Graph crossbreeding optimization algorithm-based computing power scheduling method and system
By constructing a weighted directed graph through a graph hybridization breeding optimization algorithm, computing tasks are guided to migrate and schedule among computing power nodes. This solves the problems of low efficiency and uneven resource utilization in existing computing power scheduling methods, and realizes efficient and adaptive allocation of computing resources and task scheduling.
Patent Information
- Application Number
- CN202510942979.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-09
- Publication Date
- 2025-11-07
AI Technical Summary
Existing computing power scheduling methods suffer from low scheduling efficiency and uneven resource utilization when dealing with complex tasks and dynamic environments. They also struggle to handle task dependencies. Traditional intelligent optimization algorithms have high computational complexity and slow convergence speed when dealing with large-scale, high-dimensional scheduling problems, and lack adaptability and intelligence.
A graph-based hybridization breeding optimization algorithm is adopted. By constructing a weighted directed graph, the computational tasks are guided to migrate and schedule among computing nodes. The scheduling and allocation scheme is optimized by using hybridization and self-pollination operations of maintainer lines, restorer lines and sterile lines to achieve the global optimal solution.
It improves the efficiency of computing power scheduling and resource utilization, can quickly respond to changes in system load and resource status, reduce computing costs, ensure timely task completion, and improve system performance.
Smart Images
Figure CN120909768A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of graph hybrid breeding optimization algorithm, and particularly relates to a graph hybrid breeding optimization algorithm computing power scheduling method and system. BACKGROUND
[0002] In modern computing environments, computing power scheduling is a core problem in the fields of cloud computing, big data processing, and distributed computing. As the complexity and scale of computing tasks increase, how to efficiently allocate and schedule computing resources to meet task requirements and optimize resource utilization becomes a key problem that needs to be solved. Traditional computing power scheduling methods are mainly based on rules or simple heuristic algorithms, such as round-robin scheduling and least connection scheduling. These methods are simple to implement, but when dealing with complex tasks and dynamic environments, they often exhibit low scheduling efficiency and uneven resource utilization, and are difficult to handle task dependencies. In recent years, intelligent optimization algorithms such as genetic algorithms and particle swarm optimization have been introduced into the field of computing power scheduling. These algorithms simulate natural evolution or physical processes and can optimize task scheduling schemes to some extent. However, these algorithms still have limitations such as high computational complexity, slow convergence speed, and the need for a large number of parameter adjustments and manual intervention, lacking adaptability and intelligence.
[0003] Therefore, in view of the current task dependency complexity and frequent resource dynamic changes in computing power scheduling, a new optimization method needs to be designed to cope with complex task dependencies and dynamic changes in the computing environment. SUMMARY
[0004] The present application provides a graph hybrid breeding optimization algorithm computing power scheduling method and system, which can efficiently schedule computing tasks among multiple computing nodes and achieve reasonable allocation of computing resources through optimization algorithms to maximize computing power utilization and minimize task execution time. In a first aspect, a graph hybrid breeding optimization algorithm computing power scheduling method is provided, comprising the following steps: Obtaining a plurality of computing tasks to be executed; Obtaining a plurality of computing power nodes for computing and processing all the computing tasks; Guiding each computing task to migrate and schedule in each computing power node based on a graph hybrid breeding optimization algorithm, so that each computing task is matched with an optimal computing power node, and a global final computing power scheduling result is obtained; All computing power nodes perform computing and processing on all computing tasks according to the global final computing power scheduling result.
[0005] In some embodiments, the graph hybridization breeding optimization algorithm guides the migration and scheduling of each computing task in each computing power node, so that each computing task corresponds to an optimal computing power node, and a global final computing power scheduling result is obtained, including: initially migrating each computing task to each computing power node to generate an initial computing power scheduling result including multiple scheduling allocation schemes; calculating the fitness value of each scheduling allocation scheme in the initial computing power scheduling result based on a scheduling fitness function, and sorting all scheduling allocation schemes from high to low according to the fitness value to divide into three equal parts in turn, namely a maintenance line, a recovery line and a sterile line; optimizing the scheduling allocation scheme in the maintenance line, the recovery line and the sterile line to obtain a second computing power scheduling result; iterating the second initial population until the global final computing power scheduling result is obtained when the iteration meets the termination condition.
