Decision knowledge base construction method, data center intelligent regulation and control method and device

Abstract

The application discloses a decision knowledge base construction method and a data center intelligent regulation and control method and device. The decision knowledge base construction method comprises the following steps: obtaining an initial decision corresponding to a preset moment from an initial knowledge base constructed based on historical experience of a target scene and executing; obtaining key execution data corresponding to each collection period in an evaluation window after waiting for an evaluation delay; generating a candidate knowledge item set; performing serial filtering analysis on the item set according to objective indexes; obtaining an objective passing item set and an objective passing rate; when the objective passing rate is greater than a threshold value, performing subjective analysis on the objective passing item set to obtain a subjective passing item set; then performing timeliness analysis on the item set; storing a knowledge item meeting a condition to the initial knowledge base; and taking the updated initial knowledge base as a decision knowledge base. The method can obtain safe, reliable, stable and not outdated knowledge, realize effective reuse of the knowledge and effectively improve the intelligence and accurate regulation and control of the system.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to a method for constructing a decision knowledge base, a method for intelligent control of a data center, and an apparatus. Background Technology

[0002] In industrial scenarios such as data center energy saving, flexible manufacturing, and power grid dispatching, the control objectives (PUE, yield rate, line loss) have extremely high requirements for the timeliness and precision of the strategies. Any trial and error starting from scratch may directly translate into electricity costs, scrap parts, or downtime losses.

[0003] In practical applications, the knowledge base constructed from historical experience is mainly used as a subsequent guiding measure. However, in practice, it has been found that the knowledge base constructed in this way contains a large amount of low-quality, outdated, and conflicting knowledge, as well as some cases where accidental successes are misjudged as excellent experiences. At the same time, there are problems such as the experience becoming outdated due to environmental changes. After actual adoption, it has not brought about system efficiency savings and has greatly affected the accuracy of actual control. Summary of the Invention

[0004] In view of this, the present disclosure provides a method for constructing a decision knowledge base, a method for intelligent control of a data center, and an apparatus that can obtain secure, reliable, stable, and up-to-date knowledge, realize the effective reuse of knowledge, and effectively improve the intelligent and precise control of the system.

[0005] In a first aspect, embodiments of this disclosure provide a method for constructing a decision knowledge base, including:

[0006] An initial knowledge base is constructed based on historical experience of the target scenario;

[0007] Obtain the initial decision corresponding to the target scenario at a preset time from the initial knowledge base;

[0008] After executing the initial decision and waiting for the evaluation delay, the key execution data corresponding to each collection cycle within the preset evaluation window is obtained, and a set of candidate knowledge items corresponding to the preset evaluation window is generated; wherein, the key execution data includes at least energy use efficiency, total energy consumption, and temperature data, and each collection cycle corresponds to a set of knowledge items;

[0009] The candidate knowledge item set is subjected to a series of objective indicators across several dimensions for sequential filtering analysis to obtain the objectively passed item set and the objective pass rate.

[0010] When the objective pass rate is greater than the preset pass rate threshold, a multi-dimensional analysis is performed on each group of knowledge items in the objective pass item set to obtain all items that meet the multi-dimensional analysis conditions, which are denoted as the subjective pass item set.

[0011] The timeliness analysis of each group of knowledge items in the subjective knowledge item set is performed to obtain all knowledge items that meet the conditions, and these are stored in the initial knowledge base. The updated initial knowledge base is then used as the decision knowledge base.

[0012] Optionally, obtaining the initial decision corresponding to the target scenario at a preset time from the initial knowledge base includes:

[0013] Collect initial data of the target scene at a preset time; the initial data includes at least the initial energy usage efficiency and the initial total energy consumption.

[0014] Based on the initial data, the action plan with the highest similarity is obtained from the initial knowledge base and used as the corresponding initial decision;

[0015] The preset time is either the time when the abnormal event occurs or the time determined according to a preset period.

[0016] Secondly, this disclosure also provides a data center intelligent control method, including:

[0017] The target knowledge base is obtained using the aforementioned decision knowledge base construction method;

[0018] In response to anomaly alarms in the data center, collect current on-site data in real time;

[0019] Based on the on-site data, the decision with the highest relevance is obtained from the target knowledge base and executed.

[0020] Thirdly, this disclosure also provides a computer device, which adopts the following technical solution:

[0021] The computer device includes:

[0022] At least one processor; and,

[0023] A memory communicatively connected to the at least one processor; wherein,

[0024] The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform either the decision knowledge base construction method or the data center intelligent control method described above.

[0025] Fourthly, embodiments of this disclosure also provide a computer-readable storage medium storing computer instructions; the computer instructions are used to cause a computer to execute any of the above-described decision knowledge base construction methods or data center intelligent control methods.

[0026] Fifthly, embodiments of this disclosure also provide a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of any of the methods described above.

[0027] The decision knowledge base construction method provided in this disclosure obtains the initial decision corresponding to the target scenario at a preset time from an initial knowledge base built based on the historical experience of the target scenario; executes the initial decision, and after waiting for the evaluation delay, obtains the key execution data corresponding to each collection cycle within a preset evaluation window, generating a candidate knowledge item set corresponding to the preset evaluation window; performs serial filtering analysis of the candidate knowledge item set using several dimensions of objective indicators to obtain an objectively passed item set and an objective pass rate; when the objective pass rate is greater than a preset pass rate threshold, performs multi-dimensional analysis on each group of knowledge items in the objectively passed item set to obtain all items that meet the multi-dimensional analysis conditions, which are recorded as the subjectively passed item set; Subjectively, the timeliness analysis of each group of knowledge items in the item set is performed to obtain all knowledge items that meet the conditions and store them in the initial knowledge base. The updated initial knowledge base is then used as the decision knowledge base. This application uses three layers of quality control: objective indicator analysis, multi-dimensional interpretable evaluation, and timeliness analysis. The first layer uses objective hard thresholds to prevent the existence of junk knowledge. The second layer uses multi-dimensional interpretable evaluation to screen high-quality and reusable knowledge. The third layer uses timeliness analysis to prevent knowledge from becoming outdated and expanding in scale, that is, to prevent one-time accidental results from dominating for a long time. This ensures that the knowledge obtained through the three layers of control is safe, reliable, stable, and not outdated, so that it can be directly used as an intelligent control solution on site and realize the knowledge reuse of the data center.

