A cloud computing platform task scheduling method, device and equipment

Abstract

The application provides a cloud computing platform task scheduling method, device and equipment, and the scheme selects appropriate sorting parameters based on the cluster type of the scheduling task, so that when scheduling is performed based on the sorting result, compared with the list heuristic, the parallel tasks can be processed in clusters, and the execution time is reasonably reduced; compared with the clustering heuristic, the tasks can be pre-sorted, the execution order is more reasonable, and the waste of idle time resources is reduced.

Description

Technical Field

[0001] This invention relates to the field of refrigeration equipment technology, and more specifically to a task scheduling method, apparatus, and equipment for a cloud computing platform. Background Technology

[0002] Cloud computing platform scheduling generally refers to resource scheduling and task scheduling. Resource scheduling refers to the rational and effective allocation of physical resources; task scheduling refers to the rational allocation of tasks to appropriate computing resources. Due to the variability and unpredictability of the task environment, the system's task requirements are dynamic, making it challenging to effectively schedule tasks to meet the system's fast response requirements. Meanwhile, the energy consumption of cloud computing platforms is also a crucial factor to consider when scheduling tasks. With the continuous expansion of cloud computing data centers, their energy consumption is increasing exponentially. Reducing system energy consumption has gradually become a critical challenge that cloud computing data centers urgently need to address. Furthermore, reducing data center energy consumption is also of great significance for achieving carbon neutrality.

[0003] Currently, energy consumption reduction is mainly achieved by dynamically adjusting the operating power of equipment. The drawback of this method is that it affects overall system performance, thus impacting system response speed. Due to the randomness of tasks, the system lacks prior information when performing multi-task scheduling; therefore, the scheduling system must adaptively schedule tasks. After scheduling is completed, the system's scheduling performance, such as system response speed and platform energy consumption, is evaluated and fed back to the platform scheduling center to improve performance and achieve the optimal multi-task scheduling strategy.

[0004] Among existing technical solutions, the main methods for implementing task scheduling on cloud computing platforms are:

[0005] List heuristics prioritizes and schedules tasks for execution. Clustering heuristic is another commonly used method. Its main idea is to process tasks in the same cluster, reduce parallelism by sorting parallel tasks, thereby increasing parallel execution time, and reduce communication latency between tasks by sacrificing parallelism.

[0006] List heuristics perform well with limited resources, but the need to predetermine priorities and resource selection strategies can increase task completion time and slow down response speed. Clustering heuristics can handle cases with unlimited resources and reduce communication latency, but they can lead to problems such as load imbalance and wasted resources during idle time. Summary of the Invention

[0007] In view of this, embodiments of the present invention provide a cloud computing platform task scheduling method that can comprehensively consider scheduling energy consumption and scheduling time, thereby solving the problems of slow response speed and waste of idle time resources in existing scheduling methods.

[0008] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0009] A task scheduling method for a cloud computing platform, comprising:

[0010] Obtain the target scheduling task;

[0011] Determine the cluster type of the target task, whereby the cluster type includes linear clusters and nonlinear clusters;

[0012] Extract the sorting parameters corresponding to the target task based on the cluster type of the target task;

[0013] The target tasks are sorted based on the sorting parameters using a preset sorting rule;

[0014] The sorted target tasks are scheduled based on the sorting results.

[0015] Optionally, in the above-mentioned cloud computing platform task scheduling method, the sorting parameters include duration and energy consumption. When the cluster type of the target task is a linear cluster, the duration refers to the longest execution time corresponding to the target task, and the energy consumption refers to the total energy consumption corresponding to the target task.

[0016] When the cluster type of the target task is a non-linear cluster, the duration refers to the execution duration of the target task, and the energy consumption refers to the execution energy consumption of the target task.

[0017] Optionally, in the above-mentioned cloud computing platform task scheduling method, the target tasks are sorted based on the sorting parameters using a preset sorting rule, including:

[0018] The target tasks are sorted according to their energy consumption. The lower the energy consumption, the higher the priority. For target tasks with the same energy consumption, the longer the duration, the higher the priority. When the cluster type of the target tasks is a non-linear cluster, the sorting results and the target tasks are imported into the cluster that the target tasks are adapted to.