[0006] In some embodiments, the optimization of the scheduling allocation scheme in the maintenance line, the recovery line and the sterile line to obtain a second computing power scheduling result includes: constructing a weighted directed graph for all scheduling allocation schemes, in which the scheduling allocation schemes in the maintenance line are connected with the scheduling allocation schemes in the sterile line, the scheduling allocation schemes in the recovery line are connected with each other, and the edge weight of each connection edge is set; in the weighted directed graph, the maintenance line is hybridized with the sterile line to generate new scheduling allocation schemes, and the recovery line is self-crossed to generate new scheduling allocation schemes, to obtain a second computing power scheduling result.
[0007] In some embodiments, the scheduling fitness function is as follows:
[0008] wherein,
[0009] In the formula, is a scheduling fitness function; is the longest time required for all computing power nodes to complete all computing tasks; is the mean square deviation of the load difference of each computing power node; is the priority matching degree of the actual execution of each computing power node; is an adjustable weight; is a computing task scheduled to a computing power node ; is the processing capacity required by the quantified computing task ; and task priority To guide computational tasks The scheduling priority order; To set up a computing task The time limit for completing the task; For computing power nodes The workload that can be processed per unit of time; The average load across all computing nodes; For computational tasks At computing nodes The execution time is m, where m is the number of computing nodes and n is the number of computing tasks.
[0010] In some embodiments, the method for generating a new scheduling and allocation scheme by hybridizing the maintainer line with the sterile line is shown in the following formula:
[0011]
[0012] In the formula, A new scheduling and allocation scheme is generated for hybridization; Aggregate the adjacency information for the scheduling and allocation scheme in sterile lines; for All adjacent nodes of a node; and A random value within the range of 0 to 1; For scheduling and allocation scheme To scheduling and allocation scheme Edge weights between The result after normalization.
[0013] In some embodiments, the method for the recovery system to generate a new scheduling allocation scheme through self-interaction is shown in the following equation:
[0014] In the formula, A new scheduling and allocation scheme generated by self-intersection; The optimal scheduling and allocation scheme is determined globally. To restore the scheduling and allocation scheme in the system Adjacent points; It is a random value in the range of 0 to 1.
[0015] In some embodiments, during the process of the restorer system self-crossing to generate a new scheduling allocation scheme, if the fitness value of the new scheduling allocation scheme obtained after more than a preset number of self-crossings is less than the fitness value of the original scheduling allocation scheme without self-crossing, then the method for re-generating a new scheduling allocation scheme by self-crossing the original scheduling allocation scheme is as follows:
[0016] wherein, is a new scheduling allocation scheme generated by re-selfing; is a lower bound of a search space; is an upper bound of a search space.
[0017] In a second aspect, a computing power scheduling system based on a graph hybrid breeding optimization algorithm is provided, comprising: a task acquisition module configured to acquire a plurality of computing tasks to be executed; a computing power node acquisition module configured to acquire a plurality of computing power nodes for computing and processing all the computing tasks; an optimization module in communication connection with the task acquisition module and the computing power node acquisition module, configured to guide each computing task to migrate and schedule in each computing power node based on a graph hybrid breeding optimization algorithm, so that each computing task is matched with an optimal computing power node, and a global final computing power scheduling result is obtained; and a computing and processing module in communication connection with the optimization module, configured to compute and process all the computing tasks by all the computing power nodes according to the global final computing power scheduling result.
[0018] In a third aspect, a computer readable storage medium having a computer program stored thereon is provided, the computer program being executed by a processor to implement the computing power scheduling method based on the graph hybrid breeding optimization algorithm as described above.
[0019] In a fourth aspect, an electronic device is provided, comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, the processor implementing the computing power scheduling method based on the graph hybrid breeding optimization algorithm as described above when executing the computer program.
