Computer task scheduling method and device based on artificial intelligence
Through task rating and optimization algorithms based on artificial intelligence, the inefficiency problem of traditional task scheduling methods under complex requirements and resource limitations is solved, dynamic resource allocation and efficient task processing are realized, and system adaptability and resource utilization are improved.
Patent Information
- Application Number
- CN202411216558.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-02
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-09-02
AI Technical Summary
Traditional task scheduling methods are difficult to cope with complex task requirements and resource limitations, resulting in low scheduling efficiency, undynamic resource allocation, slow response, and manual priority setting, which increases complexity and error risk.
Using a computer task scheduling method based on artificial intelligence, through task rating, sorting, building task processing triple and optimization model, and using optimization algorithms to dynamically adjust resource allocation strategies to ensure that tasks are processed according to priority and adapt to resource changes.
Improve task processing order optimization, resource utilization rate, reduce defaults and resource conflicts, and enhance the adaptability and flexibility of the system.
Smart Images

Figure CN119088524A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of task optimization scheduling, and in particular to a computer task scheduling method and device based on artificial intelligence. Background Art
[0002] Computer task scheduling refers to the process of arranging and managing task execution in a computer system. It involves allocating multiple pending tasks to computing resources according to certain strategies to optimize system performance, such as reducing response time, improving throughput, and achieving load balancing. The scheduling method can be based on priority, polling, shortest task first and other strategies, and may consider factors such as task dependencies and resource requirements.
[0003] Traditional task scheduling methods may rely on fixed priorities and simple sorting, which makes it difficult to cope with complex task requirements and resource constraints, resulting in low scheduling efficiency. Traditional task scheduling methods may not be able to dynamically adjust resource allocation, which can easily lead to resource waste or shortage. Traditional task scheduling methods are slow to respond to changes in task and resource requirements, making it difficult to flexibly adjust scheduling strategies. Traditional task scheduling methods usually require manual setting of task priorities and processing strategies, which increases the complexity of manual intervention and the risk of errors. Summary of the invention
[0004] The purpose of the present invention is to solve the problems existing in the background technology and to propose a computer task scheduling method and device based on artificial intelligence.
[0005] The technical solution of the present invention is a computer task scheduling method based on artificial intelligence, comprising:
[0006] Acquire multiple target task requests, construct a first task request queue according to the multiple target task requests, and perform a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request;
[0007] sorting the first task request queue based on the task rating table corresponding to each target task request in the first task request queue to obtain a second task request queue;
[0008] Building a task processing triplet based on the computer task processing resources required by the multiple target task requests;
[0009] Constructing a task processing model for each computer task processing node in the computer task cluster, constructing task processing constraints for each computer task node in the computer task cluster, and constructing a task processing optimization model based on the task processing model and the task processing constraints;
[0010] The task processing optimization model is solved based on an optimization algorithm to obtain a plurality of task processing strategies, and the second task request queue and the task processing triplet are matched with the plurality of task processing strategies to assign a plurality of target task requests to corresponding task processing strategies.
[0011] Preferably, performing a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request includes:
[0012] The urgency factor of each target task request in the first task request queue is calculated to obtain the urgency factor corresponding to each target task request. The urgency factor calculation formula is as follows:
[0013]
[0014] Among them, UF i (t) represents the urgency factor corresponding to the i-th target task request in the first task request queue, lag i (t) represents the task delay parameter corresponding to the i-th target task request in the first task request queue, U i Indicates the total system utilization, and lag i (t) = U i ×(ta i (t))-exec i (a i (t),t),a i (t) represents the arrival time of the i-th target task request in the first task request queue, exec i (a i (t), t) represents the workload corresponding to the i-th target task request in the first task request queue executed until time t in the actual scheduling;
[0015] The recovery time of each target task request in the first task request queue is calculated to obtain the recovery time corresponding to each target task request. The recovery time calculation formula is as follows:
[0016]
[0017] Among them, RT i (t)UF i (t) represents the recovery time corresponding to the i-th target task request in the first task request queue.
