Method, apparatus, and device for determining task scheduling information based on genetic algorithm

Abstract

The present application provides a method, apparatus, and device for determining task scheduling information based on a genetic algorithm, relating to the fields of computers and task processing technologies. The method includes: obtaining a set of task scheduling information; processing the set of task scheduling information by genetic operators of the genetic algorithm to obtain a task processing result; processing the task processing result to obtain intermediate features of the task scheduling information and the fitness of the task scheduling information; determining genetic operators for the next round of the genetic algorithm according to the intermediate features of each task scheduling information; and processing the set of task scheduling information based on the genetic operators obtained when a preset condition is reached to obtain a set of task scheduling information with higher fitness. The method of the present application can give full play to the characteristics of the genetic algorithm that takes into account both breadth and depth search, automatically adapt the genetic operators used in each round of the genetic algorithm, and improve the accuracy of scheduling task information.

Description

Technical Field

[0001] The present application relates to the fields of computers and task processing, and in particular, to a method, apparatus, and device for determining task scheduling information based on a genetic algorithm. Background Art

[0002] In the process of business processing, there are multiple tasks under the business. There is an execution order between these tasks, and it is necessary to determine the execution order between each task, and then execute these tasks to complete the business processing. That is, it is necessary to determine the task scheduling information.

[0003] In the prior art, when determining task scheduling information, a genetic algorithm can be executed based on a breadth search operator (i.e., a crossover operator), and then task scheduling information can be obtained.

[0004] However, in the prior art, only determining task scheduling information based on a breadth search operator cannot take into account the depth search method, resulting in inaccurate task scheduling information obtained. Summary of the Invention

[0005] The present application provides a method, apparatus, and device for determining task scheduling information based on a genetic algorithm to solve the problem that the obtained task scheduling information is inaccurate.

[0006] In a first aspect, the present application provides a method for determining task scheduling information based on a genetic algorithm, the method including:

[0007] Obtain a task scheduling information set, where the task scheduling information set includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task;

[0008] Repeat the following process until a preset condition is reached, where the initial value of t is 1 and t is a positive integer greater than or equal to 1: Process the task scheduling information set according to the genetic operator of the t-th round of genetic algorithm to obtain the task processing result corresponding to the t-th round of genetic algorithm, where the task processing result represents the optimized set of task scheduling information; Process the task processing result corresponding to the t-th round of genetic algorithm to obtain the intermediate features of the task scheduling information and the fitness of the task scheduling information; where the task scheduling information includes the scheduling order of each task; Determine the genetic operator of the (t + 1)-th round of genetic algorithm according to the intermediate features of each task scheduling information, and determine that the value of t is incremented by 1;

[0009] Based on the genetic operator obtained when the preset condition is reached, process the task scheduling information set to obtain a set of task scheduling information with higher fitness; where the task scheduling information represents the scheduling order between the tasks.

[0010] In a feasible implementation, the task processing result corresponding to the t-th round of genetic algorithm is processed to obtain intermediate features of the task scheduling information, including:

[0011] The task processing result corresponding to the t-th round of genetic algorithm is input into a preset multi-layer perceptron and graph neural network model for processing to obtain intermediate features of the task scheduling information.

[0012] In a feasible implementation, the task processing result corresponding to the t-th round of genetic algorithm is input into a preset multi-layer perceptron and graph neural network model for processing to obtain intermediate features of the task scheduling information, including:

[0013] The task processing result corresponding to the t-th round of genetic algorithm is input into the preset multi-layer perceptron and graph neural network model, and for the scheduling order relationship of each task in the task scheduling information in the task processing result corresponding to the t-th round of genetic algorithm, processing is performed to output intermediate features of each task scheduling information.

[0014] In a feasible implementation, according to the intermediate features of the task scheduling information, the genetic operator of the (t + 1)-th round of genetic algorithm is determined, including:

[0015] Based on a reinforcement learning model, the intermediate features of the task scheduling information are processed to obtain the genetic operator of the (t + 1)-th round of genetic algorithm.

[0016] In a feasible implementation, based on a reinforcement learning model, the intermediate features of the task scheduling information are processed to obtain the genetic operator of the (t + 1)-th round of genetic algorithm, including:

[0017] According to each task scheduling information, an objective function of the genetic algorithm is established;

[0018] Based on the objective function, the task scheduling information with the highest fitness obtained by the t-th round of genetic algorithm and the set of task scheduling information obtained by the (t - 1)-th round of genetic algorithm are processed to determine reward information; the reward information is used to indicate the selection of the genetic operator;

[0019] Based on the reinforcement learning model, the genetic operator corresponding to the intermediate features when the value of the reward information is maximized is determined as the genetic operator of the (t + 1)-th round of genetic algorithm.

[0020] In a feasible implementation, each task scheduling information has a fitness, where the fitness characterizes the quality of the task scheduling information.

[0021] In a feasible implementation, each task in the task scheduling information has task information, and the task information includes the start execution time and the task duration of the task; the method further includes:

[0022] Select the task scheduling information with the highest fitness in the task scheduling information set, perform topological sorting on the tasks in the task scheduling information to obtain the processed task scheduling information;

[0023] Determine the start time of each task according to the topological relationship between the tasks in the processed task scheduling information and the task information of each task.

