Resource scheduling method and device, computer device and storage medium

Abstract

The application relates to a resource scheduling method and device, computer equipment and a storage medium. The method comprises the following steps: obtaining a target load task to be executed by a resource scheduling center, a current environment temperature and a current controller usage rate in a current period; predicting a predicted power consumption, a temperature penalty value and an unfinished data volume generated by the resource scheduling center when executing the target load task in a future period according to the current environment temperature and the current controller usage rate; determining a penalty accumulation quantity of the resource scheduling center in the future period according to the temperature penalty value and the unfinished data volume; determining an electricity cost of the resource scheduling center in the future period according to the predicted power consumption and a preset electricity time-of-use price; and generating a target scheduling scheme of the resource scheduling center according to the penalty accumulation quantity and the electricity cost. The method can accurately determine the scheduling scheme.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a resource scheduling method, apparatus, computer equipment, and storage medium. Background Technology

[0002] With the rapid development of technologies such as cloud computing, big data, and artificial intelligence, resource scheduling centers, as core infrastructure of modern information technology, are experiencing exponential growth in scale and computing power demand. To ensure the efficient and stable operation of resource scheduling centers, resource scheduling technology has emerged, aiming to coordinate the management of computing load, cooling output, and power distribution. Currently, resource scheduling for resource scheduling centers mainly relies on multi-objective optimization algorithms and energy system management schemes.

[0003] Traditional technologies typically employ multi-objective optimization methods for load scheduling. For example, some schemes optimize population initialization strategies and elite retention mechanisms using improved non-dominated sorting genetic algorithms to balance multiple objectives such as profit, carbon emissions, service quality, and energy consumption. Other schemes utilize ant colony optimization to simulate the pheromone mechanism in ant foraging, dynamically adjusting load allocation paths. Furthermore, some studies attempt to incorporate diverse energy sources such as power grids, solar photovoltaic systems, and combined cooling, heating, and power (CCHP) systems into optimization models, using time-of-use pricing mechanisms to regulate energy acquisition methods and reduce operating costs. In terms of environmental control, some schemes establish building heat transfer models to analyze the impact of outdoor weather conditions on air conditioning loads, aiming to achieve energy-saving scheduling.

[0004] However, current resource scheduling methods oversimplify the complex physical processes of the resource scheduling center. For example, they represent the relationship between cooling system energy consumption and load as a single constant, ignoring the nonlinear characteristics of cooling efficiency as a function of load rate and ambient temperature under actual operating conditions, resulting in poor accuracy of the scheduling scheme. While other studies using methods such as computational fluid dynamics to simulate heat distribution can improve accuracy, their high computational cost makes it difficult to accurately determine the target resource scheduling scheme for the resource scheduling center. Summary of the Invention

[0005] Therefore, it is necessary to provide a resource scheduling method, apparatus, computer equipment, and storage medium that can accurately determine resource scheduling schemes to address the aforementioned technical problems.

[0006] Firstly, this application provides a resource scheduling method, including:

[0007] Obtain the target load tasks to be executed in the resource scheduling center, the current ambient temperature, and the current controller utilization rate in the current time period;

[0008] Based on the current ambient temperature and current controller utilization, predict the power consumption, temperature penalty value, and amount of incomplete data generated by the resource scheduling center in executing the target load task in the future period; the predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of the target load task;

[0009] Based on the temperature penalty value and the amount of incomplete data, determine the cumulative penalty amount for the resource scheduling center in future time periods; and,

[0010] Based on the predicted power consumption and the preset time-of-use electricity price, the electricity cost of the resource dispatch center in the future period is determined;

[0011] Based on the cumulative number of penalties and electricity costs, a target scheduling scheme for the resource scheduling center is generated.

[0012] The target scheduling scheme is used to allocate computing resources for different target load tasks in the future time period, and to allocate cooling resources and power resources for each power equipment group in the resource scheduling center in the future time period.

[0013] In one embodiment, the power-consuming equipment group includes a computing equipment group, a cooling equipment group, and other equipment groups in the resource scheduling center besides the computing equipment group and the cooling equipment group; based on the current ambient temperature and the current controller utilization rate, the predicted power consumption generated by the resource scheduling center in executing the target load task in the future time period is predicted, including:

[0014] Based on the current ambient temperature, predict the additional power consumption of other equipment groups during the target load task in the future period, and the cooling power consumption of the cooling equipment group during the target load task in the future period; and...

[0015] Based on the current controller utilization, predict the power consumption of the computing equipment group during the execution of the target load task in the future period;

[0016] The predicted power consumption is determined by summing the power consumption of computing power, cooling power, and other power consumption.

[0017] In one embodiment, based on the current controller utilization, the power consumption of the computing equipment group during the execution of the target load task in a future time period is predicted, including:

[0018] Obtain the idle power consumption of the computing equipment group under idle conditions, and the full-load power consumption under full-load conditions;

[0019] Based on the current controller utilization rate, predict the controller utilization rate of the computing equipment group during the execution of the target load task in the future period;

[0020] The computing power consumption is determined based on idle power consumption, full-load power consumption, and predicted controller utilization.

[0021] In one embodiment, based on the current ambient temperature, the prediction of additional power consumption generated by other device groups during the execution of the target load task in a future time period includes:

[0022] Obtain environmental parameters from the resource scheduling center;

[0023] Based on the current ambient temperature, predict the predicted cooling power of the refrigeration equipment group in the future period, and predict the ambient temperature of the resource scheduling center in the future period;

[0024] Determine other power consumption based on environmental parameters, predicted ambient temperature, and cooling capacity;

[0025] Among them, environmental parameters include air parameters and equipment parameters; air parameters include air heat capacity and air thermal resistance; equipment parameters include equipment heat capacity and equipment thermal resistance of the equipment group.