[0020] Compared with the prior art, the present application has the following advantages: based on a weighted directed graph, the computing tasks to be executed and their resource dependency relationships are modeled, so that the guided pairing operation of the maintainer line, the sterile line and the restorer line is guided in hybridization and selfing, avoiding the quality degradation caused by blind pairing, enabling it to quickly converge to a global optimal solution while ensuring population diversity, thereby improving the efficiency of computing power scheduling. Since in the modern computing environment, the changes in task load and resource state are normal, the graph hybrid breeding optimization algorithm can adjust the scheduling scheme in real time and quickly respond to changes in system load and resource state. By optimizing each link in the scheduling process, this algorithm not only improves the utilization rate of computing resources, but also reduces the computing cost, and improves the overall performance of the system while ensuring that tasks are completed on time. BRIEF DESCRIPTION OF DRAWINGS
[0021] Figure 1It is a flow diagram of a kind of algorithm power scheduling method based on graph hybridization breeding optimization algorithm of the application; Figure 2 It is a structure diagram of a kind of algorithm power scheduling system based on graph hybridization breeding optimization algorithm of the application. DETAILED DESCRIPTION
[0022] Reference will now be made in detail to the present application, examples of which are illustrated in the accompanying drawings. While the application will be described in conjunction with the specific embodiments, it will be understood that the application is not limited to the embodiments described. To the contrary, the application is intended to cover alternatives, modifications and equivalents, which are included within the spirit and scope of the application, as defined by the appended claims. It should be noted that the method steps described herein can all be implemented by any functional block or functional arrangement, and any functional block or functional arrangement can be implemented as a physical entity or a logical entity, or a combination of both.
[0023] To enable persons skilled in the art to better understand the present application, the present application will be further described in detail below in conjunction with the drawings and specific embodiments.
[0024] Note: the examples to be introduced next are only a specific example, and not as a limitation on the embodiments of the present application must be as follows specific steps, values, conditions, data, order, etc. Those skilled in the art can use the concept of the present application to construct more embodiments not mentioned in the present specification by reading the present specification.
[0025] Based on the problems in the background art, Figure 1 A flow chart of a kind of algorithm power scheduling method based on graph hybridization breeding optimization algorithm of one embodiment of the application is shown, and the method comprises: Step S100, a plurality of computing tasks to be executed are obtained.
[0026] Task resource modeling, the plurality of computing tasks to be executed are abstracted into a task set, as follows: Wherein, each computing task Contains task computing amount , task priority , deadline Attribute. Build task information description table, provide foundation for subsequent scheduling modeling. Each computing task in the task set records its computing amount, deadline and task priority through the task information description table, wherein, the computing amount is used to quantify the processing capacity required by the task, the deadline is used to set the time boundary of task completion, and the priority is used to guide the scheduling priority order.
[0027] Step S200, a plurality of algorithm nodes for computing processing all the computing tasks are obtained.
[0028] Modeling of computing power resources, modeling available computing power nodes as a resource set, as follows:
[0029] Each computing power node has a computing capacity and a current load attribute. Each computing power node in the computing power resource set records its computing capacity and current load status through a resource description table, wherein the computing capacity is used to describe the task load that can be processed per unit time, and the current load represents the total computing overhead of the tasks currently allocated to the node.
[0030] Step S300, based on the graph hybrid breeding optimization algorithm, guiding the migration and scheduling of each computing task in each computing power node, so that each computing task is matched with an optimal computing power node, and a global final computing power scheduling result is obtained. The specific steps include the following: S310, initially migrating and scheduling each computing task to each computing power node to generate an initial computing power scheduling result including multiple scheduling allocation schemes; Based on the modeling of computing power resources and task resources, the task scheduling problem is abstracted as an optimal resource allocation combination problem. The graph hybrid breeding optimization algorithm is used to intelligently search for the problem. In the search process, the graph structure is used to guide the migration and combination of tasks between different types of nodes, so as to obtain a scheduling result with better global optimization ability. The pros and cons of the task allocation scheme are evaluated by the scheduling fitness function defined by the present application, and a scheduling allocation scheme is defined as an individual , wherein represents that the task is allocated to the computing power , otherwise 0.