[0024] In a feasible implementation, after determining the start time of each task according to the topological relationship between the tasks in the processed task scheduling information and the task information of each task, it further includes:

[0025] Execute each task according to the topological relationship between the tasks in the processed task set and the start time of each task.

[0026] In a second aspect, the present application provides a device for determining task scheduling information based on a genetic algorithm, including:

[0027] An acquisition unit that acquires a task scheduling information set, where the task scheduling information set includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task;

[0028] A first determination unit that repeatedly executes the following process until a preset condition is reached, where the initial value of t is 1 and t is a positive integer greater than or equal to 1: Process the task scheduling information set according to the genetic operator of the t-th round of genetic algorithm to obtain the task processing result corresponding to the t-th round of genetic algorithm, where the task processing result represents the optimized set of task scheduling information; Process the task processing result corresponding to the t-th round of genetic algorithm to obtain the intermediate features of the task scheduling information and the fitness of the task scheduling information; where the task scheduling information includes the scheduling order of each task; Determine the genetic operator of the (t + 1)-th round of genetic algorithm according to the intermediate features of each task scheduling information, and determine that the value of t is incremented by 1;

[0029] A second determination unit that processes the task scheduling information set based on the genetic operator obtained when the preset condition is reached to obtain a set of task scheduling information with higher fitness; where the task scheduling information represents the scheduling order between the tasks.

[0030] In a feasible implementation manner, the first determination unit processes the task processing result corresponding to the t-th round of genetic algorithm to obtain the intermediate feature of the task scheduling information, specifically used for:

[0031] Input the task processing result corresponding to the t-th round of genetic algorithm into a preset multi-layer perceptron and graph neural network model for processing to obtain the intermediate feature of the task scheduling information.

[0032] In a feasible implementation manner, when the first determination unit inputs the task processing result corresponding to the t-th round of genetic algorithm into a preset multi-layer perceptron and graph neural network model for processing to obtain the intermediate feature of the task scheduling information, it includes:

[0033] Input the task processing result corresponding to the t-th round of genetic algorithm into the preset multi-layer perceptron and graph neural network model, and process the scheduling order relationship of each task in the task processing result corresponding to the t-th round of genetic algorithm in the task scheduling information, and output the intermediate feature of each task scheduling information.

[0034] In a feasible implementation manner, when the first determination unit determines the genetic operator of the (t + 1)-th round of genetic algorithm according to the intermediate feature of the task scheduling information, it is specifically used for:

[0035] Process the intermediate feature of the task scheduling information based on a reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of genetic algorithm.

[0036] In a feasible implementation manner, when the first determination unit processes the intermediate feature of the task scheduling information based on a reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of genetic algorithm, it is specifically used for:

[0037] Establish an objective function of the genetic algorithm according to each task scheduling information;

[0038] Based on the objective function, process the task scheduling information with the highest fitness obtained by the t-th round of genetic algorithm and the set of task scheduling information obtained by the (t - 1)-th round of genetic algorithm to determine the reward information; the reward information is used to indicate the selection of the genetic operator;

[0039] Based on the reinforcement learning model, determine that the genetic operator corresponding to the intermediate feature when the value of the reward information is the largest is the genetic operator of the (t + 1)-th round of genetic algorithm.

[0040] In a feasible implementation manner, each task scheduling information has a fitness, where the fitness characterizes the quality of the task scheduling information.

[0041] In a feasible embodiment, each task in the task scheduling information has task information, and the task information includes the start execution time and the task duration of the task; the device further includes:

[0042] A processing unit, which selects the task scheduling information with the highest fitness in the task scheduling information set, performs a topological sorting process on the tasks in the task scheduling information, and obtains the processed task scheduling information;

[0043] A third determination unit, which determines the start time of each task according to the topological relationship between the tasks in the processed task scheduling information and the task information of each task.

[0044] In a feasible embodiment, the device further includes:

[0045] An execution unit, configured to, after the third determination unit determines the start time of each task according to the topological relationship between the tasks in the processed task scheduling information and the task information of each task; execute each task according to the topological relationship between the tasks in the processed task set and the start time of each task.

[0046] In a third aspect, an embodiment of the present application provides an electronic device, including: a memory, a processor;

[0047] A memory; a memory for storing executable instructions of the processor;

[0048] Wherein, the processor is configured to execute the method described in the first aspect.

[0049] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, they are used to implement the method described in the first aspect.

[0050] In a fifth aspect, an embodiment of the present application provides a computer program product, the computer program product includes: a computer program, the computer program is stored in a readable storage medium, and at least one processor of an electronic device can read the computer program from the readable storage medium, and the at least one processor executes the computer program to enable the electronic device to execute the method described in the first aspect.