[0018] Preferably, performing a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request further includes:
[0019] The slack of each target task request in the first task request queue is calculated to obtain the slack corresponding to each target task request. The slack time calculation formula is as follows:
[0020] LAX i (t) = ARR i (t)+PER i (t)-REM i (t)-CYC i (t);
[0021] Among them, LAX i (t) represents the slack corresponding to the i-th target task request in the first task request queue, ARR i (t) represents the arrival time of the i-th target task request in the first task request queue, PER i (t) represents the working cycle corresponding to the i-th target task request in the first task request queue, REM i (t) represents the remaining execution time of the task corresponding to the i-th target task request in the first task request queue, CYC i (t) represents the current CPU clock number corresponding to the i-th target task request in the first task request queue;
[0022] The urgency factor, recovery time, and slack of each target task request in the first task request queue are combined to obtain a task rating table corresponding to each target task request in the first task request queue.
[0023] Preferably, the first task request queue is sorted based on the task rating table corresponding to each target task request in the first task request queue to obtain a second task request queue, including:
[0024] sorting each target task request in the first task request queue based on the urgency factor in the task rating table;
[0025] If the urgency factors in the task rating table corresponding to the target task requests in the first task request queue are equal, sorting the target task requests in the first task request queue based on the recovery time in the task rating table;
[0026] If the urgency factor and recovery time in the task rating table corresponding to each target task request in the first task request queue are equal, the target task requests in the first task request queue are sorted based on the slack in the task rating table to obtain a second task request queue.
[0027] Preferably, the task processing triplet is as follows:
[0028]
[0029] in, Represents the jth target task request in the first task request queue, res j Indicates the CPU and memory resources required by the i-th target task request in the first task request queue, thr j Indicates the CPU consumption threshold level required for the i-th target task request in the first task request queue.
[0030] Preferably, the task processing model includes a node reliability model, a task scheduling energy consumption model and a task transmission energy consumption model, and the node reliability model is as follows:
[0031]
[0032] Among them, l represents the reliability of the computer task processing node, r i represents the number of copies of the target task request, u represents the failure rate of the target task request, θ represents the recovery probability after the target task request, and s i represents the number of target task requests that can be processed by the computer task processing node, and I represents the number of computer task processing nodes in the computer task cluster;
[0033] The task scheduling energy consumption model is as follows:
[0034]
[0035] Among them, e s represents the energy consumption required for the target task to request task scheduling, s ij represents the amount of data requested by the target task transmitted to the computer task processing node, e so Indicates the unit energy consumption corresponding to the amount of data;
[0036] The task transmission energy consumption model is as follows:
[0037]
[0038] Among them, e t represents the energy consumption required for the target task to request the task transmission, Indicates the transmit power requested by the target task, Indicates the time taken to transfer data. represents the received power requested by the target task, Indicates the time taken to receive data. Indicates the data transmission rate, G T represents the transmit antenna gain, GR represents the receiving antenna gain, λ represents the wavelength, and d ij represents the distance of the target task request transmitted to the computer task processing node, L represents the system loss factor that is independent of propagation, Indicates the data receiving rate.
[0039] Preferably, the task processing constraints are as follows:
[0040]
[0041] in, and Respectively represent the minimum and maximum values of the transmit power requested by the target task, and They represent the minimum and maximum received power requested by the target task, respectively, and d ij_max Indicates the maximum distance of the target task request transmitted to the computer task processing node.
[0042] Preferably, the task processing optimization model is as follows:
[0043]
[0044] Among them, st represents the constraint condition.