[0028] The above description is merely an overview of the technical solution disclosed herein. In order to better understand the technical means of this disclosure and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0029] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 A flowchart illustrating the decision knowledge base construction method provided in this embodiment of the disclosure.

[0031] Figure 2 This is a flowchart illustrating a method for performing serial filtering analysis on a set of candidate knowledge entries, as provided in an embodiment of this disclosure.

[0032] Figure 3 This is a flowchart illustrating a method for obtaining a subjective access set according to an embodiment of the present disclosure.

[0033] Figure 4 A flowchart illustrating the method for updating the initial knowledge base provided in this embodiment of the disclosure.

[0034] Figure 5 This is a flowchart illustrating the intelligent control method for data centers provided in an embodiment of this disclosure.

[0035] Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure. Detailed Implementation

[0036] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0037] It should be understood that the following specific examples illustrate the implementation of this disclosure, and those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0038] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0039] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this disclosure. The drawings only show the components related to this disclosure and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0040] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0041] Reference Figure 1 This application discloses a method for constructing a decision knowledge base, including:

[0042] S100: Construct an initial knowledge base based on historical experience of the target scenario;

[0043] S200: Obtain the initial decision corresponding to the target scenario at a preset time from the initial knowledge base;

[0044] S300 executes the initial decision, and after waiting for the evaluation delay, acquires the key execution data corresponding to each collection cycle within the preset evaluation window, and generates a set of candidate knowledge items corresponding to the preset evaluation window.

[0045] S400 performs a series of filtering analyses on the candidate knowledge item set using objective indicators across several dimensions to obtain the objectively passed item set and the objective pass rate.

[0046] S500: When the objective pass rate is greater than the preset pass rate threshold, a multi-dimensional analysis is performed on each group of knowledge items in the objective pass item set to obtain all items that meet the multi-dimensional analysis conditions, which are denoted as the subjective pass item set.

[0047] S600 performs timeliness analysis on each group of knowledge items in the subjective access item set to obtain all knowledge items that meet the conditions and stores them in the initial knowledge base. The updated initial knowledge base is then used as the decision knowledge base.

[0048] The decision-making knowledge base construction method disclosed in this application employs a three-layer quality control approach: objective indicator analysis, multi-dimensional interpretable evaluation, and timeliness analysis. The first layer uses objective hard thresholds to prevent the existence of junk knowledge; the second layer uses multi-dimensional interpretable evaluation to screen for high-quality, reusable knowledge; and the third layer uses timeliness analysis to prevent knowledge from becoming outdated and expanding in scale, thus preventing a one-off, accidental result from dominating for a long time. This ensures that the knowledge obtained through these three layers of control is safe, reliable, stable, and timely, facilitating its direct use as an intelligent control solution in the field and enabling knowledge reuse in the data center. This method involves decision execution, data collection, and knowledge... The updated closed-loop process transforms the actual execution data (key execution data) of the initial decision into new candidate knowledge items, enabling the knowledge base to continuously learn from practice. It is based on historical experience and continuously optimized through execution feedback. At the same time, it ensures the hard quality of knowledge through objective indicators and mines implicit value through subjective analysis. Ultimately, it obtains a decision knowledge base that is scenario-adaptive, dynamically effective, and accurately supportive. This upgrades the knowledge base from a static historical database to a dynamic learning database, continuously adapting to changes in scenarios and avoiding decision rigidity caused by relying on outdated experience. It can provide reliable and efficient knowledge support for business decisions.

[0049] The S100 method of "building an initial knowledge base based on historical experience of the target scenario" specifically includes: sorting out historical operation records, expert experience, offline optimization / simulation results, etc., and transforming them into structured knowledge items (including scene fields, action fields, historical effect fields and necessary metadata) as the initial knowledge base.

[0050] The method for S200, "obtaining the initial decision corresponding to the target scenario at a preset time from the initial knowledge base," specifically includes:

[0051] S210, Collect initial data of the target scene at a preset time; the initial data includes at least the initial energy usage efficiency and the initial total energy consumption.

[0052] S220: Based on the initial data, obtain the action plan with the highest similarity from the initial knowledge base and use it as the corresponding initial decision.

[0053] The preset time is either the time when the abnormal event occurs or the time determined according to a preset period. In this embodiment, the trigger time t0 is the time when the system generates and issues a control decision. t0 can be determined by periodic triggering (i.e., active triggering), such as once every 15 minutes / 30 minutes / 1 hour, or by event triggering (passive triggering), such as when the temperature approaches the threshold, an alarm occurs, the load changes suddenly, or the equipment status changes.

[0054] For S300, "execute the initial decision, and after waiting for the evaluation delay, obtain the key execution data corresponding to each collection cycle within the preset evaluation window, and generate the candidate knowledge item set corresponding to the preset evaluation window". For example, when the initial decision is introduced at the preset time T0, after waiting for the evaluation delay Td (e.g., 10~30 minutes), the evaluation window W is entered for statistics. That is, after the action is executed, wait for Td first, and then enter the evaluation window [t0+Td, t0+Td+W] to perform data collection.