[0051] The method, apparatus, and device for determining task scheduling information based on a genetic algorithm provided in this application obtain a task scheduling information set, where the task scheduling information set includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task; process the task scheduling information set according to the genetic operator of the t-th round of the genetic algorithm to obtain the task processing result corresponding to the t-th round of the genetic algorithm, where the task processing result represents the optimized set of task scheduling information; process the task processing result corresponding to the t-th round of the genetic algorithm to obtain the intermediate features of the task scheduling information and the fitness of the task scheduling information; where the task scheduling information includes the scheduling order of each task; determine the genetic operator of the (t + 1)-th round of the genetic algorithm according to the intermediate features of each task scheduling information; process the task scheduling information set based on the genetic operator obtained when the preset condition is reached to obtain a set of task scheduling information with higher fitness; where the task scheduling information represents the scheduling order between tasks. Based on the method for determining task scheduling information based on a genetic algorithm provided in this embodiment, the characteristics of the genetic algorithm that combines breadth and depth search can be fully utilized, automatically adapt the genetic operator used in each round, and improve the accuracy of scheduling task information; moreover, using a preset multi-layer perceptron and a graph neural network model to process the task processing result to obtain intermediate features, and processing the intermediate features based on a reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of the genetic algorithm can, to a certain extent, avoid the problem of uneven changes in the genetic algorithm and improve the solution efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0053] Figure 1 It is a schematic flowchart of a method for determining task scheduling information based on a genetic algorithm provided by an embodiment of the present application;

[0054] Figure 2 It is a schematic flowchart of another method for determining task scheduling information based on a genetic algorithm provided by an embodiment of the present application;

[0055] Figure 3 It is a schematic structural diagram of a device for determining task scheduling information based on a genetic algorithm provided by an embodiment of the present application;

[0056] Figure 4 It is a schematic structural diagram of another device for determining task scheduling information based on a genetic algorithm provided by an embodiment of the present application;

[0057] Figure 5 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application;

[0058] Figure 6 Block diagram of an electronic device provided by an embodiment of the present application.

[0059] Through the above-mentioned drawings, specific embodiments of the present application have been shown, and more detailed descriptions will be provided hereinafter. These drawings and textual descriptions are not intended to limit the scope of the concept of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. Detailed implementation manners

[0060] Here, exemplary embodiments will be described in detail, and examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numerals in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0061] During the process of business processing, there are multiple tasks under the business. There is an execution order among these tasks, and it is necessary to determine the execution order among each task, and then execute these tasks to complete the business processing. That is, it is necessary to determine the task scheduling information. The genetic algorithm can quickly explore the set of feasible task scheduling information and screen out high-quality task scheduling information. The traditional genetic algorithm uses fixed operators and cannot achieve the best effect.

[0062] In one example, the NeuRewriter method regards the optimization problem as a rewriting problem, learns strategies to rewrite local components of the current solution, and continuously improves the solution in an iterative manner. The learning strategy is divided into two parts. The region selection strategy selects the region to be rewritten according to the current state; the rule selection strategy selects the rewriting rule. The network model is represented by a neural network and trained using the Actor-Critic method in reinforcement learning. Assume that X is the set of all feasible solutions, c: X → R is the cost function, then the goal of combinatorial optimization is to find the feasible solution with the minimum cost function, expressed as argmin x∈X c(x). y is the set of rewritten rules. Assume that is the solution obtained in the s t-th iteration, is the rewritable region corresponding to the current solution. The region selection strategy is used to select the region to be rewritten According to the selected the rule selection strategy is used to select the rewriting rule used in this iteration is the region to be rewritten, satisfying The selected rewriting rule Use in selected area Get the next solution Given an initial state (solution) g0, the goal of the strategy is to find a rewriting path Finally minimize the cost function

[0063] However, the above method is equivalent to a genetic algorithm with a population of 1. The population size is limited and only a single mutation operator (exchange) is retained. There is only breadth search, which does not fully utilize the characteristics of the genetic algorithm that takes into account both depth and breadth search. It is also a local search algorithm with low search efficiency and high training cost.

[0064] In one example, the Reinforced Genetic Algorithm Learning (REGAL) method uses reinforcement learning to improve the Biased Random Key Genetic Algorithm (BRKGA). The mutation operator in BRKGA is defined as a new individual directly initialized according to the random key. REGAL changes the random key of the mutation operator in BRKGA to the learned beta distribution, with the goal of maximizing the target reward. The training uses the Reinforce algorithm. In REGAL, the graph G = (B, D) is defined, and the graph strategy z(a|G) represents the calculation of the priority corresponding to each gene position, where a = {a b∈B}, corresponding to the specific beta distribution parameters at each gene position. Reward function Among them s (G) represents the objective function value calculated on graph G using standard BRKGA. a (G) represents the result calculated on G using REGAL.

[0065] However, among the above methods, REGAL lacks the adaptive process for the solution process and the problem, and is limited to the BRKGA paradigm and has no scalability.

[0066] The present application provides a method, device and equipment for determining task scheduling information based on a genetic algorithm, aiming to solve the above technical problems in the prior art.

[0067] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0068] Figure 1The flowchart of a method for determining task scheduling information based on a genetic algorithm provided by an embodiment of this application is shown as Figure 1 follows. The method includes:

[0069] S101. Obtain a task scheduling information set, where the task scheduling information set includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task.

[0070] Exemplarily, the execution subject of this embodiment may be an electronic device, or a terminal device, or a server, or a controller, or other devices or apparatuses that can execute this embodiment, and there is no limitation thereto.

[0071] The scheduling problem consists of multiple tasks, and the task scheduling information is the scheduling order of all tasks in the scheduling problem.