[0045] Preferably, the task processing optimization model is solved based on an optimization algorithm to obtain multiple task processing strategies, including:
[0046] Set the population size and the maximum number of iterations, and initialize the position of the optimization algorithm. The position initialization formula is as follows:
[0047]
[0048] Among them, X n+1 and X n They represent the individual positions at the 1+nth iteration and the nth iteration respectively, and P represents the preset initialization parameter;
[0049] The fitness value of each individual is calculated. If the fitness value of the individual is less than the preset fitness value threshold, the position of the individual is updated by the first position update formula. Otherwise, the position of the individual is updated by the second position update formula. The first position update formula is as follows:
[0050]
[0051] in, represents the position of the jth dimension of the i-th individual in the first stage, X i,jrepresents the initialization position of the jth dimension of the ith individual, r i,j Indicates a random number between 0 and 1, SF i,j represents the optimal individual position, I i,j Indicates 0 or 1;
[0052] The second position update formula is as follows:
[0053]
[0054] Among them, X i (t+1), X i (t) and X i (t-1) represents the individual positions at the t+1th iteration, the tth iteration, and the t-1th iteration, respectively. Represents a random number from 0 to π;
[0055] Calculate the fitness value of each individual again, and obtain the optimal solution based on the fitness value of each individual, and determine whether the current number of iterations is greater than the maximum number of iterations. If it is greater than the maximum number of iterations, output the optimal position and optimal fitness, otherwise, repeat the above operations.
[0056] The technical solution of the present invention is: a computer task scheduling device based on artificial intelligence, which is applicable to the computer task scheduling method based on artificial intelligence, comprising:
[0057] A request processing module, the request processing module is used to obtain multiple target task requests, build a first task request queue according to the multiple target task requests, and perform a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request;
[0058] A request sorting module, the request sorting module is used to sort the first task request queue based on the task rating table corresponding to each target task request in the first task request queue to obtain a second task request queue;
[0059] A resource construction module, the resource construction module is used to construct a task processing triplet based on the computer task processing resources required by the multiple target task requests;
[0060] A task processing module, the task processing module is used to construct a task processing model for each computer task processing node in the computer task cluster, construct task processing constraints for each computer task node in the computer task cluster, and construct a task processing optimization model based on the task processing model and the task processing constraints;
[0061] A request matching module, the request matching module is used to solve the task processing optimization model based on an optimization algorithm to obtain multiple task processing strategies, match the second task request queue and the task processing triplet with the multiple task processing strategies to assign multiple target task requests to corresponding task processing strategies.
[0062] Compared with the prior art, the above technical solution of the present invention has the following beneficial technical effects:
[0063] The present invention enables high-priority tasks to be processed earlier through task rating and sorting, optimizes the processing order of tasks, and ensures that tasks can be reasonably allocated according to actual resource requirements and improves resource utilization by constructing task processing triples and optimization models. In addition, the present invention combines task processing models and constraints to ensure that task scheduling is carried out while satisfying various restrictions, reduces defaults and resource conflicts, and utilizes optimization algorithms to adapt to different tasks and changes in resource requirements, dynamically adjust task processing strategies, and improve the adaptability and flexibility of the system. BRIEF DESCRIPTION OF THE DRAWINGS
[0064] Figure 1 A schematic flow chart of an overall method in an embodiment of the present invention;
[0065] Figure 2 A schematic diagram of the flow chart of the overall equipment in an embodiment of the present invention.
[0066] Figure numerals: 1. request processing module; 2. request sorting module; 3. resource construction module; 4. task processing module; 5. request matching module. DETAILED DESCRIPTION
[0067] Embodiment 1, as Figure 1 As shown, the present invention proposes a computer task scheduling method based on artificial intelligence, comprising:
[0068] S1. Acquire multiple target task requests, construct a first task request queue according to the multiple target task requests, and perform a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request;
[0069] S2, sorting the first task request queue based on the task rating table corresponding to each target task request in the first task request queue to obtain a second task request queue;
[0070] S3, constructing a task processing triplet based on the computer task processing resources required by the multiple target task requests;
[0071] S4, constructing a task processing model for each computer task processing node in the computer task cluster, constructing task processing constraints for each computer task node in the computer task cluster, and constructing a task processing optimization model based on the task processing model and the task processing constraints;
[0072] S5. Solve the task processing optimization model based on the optimization algorithm to obtain multiple task processing strategies, match the second task request queue and the task processing triplet with the multiple task processing strategies to assign the multiple target task requests to the corresponding task processing strategies.