[0026] In one embodiment, the refrigeration equipment group includes at least one refrigeration device; predicting the predicted cooling power of the refrigeration equipment group in a future time period based on the current ambient temperature includes:

[0027] Given that the cold storage equipment in the prediction resource scheduling center has been started, based on the current ambient temperature, predict the first cooling power of at least one cooling equipment in the future time period, and use the sum of different first cooling powers as the predicted cooling power.

[0028] In the case that the cold storage equipment in the prediction resource scheduling center is not started, based on the current ambient temperature, predict the cooling power of the cold storage equipment in the future period, and the second cooling power of at least one refrigeration equipment in the future period.

[0029] The sum of the cooling power and the different secondary cooling powers is used as the predicted cooling power.

[0030] In one embodiment, predicting the cooling power consumption generated by the cooling equipment group during the execution of a target load task in a future time period includes:

[0031] Obtain energy efficiency characteristic data of different refrigeration equipment in the refrigeration equipment group;

[0032] The power consumption for cooling is determined based on energy efficiency data and predicted cooling power.

[0033] Secondly, this application also provides a resource scheduling device, comprising:

[0034] The acquisition module is used to acquire the target load tasks to be executed by the resource scheduling center, the current ambient temperature and the current controller utilization rate in the current time period;

[0035] The prediction module is used to predict the predicted power consumption, temperature penalty value, and amount of uncompleted data generated by the resource scheduling center in executing the target load task in a future time period, based on the current ambient temperature and the current controller utilization rate; the predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of the target load task;

[0036] The determination module is used to determine the cumulative penalty amount of the resource scheduling center in the future time period based on the temperature penalty value and the amount of incomplete data; and to determine the electricity cost of the resource scheduling center in the future time period based on the predicted power consumption and the preset time-of-use electricity price.

[0037] The scheme module is used to generate a target scheduling scheme for the resource scheduling center based on the cumulative number of penalties and the electricity cost; wherein, the target scheduling scheme is used to allocate computing resources for different target load tasks in the future time period, and to allocate cooling resources and electricity resources for each group of electrical equipment in the resource scheduling center in the future time period.

[0038] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0039] Obtain the target load tasks to be executed in the resource scheduling center, the current ambient temperature, and the current controller utilization rate in the current time period;

[0040] Based on the current ambient temperature and current controller utilization, predict the power consumption, temperature penalty value, and amount of incomplete data generated by the resource scheduling center in executing the target load task in the future period; the predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of the target load task;

[0041] Based on the temperature penalty value and the amount of incomplete data, determine the cumulative penalty amount for the resource scheduling center in future time periods; and,

[0042] Based on the predicted power consumption and the preset time-of-use electricity price, the electricity cost of the resource dispatch center in the future period is determined;

[0043] Based on the cumulative number of penalties and electricity costs, a target scheduling scheme for the resource scheduling center is generated.

[0044] The target scheduling scheme is used to allocate computing resources for different target load tasks in the future time period, and to allocate cooling resources and power resources for each power equipment group in the resource scheduling center in the future time period.

[0045] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0046] Obtain the target load tasks to be executed in the resource scheduling center, the current ambient temperature, and the current controller utilization rate in the current time period;

[0047] Based on the current ambient temperature and current controller utilization, predict the power consumption, temperature penalty value, and amount of incomplete data generated by the resource scheduling center in executing the target load task in the future period; the predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of the target load task;

[0048] Based on the temperature penalty value and the amount of incomplete data, determine the cumulative penalty amount for the resource scheduling center in future time periods; and,

[0049] Based on the predicted power consumption and the preset time-of-use electricity price, the electricity cost of the resource dispatch center in the future period is determined;

[0050] Based on the cumulative number of penalties and electricity costs, a target scheduling scheme for the resource scheduling center is generated.

[0051] The target scheduling scheme is used to allocate computing resources for different target load tasks in the future time period, and to allocate cooling resources and power resources for each power equipment group in the resource scheduling center in the future time period.

[0052] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0053] Obtain the target load tasks to be executed in the resource scheduling center, the current ambient temperature, and the current controller utilization rate in the current time period;

[0054] Based on the current ambient temperature and current controller utilization, predict the power consumption, temperature penalty value, and amount of incomplete data generated by the resource scheduling center in executing the target load task in the future period; the predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of the target load task;

[0055] Based on the temperature penalty value and the amount of incomplete data, determine the cumulative penalty amount for the resource scheduling center in future time periods; and,

[0056] Based on the predicted power consumption and the preset time-of-use electricity price, the electricity cost of the resource dispatch center in the future period is determined;

[0057] Based on the cumulative number of penalties and electricity costs, a target scheduling scheme for the resource scheduling center is generated.

[0058] The target scheduling scheme is used to allocate computing resources for different target load tasks in the future time period, and to allocate cooling resources and power resources for each power equipment group in the resource scheduling center in the future time period.