[0072] Among them, the scheduling problem refers to allocating limited resources to several tasks over time under certain constraint conditions to obtain the optimal task execution order.

[0073] Perform multiple rounds of random sorting on the tasks of the scheduling problem to obtain multiple task scheduling information, and the multiple task scheduling information constitutes the task scheduling information set.

[0074] For example, a scheduling problem consists of m tasks π1, π2,..., π m . Perform a round of random sorting on the m tasks to obtain a task scheduling information I1, where I1 = (π1, π2,..., π m ); perform another round of random sorting on the m tasks to obtain task scheduling information I2, where I2 = (π2, π1,..., π m ). Among them, m is a positive integer greater than 1.

[0075] Perform n rounds of random sorting on the m tasks to obtain n task scheduling information, and the n task scheduling information constitutes the task scheduling information set I, where I = {I1, I2,..., I n}. Among them, n is a positive integer greater than 1.

[0076] S102. Repeat the following process until a preset condition is met. Here, the initial value of t is 1, and t is a positive integer greater than or equal to 1: Process the task scheduling information set according to the genetic operator of the t-th round of the genetic algorithm to obtain the task processing result corresponding to the t-th round of the genetic algorithm, where the task processing result represents the optimized set of task scheduling information; Process the task processing result corresponding to the t-th round of the genetic algorithm to obtain the intermediate features of the task scheduling information and the fitness of the task scheduling information; where the task scheduling information includes the scheduling order of each task; Determine the genetic operator of the (t + 1)-th round of the genetic algorithm based on the intermediate features of each task scheduling information, and determine that the value of t is incremented by 1.

[0077] Exemplarily, after obtaining the task scheduling information set, process the task scheduling information set based on the genetic algorithm to obtain the corresponding task processing result, and the corresponding task processing result is the optimized task scheduling information set; Process the task processing result to obtain the intermediate features of the task processing result and the fitness of the task scheduling information, and determine the genetic operator of the next round of the genetic algorithm based on the intermediate features of the task processing result.

[0078] Among them, the genetic algorithm is a computational model that simulates the natural selection of Darwin's biological evolution theory and the biological evolution process of genetic mechanisms, and is a method for searching for the optimal solution by simulating the natural evolution process; The genetic operator is a step of the genetic algorithm. The basic genetic algorithm includes a selection operator, a crossover operator, and a mutation operator. Among them, the selection operator is responsible for deeply searching the accumulated information, and the crossover operator and the mutation operator are responsible for widely searching new regions in the solution space.

[0079] Based on the genetic algorithm to process the task scheduling information set to obtain the corresponding task processing result, the calculation process of the t-th round of the genetic algorithm specifically includes:

[0080] Determine the initial parameters of the genetic algorithm, including the maximum number of genetic generations T, the population size O, the mating pool size S, and the number of elite individuals P, where T, O, and S are all positive integers greater than or equal to 2, and P is a positive integer greater than or equal to 1;

[0081] Among them, the maximum number of genetic generations is the preset number of times for the genetic algorithm to perform; The population is the number of task scheduling information in the obtained task scheduling information set; The mating pool is a set of relatively good task scheduling information in the task scheduling information set. For example, all scheduling tasks have a short completion time, or all scheduling tasks cost less to complete; The elite individuals are the set of the few optimal task scheduling information in the task scheduling information set;

[0082] Reserve P elite individuals from the task scheduling information set;

[0083] Selection operator, which selects a breeding pool from the set of task scheduling information, i.e., a set of S task scheduling information;

[0084] Crossover operator, which combines the task scheduling information in the breeding pool pairwise to generate offspring individuals, i.e., new task scheduling information;

[0085] Mutation operator, which performs a mutation operation on the offspring individuals obtained by the crossover operator to obtain the mutated task scheduling information;

[0086] The mutated task scheduling information is merged with the elite individuals to obtain an optimized set of task scheduling information, i.e., the task processing result.

[0087] In this embodiment, the selection operator, crossover operator, and mutation operator can select all operators suitable for the standard genetic algorithm, which has good scalability; moreover, some operators that are more practical for the scheduling problem in the genetic algorithm are adopted. For example, the selection operator can be one of the selection operators such as steady selection, tournament selection, ranking selection, random selection, roulette selection, stochastic universal selection, etc.; the crossover operator can be one of the crossover operators such as single-point exchange, preorder-preserving exchange, two-point exchange, uniform exchange, etc.; the mutation operator can be one of the mutation operators such as exchange, scramble, reverse order, random, etc.