[0055] The key execution data includes at least Power Usage Effectiveness (PUE), Total Energy Consumption (E), and temperature data, and the key execution data corresponding to each collection cycle is recorded as a set of knowledge entries; Power Usage Effectiveness (PUE) is the ratio of all energy consumed by the data center to the energy consumed by the IT load.

[0056] Furthermore, the candidate knowledge item set may include the following information sets: 1) Scenario information: outdoor temperature, IT load, time period / season, equipment configuration identifier, key sensor summary; 2) Action information: key control quantities for this control decision (e.g., number of chillers, load allocation, water supply temperature, etc.); 3) Actual effect information: PUE before and after execution, energy consumption change, energy saving, number of alarms, temperature statistics, etc.; 4) Data quality information: data integrity, sensor error rate, missing key measurement points, etc.

[0057] In this application, the candidate knowledge item set is not derived from external data or theoretical derivation, but is generated based on key data after the initial decision is executed. It is directly related to the execution results of actual business scenarios. This knowledge generation logic, which comes from execution and goes to decision-making, makes the knowledge in the knowledge base highly bound to the actual business process, avoids the problem of theoretical knowledge being disconnected from practice, and provides more targeted support for subsequent decisions.

[0058] Reference Figure 2 The method S400, which involves "continuous filtering analysis of candidate knowledge item sets using several dimensions of objective indicators to obtain the objectively passed item set and the objective pass rate," specifically includes the following:

[0059] S410, perform PUE improvement dimension analysis on the candidate knowledge item set to obtain all knowledge items that meet the PUE difference condition, which are denoted as the first dimension through the knowledge item set, and each set of knowledge items corresponds to a set of key execution data.

[0060] Specifically, the PUE improvement degree of each group of knowledge items in the candidate knowledge item set is obtained; all knowledge items with a PUE improvement degree not less than a preset improvement degree threshold are obtained, and these knowledge items that meet the PUE difference condition are recorded as the first dimension passing knowledge item set. Through this step, false improvements caused by noise can be eliminated.

[0061] Wherein, the improvement degree of PUE is ΔPUE: ΔPUE = PUE(t0) - PUE(ti), where PUE(t0) is the PUE of the current scenario when the initial decision is made (i.e., time t0), and PUE(ti) is the PUE of the current scenario when each set of knowledge items is collected (i.e., time ti). The larger the ΔPUE, the more significant the energy efficiency improvement. In this embodiment, PUE = total energy consumption of the data center / energy consumption of IT load.

[0062] S420, perform energy consumption dimension analysis on the first dimension through the knowledge item set to obtain all knowledge items that meet the energy saving conditions, which are denoted as the second dimension through the knowledge item set.

[0063] Specifically, the energy-saving data for each group of knowledge entries in the first dimension's knowledge entry set is obtained; all knowledge entries with energy-saving data not less than a preset energy-saving threshold are then identified, and these knowledge entries that meet the energy-saving condition are recorded as the second dimension's knowledge entry set. Using the energy-saving threshold as a standard in this step avoids storing entries with low operational value and susceptibility to noise.

[0064] Wherein, the energy saving is △E: △E = E(t0) - E(ti), where E(t0) is the energy consumption of the current scenario when the initial decision is made (i.e., time t0), and E(ti) is the energy consumption of the current scenario when collecting each set of knowledge items (i.e., time ti).

[0065] S430, perform temperature safety dimension analysis on the second dimension through the knowledge item set to obtain all knowledge items that meet the temperature conditions, which are denoted as the third dimension through the knowledge item set.

[0066] Specifically, for the second dimension, temperature analysis is performed on each group of knowledge entries in the knowledge entry set to obtain all knowledge entries whose temperatures meet the temperature conditions (not exceeding the corresponding preset temperature threshold). The temperature data corresponding to each group of knowledge entries may include one or more of the following: cold aisle temperature, hot aisle temperature, chilled water supply and return water temperature, etc.

[0067] By analyzing this dimension, it can be determined whether the action corresponding to each group of knowledge entries has exceeded the safe operation boundary. If any limit is exceeded or an alarm is triggered, the entry into the database will be rejected. In other words, if the entry is successful with risk, it will not be entered into the knowledge base.

[0068] S440, perform stability dimension analysis on the third dimension through the knowledge item set to obtain all knowledge items that meet the stability conditions, denoted as the fourth dimension through the knowledge item set.

[0069] Specifically, the third dimension obtains all temperature data corresponding to each group of knowledge entries in the knowledge entry set; based on the temperature data, the temperature fluctuation information corresponding to each group of knowledge entries is obtained and used as stability information; knowledge entries whose stability information does not exceed the preset threshold during the collection period are recorded as knowledge entries that meet the stability conditions.

[0070] The preferred temperature fluctuation information is the oscillation amplitude, which is the fluctuation amplitude of the temperature collected for the corresponding knowledge item relative to the temperature at the time of issuing the initial decision (i.e., time T0). It should be noted that the temperature fluctuation information can also be flexibly set to other information according to actual needs, all of which are within the protection scope of this application.

[0071] Knowledge base reuse can amplify the impact of a certain strategy. If strategies that are prone to causing oscillations are stored in the knowledge base, future reuse may cause problems such as temperature fluctuations, impact on IT security, frequent equipment adjustments, increased wear and tear, and loss of trust in the system by maintenance personnel. In this embodiment, stability analysis is used to accurately determine whether the corresponding action will cause system oscillations, frequent adjustments, or amplified fluctuations, that is, to select knowledge with better stability.