[0073] In the present invention, a task request queue is a data structure used to store pending task requests, which may come from users, systems or other services, and usually include workloads or operations that need to be completed; an optimization algorithm is a class of algorithms used to find the best solution, usually maximizing or minimizing an objective function under certain constraints, and they are used in task scheduling to determine how to allocate tasks to achieve optimal performance.
[0074] Embodiment 2, a computer task scheduling method based on artificial intelligence proposed by the present invention, compared with embodiment 1, this embodiment further includes: performing a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request, including:
[0075] A1. Calculate the urgency factor of each target task request in the first task request queue to obtain the urgency factor corresponding to each target task request. The urgency factor calculation formula is as follows:
[0076]
[0077] Among them, UF i (t) represents the urgency factor corresponding to the i-th target task request in the first task request queue, lag i (t) represents the task delay parameter corresponding to the i-th target task request in the first task request queue, U i Indicates the total system utilization, and lag i (t) = U i ×(ta i (t))-exec i (a i (t),t),a i (t) represents the arrival time of the i-th target task request in the first task request queue, exec i (a i (t), t) represents the workload corresponding to the i-th target task request in the first task request queue executed until time t in the actual scheduling;
[0078] A2. Calculate the recovery time of each target task request in the first task request queue to obtain the recovery time corresponding to each target task request. The recovery time calculation formula is as follows:
[0079]
[0080] Among them, RT i (t)UF i (t) represents the recovery time corresponding to the i-th target task request in the first task request queue.
[0081] In an optional embodiment, performing a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request further includes:
[0082] A3. Calculate the slack of each target task request in the first task request queue to obtain the slack corresponding to each target task request. The slack time calculation formula is as follows:
[0083] LAX i (t) = ARR i (t)+PER i (t)-REM i (t)-CYC i (t);
[0084] Among them, LAX i (t) represents the slack corresponding to the i-th target task request in the first task request queue, ARR i (t) represents the arrival time of the i-th target task request in the first task request queue, PER i (t) represents the working cycle corresponding to the i-th target task request in the first task request queue, REM i (t) represents the remaining execution time of the task corresponding to the i-th target task request in the first task request queue, CYC i (t) represents the current CPU clock number corresponding to the i-th target task request in the first task request queue;
[0085] A4. Combine the urgency factor, recovery time, and slack of each target task request in the first task request queue to obtain a task rating table corresponding to each target task request in the first task request queue.
[0086] In an optional embodiment, the first task request queue is sorted based on the task rating table corresponding to each target task request in the first task request queue to obtain a second task request queue, including:
[0087] B1. Sort each target task request in the first task request queue based on the urgency factor in the task rating table;
[0088] B2. If the urgency factors in the task rating table corresponding to each target task request in the first task request queue are equal, sorting each target task request in the first task request queue based on the recovery time in the task rating table;
[0089] B3. If the urgency factor and recovery time in the task rating table corresponding to each target task request in the first task request queue are equal, sort each target task request in the first task request queue based on the slack in the task rating table to obtain a second task request queue.
[0090] In an optional embodiment, the task processing triple is as follows:
[0091]
[0092] in, Represents the jth target task request in the first task request queue, res j Indicates the CPU and memory resources required by the i-th target task request in the first task request queue, thr j Indicates the CPU consumption threshold level required for the i-th target task request in the first task request queue.
[0093] In an optional embodiment, the task processing model includes a node reliability model, a task scheduling energy consumption model and a task transmission energy consumption model, and the node reliability model is as follows:
[0094]
[0095] Among them, l represents the reliability of the computer task processing node, r i represents the number of copies of the target task request, u represents the failure rate of the target task request, θ represents the recovery probability after the target task request, and s i represents the number of target task requests that can be processed by the computer task processing node, and I represents the number of computer task processing nodes in the computer task cluster;
[0096] The task scheduling energy consumption model is as follows:
[0097]
[0098] Among them, e s represents the energy consumption required for the target task to request task scheduling, s ij represents the amount of data requested by the target task transmitted to the computer task processing node, e soIndicates the unit energy consumption corresponding to the amount of data;
[0099] The task transmission energy consumption model is as follows:
[0100]
[0101] Among them, e t represents the energy consumption required for the target task to request the task transmission, Indicates the transmit power requested by the target task, Indicates the time taken to transfer data. represents the received power requested by the target task, Indicates the time taken to receive data. Indicates the data transmission rate, G T represents the transmit antenna gain, G R represents the receiving antenna gain, λ represents the wavelength, and d ij represents the distance of the target task request transmitted to the computer task processing node, L represents the system loss factor that is independent of propagation, Indicates the data receiving rate.