[0059] The aforementioned resource scheduling methods, devices, computer equipment, and storage media, by comprehensively considering the current ambient temperature and controller utilization rate, accurately predict the power consumption, temperature penalty value, and amount of uncompleted data for future periods. Based on this, they integrate time-of-use pricing mechanisms and cumulative penalty amounts to generate a globally optimized scheduling scheme that balances electricity costs and operational constraints. This achieves dynamic and coordinated allocation of computing resources, cooling resources, and electricity resources, effectively avoiding resource mismatch and energy waste caused by independent decisions of each subsystem in traditional schemes. Under the premise of ensuring compliant operation of data center temperature and load tasks, it significantly improves the economic efficiency and resource utilization efficiency of system operation. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 This is an application environment diagram of a resource scheduling method provided in this embodiment;

[0062] Figure 2 This is a flowchart illustrating a resource scheduling method provided in this embodiment;

[0063] Figure 3 This is a flowchart illustrating a step for determining predicted power consumption in this embodiment.

[0064] Figure 4 This is a flowchart illustrating the steps for determining the power consumption of computing power in this embodiment.

[0065] Figure 5 This is a schematic flowchart illustrating a temperature power prediction step provided in this embodiment;

[0066] Figure 6This is a flowchart illustrating the steps for determining the power consumption for refrigeration, as provided in this embodiment.

[0067] Figure 7 This is a schematic diagram of a resource scheduling device provided in this embodiment;

[0068] Figure 8 This is an internal structural diagram of a computer device provided in this embodiment. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0070] The resource scheduling method provided in this application embodiment can be applied to, for example, Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104, or it can be located in the cloud or on another network server. The computer equipment acquires the target load tasks to be executed by the resource scheduling center, the current ambient temperature, and the current controller utilization rate in the current time period. Based on the current ambient temperature and current controller utilization rate, it predicts the predicted power consumption, temperature penalty value, and amount of unfinished data generated by the resource scheduling center in executing the target load tasks in the future time period. The predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of the target load tasks. Based on the temperature penalty value and the amount of unfinished data, it determines the cumulative penalty amount of the resource scheduling center in the future time period. Furthermore, based on the predicted power consumption and the preset time-of-use electricity price, it determines the electricity cost of the resource scheduling center in the future time period. Based on the cumulative penalty amount and electricity cost, it generates a target scheduling plan for the resource scheduling center. The target scheduling plan is used to allocate computing resources to different target load tasks in the future time period, and to allocate cooling resources and electricity resources to each group of electrical equipment in the resource scheduling center in the future time period. The computer equipment can be a terminal or a server; the terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. Portable wearable devices can be smartwatches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented using a standalone server or a server cluster composed of multiple servers.

[0071] In one exemplary embodiment, such as Figure 2As shown, a resource scheduling method is provided, which can be applied to... Figure 1 Taking a computer device as an example, the explanation includes the following steps S201 to S204. Wherein:

[0072] S201 obtains the target load tasks to be executed by the resource scheduling center, the current ambient temperature and the current controller utilization rate in the current time period.

[0073] In some embodiments, the target load tasks to be executed by the resource scheduling center, the current ambient temperature and the current controller utilization rate are obtained directly.

[0074] Based on the current ambient temperature and current controller utilization, S202 predicts the power consumption, temperature penalty value, and amount of incomplete data generated by the resource scheduling center in the future when executing target load tasks.

[0075] Among them, the predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of target load tasks.

[0076] In some embodiments, based on a prediction model, the predicted power consumption, temperature penalty value, and amount of incomplete data generated by the resource scheduling center in executing target load tasks in future time periods are predicted according to the current ambient temperature and the current controller utilization.

[0077] In one alternative embodiment, the method for predicting the temperature penalty value and the amount of incomplete data in the future time period of the resource scheduling center is as follows: the target load tasks to be executed by the resource scheduling center are often distinguished into latency-tolerant loads and latency-sensitive loads.

[0078] Among them, latency-sensitive loads include service requests with high priority and real-time requirements, which require the workload to be processed in a timely and real-time manner and do not have scheduling space; these are denoted as a set. For latency-sensitive loads, the scheduling capability is described as follows: In the formula, This refers to load tasks. exist The amount of data processed at any given moment; For load tasks exist The amount of data required at any given moment; This refers to load tasks. The state indicator at time t is 0-1. If it is 1, the task is being processed; if it is 0, the task is in a suspended state.

[0079] Delay-tolerant loads, on the other hand, only need to be completed before the specified deadline to meet the requirements; these are denoted as a set. For latency-tolerant loads, the scheduling capability is described as follows: In the formula, This refers to load tasks. exist The amount of data processed at any given moment; The system can be configured for load tasks at time t. Maximum amount of data processing resources provided; For load tasks The minimum data processing resource requirement per unit time during processing; This refers to load tasks. The state indicator at time t is 0-1. If it is 1, the task is being processed; if it is 0, the task is in a suspended and unprocessed state. That is, used for marking time. Is it due to load tasks? Processing time window set If it is within the processing window, it is 1; otherwise, it is 0.

[0080] For any load task Regardless of whether it is a latency-sensitive task or a latency-tolerant task, it has the following constraints: In the formula, The load task as of time t Cumulative amount of completed data processing; For load tasks The total amount of task data that the entire task needs to process; This is a load task. After the processing time window has expired The amount of data that has not yet been processed at any given time; This is the task completion indicator; 1 represents that the current task has been completed, and 0 represents the opposite. This is an auxiliary constant for the algorithm, representing the maximum value. Auxiliary constants for the algorithm, representing the minimum value; load tasks. Processing time window set .