[0019] Optionally, in the above-mentioned cloud computing platform task scheduling method, the target tasks are sorted based on the sorting parameters using a preset sorting rule, including:

[0020] The target tasks are sorted based on their duration to obtain a first sorting result;

[0021] The target tasks are sorted based on energy consumption to obtain a second sorting result;

[0022] The target tasks are sorted by the average of the first and second sorting results to obtain a third sorting result;

[0023] If two target tasks have the same third ranking result, the target task with higher priority in the second ranking result will have higher priority in the third ranking result.

[0024] When the cluster type of the target task is a non-linear cluster, the sorting results and the target task are imported into the cluster that the target task is adapted to.

[0025] Optionally, in the above-mentioned cloud computing platform task scheduling method, scheduling the sorted target tasks based on the sorting results includes: using a genetic algorithm to schedule tasks within the current scheduling period based on the sorting results.

[0026] Optionally, in the above-mentioned task scheduling method for cloud computing platforms, the benefit function of the genetic algorithm is:

[0027] F(j)=FitnessTime(j)+1 / s(j);

[0028] in, The subTask j，i The energy consumption of the i-th scheduling task allocated to the j-th physical resource is given by avgTask, where avgTask is the average energy consumption of all scheduling tasks, and n is the total number of scheduling tasks corresponding to the chromosome.

[0029] FitnessTime is a time-dependent chromosome benefit function.

[0030] A task scheduling device for a cloud computing platform, comprising:

[0031] The scheduling task acquisition unit is used to acquire target scheduling tasks;

[0032] A cluster type determination unit is used to determine the cluster type of the target task, wherein the cluster type includes linear clusters and nonlinear clusters;

[0033] A sorting unit is used to extract sorting parameters corresponding to the target task based on the cluster type of the target task; and to sort the target task based on the sorting parameters using a preset sorting rule;

[0034] The scheduling unit is used to schedule the sorted target tasks based on the sorting results.

[0035] Optionally, in the above-mentioned cloud computing platform task scheduling device, the sorting parameters include duration and energy consumption;

[0036] When the cluster type of the target task is a linear cluster, the duration refers to the longest execution time corresponding to the target task, and the energy consumption refers to the total energy consumption corresponding to the target task;

[0037] When the cluster type of the target task is a non-linear cluster, the duration refers to the execution duration of the target task, and the energy consumption refers to the execution energy consumption of the target task.

[0038] Optionally, in the above-mentioned cloud computing platform task scheduling device, when the sorting unit sorts the target tasks based on the sorting parameters using preset sorting rules, it is specifically used for:

[0039] The target tasks are sorted according to their energy consumption. The lower the energy consumption, the higher the priority. For target tasks with the same energy consumption, the longer the duration, the higher the priority. When the cluster type of the target tasks is a non-linear cluster, the sorting results and the target tasks are imported into the cluster that the target tasks are adapted to.

[0040] A task scheduling device for a cloud computing platform, comprising:

[0041] Memory and processor;

[0042] The memory is used to store programs;

[0043] The processor is used to execute the program to implement each step of the cloud computing platform task scheduling method described above.

[0044] Based on the above technical solutions, the solutions provided in the embodiments of the present invention, by selecting appropriate sorting parameters based on the cluster type of the scheduling tasks, enable the cluster processing of parallel tasks and reasonably reduce the execution time when scheduling based on the sorting results, compared with the list heuristic; compared with the clustering heuristic, the tasks can be pre-sorted, making their execution order more reasonable and reducing the waste of idle time resources. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0046] Figure 1 This is a flowchart illustrating a task scheduling method for a cloud computing platform disclosed in an embodiment of this application;

[0047] Figure 2 This is a schematic diagram of the structure of a cloud computing platform task scheduling device disclosed in an embodiment of this application;

[0048] Figure 3 This is a schematic diagram of the scheduling process of a genetic algorithm;

[0049] Figure 4 This is a schematic diagram of the structure of a cloud computing platform task scheduling device disclosed in an embodiment of this application. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] First, let me explain some of the technical terms used in this application:

[0052] Nonlinear cluster: If there are two independent neighboring tasks in the same cluster, it is called a nonlinear cluster.

