Electric power scientific calculation operator optimization method and system based on artificial intelligence

Abstract

The invention discloses an electric power scientific calculation operator optimization method and system based on artificial intelligence, and the method comprises the steps: monitoring the electric power scientific calculation performance in real time, and the electric power scientific calculation performance specifically comprises the electric power processor performance, the electric power memory performance and the electric power storage performance. The power processor performance, the power memory performance and the power storage performance are automatically monitored in real time, calculated and converted into numerical values, the numerical values are substituted into a formula according to a power performance health degree calculation formula, and the current health degree of the power scientific calculation performance is calculated; if the current electric power health degree value is too small, an alarm is given out immediately, scheduling processing is executed, then resources needed by electric power calculation are intelligently distributed through the calculation state in the electric power optimization state, and it is guaranteed that the performance of electric power calculation is optimal.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an artificial intelligence-based power scientific computing operator optimization method and system. Background Art

[0002] Cloud computing and big data are major changes and inevitable trends in the development of informatization. They are the commanding heights of international competition in the information age and the fuel for new economic development momentum. In power science computing, operators refer to mathematical objects that operate on certain power system-related data, models or algorithms.

[0003] In order to meet the service quality requirements of users, the existing power science computing platform has to adopt the method of over-provisioning resources, which has caused a huge waste of resources. With the development of quantum computing technology, quantum computing has shown great potential in solving complex optimization problems. Quantum computing can have a natural advantage in searching large-scale solution spaces by utilizing the parallel computing of quantum bits and the properties of quantum superposition states.

[0004] However, existing power science computing network optimization methods still face some challenges. On the one hand, quantum computing is still in the development stage in terms of hardware and algorithms, which limits its application in practical problems. On the other hand, quantum computing may require more advanced algorithms and methods for the complexity and diversity of power networks. Summary of the invention

[0005] In order to solve the above technical problems, an artificial intelligence-based power scientific computing operator optimization method and system are provided. This technical solution solves the problems that the existing power scientific computing network optimization method proposed in the above background technology still faces some challenges. On the one hand, quantum computing is still in the development stage in terms of hardware and algorithms, which limits its application in practical problems. On the other hand, quantum computing may require more advanced algorithms and methods for the complexity and diversity of power networks.

[0006] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0007] An artificial intelligence-based power scientific computing operator optimization method, comprising:

[0008] Real-time monitoring of power scientific computing performance, wherein the power scientific computing performance specifically includes power processor performance, power memory performance, and power storage performance;

[0009] Based on the power scientific computing performance, the power processor performance, power memory performance and power storage performance are calculated and converted into numerical values;

[0010] Establish a power computing performance health calculation formula, substitute the converted values ​​of power processor performance, power memory performance and power storage performance into the performance health calculation formula, and obtain the power computing health value;

[0011] Compare the power calculation health value with the preset threshold. If the power calculation health value is greater than or equal to the preset threshold, it is necessary to optimize the power scientific calculation performance configuration, otherwise it is not necessary.

[0012] Based on the confirmation of the optimized power scientific computing performance configuration, the required power scientific computing resource configuration is scheduled according to the optimal computing state.

[0013] Preferably, the step of calculating the power processor performance and converting it into a numerical value specifically comprises the following steps:

[0014] Get the number of cycles executed per second by the power processor and the number of cores of the power processor during calculation;

[0015] Based on the number of cycles executed by the power processor per second, obtaining the number of power calculation instructions executed by the power processor in each clock cycle;

[0016] The power processor performance value is obtained by performing calculations based on the number of cycles executed per second by the power processor, the number of cores, and the number of power computing instructions executed.

[0017] Preferably, the calculation formula of the power processor performance value is:

[0018] M a =T a ×N a ×C a

[0019] Where M a is the processor performance value, T a is the number of cycles executed per second by the power processor, N a The number of instructions executed is calculated as the power consumed, C a is the number of cores.