[0088] For example, the obtained set of task scheduling information I = {I1, I2,..., I n}, and the breeding pool to be selected is S, and the size of the breeding pool is n s where n s is a positive integer greater than 1, and the processing process of the selection operator is as follows:

[0089] Steady selection, randomly select n s task scheduling information to form a new set of task scheduling information S, S = {I i |i = 1, 2,..., n s};

[0090] Tournament selection, each time randomly select a certain number of task scheduling information from I, put the task scheduling information with the highest fitness into the breeding pool, and perform n s rounds of operations to select n s task scheduling information to form a new set of task scheduling information S, S = {I i |i = 1, 2,..., n s}; n s is a positive integer greater than 1;

[0091] Ranking selection, sort the task scheduling information in I in descending order of fitness, where the probability that I i is selected is Among them, n is the number of task scheduling information in the task scheduling information set I, n is a positive integer greater than 1, and i is the ranking of the fitness of i in the task scheduling information set; perform n s rounds of operations to select n s task scheduling information to form a new task scheduling information set S, S = {I i | i = 1, 2,..., n s}; n s is a positive integer greater than 1;

[0092] Random selection, the probability of the task scheduling information in I being selected is the same, all being Among them, n is the number of task scheduling information in the task scheduling information set I, n is a positive integer greater than 1, perform n s rounds of operations to select n s task scheduling information to form a new task scheduling information set S, S = {I i | i = 1, 2,..., n s}; n s is a positive integer greater than 1;

[0093] Roulette wheel selection, the probability of the task scheduling information in I being selected is proportional to the fitness of the task scheduling information, then the probability of I i being selected is Among them, j is a positive integer greater than or equal to 1 and less than or equal to n, perform n s rounds of operations to select n s task scheduling information to form a new task scheduling information set S, S = {I i | i = 1, 2,..., n s}; n s is a positive integer greater than 1;

[0094] Stochastic universal selection, an operator improved based on roulette wheel selection, select n s task scheduling information, generate n s equally spaced marker pointer positions according to the probability, define the distance between the two marker pointers corresponding to I i as [L i , U i ], where, L i represents the upper bound of the distance, U i represents the lower bound of the distance, L1 = 0, for i > 1, L i+1 = U i , where, j is a positive integer greater than or equal to 1 and less than or equal to n, the positions of each marker pointer Among them According to the position of the pointer, select I i ​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​s.t.p i ∈ [L i , U i ), where s.t. is a mathematical symbol indicating that the following content holds; the new task scheduling information set S = {I i | i = 1, 2,..., n s}; n s is a positive integer greater than 1.

[0095] For example, randomly select two task scheduling information I i and I i+1 from the breeding pool. Their task scheduling orders are respectively, I i =(π1, π2,..., π i ,..., π j , π j+1 ,..., π m-1 , π m ), I i+1 =(π2, π1,..., π i ,..., π j+1 , π j ,..., π m , π m-1 ). Perform a mutation operation on I i and I i+1 to generate new individuals I j and I j+1 . The processing process of the mutation operator is as follows:

[0096] Single-point exchange: Randomly select any task in I i and I i+1 as the crossover point, and swap the tasks on the right side of the crossover point. If the selected crossover point is π i , then I j =(π1, π2,..., π i ,..., π j+1 , π j ,..., π m , π m-1 ), I j+1 =(π2, π1,..., π i ,..., π i , π j+1 ,..., π m-1 , π m );

[0097] Preorder-preserving exchange: Randomly select a task π i , splice I i and I i+1 from the selected π i , and iThe latter part of the task is rearranged according to the relative order of tasks in I i+1 That is, the newly generated task scheduling information is I j =(π1, π2,..., π i ,..., π j , π j+1 ,..., π m-1 , π m );

[0098] Two-point exchange: Randomly select two tasks from the task scheduling information I i and I i+1 as the crossover points, and exchange the tasks in the middle of the two tasks. If the second task and the (m - 1)-th task in I i and I i+1 are selected as the crossover points, where m is a positive integer greater than 1, then I j =(π1, π2,..., π i ,..., π j+1 , π j ,..., π m , π m-1 ), I j+1 =(π2,..., π i ,..., π i , π j+1 ,..., π m-1 , π m );

[0099] Uniform exchange: Randomly select tasks from I j and I j+1 with equal probability from each task in I i and I i+1 .

[0100] For example, for the newly generated task scheduling information I j =(π1, π2,..., π i ,..., π j+1 , π j ,..., π m , π m-1 ), process I j . The processing process of the mutation operator is as follows:

[0101] Exchange: Randomly select two tasks on I j for exchange. If π i and π j are selected for exchange, then I j =(π1, π2,..., π j ,..., π j+1 , π i,..., π m , π m-1 );

[0102] Shuffle and select I j For the partial tasks on, randomly arrange them. If the tasks between the second task and the (m - 1)-th task are randomly arranged, where m is a positive integer greater than 1, then I j can be, I j =(π1, π2,..., π j ,..., π i , π j+1 , π i ,..., π m , π m-1 );

[0103] Reverse the order and select I j Two points on, I j is divided into three parts, and reverse the order of the middle part. If the second task and the (m - 1)-th task are selected as the demarcation points, m is a positive integer greater than 1, then I j =(π1, π2,..., π j , π j+1 ,..., π i ,..., π m , π m-1 ).

[0104] This embodiment also selects other operators in the genetic algorithm that are more practical for the scheduling problem, such as adaptive mutation. Assume that after the selection operator, crossover operator, and mutation operator of the t-th round of the genetic algorithm are executed, the And, it has been sorted in descending order of fitness, and the corresponding fitness The mutation probability is p m , for If Then The mutation probability of is set to Otherwise, it is p m .