[0081] Since some loads may have characteristics that cannot be interrupted during processing, the following constraints are used to describe the uninterruptible characteristics of the load: In the formula, This refers to load tasks. exist The task status (compiling / suspended) is constantly being switched between 0 and 1 indicators; This refers to load tasks. During the execution time window Total number of internal switching counts. This constitutes the set of uninterruptible tasks.

[0082] Load tasks There may be a series of preload tasks, the set of which is: Completing the preceding load tasks is the load task. The preconditions for startup are described as follows: Ultimately, given the current resource scheduling center computing unit... Total set of tasks to be scheduled Then the following constraints apply: In the formula, i.e., computing unit exist Total data processing volume at any given time; This refers to load tasks. exist The amount of data processed at any given moment; for Time Calculation Unit The amount of data processing corresponding to the total available computing resources.

[0083] It is important to note that, given that power system energy dispatch is often conducted on a minute-by-minute basis, there may be a mismatch between this and the actual load fluctuation rate at the resource dispatch center. Therefore, the load tasks for the resource dispatch center in this method do not refer to specific actual computational tasks, but rather to a set of load tasks that are abstracted and aggregated based on actual computational task predictions. The computational resource allocation plan obtained through algorithm optimization aims to guide the allocation of upper limits of computational resources within the predicted load task boundaries.

[0084] S203 determines the cumulative penalty amount for the resource dispatch center in future periods based on the temperature penalty value and the amount of incomplete data; and determines the electricity cost of the resource dispatch center in future periods based on the predicted power consumption and the preset time-of-use electricity price.

[0085] In some embodiments, the cumulative penalty amount for the resource scheduling center in future time periods is determined based on the temperature penalty value and the amount of incomplete data as follows: In the formula, To penalize the cumulative number, This is a load task. After the processing time window has expired The amount of incomplete data at the current moment. Its corresponding penalty coefficient; That is Temperature penalty value at any time The corresponding penalty coefficient is given.

[0086] In some embodiments, the method for determining the electricity cost of the resource dispatch center in future time periods based on predicted power consumption and preset time-of-use electricity prices is as follows: In the formula, For electricity costs, This is the predicted power consumption at the grid connection point; The time-of-use electricity price is preset.

[0087] S204 generates the target scheduling scheme for the resource scheduling center based on the cumulative number of penalties and electricity costs.

[0088] The target scheduling scheme is used to allocate computing resources for different target load tasks in the future time period, and to allocate cooling resources and power resources for each power equipment group in the resource scheduling center in the future time period.

[0089] In some embodiments, the target scheduling scheme of the resource scheduling center is generated based on the cumulative penalty amount and electricity cost as follows: ;in, To penalize the cumulative number; The penalty mileage unit coefficient corresponds to the cumulative number of penalties. For electricity costs, This is the coefficient corresponding to the electricity cost.

[0090] In the above embodiments, by comprehensively considering the current ambient temperature and controller utilization, the power consumption, temperature penalty value, and amount of incomplete data in future periods are accurately predicted. Based on this, the time-of-use pricing mechanism and the cumulative penalty amount are integrated to generate a global optimized scheduling scheme that takes into account both power costs and operational constraints. This achieves dynamic and coordinated allocation of computing resources, cooling resources, and power resources, effectively avoiding resource mismatch and energy waste caused by independent decision-making of each subsystem in traditional schemes. Under the premise of ensuring the compliant operation of data center temperature and load tasks, the system's economic efficiency and resource utilization efficiency are significantly improved.

[0091] Figure 3 This is a flowchart illustrating the steps for determining predicted power consumption in one embodiment. In this embodiment, the power-consuming equipment group includes a computing equipment group, a cooling equipment group, and other equipment groups in the resource scheduling center besides the computing equipment group and the cooling equipment group. Based on the current ambient temperature and the current controller utilization rate, the predicted power consumption generated by the resource scheduling center executing the target load task in the future time period is predicted, including the following steps:

[0092] Based on the current ambient temperature, S301 predicts the other power consumption generated by other equipment groups during the execution of the target load task in the future period, as well as the cooling power consumption generated by the cooling equipment group during the execution of the target load task in the future period; and based on the current controller utilization rate, predicts the computing power consumption generated by the computing equipment group during the execution of the target load task in the future period.

[0093] In some embodiments, based on a first prediction model, and according to the current ambient temperature, the power consumption of other equipment groups during the execution of the target load task in a future period, as well as the cooling power consumption of the cooling equipment group during the execution of the target load task in a future period, are predicted. Based on a second prediction model, and according to the current controller utilization rate, the computing power consumption of the computing equipment group during the execution of the target load task in a future period is predicted.

[0094] S302 determines the predicted power consumption based on the sum of the power consumption for computing, cooling, and other power consumption.

[0095] In some embodiments, the predicted power consumption is determined based on the sum of computing power consumption, cooling power consumption, and other power consumption as follows: the power consumption at the grid connection point satisfies the following constraints: In the formula, This refers to the combined power of the grid connection point; This is the power consumption calculated for the data center load after considering UPS power supply efficiency; This refers to the electrical power consumption of the refrigeration system. For other auxiliary power consumption of the data center; This refers to the available power distribution capacity at the grid connection point.

[0096] In the above embodiments, the power consumption equipment in the resource scheduling center is subdivided into computing power equipment groups, cooling equipment groups, and other equipment groups. The power consumption of computing power is predicted based on the current controller utilization rate, and the power consumption of cooling and other equipment is predicted based on the current ambient temperature. These predictions are then summed to obtain a high-precision total predicted power consumption. This fine-grained decoupled prediction mechanism accurately characterizes the differentiated impact of different types of loads and the external environment on energy consumption. It avoids the prediction bias caused by conflating the energy consumption of all equipment in traditional solutions, providing a more reliable data foundation for subsequent resource optimization and scheduling. This effectively supports the refined collaborative allocation of computing power, cooling, and power resources.