[0020] Preferably, the step of calculating the power memory performance and converting it into a numerical value specifically comprises the following steps:

[0021] In the process of calculating the power memory performance, an N×N matrix is ​​established, where N is the total number of nodes in the power system;

[0022] Get the memory size of a single matrix element;

[0023] The power memory performance value is calculated based on the total number of nodes in the power system and the memory size of a single matrix element.

[0024] Preferably, the calculation formula of the power memory performance value is:

[0025] M b =N 2 ×S

[0026] Where M b is the power memory performance value, N is the total number of nodes in the power system, and S is the memory size of a single matrix element.

[0027] Preferably, the step of calculating the power storage performance and converting it into a numerical value specifically comprises the following steps:

[0028] Set standard time intervals;

[0029] Real-time monitoring of the amount of data transmitted within a set standard time interval;

[0030] Calculate based on standard time interval and amount of data transmitted to obtain power storage performance value;

[0031] The calculation formula of the power storage performance value is:

[0032]

[0033] Where M C is the power storage performance value, T1 is the set standard time interval, and Y is the amount of data transmitted.

[0034] Preferably, the power calculation health value calculation formula is:

[0035]

[0036] In the formula, G is the power calculation health value;

[0037] M e Calculate indicators of abnormalities in power processor performance;

[0038] M n Calculate indicators for abnormal power memory performance;

[0039] M d Calculate indicators for abnormal performance of power storage;

[0040] α, β1 and β2 are coefficients of performance health.

[0041] Preferably, the calculation formula of the abnormal operation index is:

[0042] M=∑Δ 偏 ×t

[0043] In the formula, Δ 偏is the offset of the operating data relative to the standard operating data, and t is the time in Δ 偏 The running time in the state.

[0044] Preferably, the calculation method of α, β1 and β2 is:

[0045] Classify the historical resource operation data according to whether the power scientific computing performance configuration needs to be optimized in the end, and obtain several groups of historical resource data that need to be optimized in the power scientific computing performance configuration and several groups of historical resource data that do not need to be optimized for cloud computing resources;

[0046] According to the historical resource data that need to be optimized in the power scientific computing performance configuration and the historical resource data of several groups of cloud computing resources that do not need to be optimized, α, β1 and β2 are calculated by the maximum likelihood method;

[0047] Test the significance of α, β1 and β2 on the parameters of power performance health, and determine whether α, β1 and β2 meet the significance requirements.

[0048] An artificial intelligence-based power scientific computing operator optimization system, comprising:

[0049] A monitoring module, the monitoring module is used to monitor the power scientific computing performance in real time, the power scientific computing performance specifically includes power processor performance, power memory performance and power storage performance;

[0050] A first calculation module, the calculation module is used to calculate the power processor performance, the power memory performance and the power storage performance based on the power scientific calculation performance and convert them into numerical values;

[0051] A second calculation module, the second calculation module is used to establish a power calculation performance health calculation formula, substitute the converted values ​​of power processor performance, power memory performance and power storage performance into the performance health calculation formula, and obtain the power calculation health value;

[0052] A comparison module, the comparison module is used to compare the power calculation health value with a preset threshold;

[0053] The optimization module is used to schedule the required power scientific computing resource configuration according to the optimal computing state based on the confirmation of the optimized power scientific computing performance configuration.

[0054] Compared with the prior art, the present invention provides an artificial intelligence-based power scientific computing operator optimization method and system, which has the following beneficial effects:

[0055] In power scientific computing, an operator refers to a mathematical object that operates on certain power system-related data, models or algorithms. The power scientific computing performance is related to the power processor performance, power memory performance and power storage performance. Therefore, the present invention automatically monitors the power processor performance, power memory performance and power storage performance in real time, and converts them into numerical values. According to the power performance health calculation formula, the above numerical values ​​are substituted into the formula to calculate the current health of the power scientific computing performance. If the current power health value is too large, it means that the power scientific computing is in a healthy state and does not need to be scheduled. If the current power health value is too small, an alarm is immediately issued and scheduling is performed. Then, the computing state under the power optimization state is used to intelligently allocate the resources required for power computing to ensure the optimal performance of power computing and meet the needs of the staff. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the power scientific computing operator optimization method in the present invention;