[0105] Process the task processing results of the t-th round of the genetic algorithm to obtain the intermediate features of the task processing results; at the same time, use the objective function of the genetic algorithm to calculate the fitness of each task scheduling information, and determine the genetic operator of the next round of the genetic algorithm according to the intermediate features of the task processing results, where the initial value of t is 1, and t is a positive integer greater than or equal to 1 and less than T.

[0106] Among them, the fitness of each task scheduling information is calculated by the objective function of the genetic algorithm, and the objective function of the genetic algorithm is related to the completion time of the scheduling task.

[0107] The determined genetic operators are a selection operator, a crossover operator, and a mutation operator, which are used for the calculation of the (t + 1)-th round of genetic algorithm.

[0108] S103. Process the task scheduling information set based on the genetic operators obtained when a preset condition is reached, to obtain a set of task scheduling information with higher fitness; wherein, the task scheduling information represents the scheduling order between tasks.

[0109] Exemplarily, process the task scheduling information set based on the genetic operators obtained when the genetic algorithm reaches the maximum number of genetic generations, to obtain a set of task scheduling information with higher fitness.

[0110] In this embodiment, by obtaining a task scheduling information set, where the task scheduling information set includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task; process the task scheduling information set according to the genetic operators of the t-th round of genetic algorithm, to obtain the task processing result corresponding to the t-th round of genetic algorithm, where the task processing result represents the optimized set of task scheduling information; process the task processing result corresponding to the t-th round of genetic algorithm, to obtain the intermediate features of the task scheduling information and the fitness of the task scheduling information; wherein, the task scheduling information includes the scheduling order of each task; determine the genetic operators of the (t + 1)-th round of genetic algorithm according to the intermediate features of each task scheduling information; process the task scheduling information set based on the genetic operators obtained when a preset condition is reached, to obtain a set of task scheduling information with higher fitness; wherein, the task scheduling information represents the scheduling order between tasks. Based on the method for determining task scheduling information based on genetic algorithm provided in this embodiment, the characteristics of the genetic algorithm that takes into account both breadth and depth search can be fully utilized, the genetic operators used in each round can be automatically adapted, and the accuracy of scheduling task information can be improved.

[0111] Figure 2 For another flowchart of the method for determining task scheduling information based on genetic algorithm provided in the embodiments of the present application, as Figure 2 shown, the method includes:

[0112] S201. Obtain a task scheduling information set, where the task scheduling information set includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task.

[0113] Exemplarily, the execution subject of this embodiment may be an electronic device, or a terminal device, or a server, or a controller, or other devices or apparatuses that can execute this embodiment, and no limitation is made thereto.

[0114] This step may refer to the above step S101 and will not be elaborated herein.

[0115] S202. Repeat the following process until a preset condition is met. Here, the initial value of t is 1, and t is a positive integer greater than or equal to 1: Process the task scheduling information set according to the genetic operator of the t-th round of the genetic algorithm to obtain the task processing result corresponding to the t-th round of the genetic algorithm, where the task processing result represents the optimized set of task scheduling information; Input the task processing result corresponding to the t-th round of the genetic algorithm into a preset multi-layer perceptron and graph neural network model for processing to obtain the intermediate features of the task scheduling information; Process the intermediate features of the task scheduling information based on the reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of the genetic algorithm, and determine that the value of t is incremented by 1.

[0116] In one example, for the process of "Inputting the task processing result corresponding to the t-th round of the genetic algorithm into a preset multi-layer perceptron and graph neural network model for processing to obtain the intermediate features of the task scheduling information" in the above process, it includes: Inputting the task processing result corresponding to the t-th round of the genetic algorithm into a preset multi-layer perceptron and graph neural network model, and processing the scheduling order relationship of each task in the task scheduling information in the task processing result corresponding to the t-th round of the genetic algorithm, and outputting the intermediate features of each task scheduling information.

[0117] In one example, for the process of "Processing the intermediate features of the task scheduling information based on the reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of the genetic algorithm" in the above process, it includes: Establishing an objective function of the genetic algorithm according to each task scheduling information; Based on the objective function, processing the task scheduling information with the highest fitness obtained in the t-th round of the genetic algorithm and the set of task scheduling information obtained in the (t - 1)-th round of the genetic algorithm to determine the reward information; The reward information is used to indicate the selection of the genetic operator; Based on the reinforcement learning model, determining the genetic operator corresponding to the intermediate features when the value of the reward information is the largest as the genetic operator of the (t + 1)-th round of the genetic algorithm.

[0118] In one example, each task scheduling information has a fitness, where the fitness represents the quality of the task scheduling information.

[0119] Exemplarily, after obtaining the task scheduling information set, process the task scheduling information set based on the genetic algorithm to obtain the corresponding task processing result, and the corresponding task processing result is the optimized set of task scheduling information; Based on a preset multi-layer perceptron and graph neural network model, process the task processing result to obtain the intermediate features of the task processing result and the fitness of the task scheduling information; Based on the reinforcement learning model, determine the genetic operator of the next round of the genetic algorithm according to the intermediate features of the task processing result.

[0120] For the processing of the task scheduling information set by the genetic algorithm, reference can be made to the introduction in S102, which will not be elaborated here.