[0072] S450, perform data quality dimension analysis on the fourth dimension through the knowledge item set to obtain all knowledge items that meet the data quality conditions, which is denoted as the fifth dimension through the knowledge item set.

[0073] Specifically, the fourth dimension is obtained by obtaining the expected sampling number and the effective sampling number corresponding to each group of knowledge items in the knowledge item set. Then, the quotient of the effective sampling number and the expected sampling number is used as the data quality information corresponding to each group of knowledge items. That is, in this embodiment, the analysis of the data quality dimension is the analysis of data integrity.

[0074] The expected number of samples is the quotient of the preset evaluation window and the sampling period (e.g., 1 minute / 5 minutes); the effective number of samples is the difference between the expected number of samples and the abnormal data in the corresponding collection period, i.e., the abnormal data collected is removed.

[0075] S460: Based on the knowledge item set and candidate knowledge item set of the fifth dimension, obtain the objective pass rate, and use the knowledge item set of the fifth dimension as the objective pass item set.

[0076] The objective pass rate is the ratio of the number of knowledge items in the objective pass set to the number of candidate knowledge items.

[0077] In this embodiment, the number of points passed in the previous level becomes the number of points to be entered in the next level.

[0078] It should be noted that before performing serial filtering analysis on the candidate knowledge item set, the process includes: determining several objective indicators and their priorities corresponding to the target scenario. Then, in S400, serial filtering analysis of the candidate knowledge item set based on the indicator priorities is performed on the objective indicators to obtain the objectively passed item set and the objective pass rate.

[0079] In this application, if any item in each set of knowledge items fails to meet the requirements, a rejection and a structured reason for rejection can be generated and archived for future reference as historical experience.

[0080] In this embodiment, several objective indicators (such as data integrity, performance achievement rate, and logical consistency) are used for sequential filtering to quickly eliminate obviously invalid entries (such as records with missing data or abnormal cases where the performance did not meet the standards), reducing the resource consumption of subsequent analysis and ensuring that the entries entering subjective analysis meet the basic quality requirements. For example, the objective pass rate threshold can screen out low-quality candidate sets, avoiding the waste of subjective analysis on obviously invalid knowledge.

[0081] Reference Figure 3 The method of S500, which states that "when the objective pass rate is greater than the preset pass rate threshold, a multi-dimensional analysis is performed on each group of knowledge items in the objective pass item set to obtain all items that meet the multi-dimensional analysis conditions, denoted as the subjective pass item set," specifically includes:

[0082] S510 analyzes each group of knowledge items in the objective knowledge item set from the perspectives of effectiveness, applicability, innovation, stability, and security, thereby obtaining multi-dimensional evaluation information.

[0083] The multidimensional evaluation information includes effectiveness score, applicability score, innovativeness score, stability score, and safety score.

[0084] S520 obtains a comprehensive score for each group of knowledge items based on effectiveness score, applicability score, innovativeness score, stability score, and security score.

[0085] The overall score is the sum of the effectiveness score, applicability score, innovativeness score, stability score, and safety score. Alternatively, the overall score for each knowledge item is obtained by weighted summation of the effectiveness score, applicability score, innovativeness score, stability score, and safety score, where the weight of the effectiveness score is greater than that of the applicability score, the applicability score is greater than that of the innovativeness score, the innovativeness score is greater than that of the stability score, the stability score is greater than that of the safety score, and the sum of the weights of these five dimensions is 1.

[0086] S530: Obtain all knowledge items whose comprehensive scores meet the preset scoring thresholds and store them in the subjective pass item set.

[0087] In this embodiment, the set of all knowledge items whose comprehensive score is not less than a preset score threshold is preferably denoted as the subjective pass item set.

[0088] Specifically, the method S510, which involves "analyzing each group of knowledge items in the objective knowledge item set from the dimensions of effectiveness, applicability, innovativeness, stability, and security to obtain multi-dimensional evaluation information," includes:

[0089] S511, determine the effectiveness score of each knowledge item based on the PUE improvement and energy saving corresponding to each knowledge item.

[0090] Specifically, if the analysis is conducted using PUE improvement ΔPUE and energy saving ΔE as the objects, the total mapping score for PUE improvement can be set to 60 points, and the total mapping score for energy saving can be set to 40 points. Then, based on the pre-set table of PUE improvement and mapping scores for different ranges, the first mapping score corresponding to the actual PUE improvement ΔPUE can be obtained; based on the pre-set table of energy saving and mapping scores for different ranges, the second mapping score corresponding to the actual energy saving ΔE can be obtained; the effectiveness score is (first mapping score + second mapping score) / 100, which can be flexibly set according to actual needs.

[0091] Furthermore, the effectiveness score of the corresponding knowledge item can also be obtained by considering PUE improvement, energy saving, cost saving, and target achievement rate. This score can be flexibly set according to actual needs and is within the scope of protection of this application.

[0092] S512, determine the applicability score for each group of knowledge items.

[0093] Specifically, each group of knowledge items can be analyzed for scenario field completeness, boundary clarity, and generality to obtain a corresponding applicability score. The applicability score is calculated as follows: (Scenario Field Completeness Score + Boundary Clarity Score + Generality Analysis Score) / 3. The scenario field completeness score can be assigned based on the coverage of required fields. The boundary clarity score can be assigned based on whether the variable range / device version / policy version is given. The generality analysis score can be assigned based on the frequency of occurrence of scenario conditions in the corresponding system. It can be stipulated that the maximum score for each item shall not exceed 100 points, and the specific settings can be flexibly configured according to actual needs.