[0097] Figure 4 This is a flowchart illustrating the steps for determining computing power consumption in one embodiment. In this embodiment, based on the current controller utilization rate, the computing power consumption generated by the computing equipment group during the execution of the target load task in a future time period is predicted, including the following steps:

[0098] S401 obtains the idle power consumption of the computing equipment group under idle conditions, and the full-load power consumption under full-load conditions.

[0099] In some embodiments, the idle power consumption of the computing power equipment group under idle conditions and the full-load power consumption under full-load conditions are directly obtained.

[0100] Based on the current controller utilization rate, S402 predicts the controller utilization rate of the computing equipment group during the execution of the target load task in the future time period.

[0101] In some embodiments, the predicted controller utilization rate of the computing equipment group during the execution of the target load task in a future time period is predicted based on the current controller utilization rate as follows: using the predicted controller utilization rate as the main reference quantity for the machine activity level, the following calculation unit is formed. Power consumption expression: In the formula, That is, the calculation unit at time t. Total power consumption for data processing; This refers to the machine's power consumption during idle operation. This refers to the machine's electrical power consumption under full load conditions. The CPU utilization of unit u at time t; These are calibration parameters.

[0102] The above relationship is difficult to express directly in mixed integer programming (MILP) problems. Therefore, this method treats the relationship as a piecewise linear function, such as a three-segment example as follows: In the formula, That is, the total power consumption of the computing unit for data processing at time t; , It is a linear fitting constant; , This is the fitting segment threshold; Let t be the total available computing resources in the data center at time t. Based on the above formula, the Big M method can be reasonably applied to fit the power consumption level of the data center load in the MILP problem, and then incorporate it into the problem optimization.

[0103] S403 determines the computing power consumption based on idle power consumption, full-load power consumption, and predicted controller utilization.

[0104] In some embodiments, the computing power consumption is determined based on idle power consumption, full-load power consumption, and predicted controller utilization as follows:

[0105] The power supply efficiency under different load conditions is incorporated into the model. Considering that UPS power supply efficiency typically exhibits a characteristic of being low at both ends and high in the middle, this method also uses a three-segment piecewise function to fit its power supply efficiency, as shown in the following formula:

[0106]

[0107] in, This represents the total power consumption on the secondary side. This refers to the primary power consumption of the UPS. , All are fitted constants; , This is the fitting segment threshold; This represents the power consumption of the data center load after considering the UPS power supply efficiency. Based on the above formula, the power supply efficiency can be fitted in the MILP problem by reasonably applying the Big M method, and then incorporated into the problem optimization.

[0108] In the above embodiments, by introducing the baseline power consumption of the computing power equipment group under two extreme operating conditions—idle and full load—and combining it with the future time period prediction of the controller utilization rate dynamically predicted by the current controller utilization rate, a precise linearization or piecewise linearization fitting of the computing power consumption is achieved. This prediction method based on the operating condition baseline effectively captures the nonlinear relationship between computing load and energy consumption, avoiding the systematic errors introduced by the traditional scheme that simply regards computing power energy consumption as a linear model proportional to the load rate. This significantly improves the estimation accuracy of computing power consumption under dynamic load, providing a more reliable decision-making basis for subsequent computing resource allocation and overall power cost optimization.

[0109] Figure 5 This is a flowchart illustrating the temperature power prediction step in one embodiment. In this embodiment, based on the current ambient temperature, the prediction of other power consumption generated by other equipment groups during the execution of the target load task in a future time period includes the following steps:

[0110] S501 obtains the environmental parameters of the resource scheduling center.

[0111] In some embodiments, the environmental parameters of the resource scheduling center can be obtained directly.

[0112] Based on the current ambient temperature, S502 predicts the predicted cooling power of the refrigeration equipment group in the future period, as well as the predicted ambient temperature of the resource scheduling center in the future period.

[0113] In some embodiments, if the cold storage device in the prediction resource scheduling center has been activated, a first cooling power corresponding to at least one refrigeration device in a future time period is predicted based on the current ambient temperature, and the sum of different first cooling powers is used as the predicted cooling power; if the cold storage device in the prediction resource scheduling center has not been activated, the cooling power released by the cold storage device in a future time period is predicted based on the current ambient temperature, and a second cooling power corresponding to at least one refrigeration device in a future time period is also predicted; the sum of the cooling power released and different second cooling powers is used as the predicted cooling power.

[0114] For example, in scenarios without cold storage and with direct air conditioning, the power consumption of air conditioning mainly includes the power consumption of the chiller, the circulation system, and the terminal fans. As for air conditioning units... In fact, its operation must meet the following constraints: In the formula, For air conditioning units At Cooling power at any given time; , air conditioning units Minimum and maximum cooling power when the unit is powered on; For air conditioning units The status indicator is 0-1, where 0 indicates the unit is stopped and 1 indicates the unit is running. Total system cooling capacity. The following constraints must be satisfied: In the formula, For air conditioning units The cooling capacity; This refers to a collection of air conditioning units for data centers.

[0115] For example, in scenarios involving cold storage and dual-condition chillers, for air conditioning units... In fact, its operation must meet the following constraints: In the formula, For air conditioning units At Cooling power at any given time; , air conditioning units Minimum and maximum cooling power when the unit is powered on; For air conditioning units The status is indicated by 0-1, where 0 indicates the unit is stopped and 1 indicates the unit is running.