[0121] Before processing the task processing results corresponding to the genetic algorithm, it is first necessary to obtain a multi-layer perceptron and a graph neural network. The multi-layer perceptron and the graph neural network process the task processing results obtained by the genetic algorithm to obtain intermediate features of task scheduling information; and a reinforcement learning model is obtained. The reinforcement learning model determines the genetic operator used in the next round of the genetic algorithm based on the intermediate features of the task scheduling information. Among them, the reinforcement learning model is trained using the standard proximal policy optimization algorithm, or any standard training algorithm is used, such as the asynchronous advantage actor-critic algorithm.

[0122] Among them, the graph neural network is an algorithm that aims to represent the vertices in the graph as low-dimensional vectors by retaining the network topology structure and node content information of the graph, so as to be processed using simple machine learning algorithms; reinforcement learning is one of the paradigms and methodologies of machine learning, used to describe and solve the problem that an intelligent agent maximizes the reward or achieves a specific goal by learning strategies during the interaction with the environment; the reinforcement learning environment consists of four parts: state, action, reward, and transition.

[0123] In this embodiment, the state of reinforcement learning is composed of the scheduling problem and the set of scheduling task information; the action of reinforcement learning is to select the genetic operator for the next round of the genetic algorithm; the transition of reinforcement learning refers to the transition of the reinforcement learning state. After the genetic operator of the genetic algorithm performs one round of operations to obtain an optimized set of task scheduling information, the state of reinforcement learning will also change; the reward of reinforcement learning is the feedback signal obtained by reinforcement learning after performing one round of operations of the genetic algorithm.

[0124] In the process of "inputting the task processing results corresponding to the t-th round of the genetic algorithm into a preset multi-layer perceptron and graph neural network model for processing to obtain intermediate features of task scheduling information", the task processing results obtained for the t-th round of the genetic algorithm are For I t Establish G=(V, E), where V represents the information of the tasks, including the start execution time and duration of the tasks, and E represents the precedence relationship between the tasks. The calculation process of the multi-layer perceptron and the graph neural network model is as follows:

[0125]

[0126]

[0127]

[0128] Among them, x v represents each task in the task scheduling information; MLP represents the multi-layer perceptron, MLP0 represents the 0th layer of the multi-layer perceptron, MLPt Denote the t-th layer of the multi-layer perceptron; and respectively represent the results obtained by passing each task in the task scheduling information through the 0-th layer and the k-th layer of the multi-layer perceptron; (u, v) ∈ E indicates that task v is a prerequisite for task u, and E represents the precedence relationship between tasks. Denote the summation calculation of the graph neural network; h represents the intermediate feature of the task scheduling information, which is the calculation result of each task through the multi-layer perceptron and the graph neural network.

[0129] h i Denote h obtained through the calculation of the graph neural network, f i Denote the fitness of, then the state s of the reinforcement learning t is:

[0130]

[0131] In "processing the intermediate features of the task scheduling information based on the reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of the genetic algorithm", according to each task scheduling information, establish the objective function f(I) of the genetic algorithm, and use the objective function to process the task scheduling information with the highest fitness obtained in the t-th round of the genetic algorithm to obtain f(I t ), and process the set of task scheduling information obtained in the (t - 1)-th round of the genetic algorithm to obtain f(I t-1 ), where the reward information r t = f(I t ) - f(I t-1 ), and the reward information is used to indicate the selection of the genetic operator; the reinforcement learning model determines the genetic operator that maximizes the value of r t according to the intermediate features of the task processing results and the fitness of the task processing results, as the genetic operator of the (t + 1)-th round of the genetic algorithm.

[0132] Among them, the fitness of each task scheduling information is calculated by the objective function of the genetic algorithm, and the objective function of the genetic algorithm is related to the completion time of the scheduling tasks.

[0133] For example, the higher the fitness of the task scheduling information, the shorter the completion time or the less cost it means to execute all tasks according to the task scheduling order of the task scheduling information.

[0134] S203. Process the set of task scheduling information based on the genetic operator obtained when reaching the preset condition to obtain a set of task scheduling information with higher fitness; where the task scheduling information represents the scheduling order between tasks.

[0135] Exemplarily, this step can refer to the above step S103 and will not be elaborated further.

[0136] After repeatedly executing S201 - S203 multiple times, when the preset condition is determined, the genetic operator corresponding to the maximum genetic algebra of the genetic algorithm can be determined. Among them, the preset condition is that the value of t reaches T + 1.

[0137] After reaching the preset condition, based on the genetic operator corresponding to the maximum genetic algebra of the genetic algorithm, the task scheduling information set is optimized to obtain the final task processing result.

[0138] S204. Select the task scheduling information with the highest fitness in the task scheduling information set, and perform topological sorting on the tasks in the task scheduling information to obtain the processed task scheduling information.

[0139] Exemplarily, select the task scheduling information with the highest fitness in the task processing result obtained by the genetic algorithm. For example, according to the task scheduling order corresponding to this task scheduling information, the time taken to complete all tasks is the shortest, or according to the task scheduling order corresponding to this task scheduling information, the cost of completing all tasks is the least; perform topological sorting on the tasks in this task scheduling information to obtain the processed task scheduling information.