[0094] S513, obtain the similarity between each group of knowledge items and all decisions in the initial knowledge base, and determine the innovation score corresponding to the knowledge item based on the similarity.

[0095] Specifically, the method for obtaining the innovation score includes: 1) Based on the maximum similarity with existing knowledge, the greater the maximum similarity, the smaller the difference with existing actions, the lower the novelty, and the lower the corresponding novelty score; 2) Obtain the number of existing decisions with a similarity greater than 50% with all decisions in the initial knowledge base. The smaller this number, the scarcer the corresponding knowledge item, and the higher the corresponding exploratory score; 3) Obtain the innovation score corresponding to each group of knowledge items based on the novelty score and the exploratory score, i.e., innovation score = (novelty score + exploratory score) / 2. In this embodiment, the score for each item is preferably no more than 100 points, but can be flexibly set according to actual needs.

[0096] Suppose that in the same scenario, only knowledge about a water supply temperature of 8℃ and two chillers is consistently recommended. The system might miss strategies that are more energy-efficient under certain humidity / nighttime conditions, such as a water supply temperature of 9℃ and two chillers. By introducing innovation, when the 9℃ strategy is within the safety / stability threshold and meets the ΔPUE / ΔE criteria, even with a small sample size, it can be labeled as an exploratory strategy and added to the database (more cautious reuse strategies can be set, such as initial A / B small-flow trial runs), thereby gradually accumulating evidence and preventing the knowledge base from being fixed on a single strategy.

[0097] S514. Determine the stability score based on the analysis results of the stability dimension corresponding to each group of knowledge items.

[0098] Specifically, the better the analysis results of the stability dimension, that is, the smaller the fluctuation, the higher the stability score is assigned. The specific score can be flexibly set according to actual needs.

[0099] S515 determines the safety score based on the temperature safety dimension analysis corresponding to each knowledge item and the corresponding number of alarms.

[0100] Specifically, the better the results of the temperature safety dimension analysis, the safer the system, and the fewer the alarms, the higher the safety score will be. The specific settings can be flexibly configured according to actual needs.

[0101] In one specific embodiment, assuming the preset evaluation window is 6 months, the number of knowledge items in the candidate knowledge item set corresponding to the preset evaluation window is 5400; PUE improvement dimension analysis is performed on these 5400 knowledge items, and 4320 knowledge items that meet the PUE difference condition are obtained, and the remaining 1080 knowledge items that do not meet the PUE difference condition are deleted.

[0102] Next, an energy consumption dimension analysis was performed on the 4,320 knowledge items that met the PUE difference condition. 3,888 knowledge items met the energy saving condition, and the remaining 432 knowledge items that did not meet the energy saving condition were deleted.

[0103] Then, a temperature safety dimension analysis was performed on the 3,888 knowledge items that met the energy-saving conditions, and 3,693 knowledge items that met the temperature conditions were obtained. The remaining 195 knowledge items that did not meet the temperature conditions were deleted.

[0104] Then, a stability dimension analysis was performed on the 3693 knowledge items that met the temperature conditions, and 3540 knowledge items were found to meet the stability conditions. The remaining 153 knowledge items that did not meet the stability conditions were then deleted.

[0105] Then, a data quality dimension analysis was performed on the 3540 knowledge items that met the stability conditions. 3240 knowledge items met the data quality conditions, with an objective pass rate of 3240÷5400×100%=60%. The remaining 300 knowledge items that did not meet the data quality conditions were then deleted.

[0106] In this embodiment, the 3240 knowledge entries obtained meet the first-level quality control (i.e., objective indicator verification). Then, a second-level quality control analysis is performed on these knowledge entries. This layer does not emphasize the model used, but rather the quality assessment task to be completed. That is, the candidate knowledge that passes the first level is comprehensively evaluated from the perspective of interpretability, and truly high-quality, reusable, and scalable experiences are selected. Structured comments are output for traceability. Specifically, a traceable evaluation report can be output, which may include: total score, sub-scores, strengths, weaknesses, and recommended tags (e.g., "robust strategy / exploratory strategy / only for specific devices"). This layer of analysis can effectively prevent the knowledge base from remaining at a local optimum for a long time. The innovation evaluation and safety evaluation are carried out in parallel, which effectively avoids high-risk attempts being mistakenly absorbed due to novelty.

[0107] Specifically, taking knowledge ID CAND_2024_0715_001, scenario of summer high-load cooling optimization, and decision of dual-machine operation (75%+70%) + water supply temperature of 8°C as an example, the obtained effectiveness score is 25.2, applicability score is 21.9, innovation score is 14.0, stability score is 13.5, and safety score is 9.5. The comprehensive score corresponding to this group of knowledge items is 84.1. If the preset scoring threshold is 80 points, then this group of knowledge items is greater than the preset scoring threshold, indicating that it is an excellent decision. That is, when this knowledge item is applied to the system, its effectiveness is outstanding, the PUE is significantly improved, the system stability is excellent, there is no oscillation, and the safety is extremely high. Therefore, it can be stored in the knowledge base as a new empirical decision and can be marked as a successfully verified robust strategy.

[0108] In this embodiment, for items that pass the objective test, further multi-dimensional analysis (such as business relevance, strategic alignment, implicit experience association, etc.) is conducted to uncover soft value that cannot be covered by objective indicators. This combination of objective screening and subjective depth mining ensures that knowledge meets both hard standards and fits the complex needs of the scenario, avoiding the one-sidedness of relying solely on data or pure experience.