[0053] Linear clusters: A concept relative to nonlinear clusters. A linear cluster is a group of tasks in which no two neighboring tasks are independent.

[0054] This application proposes a prediction-based task scheduling algorithm to balance system response speed and platform energy consumption by allocating tasks to reasonable computing resources. The basic scheme consists of two parts: task preprocessing and adaptive scheduling. For task preprocessing, for linear clusters, the maximum task execution time and energy consumption are calculated; for nonlinear clusters, priority sorting is used to reduce parallelism. A list of tasks is obtained and sorted. After obtaining the ordered task sequence through task preprocessing, a benefit function that minimizes energy consumption is constructed, and a genetic algorithm is used to schedule and execute the tasks, achieving a balance between system response speed and platform energy consumption.

[0055] For details, see Figure 1 The specific process of task scheduling on the cloud computing platform disclosed in this application embodiment may include:

[0056] Step S101: Obtain the target scheduling task;

[0057] The target task is all the tasks that need to be scheduled.

[0058] Step S102: Determine the cluster type of the target task, where the cluster type includes linear clusters and nonlinear clusters.

[0059] After obtaining the target task, its cluster type is determined, specifically whether it is a linear or non-linear cluster. This determination can be made by checking if the target task has neighboring tasks. If the target task has neighboring tasks, its cluster type is linear; otherwise, it is non-linear. In other words, if the target task has neighboring tasks, and the number of neighboring tasks is greater than or equal to one, the target task's cluster type is linear; otherwise, it is non-linear.

[0060] Step S103: Extract the sorting parameters corresponding to the target task based on the cluster type of the target task.

[0061] In the technical solutions disclosed in the embodiments of this application, the cluster types of the target tasks are different, and the sorting parameters used when sorting the target tasks are different. The different values ​​of the sorting parameters for different target tasks will result in different priorities of the target tasks in the sorting results.

[0062] In the technical solutions disclosed in the embodiments of this application, the sorting parameters may include duration and energy consumption. When the cluster type of the target task is a linear cluster and a nonlinear cluster, the specific referents of duration and energy consumption are different. Specifically, when the cluster type of the target task is a linear cluster, the duration refers to the longest execution time corresponding to the target task, and the energy consumption refers to the total energy consumption corresponding to the target task. The longest execution time and total energy consumption refer to the execution time and energy consumption of a single scheduled task. When the cluster type of the target task is a nonlinear cluster, the duration refers to the execution duration of the target task, and the energy consumption refers to the execution energy consumption of the target task.

[0063] Step S104: Sort the target tasks according to the sorting parameters using a preset sorting rule.

[0064] When the cluster type of the target task is a linear cluster, this step specifically involves calculating the longest execution time and total energy consumption corresponding to the target task, sorting them according to a preset sorting rule, and then executing them in the cluster.

[0065] When the cluster type of the target task is a non-linear cluster, this step specifically involves sorting the target tasks according to a preset sorting rule based on their execution time and energy consumption to obtain an ordered task sequence.

[0066] Step S105: Schedule the sorted target tasks based on the sorting results.

[0067] In this scheme, steps S101-S104 are the preprocessing part of this scheme, which is performed in the preprocessing module, and step S105 is the adaptive scheduling part of this application, which is performed in the adaptive scheduling module.

[0068] In this step, after the target tasks are sorted, the adaptive scheduling module divides the time into time slices to process the target tasks to be scheduled. Within the scheduling period, the corresponding tasks are scheduled using relevant scheduling algorithms. For tasks that cannot be processed within the current scheduling period, the delayed tasks are stored in the delayed execution matrix and given priority consideration in the next scheduling period.