[0031] The task set in the scheduling problem is mapped to the computing power resource set to generate an initial computing power scheduling result of the initial N scheduling allocation scheme individuals Each individual is a task-resource scheduling allocation scheme and is represented as:
[0032] The generation method of the initial computing power scheduling result includes one of the following three strategies or a combination thereof: (1) balanced allocation strategy: preferentially allocating tasks to computing power nodes with the smallest current load; (2) proximity strategy: preferentially selecting nodes with the shortest communication delay or network distance; (3) random initialization strategy: randomly selecting nodes from the available computing power resource set to improve the diversity of the initial population.
[0033] S320, calculate the fitness value of each scheduling allocation scheme in the initial computing power scheduling result based on the scheduling fitness function, sort all scheduling allocation schemes from high to low according to the fitness value, and divide them into three equal parts: maintainer line, restorer line and sterile line.
[0034] That is, the initial computing power scheduling results Each individual in By scheduling fitness function An evaluation was conducted, and the population was ranked from high to low fitness, dividing individuals into three categories: the top 1 / 3 of the best individuals were designated as maintainer lines, the middle 1 / 3 as restorer lines, and the bottom 1 / 3 of individuals with poor fitness were designated as sterile lines.
[0035] The scheduling fitness function takes into account the following factors: 1. Maximum completion time:
[0036] This represents the longest time required for all computing nodes to complete all assigned tasks, with the goal of minimizing the overall completion time.
[0037] 2. Resource load balancing:
[0038]
[0039] 3. Task priority matching degree:
[0040] in, Represents computational task At computing nodes The execution time is measured. This metric assesses whether the actual execution of a task reasonably matches its priority and urgency. The goal is to penalize high-priority tasks that are assigned to slow nodes and consume excessive processing time.
[0041] In summary, the fitness function is defined as follows:
[0042] in, For adjustable weights, satisfying It is used to balance various objectives.
[0043] In the above expression, The fitness function is used for scheduling. The longest time required for all computing nodes to complete all computing tasks; The mean squared error is used to measure the load difference between each computing node; Priority matching degree for the actual execution status of each computing node; For quantization computation tasks Required processing power; task priority To guide computational tasks The scheduling priority order; To set up a computing task The time limit for completing the task; For computing power nodes The workload that can be processed per unit of time; The average load across all computing nodes; For computational tasks At computing nodes The execution time is m, where m is the number of computing nodes and n is the number of computing tasks.
[0044] S330, optimize the scheduling allocation scheme in the maintainer line, the restorer line, and the sterile line to obtain the second computing power scheduling result; this step includes the following: A weighted directed graph is constructed for all scheduling and allocation schemes. In the weighted directed graph, the scheduling and allocation schemes in the maintenance line are connected to the scheduling and allocation schemes in the sterile line, and the scheduling and allocation schemes in the restorer line are connected to each other. The edge weight of each connection edge is set. In the weighted directed graph, the maintainer line and the sterile line are crossbred to generate a new scheduling allocation scheme, and the restorer line is self-crossed to generate a new scheduling allocation scheme, thus obtaining the second computing power scheduling result.
[0045] Specifically, in this embodiment, in the construction of the weighted directed graph structure, for each individual Construct a weighted graph ,in, All individuals are nodes in the graph, and each sterile line node... With the diagram The node with the greatest "performance complementarity" Connected, performance complementarity is calculated through fitness differences. Each restorer node... With the diagram Each edge is connected to a node that is also a recovery system. Additionally, an edge weight is assigned to each edge, defined as follows:
[0046] in, Represents an individual To individuals The edge weights. Because from individual to The sum of the edge weights of the individual elements is not equal to 1, so normalization is required. The specific formula is as follows:
[0047] The above For All the adjacent nodes of the node.
[0048] And aggregate the neighbor information of each sterile line node, denoted as:
[0049] In the graph-guided hybridization, the hybridization operation occurs between the maintainer and the sterile line, aiming to produce new individuals superior to the original sterile line individuals. The hybridization operation is denoted as: And Neighbor node aggregation information Hybridization is carried out, and the new population obtained by hybridization is:
[0050] In the formula, And is a random value in the range of 0 to 1.