[0109] Furthermore, by designing objective serial filtering (S400) and objective pass rate threshold (S500), a funnel-shaped optimization of knowledge screening is achieved. First, a large number of low-quality and obviously invalid entries (such as candidate knowledge with data anomalies or logical contradictions) are quickly filtered out through objective indicators, reducing the workload of subsequent subjective analysis. Only when the objective pass rate meets the standard (i.e., the overall quality of the candidate set is high) is the more resource-intensive subjective multi-dimensional analysis initiated, avoiding the waste of manpower / computing power on low-quality candidate sets. This hierarchical screening mechanism reduces the resource input of knowledge management (such as manual review costs and computing resources) while improving the efficiency of knowledge updates, enabling the knowledge base to iterate rapidly while ensuring quality.

[0110] Reference Figure 4 The method S600, which "conducts a timeliness analysis on each group of knowledge items in the subjective knowledge item set to obtain all knowledge items that meet the conditions, stores them in the initial knowledge base, and uses the updated initial knowledge base as the decision knowledge base," specifically includes the following:

[0111] S610, obtain the lifecycle of each group of knowledge items in the subjective entry set.

[0112] The lifecycle is △T, where △T = T1 - T0, T1 is the current time, and T0 is the collection time of the corresponding knowledge item.

[0113] S620 determines the time decay factor for each group of knowledge items based on the life cycle and the preset half-life.

[0114] The time decay factor is α: , The preset half-life.

[0115] By quantifying the value decay of knowledge over its life cycle through time decay factors (such as policy expiration, technological iteration, and changes in the market environment leading to knowledge obsolescence), outdated and obsolete knowledge items (such as old regulations, obsolete technical standards, and expired data) can be effectively excluded. This effectively prevents such knowledge from misleading decisions after entering the decision-making knowledge base (such as formulating plans based on old policies or adopting outdated technical parameters), thereby reducing decision-making errors caused by information lag and improving the accuracy of decision-making basis.

[0116] S630, determine the usage frequency bonus for each group of knowledge items.

[0117] Frequency of use is added as follows: ,in, The upper limit for frequency of use is set (preferably 1.2). This is the addition factor (preferably 1.0). This represents the number of times the device can be reused.

[0118] Introducing a frequency-based evaluation system effectively avoids the indiscriminate elimination of old knowledge: For knowledge that has been around for a long time but is still frequently used (such as classic theories, core processes, and frequently reused tools), its value is preserved through positive frequency enhancement; while knowledge that is both outdated and infrequently used (such as obsolete processes and replaced technologies) is filtered out through a combination of attenuation factors and low frequency. This design allows the knowledge base to dynamically update and eliminate invalid knowledge while retaining long-term effective and frequently relied-upon core knowledge, achieving a balance between updating and stability.

[0119] S640 determines the timeliness score for each knowledge item based on the time decay factor, usage frequency bonus, and comprehensive score.

[0120] Timeliness score = Overall score × Time decay factor × Frequency of use bonus.

[0121] Traditional knowledge screening relies on subjective human judgment (such as judging whether something is outdated based on experience), which is easily affected by individual cognition and inconsistent standards, leading to unstable screening results. This method, through the quantitative integration of time decay factors, usage frequency bonuses, and comprehensive scores, transforms timeliness assessment into calculable scoring indicators, achieving data-driven screening. This mechanism reduces subjective bias, ensures consistency in knowledge screening standards across different batches and scenarios, and improves the objectivity and reliability of knowledge management.

[0122] S650, obtain the knowledge items corresponding to the timeliness scores that meet the preset timeliness threshold.

[0123] By filtering through preset thresholds, only knowledge items that meet timeliness requirements are retained in the decision knowledge base, avoiding redundant and low-value knowledge from occupying storage resources and retrieval space. When using the decision knowledge base, users do not need to filter through massive amounts of outdated or invalid information, and can quickly locate high-timeliness and high-value knowledge, reducing information overload and improving knowledge retrieval efficiency. At the same time, storage resources are focused on effective knowledge, reducing system maintenance costs (such as storage, backup, and update costs).

[0124] In this embodiment, the calculation of timeliness score depends on the life cycle (dynamic change) and usage frequency (real-time statistics) of knowledge. Therefore, the screening results of the knowledge base can be automatically updated with changes in business scenarios and user behavior: when a piece of knowledge changes from high-frequency use to low-frequency use, or exceeds a preset half-life, its timeliness score decreases and it will be eliminated in the next screening; while newly generated high-value knowledge (such as new regulations and new technologies) will be included due to high comprehensive scores and low decay factors. This dynamic iteration capability enables the decision-making knowledge base to continuously adapt to changes in the external environment (such as industry policies and technological trends) and internal needs (such as adjustments in business priorities) and maintain long-term effectiveness.

[0125] Timeliness scoring does not solely rely on time or frequency, but combines time decay (objective timeliness), usage frequency (actual value), and comprehensive scoring (such as accuracy and relevance in preliminary assessments). This multi-dimensional integration ensures that the knowledge selected for the decision-making knowledge base not only meets timeliness standards but also considers core quality dimensions such as accuracy and relevance. It avoids the inclusion of items that are only new but not excellent (e.g., the latest knowledge but with low accuracy) or only frequently used but irrelevant (e.g., frequently used knowledge but unrelated to decision-making objectives). Ultimately, this forms a foundation for decision support that is both timely and high-quality. This method, through quantitative evaluation, dynamic screening, and multi-dimensional integration mechanisms, transforms the timeliness of knowledge from a vague subjective judgment into a calculable and iterative objective indicator. Ultimately, it constructs a dynamically updated, accurate, and effective knowledge base that adapts to decision-making needs, providing reliable support for efficient and accurate decision-making while optimizing the efficiency of knowledge management and resource allocation.