[0116] For cold storage tanks, the following constraints must be met to manage their cold storage capacity, cold release capacity, and corresponding operating conditions:

[0117]

[0118] In the formula, , These are the minimum cold storage capacity and minimum cold release capacity of the cold storage tank, respectively. , These are the 0-1 indicator variables for the cold storage and cold release conditions of the cold storage tank, respectively. , These are the cold storage capacity and the cold release capacity of the cold storage tank, respectively. , These represent the maximum cold storage capacity and the maximum cold release capacity of the cold storage tank, respectively.

[0119] In addition to the constraints mentioned above, the operating conditions of the cold storage tank also consider the following modes of operation: the refrigeration unit does not provide cooling during cold release, the cold storage tank relies on the refrigeration unit for cooling during cold storage, and the cold storage tank does not supply cooling to the terminal during cold storage, resulting in the following formula:

[0120]

[0121] In the above formula, , These are the 0-1 indicator variables for the cold storage and cold release conditions of the cold storage tank, respectively. For air conditioning units The status indicator is 0-1, where 0 indicates the unit is stopped and 1 indicates the unit is running. This refers to a collection of air conditioning units for data centers; The variable is 0-1, indicating the state of the data center air conditioning system when it is directly cooled by the refrigeration unit. 1 indicates direct cooling, and 0 indicates otherwise.

[0122] Furthermore, establish the correlation between cold storage in the cold storage tank, refrigeration power of the refrigeration machine, and system operating conditions:

[0123]

[0124] In the above formula, For air conditioning units At Cooling power at any given time; This refers to a collection of air conditioning units for data centers; The variable is 0-1, indicating the state of the data center air conditioning system when it is directly cooled by the refrigeration unit. 1 indicates direct cooling, and 0 indicates otherwise. The cold storage capacity of the cold storage tank.

[0125] The cold storage capacity of the cold storage tank satisfies the following constraints:

[0126]

[0127] In the formula, Let t be the amount of cold storage in the cold storage pool. The self-cooling coefficient per unit time of the cold storage tank; , These are the cold storage capacity and the cold release capacity of the cold storage tank, respectively. , These refer to the cold storage and cooling efficiency of the cold storage tank, respectively. , These represent the maximum and minimum cold storage capacity of the cold storage tank.

[0128] Ultimately, the total cooling capacity of the air conditioning system in scenarios with cold storage and dual-condition refrigeration units can be obtained. Satisfy the following equation: ;in, These are the cooling power of the cold storage tank, respectively; For air conditioning units The cooling capacity; This refers to a collection of air conditioning units for data centers.

[0129] S503 determines other power consumption based on environmental parameters, predicted ambient temperature, and cooling capacity.

[0130] Among them, environmental parameters include air parameters and equipment parameters; air parameters include air heat capacity and air thermal resistance; equipment parameters include equipment heat capacity and equipment thermal resistance of the equipment group.

[0131] In some embodiments, based on the second-order ETP (equivalent thermal parameter) model, the correlation between equipment heat dissipation, indoor air temperature changes, and air conditioning output cooling capacity demand caused by load processing in the data center is established, thereby realizing data center computing-cooling collaborative scheduling that takes into account indoor temperature constraints.

[0132] The second-order ETP model is shown in the following formula:

[0133]

[0134] In the formula, , That is Real-time indoor air temperature and indoor solid temperature; , These are the indoor air heat capacity and thermal resistance, respectively. , These are the indoor solid heat capacity and thermal resistance, respectively; Outdoor air temperature; Indoor heat source heat dissipation power; This refers to the overall cooling capacity of the system.

[0135] It should be noted that, in order to ensure the operation of data center IT equipment, a preset acceptable range for data center indoor temperature is established. Therefore, the following constraints apply: In the above formula That is The penalty count for exceeding the temperature limit at any given time.

[0136] For facilities like data centers, the primary source of internal heat is their IT equipment. The heat dissipation of this IT equipment and its power supply can be considered equivalent to the electrical energy it consumes. .

[0137] Based on the above model, the optimized time-of-use cooling power requirements of the data center can be obtained and the corresponding power consumption can be quantitatively assessed. In the actual scheduling of the cooling system, the specific control methods of the cooling equipment can be combined with the time-of-use cooling power requirements obtained from the model optimization, and methods such as controlling the supply / return air temperature difference at the terminal, the supply and return water temperature difference of chilled water, and the frequency of chilled water pumps can be used to guide the actual cooling output control of the cooling system.

[0138] In this embodiment, by introducing refined environmental parameters including air heat capacity and thermal resistance, as well as equipment heat capacity and thermal resistance, a predictive model based on thermodynamic mechanisms is constructed. This model enables the joint calculation of predicted ambient temperature, predicted cooling power, and power consumption of other equipment groups for future periods. This solution can quantitatively analyze the dynamic coupling relationship and thermal inertia effect between internal heat source heat dissipation, cooling output, and the indoor thermal environment. This allows for the accurate identification and calculation of the energy consumption components of other auxiliary equipment while ensuring compliance with equipment operating temperature regulations. Compared to traditional solutions that treat the energy consumption of other equipment as a constant or simply correlate it with the main equipment, this solution significantly improves the accuracy of characterizing auxiliary energy consumption in the complex thermal environment of data centers. This lays a solid quantitative foundation for subsequent refined collaborative scheduling of computing power, cooling, and auxiliary power consumption across the entire chain, effectively avoiding over-cooling or inaccurate estimation of auxiliary equipment energy consumption due to misjudgment of the thermal environment.