[0102] In an optional embodiment, the task processing constraints are as follows:
[0103]
[0104] in, and Respectively represent the minimum and maximum values of the transmit power requested by the target task, and They represent the minimum and maximum received power requested by the target task, respectively, and d ij_max Indicates the maximum distance of the target task request transmitted to the computer task processing node.
[0105] In an optional embodiment, the task processing optimization model is as follows:
[0106]
[0107] Among them, st represents the constraint condition.
[0108] In an optional embodiment, the task processing optimization model is solved based on an optimization algorithm to obtain multiple task processing strategies, including:
[0109] C1. Set the population size and the maximum number of iterations, and initialize the position of the optimization algorithm. The position initialization formula is as follows:
[0110]
[0111] Among them, X n+1 and X n They represent the individual positions at the 1+nth iteration and the nth iteration respectively, and P represents the preset initialization parameter;
[0112] C2. Calculate the fitness value of each individual. If the fitness value of the individual is less than the preset fitness value threshold, the position of the individual is updated by the first position update formula. Otherwise, the position of the individual is updated by the second position update formula. The first position update formula is as follows:
[0113]
[0114] in, represents the position of the jth dimension of the i-th individual in the first stage, X i,j represents the initialization position of the jth dimension of the ith individual, r i,j Indicates a random number between 0 and 1, SF i,j represents the optimal individual position, I i,j Indicates 0 or 1;
[0115] The second position update formula is as follows:
[0116]
[0117] Among them, X i (t+1), X i (t) and X i (t-1) represents the individual positions at the t+1th iteration, the tth iteration, and the t-1th iteration, respectively. Represents a random number from 0 to π;
[0118] C3. Calculate the fitness value of each individual again, and obtain the optimal solution based on the fitness value of each individual, and determine whether the current number of iterations is greater than the maximum number of iterations. If it is greater than the maximum number of iterations, output the optimal position and optimal fitness, otherwise, repeat the above operations.
[0119] Embodiment three, as Figure 2 As shown, the present invention proposes a computer task scheduling device based on artificial intelligence, which is applicable to the computer task scheduling method based on artificial intelligence, including:
[0120] Request processing module 1, request processing module 1 is used to obtain multiple target task requests, build a first task request queue according to the multiple target task requests, and perform a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request;
[0121] The request sorting module 2 is used to sort the first task request queue based on the task rating table corresponding to each target task request in the first task request queue to obtain a second task request queue;
[0122] Resource construction module 3, resource construction module 3 is used to construct task processing triples based on computer task processing resources required by multiple target task requests;
[0123] Task processing module 4, task processing module 4 is used to construct a task processing model for each computer task processing node in the computer task cluster, construct task processing constraints for each computer task node in the computer task cluster, and construct a task processing optimization model based on the task processing model and the task processing constraints;
[0124] The request matching module 5 is used to solve the task processing optimization model based on the optimization algorithm to obtain multiple task processing strategies, match the second task request queue and the task processing triplet with the multiple task processing strategies, and assign multiple target task requests to corresponding task processing strategies.
[0125] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited thereto, and various changes can be made within the knowledge scope of technicians in the relevant technical field without departing from the purpose of the present invention.