[0069] The technical solutions disclosed in the above embodiments of this application, by selecting appropriate sorting parameters based on the cluster type of the scheduling task, enable the cluster processing of parallel tasks and reasonably reduce execution time when scheduling based on the sorting results, compared with the list heuristic; compared with the clustering heuristic, the tasks can be pre-sorted, making their execution order more reasonable and reducing the waste of idle time resources.

[0070] Specifically, in this scheme, when scheduling the sorted target tasks based on the ranking results, the scheduling algorithm used can be a genetic algorithm. The Genetic Algorithm (GA) was first proposed by John Holland in the 1970s. This algorithm is designed based on the evolutionary laws of organisms in nature. It is a computational model that simulates the biological evolutionary process of natural selection and genetic mechanisms in Darwin's theory of evolution, and is a method for searching for optimal solutions by simulating the natural evolutionary process. This algorithm uses mathematical methods and computer simulation to transform the problem-solving process into processes similar to the crossover and mutation of chromosome genes in biological evolution. When solving relatively complex combinatorial optimization problems, it can usually obtain better optimization results faster than some conventional optimization algorithms. Genetic algorithms have been widely used in combinatorial optimization, machine learning, signal processing, adaptive control, and artificial life. Using a genetic algorithm to schedule target tasks within the current scheduling cycle can comprehensively consider task execution time and energy consumption through the benefit function of the genetic algorithm, achieving a balance between system response speed and energy consumption.

[0071] See Figure 2 The scheduling process of the genetic algorithm for the target task can be summarized as follows:

[0072] Step S201: Chromosome encoding.

[0073] Chromosomes are encoded using a resource-target task correspondence method.

[0074] Step S202: Initial population generation.

[0075] Assuming a population size of S, a target task number of M to be scheduled in the current scheduling cycle, and a resource number of R, then initialize and generate S chromosomes with a length of M and gene values ​​as random integers between [1, R].

[0076] Step S203: Fitness function.

[0077] In this step, the standard deviation of the energy consumption of tasks allocated across various resources on the chromosome is used to measure the energy consumption problem. The fitness function for the standard deviation of the energy consumption of tasks allocated to physical resources on the chromosome is:

[0078]

[0079] Wherein, subTaskj,i is the energy consumption of target task i allocated to the j-th physical resource, j represents the physical resource, avgTask is the average energy consumption of all scheduled tasks (the target task to be scheduled) on the chromosome, and n is the total number of scheduled tasks corresponding to the chromosome.

[0080] The benefit function of the genetic algorithm is defined as follows:

[0081] F(j) = FitnessTime(j) + 1 / s(j)

[0082] Where FitnessTime is a time-dependent chromosome benefit function.

[0083] The benefit function comprehensively considers task execution time and energy consumption to derive the optimal or near-optimal solution. The "time-related chromosome benefit function" refers to a function that uses task execution time as the independent variable and reflects how task execution time affects chromosome performance. The general principle is that the shorter the execution time, the higher the benefit. It can be defined as the reciprocal of the execution times of all tasks (ranked from shortest to longest).

[0084] In principle, the benefit function is oriented towards shorter task duration and lower energy consumption relative to average energy consumption, thus resulting in a larger benefit function. The genetic algorithm's concept originates from the natural selection principle of survival of the fittest in biological evolution, searching for optimal or near-optimal solutions through evolution. Since task duration and energy consumption are coupled, with longer durations often resulting in higher energy consumption, the benefit function can comprehensively consider both task execution duration and energy consumption, achieving a balance between system response speed and energy consumption.

[0085] Step S204: Crossover Mutation

[0086] Based on the fitness values ​​of each individual, superior individuals are selected from the previous generation and passed on to the next generation according to certain rules or methods. KTLBCGA uses roulette wheel selection as the selection operator, and the iteration of the genetic algorithm terminates after a certain number of generations.

[0087] The technical solution disclosed in this embodiment provides two preset sorting rules, which can be used to sort the target task. One of the preset sorting rules is:

[0088] The target tasks are sorted according to their energy consumption. The lower the energy consumption, the higher the priority. For target tasks with the same energy consumption, the longer the duration, the higher the priority. When the cluster type of the target tasks is a non-linear cluster, the sorting results and the target tasks are imported into the cluster that the target tasks are adapted to.