[0051] In the graph-guided self-interaction, the self-interaction operation occurs between the maintainer individuals, where the global optimal individual is incorporated, aiming to produce new individuals superior to the original maintainer individuals. The individuals in the maintainer are traversed in turn, and the self-interaction times that still cannot be superior to the original individuals after updating also need to be saved for subsequent updating operations. The specific formula of self-interaction is as follows:
[0052] In the formula, is the new scheduling allocation scheme generated by self-interaction; is the global optimal scheduling allocation scheme; is the adjacent node of the scheduling allocation scheme in the maintainer; is a random value in the range of 0 to 1.
[0053] For the maintainer individuals exceeding the preset upper limit of self-interaction times, they are reinitialized as new individuals through the updating operation, and the self-interaction times are reset to zero. The purpose of this step is to expand the distribution range of individuals in the entire solution space, and also improve the randomness of the algorithm. The updating operation formula is as follows:
[0054] In the formula, is the new scheduling allocation scheme generated by re-self-interaction; is the lower bound of the search space; is the upper bound of the search space, the search space refers to the feasible solution range of all possible values of the scheduling allocation scheme, and the upper and lower bounds are obtained by sorting in ascending order through the fitness function.
[0055] S340, iterating the second initial population until a termination condition is met to obtain a global final computing power scheduling result, wherein the termination condition is a preset maximum number of running iterations.
[0056] Step S400, all computing power nodes perform computing processing on all computing tasks according to the global final computing power scheduling result.
[0057] Specifically, in the embodiment, the global final computing power scheduling result obtained in step S300 is used to bind and allocate computing power tasks and nodes, and finally scheduling mapping is realized. The global final computing power scheduling result is represented as The scheduling solution defines the optimal binding relationship between each computing task in the task set and a computing power node in the computing power resource set , that is, there is a mapping function as follows:
[0058] Wherein, the above indicates that the task is allocated to the resource node . According to the mapping, the task is allocated to the corresponding computing power node, and the task running parameters are configured for each node to generate a global scheduling execution table.
[0059] According to the above mapping, each task is sequentially transmitted to the corresponding computing power node and its running parameters are configured, at the same time, the scheduling system establishes a task scheduling queue according to the task priority and resource use sequence, controls the parallel start and resource scheduling rhythm. The system monitors the task execution state, resource use condition and network transmission state, and feeds back the real-time information to the scheduling controller. If a resource bottleneck or task failure is detected during running, the system will trigger a rescheduling mechanism to restart the graph hybrid optimization process to optimize the current allocation.
[0060] In summary, the present application proposes a computing power scheduling method and system based on a graph hybrid breeding optimization algorithm, which optimizes the allocation and scheduling of computing power resources to improve the processing efficiency of large-scale computing tasks. This method not only can dynamically adjust the computing resources according to the actual demand, but also can fully consider the maximum completion time, resource load balancing and task priority matching degree factors, so as to realize the global optimization in the computing power scheduling process.
[0061] Referring to Figure 2 , the present application embodiment provides a computing power scheduling system based on a graph hybrid breeding optimization algorithm, which comprises: A task acquisition module is used to acquire a plurality of computing tasks to be executed; The computing power node acquisition module is configured to acquire a plurality of computing power nodes for computing and processing all the computing tasks. The optimization module is in communication connection with the task acquisition module and the computing power node acquisition module, and is configured to guide the migration and scheduling of each computing task in each computing power node based on a graph hybrid breeding optimization algorithm, so that each computing task is matched with an optimal computing power node, and a global final computing power scheduling result is obtained. The computing and processing module is in communication connection with the optimization module, and is configured to perform computing and processing on all the computing tasks by all the computing power nodes according to the global final computing power scheduling result.