[0139] Figure 6 This is a flowchart illustrating the steps for determining cooling power consumption in one embodiment. In this embodiment, predicting the cooling power consumption generated by the cooling equipment group during the execution of a target load task in a future time period includes the following steps:

[0140] S601 acquires energy efficiency characteristic data of different refrigeration equipment in the refrigeration equipment group.

[0141] In some embodiments, energy efficiency characteristic data of different refrigeration devices in the refrigeration equipment group are directly obtained.

[0142] S602 determines the power consumption for cooling based on energy efficiency data and predicted cooling power.

[0143] In some embodiments, the cooling power consumption is determined based on energy efficiency characteristic data and predicted cooling power as follows: the cooling power consumption satisfies the following formula: In the formula, This refers to the total electrical power consumption of the refrigeration system, where the power consumption of the chiller can be based on the unit's COP-PLR characteristic curve. and units Real-time cooling power calculation , This is a function of the power consumption of the circulation system, terminal fans, and other equipment, relative to the total cooling power of the system. In this method, , All of them used piecewise linear fitting, which will not be elaborated here.

[0144] In the above embodiments, by introducing energy efficiency characteristic data (such as COP-PLR characteristic curves) of different refrigeration equipment and combining them with the predicted system cooling power, accurate quantification of cooling power consumption is achieved. This scheme can convert abstract cooling power demand into specific power consumption values ​​based on the actual energy efficiency levels of different refrigeration equipment under specific load rates, effectively capturing the nonlinear energy consumption characteristics of refrigeration units under different operating conditions. Compared to the traditional scheme that simply considers cooling energy consumption as a linear model proportional to cooling capacity, this scheme significantly improves the estimation accuracy of overall power consumption under the combined operation of multiple types of refrigeration equipment, providing crucial quantitative support for the subsequent scientific allocation of cooling tasks among different energy-efficient equipment and optimization of overall power costs while ensuring temperature constraints.

[0145] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0146] Based on the same inventive concept, this application also provides a resource scheduling apparatus for implementing the resource scheduling method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more resource scheduling apparatus embodiments provided below can be found in the limitations of the resource scheduling method described above, and will not be repeated here.

[0147] In one exemplary embodiment, such as Figure 7 As shown, a resource scheduling device is provided, including: an acquisition module 10, a prediction module 11, a determination module 12, and a scheme module 13, wherein:

[0148] Module 10 is used to obtain the target load tasks to be executed in the resource scheduling center, the current ambient temperature and the current controller utilization rate in the current time period;

[0149] Prediction module 11 is used to predict the predicted power consumption, temperature penalty value, and amount of uncompleted data generated by the resource scheduling center in executing the target load task in a future time period based on the current ambient temperature and the current controller utilization rate; the predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of the target load task;

[0150] The determining module 12 is used to determine the cumulative penalty amount of the resource scheduling center in the future time period based on the temperature penalty value and the amount of incomplete data; and to determine the electricity cost of the resource scheduling center in the future time period based on the predicted power consumption and the preset time-of-use electricity price.

[0151] Scheme module 13 is used to generate a target scheduling scheme for the resource scheduling center based on the cumulative number of penalties and the electricity cost; wherein, the target scheduling scheme is used to allocate computing resources for different target load tasks in the future time period, and to allocate cooling resources and electricity resources for each power-consuming equipment group in the resource scheduling center in the future time period.

[0152] In some embodiments, the prediction module 11 is further configured to predict, based on the current ambient temperature, other power consumption generated by other equipment groups during the execution of the target load task in a future period, and cooling power consumption generated by the cooling equipment group during the execution of the target load task in a future period; and computing power consumption generated by the computing equipment group during the execution of the target load task in a future period based on the current controller utilization rate; and to determine the predicted power consumption based on the sum of computing power consumption, cooling power consumption, and other power consumption.

[0153] In some embodiments, the prediction module 11 is further configured to obtain the idle power consumption of the computing power equipment group under idle conditions and the full-load power consumption under full-load conditions; predict the predicted controller utilization rate of the computing power equipment group during the execution of the target load task in a future period based on the current controller utilization rate; and determine the computing power consumption based on the idle power consumption, the full-load power consumption and the predicted controller utilization rate.

[0154] In some embodiments, the prediction module 11 is further configured to obtain environmental parameters of the resource scheduling center; predict the predicted cooling power of the cooling equipment group in a future time period based on the current ambient temperature, and the predicted ambient temperature of the resource scheduling center in the future time period; and determine other power consumption based on the environmental parameters, the predicted ambient temperature, and the cooling power; wherein the environmental parameters include air parameters and equipment parameters; the air parameters include air heat capacity and air thermal resistance; and the equipment parameters include the equipment heat capacity and equipment thermal resistance of the equipment group.

[0155] In some embodiments, the prediction module 11 is further configured to, when the cold storage device in the prediction resource scheduling center has been started, predict the first cooling power corresponding to at least one refrigeration device in a future time period based on the current ambient temperature, and use the sum of different first cooling powers as the predicted cooling power; when the cold storage device in the prediction resource scheduling center has not been started, predict the cooling power released by the cold storage device in a future time period based on the current ambient temperature, and the second cooling power corresponding to at least one refrigeration device in a future time period; and use the sum of the cooling power released and different second cooling powers as the predicted cooling power.

