing power request receiving and preliminary demand analysis module, used to receive multiple computing power requests and corresponding parameters, and analyze and determine the preliminary computing power demand based on the received first computing power request and parameters; A preliminary computing power resource configuration module, configured to select at least a portion of computing power resources from a preset computing power pool as first candidate computing power resources according to the preliminary computing power demand; A computing power requirement update and resource adjustment module, configured to form a second computing power requirement based on the first candidate computing power resource, in combination with the second computing power request and the parameter update computing power requirement, and adjust the first candidate computing power resource based on the updated computing power requirement to obtain a second candidate computing power resource set; an optimal computing power allocation and scheduling module, configured to use the second candidate computing power resource set to calculate an optimal computing power allocation scheme that matches the third computing power request and parameters, allocate the second candidate computing power resource set according to the optimal computing power allocation scheme, and generate an actual scheduling instruction; The scheduling execution and feedback optimization module is used to execute the actual scheduling instructions, monitor the execution effect to evaluate the scheduling results, and optimize the subsequent computing power request processing logic based on the scheduling result feedback.

Images