[0062] Therefore, the present application proposes a computing power scheduling method based on a graph hybrid breeding optimization algorithm. As a new optimization method combining graph modeling and evolutionary computation, the graph hybrid breeding optimization algorithm shows superior scheduling adaptability and search ability. The algorithm models the to-be-scheduled tasks and their resource dependency relationships based on a graph structure, in which the nodes in the graph represent scheduling individuals, and the edge weights reflect the similarity between the scheduling individuals, thereby guiding the pairing operation of the maintainer line, the sterile line and the restorer line in hybridization and selfing to be more guided, avoiding the quality degradation caused by blind pairing, so that it can quickly converge to the global optimal solution while ensuring the diversity of the population, thereby improving the efficiency of computing power scheduling. Compared with traditional scheduling methods, the graph hybrid breeding optimization algorithm can better cope with complex task dependency relationships and dynamic changes in the computing environment. In the modern computing environment, changes in task load and resource state are the norm, and the graph hybrid breeding optimization algorithm can adjust the scheduling scheme in real time and quickly respond to changes in system load and resource state. By optimizing each link in the scheduling process, the algorithm not only improves the utilization of computing resources, but also reduces the computing cost, and improves the overall performance of the system while ensuring that the tasks are completed on time.
[0063] Specifically, the present embodiment corresponds to the above-mentioned method embodiment one by one, and the functions of each module have been described in detail in the corresponding method embodiment, so they will not be repeated here.
[0064] Based on the same inventive concept, the embodiments of the present application also provide a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement all method steps or part of the method steps of the above-mentioned method.
[0065] The present application implements all or part of the processes in the above method, and can also be completed by a computer program instructing related hardware. The computer program can be stored in a computer readable storage medium, and the computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution medium, etc. It should be noted that the content of the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.
[0066] Based on the same inventive concept, the embodiments of the present application also provide an electronic device, including a memory and a processor, the memory stores a computer program running on the processor, and the processor implements all method steps or part of the method steps of the above method when executing the computer program.
[0067] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The processor is the control center of the computer device, and connects all parts of the computer device through various interfaces and lines.
[0068] The memory can be used to store computer programs and / or modules, and the processor realizes various functions of the computer device by running or executing the computer programs and / or modules stored in the memory, and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application program required by a function (for example, a sound playing function, an image playing function, etc.); and the data storage area can store data created according to use of the mobile phone (for example, audio data, video data, etc.). In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory devices.
[0069] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, a server or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer usable program codes.
[0070] The present application is described with reference to flowcharts and / or block diagrams of methods, devices (systems), servers and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device that implements the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that carries out the functions specified in one or more blocks or flows.
[0071] These computer program instructions can also be stored in a computer readable memory capable of directing the computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including instruction apparatus, which realizes the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that carries out the functions specified in one or more blocks or flows.
[0072] These computer program instructions can also be loaded into a computer or other programmable data processing devices, so that a series of operational steps are performed on the computer or other programmable devices to generate computer-implemented processes, thus the instructions executed on the computer or other programmable devices provide processes for implementing the flow Figure 1 one or more flows and / or blocks Figure 1 one or more blocks or steps of the functions specified in the flow(s) or block(s).
[0073] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.
Claims
1. A method for scheduling computing power based on graph hybrid breeding optimization algorithm, characterized in that, The method comprises the steps of: obtaining a plurality of computing tasks to be executed; obtaining a plurality of computing power nodes for performing computing processing on all the computing tasks; based on a graph hybrid breeding optimization algorithm, guiding the migration scheduling of each computing task in each computing power node, so that each computing task corresponds to match an optimal computing power node, and obtaining a global final computing power scheduling result; all computing power nodes perform computing processing on all computing tasks according to the global final computing power scheduling result.
2. The method of claim 1, wherein, The method based on the graph hybrid breeding optimization algorithm guides the migration scheduling of each computing task in each computing power node, so that each computing task corresponds to match an optimal computing power node, and obtains a global final computing power scheduling result, which comprises the steps of: initially migrating each computing task to each computing power node to generate an initial computing power scheduling result comprising a plurality of scheduling allocation schemes; calculating the fitness value of each scheduling allocation scheme in the initial computing power scheduling result based on a scheduling fitness function, and sorting all scheduling allocation schemes from high to low according to the fitness value to obtain three equal parts, which are a maintenance line, a recovery line and a sterile line in turn; optimizing the scheduling allocation scheme in the maintenance line, the recovery line and the sterile line to obtain a second computing power scheduling result; iterating the second initial population until the iteration is terminated to obtain a global final computing power scheduling result.