[0089] Another preset sorting rule includes:

[0090] The target tasks are sorted based on their duration to obtain a first sorting result;

[0091] The target tasks are sorted based on energy consumption to obtain a second sorting result;

[0092] The target tasks are sorted by calculating the average of the first and second sorting results to obtain a third sorting result. The average of the first sorting results may refer to the average of the sorted names of the target tasks in the first and second sorting results.

[0093] If two target tasks have the same third ranking result, the target task with higher priority in the second ranking result will have higher priority in the third ranking result.

[0094] When the cluster type of the target task is a linear cluster, the sorting results and the target task are imported into the cluster that the target task is adapted to.

[0095] This embodiment discloses a task scheduling device for a cloud computing platform. For the specific working content of each unit in the device, please refer to the content of the above method embodiment.

[0096] The following describes the cloud computing platform task scheduling device provided in the embodiments of the present invention. The cloud computing platform task scheduling device described below can be referred to in correspondence with the cloud computing platform task scheduling method described above.

[0097] See Figure 3 The cloud computing platform task scheduling device disclosed in this application embodiment may include:

[0098] Scheduling task acquisition unit A is used to acquire target scheduling tasks;

[0099] Cluster type determination unit B is used to determine the cluster type of the target task, wherein the cluster type includes linear clusters and nonlinear clusters;

[0100] Sorting unit C is used to extract sorting parameters corresponding to the target task based on the cluster type of the target task; and sort the target task based on the sorting parameters using a preset sorting rule;

[0101] The scheduling unit D is used to schedule the sorted target tasks based on the sorting results.

[0102] Corresponding to the above method, when the sorting unit sorts the target tasks based on the sorting parameters using preset sorting rules, it is specifically used for:

[0103] The target tasks are sorted according to their energy consumption. The lower the energy consumption, the higher the priority. For target tasks with the same energy consumption, the longer the duration, the higher the priority. When the cluster type of the target tasks is a non-linear cluster, the sorting results and the target tasks are imported into the cluster that the target tasks are adapted to.

[0104] Figure 4 For a hardware structure diagram of the cloud computing platform task scheduling device provided in this embodiment of the invention, please refer to... Figure 4 As shown, it may include: at least one processor 100, at least one communication interface 200, at least one memory 300 and at least one communication bus 400;

[0105] In this embodiment of the invention, the number of processor 100, communication interface 200, memory 300, and communication bus 400 is at least one, and the processor 100, communication interface 200, and memory 300 communicate with each other through communication bus 400; obviously, Figure 4 The communication connections shown for the processor 100, communication interface 200, memory 300, and communication bus 400 are optional.

[0106] Optionally, the communication interface 200 can be an interface of a communication module, such as the interface of a GSM module;

[0107] Processor 100 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0108] The memory 300 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0109] Specifically, processor 100 is used for:

[0110] Obtain the target scheduling task;

[0111] Determine the cluster type of the target task, whereby the cluster type includes linear clusters and nonlinear clusters;

[0112] Extract the sorting parameters corresponding to the target task based on the cluster type of the target task;

[0113] The target tasks are sorted based on the sorting parameters using a preset sorting rule;

[0114] The sorted target tasks are scheduled based on the sorting results.

[0115] The processor is also used to perform other steps disclosed in the above method embodiments, which will not be repeated here.

[0116] For ease of description, the above system is described by dividing it into various modules based on their functions. Of course, in implementing this invention, the functions of each module can be implemented in one or more software and / or hardware components.