[0156] In some embodiments, the prediction module 11 is further configured to acquire energy efficiency characteristic data of different refrigeration devices in the refrigeration equipment group; and determine the power consumption for refrigeration based on the energy efficiency characteristic data and the predicted refrigeration power.

[0157] Each module in the aforementioned resource scheduling device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0158] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 8 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements a resource scheduling method.

[0159] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0160] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0161] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0162] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0163] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0164] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0165] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0166] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A resource scheduling method, characterized in that, The method includes: Obtain the target load tasks to be executed in the resource scheduling center, the current ambient temperature, and the current controller utilization rate in the current time period; Based on the current ambient temperature and the current controller utilization rate, predict the predicted power consumption, temperature penalty value, and amount of incomplete data generated by the resource scheduling center in executing the target load task in the future time period; the predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of the target load task; Based on the temperature penalty value and the amount of incomplete data, determine the cumulative penalty amount for the resource scheduling center in future time periods; and, Based on the predicted power consumption and the preset time-of-use electricity price, the electricity cost of the resource dispatch center in the future time period is determined; Based on the cumulative number of penalties and the electricity cost, a target scheduling scheme for the resource scheduling center is generated. The target scheduling scheme is used to allocate computing resources for different target load tasks in future time periods, and to allocate cooling resources and power resources for each power equipment group in the resource scheduling center in future time periods.

2. The method according to claim 1, characterized in that, The power-consuming equipment group includes a computing power equipment group, a cooling equipment group, and other equipment groups in the resource scheduling center besides the computing power equipment group and the cooling equipment group; the step of predicting the predicted power consumption generated by the resource scheduling center in the future time period when executing the target load task based on the current ambient temperature and the current controller utilization includes: Based on the current ambient temperature, predict the additional power consumption generated by the other equipment group during the execution of the target load task in the future time period, and the cooling power consumption generated by the cooling equipment group during the execution of the target load task in the future time period; and Based on the current controller utilization rate, predict the computing power consumption of the computing equipment group during the execution of the target load task in the future time period; The predicted power consumption is determined based on the sum of the computing power consumption, the cooling power consumption, and the other power consumption.

3. The method according to claim 2, characterized in that, The step of predicting the power consumption of the computing equipment group during the execution of the target load task in a future time period based on the current controller utilization includes: Obtain the idle power consumption of the computing power equipment group under idle conditions, and the full-load power consumption under full-load conditions; Based on the current controller utilization rate, predict the predicted controller utilization rate of the computing power equipment group during the execution of the target load task in a future time period; The computing power consumption is determined based on the idle power consumption, the full-load power consumption, and the predicted controller utilization rate.

4. The method according to claim 2, characterized in that, Based on the current ambient temperature, predict the additional power consumption generated by the other equipment group during the execution of the target load task in the future, including: Obtain the environmental parameters of the resource scheduling center; Based on the current ambient temperature, predict the predicted cooling power of the refrigeration equipment group in the future time period, and predict the ambient temperature of the resource scheduling center in the future time period; The other power consumption is determined based on the environmental parameters, the predicted ambient temperature, and the cooling power. The environmental parameters include air parameters and equipment parameters; the air parameters include air heat capacity and air thermal resistance; the equipment parameters include the equipment heat capacity and equipment thermal resistance of the equipment group.

5. The method according to claim 4, characterized in that, The refrigeration equipment group includes at least one refrigeration device; based on the current ambient temperature, predicting the predicted refrigeration power of the refrigeration equipment group for a future time period includes: If it is predicted that the cold storage equipment in the resource scheduling center has been started, based on the current ambient temperature, the first cooling power of at least one cooling equipment in the future time period is predicted, and the sum of different first cooling powers is taken as the predicted cooling power. If it is predicted that the cold storage equipment in the resource scheduling center will not be started, the cold power of the cold storage equipment in the future period is predicted based on the current ambient temperature, and the second cooling power of at least one refrigeration equipment in the future period is predicted. The sum of the cooling power and the different second cooling powers is taken as the predicted cooling power.

6. The method according to any one of claims 4-5, characterized in that, Predicting the cooling power consumption generated by the cooling equipment group during the execution of the target load task in the future, including: Obtain energy efficiency characteristic data of different refrigeration devices in the refrigeration equipment group; The power consumption for cooling is determined based on the energy efficiency characteristic data and the predicted cooling power.

7. A resource scheduling device, characterized in that, The device includes: The acquisition module is used to acquire the target load tasks to be executed by the resource scheduling center, the current ambient temperature and the current controller utilization rate in the current time period; The prediction module is used to predict the predicted power consumption, temperature penalty value, and amount of uncompleted data generated by the resource scheduling center in executing the target load task in a future time period, based on the current ambient temperature and the current controller utilization rate; the predicted power consumption is the power consumption generated by different groups of electrical equipment in the resource scheduling center during the execution of the target load task; The determination module is used to determine the cumulative penalty amount of the resource scheduling center in the future time period based on the temperature penalty value and the amount of incomplete data; and to determine the electricity cost of the resource scheduling center in the future time period based on the predicted power consumption and the preset time-of-use electricity price. The scheme module is used to generate a target scheduling scheme for the resource scheduling center based on the cumulative number of penalties and the electricity cost; wherein, the target scheduling scheme is used to allocate computing resources for different target load tasks in the future time period, and to allocate cooling resources and electricity resources for each group of electrical equipment in the resource scheduling center in the future time period.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images