[0126] In this embodiment, the timeliness of subjectively approved knowledge items is assessed using quantitative indicators such as time decay factors and usage frequency bonuses. This ensures that the knowledge entering the knowledge base is dynamically updated over time, eliminating outdated knowledge, such as old compliance requirements due to policy changes or obsolete processes after technological iterations. The time decay factor reduces their scores to avoid misleading decisions. Conversely, long-term knowledge, such as classic management models and frequently reused core processes, is retained. The usage frequency bonuses maintain their value, avoiding a one-size-fits-all approach to elimination. This mechanism ensures that the decision-making knowledge base always focuses on currently effective knowledge, making it particularly suitable for rapidly changing scenarios and reducing decision-making risks caused by knowledge lag, such as developing compliance plans based on old policies.

[0127] Table 1 Comparison before and after implementation (6 months of data)

[0128]

[0129] As can be seen from Table 1, the knowledge base obtained through the method disclosed in this application has a significantly higher proportion of high-quality knowledge and a significantly lower knowledge base contamination rate. Using this knowledge base can effectively improve the intelligent management and control efficiency of the data center and reduce manual maintenance time.

[0130] This solution is not only applicable to data centers, but can also be extended to: 1) the industrial manufacturing sector, with corresponding scenarios including production parameter optimization (temperature, pressure, formula, etc.), which can accumulate optimal process parameters and avoid defective products; 2) the intelligent transportation sector, with corresponding scenarios including traffic light optimization strategies, which can accumulate optimal timing schemes under different time periods / weather conditions; 3) the financial risk control sector, with corresponding scenarios including risk assessment rule bases, which can accumulate effective risk identification patterns and eliminate outdated rules; 4) the medical diagnostics sector, with corresponding scenarios including clinical decision support systems, which can accumulate validated treatment plans to ensure medical safety.

[0131] Reference Figure 5 Secondly, this application discloses a data center intelligent control method, comprising:

[0132] S10, obtain the target knowledge base corresponding to the data center.

[0133] The target knowledge base is obtained by conducting tests, analyses, and updates within a preset period using the decision knowledge base construction method disclosed in the first aspect of this application.

[0134] S20, responding to anomaly alarms in the data center, collects current on-site data in real time;

[0135] S30: Based on the on-site data, retrieve the decision with the highest relevance from the target knowledge base and execute it.

[0136] A computer device according to embodiments of the present disclosure includes a memory and a processor. The memory is used to store non-transitory computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0137] The processor may be a central processing unit (CPU) or other processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the computer device to perform desired functions. In one embodiment of this disclosure, the processor is used to execute computer-readable instructions stored in the memory, causing the computer device to perform all or part of the steps of the decision knowledge base construction method or data center intelligent control method described in the foregoing embodiments of this disclosure.

[0138] Those skilled in the art will understand that, in order to solve the technical problem of how to achieve a good user experience, this embodiment may also include well-known structures such as communication buses and interfaces, and these well-known structures should also be included within the protection scope of this disclosure.

[0139] like Figure 6 This is a schematic diagram of a computer device provided for an embodiment of the present disclosure. It illustrates a structural schematic diagram suitable for implementing the computer device in the embodiments of the present disclosure. Figure 6 The computer device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0140] like Figure 6 As shown, a computer device may include a processor (such as a central processing unit, graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) or programs loaded from storage devices into random access memory (RAM). The RAM also stores various programs and data required for the operation of the computer device. The processor, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0141] Typically, the following devices can be connected to the I / O interface: input devices, such as sensors or visual information acquisition devices; output devices, such as displays; storage devices, such as magnetic tapes or hard drives; and communication devices. Communication devices allow the computer device to communicate wirelessly or wiredly with other devices (such as edge computing devices) to exchange data. Although Figure 6 A computer apparatus with various devices is shown, but it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or included alternatively.

[0142] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processor, all or part of the steps of the decision knowledge base construction method or the data center intelligent control method of embodiments of this disclosure are performed.

[0143] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0144] A computer-readable storage medium according to embodiments of the present disclosure stores non-transitory computer-readable instructions. When the non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the decision knowledge base construction method or data center intelligent control method described in the foregoing embodiments of the present disclosure are performed.

[0145] The aforementioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or portable hard drive), media with built-in rewritable non-volatile memory (e.g., memory card), and media with built-in ROM (e.g., ROM cartridge).

[0146] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0147] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.

[0148] In this disclosure, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The block diagrams of devices, apparatuses, devices, and systems involved in this disclosure are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as "comprising," "including," "having," etc., are open-ended terms meaning "including but not limited to," and are used interchangeably with them. The terms "or" and "and" as used herein refer to the terms "and / or," and are used interchangeably with them unless the context clearly indicates otherwise. The term "such as" as used herein refers to the phrase "such as but not limited to," and is used interchangeably with it.

[0149] Additionally, as used herein, the "or" used in a list of items beginning with "at least one" indicates a separate list, such that a list of, for example, "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not imply that the described example is preferred or better than other examples.

[0150] It should also be noted that in the systems and methods of this disclosure, the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions to this disclosure.