[0057] Figure 2 A schematic diagram of a method for calculating and converting the performance of a power processor into a numerical value in the present invention;

[0058] Figure 3 A schematic diagram of a method for calculating and converting power memory performance into numerical values ​​in the present invention;

[0059] Figure 4 It is a schematic diagram of the method for calculating the power storage performance and converting it into a numerical value in the present invention. DETAILED DESCRIPTION

[0060] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0061] Example 1

[0062] Please refer to Figure 1-Figure 4 As shown, an artificial intelligence-based power scientific computing operator optimization method and system include:

[0063] Real-time monitoring of power scientific computing performance, which includes power processor performance, power memory performance, and power storage performance;

[0064] Based on the power scientific computing performance, the power processor performance, power memory performance and power storage performance are calculated and converted into numerical values;

[0065] Establish a power computing performance health calculation formula, substitute the converted values ​​of power processor performance, power memory performance and power storage performance into the performance health calculation formula, and obtain the power computing health value;

[0066] Compare the power calculation health value with the preset threshold. If the power calculation health value is greater than or equal to the preset threshold, it is necessary to optimize the power scientific calculation performance configuration, otherwise it is not necessary.

[0067] Based on the confirmation of the optimized power scientific computing performance configuration, the required power scientific computing resource configuration is scheduled according to the optimal computing state.

[0068] It can be understood by those skilled in the art that, in power scientific computing, an operator refers to a mathematical object that operates on certain power system-related data, models or algorithms. The power scientific computing performance is related to the power processor performance, the power memory performance and the power storage performance. Therefore, the present invention automatically monitors the power processor performance, the power memory performance and the power storage performance in real time, and converts them into numerical values. According to the power performance health calculation formula, the above numerical values ​​are substituted into the formula to calculate the current health of the power scientific computing performance. If the current power health value is too large, it means that the power scientific computing is in a healthy state and does not need to be scheduled. If the current power health value is too small, an alarm is immediately issued and scheduling is performed. The computing state under the power optimization state is then used to intelligently allocate the resources required for power computing to ensure the optimal performance of power computing and meet the needs of the staff.

[0069] Calculating the power processor performance and converting it into a numerical value specifically includes the following steps:

[0070] Get the number of cycles executed per second by the power processor and the number of cores of the power processor during calculation;

[0071] Based on the number of cycles executed by the power processor per second, obtaining the number of power calculation instructions executed by the power processor in each clock cycle;

[0072] The power processor performance value is obtained by performing calculations based on the number of cycles executed per second by the power processor, the number of cores, and the number of power computing instructions executed.

[0073] The calculation formula of the power processor performance value is:

[0074] M a =T a ×N a ×C a

[0075] Where M a is the processor performance value, Ta is the number of cycles executed per second by the power processor, N a The number of instructions executed is calculated as the power consumed, C a is the number of cores.

[0076] The specific steps of calculating the power memory performance and converting it into a numerical value include the following:

[0077] In the process of calculating the power memory performance, an N×N matrix is ​​established, where N is the total number of nodes in the power system;

[0078] Get the memory size of a single matrix element;

[0079] The power memory performance value is calculated based on the total number of nodes in the power system and the memory size of a single matrix element.

[0080] The calculation formula for the power memory performance value is:

[0081] M b =N 2 ×S

[0082] Where M b is the power memory performance value, N is the total number of nodes in the power system, and S is the memory size of a single matrix element.

[0083] The calculation of power storage performance and conversion into numerical values ​​specifically includes the following steps:

[0084] Set standard time intervals;

[0085] Real-time monitoring of the amount of data transmitted within a set standard time interval;

[0086] Calculate based on standard time interval and amount of data transmitted to obtain power storage performance value;

[0087] The calculation formula for the power storage performance value is:

[0088]

[0089] Where M C is the power storage performance value, T1 is the set standard time interval, and Y is the amount of data transmitted.