Claims
1. A computer task scheduling method based on artificial intelligence, characterized in that: include: Acquire multiple target task requests, construct a first task request queue according to the multiple target task requests, and perform a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request; sorting the first task request queue based on the task rating table corresponding to each target task request in the first task request queue to obtain a second task request queue; Building a task processing triplet based on the computer task processing resources required by the multiple target task requests; Constructing a task processing model for each computer task processing node in the computer task cluster, constructing task processing constraints for each computer task node in the computer task cluster, and constructing a task processing optimization model based on the task processing model and the task processing constraints; The task processing optimization model is solved based on an optimization algorithm to obtain a plurality of task processing strategies, and the second task request queue and the task processing triplet are matched with the plurality of task processing strategies to assign a plurality of target task requests to corresponding task processing strategies.
2. The computer task scheduling method based on artificial intelligence according to claim 1, characterized in that: Performing a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request includes: The urgency factor of each target task request in the first task request queue is calculated to obtain the urgency factor corresponding to each target task request. The urgency factor calculation formula is as follows: Among them, UF i (t) represents the urgency factor corresponding to the i-th target task request in the first task request queue, lag i (t) represents the task delay parameter corresponding to the i-th target task request in the first task request queue, U i Indicates the total system utilization, and lag i (t) = U i ×(ta i (t))-exec i (a i (t),t),a i (t) represents the arrival time of the i-th target task request in the first task request queue, exec i (a i (t), t) represents the workload corresponding to the i-th target task request in the first task request queue executed until time t in the actual scheduling; The recovery time of each target task request in the first task request queue is calculated to obtain the recovery time corresponding to each target task request. The recovery time calculation formula is as follows: Among them, RT i (t)UF i (t) represents the recovery time corresponding to the i-th target task request in the first task request queue.
3. The computer task scheduling method based on artificial intelligence according to claim 2 is characterized in that: Performing a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request, further comprising: The slack of each target task request in the first task request queue is calculated to obtain the slack corresponding to each target task request. The slack time calculation formula is as follows: LAX i (t)=ARR i (t)+PER i (t)-REM i (t)-CYC i (t); Among them, LAX i (t) represents the slack corresponding to the i-th target task request in the first task request queue, ARR i (t) represents the arrival time of the i-th target task request in the first task request queue, PER i (t) represents the working cycle corresponding to the i-th target task request in the first task request queue, REM i (t) represents the remaining execution time of the task corresponding to the i-th target task request in the first task request queue, CYC i (t) represents the current CPU clock number corresponding to the i-th target task request in the first task request queue; The urgency factor, recovery time, and slack of each target task request in the first task request queue are combined to obtain a task rating table corresponding to each target task request in the first task request queue.
4. The computer task scheduling method based on artificial intelligence according to claim 3 is characterized in that: The first task request queue is sorted based on the task rating table corresponding to each target task request in the first task request queue to obtain a second task request queue, including: sorting each target task request in the first task request queue based on the urgency factor in the task rating table; If the urgency factors in the task rating table corresponding to the target task requests in the first task request queue are equal, sorting the target task requests in the first task request queue based on the recovery time in the task rating table; If the urgency factor and recovery time in the task rating table corresponding to each target task request in the first task request queue are equal, the target task requests in the first task request queue are sorted based on the slack in the task rating table to obtain a second task request queue.
5. The computer task scheduling method based on artificial intelligence according to claim 1 is characterized in that: The task processing triples are as follows: in, Represents the jth target task request in the first task request queue, res j Indicates the CPU and memory resources required by the i-th target task request in the first task request queue, thr j Indicates the CPU consumption threshold level required for the i-th target task request in the first task request queue.