[0151] Various changes, substitutions, and modifications can be made to the technology described herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0152] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0153] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A method for constructing a decision knowledge base, characterized in that, Intelligent control for data centers, including: An initial knowledge base is constructed based on historical experience of the target scenario; Obtain the initial decision corresponding to the target scenario at a preset time from the initial knowledge base; After executing the initial decision and waiting for the evaluation delay, the key execution data corresponding to each collection cycle within the preset evaluation window is obtained, and a set of candidate knowledge items corresponding to the preset evaluation window is generated; wherein, the key execution data includes at least energy use efficiency, total energy consumption, and temperature data, and each collection cycle corresponds to a set of knowledge items; The candidate knowledge item set is subjected to a series of objective indicators across several dimensions for sequential filtering analysis to obtain the objectively passed item set and the objective pass rate; wherein, the passed knowledge item set of the previous dimension serves as the input set for the next dimension. When the objective pass rate is greater than the preset pass rate threshold, a multi-dimensional analysis is performed on each group of knowledge items in the objective pass item set to obtain all items that meet the multi-dimensional analysis conditions, which are recorded as the subjective pass item set, and a comprehensive score corresponding to each group of knowledge items is obtained at the same time. The timeliness analysis of each group of knowledge items in the subjective entry set is performed to obtain all knowledge items that meet the conditions and store them in the initial knowledge base. The updated initial knowledge base is then used as the decision knowledge base. The step of performing timeliness analysis on each group of knowledge items in the subjective pass item set to obtain all knowledge items that meet the conditions includes: obtaining the lifecycle of each group of knowledge items in the subjective pass item set; determining the time decay factor of each group of knowledge items based on the lifecycle and a preset half-life; determining the usage frequency bonus of each group of knowledge items; determining the timeliness score of each group of knowledge items based on the time decay factor, the usage frequency bonus, and the comprehensive score; and obtaining the knowledge items corresponding to the timeliness scores that meet the preset timeliness threshold; wherein, the timeliness score = comprehensive score × time decay factor × usage frequency bonus; the time decay factor is α: , The preset half-life is defined as follows: ΔT = T1 - T0, where T1 is the current time and T0 is the collection time of the corresponding knowledge item; the frequency of use is added together. : , The upper limit is added to the frequency of use. This is the addition factor. This represents the number of times the device can be reused.

2. The method for constructing a decision knowledge base according to claim 1, characterized in that, The step of obtaining the initial decision corresponding to the target scenario at a preset time from the initial knowledge base includes: Collect initial data of the target scene at a preset time; the initial data includes at least the initial energy usage efficiency and the initial total energy consumption. Based on the initial data, the action plan with the highest similarity is obtained from the initial knowledge base and used as the corresponding initial decision; The preset time is either the time when the abnormal event occurs or the time determined according to a preset period.

3. The method for constructing a decision knowledge base according to claim 2, characterized in that, The sequential filtering analysis of the candidate knowledge item set using several dimensions of objective indicators to obtain the objectively passed item set and the objective pass rate includes: Perform a PUE improvement dimension analysis on the candidate knowledge item set to obtain all knowledge items that meet the PUE difference condition, which are denoted as the first dimension pass knowledge item set. Each group of knowledge items corresponds to a group of key execution data. Energy consumption dimension analysis is performed on the first dimension through the knowledge item set to obtain all knowledge items that meet the energy-saving conditions, which are denoted as the second dimension through the knowledge item set. The second dimension is analyzed using the knowledge item set to obtain all knowledge items that meet the temperature conditions, which are denoted as the third dimension using the knowledge item set. The third dimension is analyzed for stability through the knowledge item set to obtain all knowledge items that meet the stability conditions, which are denoted as the fourth dimension through the knowledge item set. The fourth dimension is analyzed for data quality through the knowledge item set to obtain all knowledge items that meet the data quality conditions, which are denoted as the fifth dimension through the knowledge item set. The objective pass rate is obtained based on the knowledge entry set and the candidate knowledge entry set of the fifth dimension, and the knowledge entry set of the fifth dimension is used as the objective pass entry set.

4. The method for constructing a decision knowledge base according to claim 3, characterized in that, The PUE improvement dimension analysis is performed on the candidate knowledge item set to obtain all knowledge items that satisfy the PUE difference condition, including: Based on the initial energy usage efficiency of the target scenario at a preset time, the PUE improvement rate is obtained for each collection cycle within a preset evaluation window. Obtain the knowledge entries corresponding to all collection cycles where the PUE improvement is not less than the first preset threshold.

5. The method for constructing a decision knowledge base according to claim 4, characterized in that, The process involves performing multi-dimensional analysis on each group of knowledge items in the objective pass item set to obtain all items that meet the multi-dimensional analysis conditions, denoted as the subjective pass item set, including: Each set of knowledge items in the objective knowledge item set is analyzed from the dimensions of effectiveness, applicability, innovation, stability, and security to obtain multidimensional evaluation information, which includes effectiveness score, applicability score, innovation score, stability score, and security score. A comprehensive score is obtained for each group of knowledge items based on the effectiveness score, applicability score, innovativeness score, stability score, and security score. Obtain all knowledge items whose comprehensive scores meet the preset scoring threshold and store them in the subjective pass item set.

6. A data center intelligent control method, characterized in that, include: The target knowledge base is obtained by using the decision knowledge base construction method according to any one of claims 1-5; In response to anomaly alarms in the data center, collect current on-site data in real time; Based on the on-site data, the decision with the highest relevance is obtained from the target knowledge base and executed.

7. A computer device, characterized in that, The computer device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the decision knowledge base construction method according to any one of claims 1-5 or the data center intelligent control method according to claim 6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions; the computer instructions are used to cause the computer to execute the decision knowledge base construction method according to any one of claims 1-5 or the data center intelligent control method according to claim 6.

9. A computer program product comprising computer instructions, characterized in that, When executed by a processor, the computer instructions implement the steps of the decision knowledge base construction method according to any one of claims 1-5 or the data center intelligent control method according to claim 6.

Images