[0090] The calculation formula for the power calculation health value is:

[0091]

[0092] In the formula, G is the power calculation health value;

[0093] M e Calculate indicators of abnormalities in power processor performance;

[0094] M n Calculate indicators for abnormal power memory performance;

[0095] M d Calculate indicators for abnormal performance of power storage;

[0096] α, β1 and β2 are coefficients of performance health.

[0097] The calculation formula of abnormal operation index is:

[0098] M=∑Δ 偏 ×t

[0099] In the formula, Δ 偏 is the offset of the operating data relative to the standard operating data, and t is the time in Δ 偏 The running time in the state.

[0100] The calculation method of α, β1 and β2 is:

[0101] Classify the historical resource operation data according to whether the power scientific computing performance configuration needs to be optimized in the end, and obtain several groups of historical resource data that need to be optimized in the power scientific computing performance configuration and several groups of historical resource data that do not need to be optimized for cloud computing resources;

[0102] According to the historical resource data that need to be optimized in the power scientific computing performance configuration and the historical resource data of several groups of cloud computing resources that do not need to be optimized, α, β1 and β2 are calculated by the maximum likelihood method;

[0103] Test the significance of α, β1 and β2 on the parameters of power performance health, and determine whether α, β1 and β2 meet the significance requirements.

[0104] An artificial intelligence-based power scientific computing operator optimization system, comprising:

[0105] A monitoring module, which is used to monitor the power scientific computing performance in real time. The power scientific computing performance specifically includes the power processor performance, the power memory performance and the power storage performance;

[0106] A first calculation module, the calculation module is used to calculate the power processor performance, the power memory performance and the power storage performance based on the power scientific calculation performance and convert them into numerical values;

[0107] A second calculation module, the second calculation module is used to establish a power calculation performance health calculation formula, substitute the converted values ​​of power processor performance, power memory performance and power storage performance into the performance health calculation formula, and obtain the power calculation health value;

[0108] A comparison module, the comparison module is used to compare the power calculation health value with a preset threshold;

[0109] The optimization module is used to schedule the required power scientific computing resource configuration according to the optimal computing state based on the confirmation of the optimized power scientific computing performance configuration.

[0110] To sum up, in power scientific computing, an operator refers to a mathematical object that operates on certain power system-related data, models or algorithms. The power scientific computing performance is related to the power processor performance, power memory performance and power storage performance. Therefore, the present invention automatically monitors the power processor performance, power memory performance and power storage performance in real time, and calculates and converts them into numerical values. According to the power performance health calculation formula, the above numerical values ​​are substituted into the formula to calculate the current health of the power scientific computing performance. If the current power health value is too large, it means that the power scientific computing is in a healthy state and does not need to be scheduled. If the current power health value is too small, an alarm is immediately issued and scheduling is performed. The computing state under the power optimization state is used to intelligently allocate the resources required for power calculation to ensure the optimal performance of power calculation and meet the needs of the staff.

[0111] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. An artificial intelligence-based power scientific computing operator optimization method, characterized in that: include: Real-time monitoring of power scientific computing performance, wherein the power scientific computing performance specifically includes power processor performance, power memory performance, and power storage performance; Based on the power scientific computing performance, the power processor performance, power memory performance and power storage performance are calculated and converted into numerical values; Establish a power computing performance health calculation formula, substitute the converted values ​​of power processor performance, power memory performance and power storage performance into the performance health calculation formula, and obtain the power computing health value; Compare the power calculation health value with the preset threshold. If the power calculation health value is greater than or equal to the preset threshold, it is necessary to optimize the power scientific calculation performance configuration, otherwise it is not necessary. Based on the confirmation of the optimized power scientific computing performance configuration, the required power scientific computing resource configuration is scheduled according to the optimal computing state.

2. The power scientific computing operator optimization method based on artificial intelligence according to claim 1 is characterized in that: The calculation of the power processor performance and conversion into a numerical value specifically comprises the following steps: Get the number of cycles executed per second by the power processor and the number of cores of the power processor during calculation; Based on the number of cycles executed by the power processor per second, obtaining the number of power calculation instructions executed by the power processor in each clock cycle; The power processor performance value is obtained by performing calculations based on the number of cycles executed per second by the power processor, the number of cores, and the number of power computing instructions executed.