6. The method for scheduling computer tasks based on artificial intelligence according to claim 1, characterized in that: The task processing model includes a node reliability model, a task scheduling energy consumption model and a task transmission energy consumption model. The node reliability model is as follows: Among them, l represents the reliability of the computer task processing node, r i represents the number of copies of the target task request, u represents the failure rate of the target task request, θ represents the recovery probability after the target task request, and s i represents the number of target task requests that can be processed by the computer task processing node, and I represents the number of computer task processing nodes in the computer task cluster; The task scheduling energy consumption model is as follows: Among them, e s represents the energy consumption required for the target task to request task scheduling, s ij represents the amount of data requested by the target task transmitted to the computer task processing node, e so Indicates the unit energy consumption corresponding to the amount of data; The task transmission energy consumption model is as follows: Among them, e t represents the energy consumption required for the target task to request the task transmission, Indicates the transmit power requested by the target task, Indicates the time taken to transfer data. represents the received power requested by the target task, Indicates the time taken to receive data. Indicates the data transmission rate, G T represents the transmit antenna gain, G R represents the receiving antenna gain, λ represents the wavelength, and d ij represents the distance of the target task request transmitted to the computer task processing node, L represents the system loss factor that is independent of propagation, Indicates the data receiving rate.
7. The computer task scheduling method based on artificial intelligence according to claim 6 is characterized in that: The task processing constraints are as follows: in, and Respectively represent the minimum and maximum values of the transmit power requested by the target task, and They represent the minimum and maximum received power requested by the target task, respectively, and d ij_max Indicates the maximum distance of the target task request transmitted to the computer task processing node.
8. The computer task scheduling method based on artificial intelligence according to claim 7 is characterized in that: The task processing optimization model is as follows: Among them, st represents the constraint condition.
9. The computer task scheduling method based on artificial intelligence according to claim 8, characterized in that: The task processing optimization model is solved based on an optimization algorithm to obtain multiple task processing strategies, including: Set the population size and the maximum number of iterations, and initialize the position of the optimization algorithm. The position initialization formula is as follows: Among them, X n+1 and X n They represent the individual positions at the 1+nth iteration and the nth iteration respectively, and P represents the preset initialization parameter; The fitness value of each individual is calculated. If the fitness value of the individual is less than the preset fitness value threshold, the position of the individual is updated by the first position update formula. Otherwise, the position of the individual is updated by the second position update formula. The first position update formula is as follows: in, represents the position of the jth dimension of the i-th individual in the first stage, X i,j represents the initialization position of the jth dimension of the ith individual, r i,j Represents a random number between 0 and 1, SF i,j represents the optimal individual position, I i,j Indicates 0 or 1; The second position update formula is as follows: Among them, X i (t+1), X i (t) and X i (t-1) represents the individual positions at the t+1th iteration, the tth iteration, and the t-1th iteration, respectively. Represents a random number from 0 to π; Calculate the fitness value of each individual again, and obtain the optimal solution based on the fitness value of each individual, and determine whether the current number of iterations is greater than the maximum number of iterations. If it is greater than the maximum number of iterations, output the optimal position and optimal fitness, otherwise, repeat the above operations.
10. A computer task scheduling device based on artificial intelligence, which is applicable to a computer task scheduling method based on artificial intelligence according to any one of claims 1 to 9, characterized in that: include: A request processing module (1), the request processing module (1) is used to obtain a plurality of target task requests, construct a first task request queue according to the plurality of target task requests, and perform a task rating table on each target task request in the first task request queue to obtain a task rating table corresponding to each target task request; A request sorting module (2), the request sorting module (2) being used to sort the first task request queue based on the task rating table corresponding to each target task request in the first task request queue to obtain a second task request queue; A resource construction module (3), the resource construction module (3) is used to construct a task processing triplet based on the computer task processing resources required by the multiple target task requests; A task processing module (4), the task processing module (4) is used to construct a task processing model for each computer task processing node in the computer task cluster, construct task processing constraints for each computer task node in the computer task cluster, and construct a task processing optimization model based on the task processing model and the task processing constraints; A request matching module (5), the request matching module (5) is used to solve the task processing optimization model based on an optimization algorithm to obtain multiple task processing strategies, match the second task request queue and the task processing triplet with the multiple task processing strategies, and allocate multiple target task requests to corresponding task processing strategies.
Citation Information
Patent Citations
Multi-target dynamic task scheduling method and system based on improved ant colony algorithm
CN114968510A
Task unloading method in space-air-ground network based on multi-target depth Q network
CN116431240A
Multi-task processing method and device, equipment and storage medium
CN118093146A