[0117] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and relevant parts can be referred to the descriptions in the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0118] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0119] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0120] It should also be noted that, in this document, relational terms such as "first" and "second" are used only 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. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0121] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A cloud computing platform task scheduling method, characterized in that, include: Obtain the target scheduling task; The cluster type of the target task is determined. The cluster type includes linear clusters and nonlinear clusters. A nonlinear cluster is defined as having two independent neighbor tasks in the same cluster, while a linear cluster is defined as having no two independent neighbor tasks in the same cluster. Extract the sorting parameters corresponding to the target task based on the cluster type of the target task; The target tasks are sorted based on the sorting parameters using a preset sorting rule; The sorted target tasks are scheduled based on the sorting results; The sorting parameters include duration and energy consumption. When the cluster type of the target task is a linear cluster, the duration refers to the longest execution time corresponding to the target task, and the energy consumption refers to the total energy consumption corresponding to the target task. When the cluster type of the target task is a non-linear cluster, the duration refers to the execution duration of the target task, and the energy consumption refers to the execution energy consumption of the target task; The scheduling of the sorted target tasks based on the sorting results includes: using a genetic algorithm to schedule tasks within the current scheduling period based on the sorting results; The benefit function of the genetic algorithm is: ； in, The subTask j，i The energy consumption of the i-th scheduling task allocated to the j-th physical resource is given by avgTask, where avgTask is the average energy consumption of all scheduling tasks, and n is the total number of scheduling tasks corresponding to the chromosome. FitnessTime is a time-dependent chromosome benefit function.

2. The task scheduling method for a cloud computing platform according to claim 1, characterized in that, Sort the target tasks according to the sorting parameters using a preset sorting rule, including: The target tasks are sorted according to their energy consumption. The lower the energy consumption, the higher the priority. For target tasks with the same energy consumption, the longer the duration, the higher the priority. When the cluster type of the target tasks is a linear cluster, the sorting results and the target tasks are imported into the cluster that the target tasks are adapted to.

3. The task scheduling method for a cloud computing platform according to claim 1, characterized in that, Sort the target tasks according to the sorting parameters using a preset sorting rule, including: The target tasks are sorted based on their duration to obtain a first sorting result; The target tasks are sorted based on energy consumption to obtain a second sorting result; The target tasks are sorted by the average of the first and second sorting results to obtain a third sorting result; If two target tasks have the same third ranking result, the target task with higher priority in the second ranking result will have higher priority in the third ranking result. When the cluster type of the target task is a linear cluster, the sorting results and the target task are imported into the cluster that the target task is adapted to.

4. A task scheduling device for a cloud computing platform, characterized in that, include: The scheduling task acquisition unit is used to acquire target scheduling tasks; A cluster type determination unit is used to determine the cluster type of the target task, wherein the cluster type includes linear clusters and nonlinear clusters; A sorting unit is used to extract sorting parameters corresponding to the target task based on the cluster type of the target task; and to sort the target task based on the sorting parameters using a preset sorting rule; The scheduling unit is used to schedule the sorted target tasks based on the sorting results. The sorting parameters include duration and energy consumption; When the cluster type of the target task is a linear cluster, the duration refers to the longest execution time corresponding to the target task, and the energy consumption refers to the total energy consumption corresponding to the target task; When the cluster type of the target task is a non-linear cluster, the duration refers to the execution duration of the target task, and the energy consumption refers to the execution energy consumption of the target task; Specifically, the scheduling unit includes: using a genetic algorithm to schedule tasks within the current scheduling period based on the sorting results; The benefit function of the genetic algorithm is: ； in, The subTask j，i The energy consumption of the i-th scheduling task allocated to the j-th physical resource is given by avgTask, where avgTask is the average energy consumption of all scheduling tasks, and n is the total number of scheduling tasks corresponding to the chromosome. FitnessTime is a time-dependent chromosome benefit function.

5. The cloud computing platform task scheduling device according to claim 4, characterized in that, When the sorting unit sorts the target tasks based on the sorting parameters using a preset sorting rule, it is specifically used for: The target tasks are sorted according to their energy consumption. The lower the energy consumption, the higher the priority. For target tasks with the same energy consumption, the longer the duration, the higher the priority. When the cluster type of the target tasks is a non-linear cluster, the sorting results and the target tasks are imported into the cluster that the target tasks are adapted to.

6. A task scheduling device for a cloud computing platform, characterized in that, include: Memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the cloud computing platform task scheduling method according to any one of claims 1-3.

Images