3. The power scientific computing operator optimization method based on artificial intelligence according to claim 1 is characterized in that: The calculation formula of the power processor performance value is: M a =T a ×N a ×C a Where M a is the processor performance value, T a is the number of cycles executed per second by the power processor, N a The number of instructions executed is calculated as the power consumed, C a is the number of cores.

4. The power scientific computing operator optimization method based on artificial intelligence according to claim 1 is characterized in that: The calculation of the power memory performance and conversion into a numerical value specifically comprises the following steps: In the process of calculating the power memory performance, an N×N matrix is ​​established, where N is the total number of nodes in the power system; Get the memory size of a single matrix element; The power memory performance value is calculated based on the total number of nodes in the power system and the memory size of a single matrix element.

5. The method for optimizing power scientific computing operators based on artificial intelligence according to claim 1, characterized in that: The calculation formula of the power memory performance value is: M b =N 2 ×S Where M b is the power memory performance value, N is the total number of nodes in the power system, and S is the memory size of a single matrix element.

6. The method for optimizing power scientific computing operators based on artificial intelligence according to claim 1, characterized in that: The calculation of the power storage performance and conversion into a numerical value specifically comprises the following steps: Set standard time intervals; Real-time monitoring of the amount of data transmitted within a set standard time interval; Calculate based on standard time interval and amount of data transmitted to obtain power storage performance value; The calculation formula of the power storage performance value is: Where M C is the power storage performance value, T1 is the set standard time interval, and Y is the amount of data transmitted.

7. The method for optimizing power scientific computing operators based on artificial intelligence according to claim 1, characterized in that: The calculation formula for the power calculation health value is: In the formula, G is the power calculation health value; M e Calculate indicators of abnormalities in power processor performance; M n Calculate indicators for abnormal power memory performance; M d Calculate indicators for abnormal performance of power storage; α, β1 and β2 are coefficients of performance health.

8. The method for optimizing power scientific computing operators based on artificial intelligence according to claim 1, characterized in that: The calculation formula of the abnormal operation index is: M=∑Δ 偏 ×t In the formula, Δ 偏 is the offset of the operating data relative to the standard operating data, and t is the time in Δ 偏 The running time in the state.

9. The method for optimizing power scientific computing operators based on artificial intelligence according to claim 1, characterized in that: The calculation method of α, β1 and β2 is: Classify the historical resource operation data according to whether the power scientific computing performance configuration needs to be optimized in the end, and obtain several groups of historical resource data that need to be optimized in the power scientific computing performance configuration and several groups of historical resource data that do not need to be optimized for cloud computing resources; According to the historical resource data that need to be optimized in the power scientific computing performance configuration and the historical resource data of several groups of cloud computing resources that do not need to be optimized, α, β1 and β2 are calculated by the maximum likelihood method; Test the significance of α, β1 and β2 on the parameters of power performance health, and determine whether α, β1 and β2 meet the significance requirements.

10. An artificial intelligence-based power scientific computing operator optimization system, used to implement an artificial intelligence-based power scientific computing operator optimization method as described in any one of claims 1 to 9, characterized in that: include: A monitoring module, the monitoring module is used to monitor the power scientific computing performance in real time, the power scientific computing performance specifically includes power processor performance, power memory performance and power storage performance; A first calculation module, the calculation module is used to calculate the power processor performance, the power memory performance and the power storage performance respectively based on the power scientific calculation performance and convert them into numerical values; A second calculation module, the second calculation module is used to establish a power calculation performance health calculation formula, substitute the converted values ​​of the power processor performance, the power memory performance and the power storage performance into the performance health calculation formula, and obtain the power calculation health value; A comparison module, the comparison module is used to compare the power calculation health value with a preset threshold; The optimization module is used to schedule the required power scientific computing resource configuration according to the optimal computing state based on the confirmation of the optimized power scientific computing performance configuration.

Images