3. The method of claim 1, wherein, The method of optimizing the scheduling allocation scheme in the maintenance line, the recovery line and the sterile line to obtain a second computing power scheduling result comprises the steps of: constructing a weighted directed graph for all scheduling allocation schemes, in which the scheduling allocation schemes in the maintenance line are connected with the scheduling allocation schemes in the sterile line, the scheduling allocation schemes in the recovery line are connected with each other, and the edge weight of each connection edge is set; in the weighted directed graph, the maintenance line is hybridized with the sterile line to generate a new scheduling allocation scheme, and the recovery line is self-crossed to generate a new scheduling allocation scheme to obtain a second computing power scheduling result.
4. The method of claim 2, wherein, The scheduling fitness function is as follows: wherein, In the formula, is a scheduling fitness function; is the longest time required for all computing nodes to complete all computing tasks; is the mean square deviation of the load difference of each computing node; is the priority matching degree of the actual execution of each computing node; is an adjustable weight; is a computing task scheduling to a computing node ; is the processing capacity required by a quantified computing task task priority is the scheduling priority order of the computing task ; is the time boundary for completing the computing task ; is the task load that can be processed by a computing node in a unit of time; is the average load of all computing nodes; is the execution time of a computing task on a computing node ; m is the number of computing nodes; and n is the number of computing tasks.
5. The method of claim 3, wherein, The method of hybridizing the maintenance line with the sterile line to generate a new scheduling allocation scheme is as follows: wherein, a new schedule allocation scheme resulting from the hybridization; a contiguous point aggregation information of the schedule allocation scheme in the sterile line; a new schedule allocation scheme resulting from the hybridization; all contiguous points of the node; and a random value in the range of 0 to 1; a schedule allocation scheme a schedule allocation scheme an edge weight between the schedule allocation scheme a normalized result.
6. The method of claim 3, wherein the method is based on a graph hybridization breeding optimization algorithm. The method of self-crossing the recovery line to generate a new scheduling allocation scheme is as follows: wherein a new scheduling allocation scheme generated from the self-interaction; a globally optimal scheduling allocation scheme; a scheduling allocation scheme for restoring the system in the event of a failure a neighboring point of the scheduling allocation scheme a random value in the range 0 to 1.
7. The method of claim 3, wherein the method is based on a graph hybridization breeding optimization algorithm. In the process of self-crossing the recovery line to generate a new scheduling allocation scheme, when the fitness value of the new scheduling allocation scheme obtained by exceeding the preset self-crossing times is less than the fitness value of the original scheduling allocation scheme which is not self-crossed, the method of self-crossing the original scheduling allocation scheme to generate a new scheduling allocation scheme is as follows: wherein is a newly generated scheduling assignment resulting from re-breeding; is a lower bound of the search space; is an upper bound of the search space.
8. A system for scheduling computing power based on a graph hybrid breeding optimization algorithm, characterized in that, The method comprises the steps of: a task acquisition module for obtaining a plurality of computing tasks to be executed; a computing power node acquisition module for obtaining a plurality of computing power nodes for performing computing processing on all the computing tasks; an optimization module in communication connection with the task acquisition module and the computing power node acquisition module, for guiding the migration scheduling of each computing task in each computing power node based on a graph hybrid breeding optimization algorithm, so that each computing task corresponds to match an optimal computing power node, and obtaining a global final computing power scheduling result; and A computing processing module, in communication connection with the optimization module, is configured to perform computing processing on all computing tasks by all computing nodes according to the global final computing power scheduling result.
9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by a processor, implements the computing power scheduling method based on the graph hybridization breeding optimization algorithm according to any one of claims 1 to 7.
10. An electronic device comprising a storage medium, a processor, and a computer program stored in the storage medium and operable on the processor, characterized in that, The processor, when running the computer program, implements the computing power scheduling method based on the graph hybridization breeding optimization algorithm according to any one of claims 1 to 7.