[0140] Each task in the task scheduling information has task information, and the task information includes the start execution time and task duration of the task; according to the topological relationship between tasks in the processed task scheduling information and the task information of each task, determine the start time of each task.

[0141] Exemplarily, each task in the task scheduling information has a start execution time and a task duration. According to the topological relationship between tasks in the task scheduling information after topological sorting and the task information, determine the start time of the task.

[0142] S206. According to the topological relationship between tasks in the processed task set and the start time of each task, execute each task.

[0143] Exemplarily, according to the topological relationship between tasks in the task scheduling information after topological sorting and the task information, determine the start time of the task, and execute each task to generate a scheduling plan for the scheduling problem.

[0144] In this embodiment, by obtaining a set of task scheduling information, where the set of task scheduling information includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task; processing the set of task scheduling information according to the genetic operator of the t-th round of genetic algorithm to obtain the task processing result corresponding to the t-th round of genetic algorithm, where the task processing result represents the optimized set of task scheduling information; processing the task processing result corresponding to the t-th round of genetic algorithm to obtain the intermediate feature of the task scheduling information and the fitness of the task scheduling information; where the task scheduling information includes the scheduling order of each task; determining the genetic operator of the (t + 1)-th round of genetic algorithm according to the intermediate features of each task scheduling information; based on the genetic operator obtained when reaching the preset condition, processing the set of task scheduling information to obtain a set of task scheduling information with higher fitness; where the task scheduling information represents the scheduling order between tasks. Based on the method for determining task scheduling information based on genetic algorithm provided in this embodiment, the characteristics of the genetic algorithm that takes into account both breadth and depth search can be fully utilized, automatically adapting the genetic operator used in each round, and improving the accuracy of scheduling task information; moreover, using the preset multi-layer perceptron and graph neural network model to process the task processing result to obtain the intermediate feature, and processing the intermediate feature based on the reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of genetic algorithm can, to a certain extent, avoid the problem of uneven changes in the genetic algorithm and improve the solution efficiency.

[0145] Figure 3 FIG. is a schematic structural diagram of a device for determining task scheduling information based on genetic algorithm provided in an embodiment of the present application, as Figure 3 shown, the device includes:

[0146] An obtaining unit 31, which obtains a set of task scheduling information, where the set of task scheduling information includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task.

[0147] A first determining unit 32, which repeatedly executes the following process until the preset condition is reached, where the initial value of t is 1 and t is a positive integer greater than or equal to 1: Processing the set of task scheduling information according to the genetic operator of the t-th round of genetic algorithm to obtain the task processing result corresponding to the t-th round of genetic algorithm, where the task processing result represents the optimized set of task scheduling information; processing the task processing result corresponding to the t-th round of genetic algorithm to obtain the intermediate feature of the task scheduling information and the fitness of the task scheduling information; where the task scheduling information includes the scheduling order of each task; determining the genetic operator of the (t + 1)-th round of genetic algorithm according to the intermediate features of each task scheduling information, and determining that the value of t is incremented by 1.

[0148] The second determination unit 33 processes the task scheduling information set based on the genetic operator obtained when a preset condition is reached, and obtains a set of task scheduling information with higher fitness; wherein, the task scheduling information represents the scheduling order between tasks.

[0149] The device in this embodiment can execute the technical solutions in the above method, and its specific implementation process and technical principle are the same, which will not be elaborated here.

[0150] Figure 4 As shown in the structural schematic diagram of another device for determining task scheduling information based on a genetic algorithm provided by an embodiment of the present application, Figure 4 as shown, the device includes:

[0151] An acquisition unit 41 acquires a task scheduling information set, where the task scheduling information set includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task.

[0152] A first determination unit 42 repeatedly executes the following process until a preset condition is reached, where the initial value of t is 1 and t is a positive integer greater than or equal to 1: Process the task scheduling information set according to the genetic operator of the t-th round of the genetic algorithm to obtain the task processing result corresponding to the t-th round of the genetic algorithm, where the task processing result represents an optimized set of task scheduling information; Process the task processing result corresponding to the t-th round of the genetic algorithm to obtain the intermediate features of the task scheduling information and the fitness of the task scheduling information; wherein, the task scheduling information includes the scheduling order of each task; Determine the genetic operator of the (t + 1)-th round of the genetic algorithm according to the intermediate features of each task scheduling information, and determine that the value of t is incremented by 1.

[0153] The second determination unit 43 processes the task scheduling information set based on the genetic operator obtained when a preset condition is reached, and obtains a set of task scheduling information with higher fitness; wherein, the task scheduling information represents the scheduling order between tasks.

[0154] In one example, when the first determination unit 42 processes the task processing result corresponding to the t-th round of the genetic algorithm to obtain the intermediate features of the task scheduling information, it is specifically used for:

[0155] Input the task processing result corresponding to the t-th round of the genetic algorithm into a preset multi-layer perceptron and graph neural network model for processing to obtain the intermediate features of the task scheduling information.

[0156] In one example, when the first determination unit 42 inputs the task processing result corresponding to the t-th round of the genetic algorithm into a preset multi-layer perceptron and graph neural network model for processing to obtain the intermediate features of the task scheduling information, it includes:

[0157] Input the task processing result corresponding to the t-th round of the genetic algorithm into a preset multi-layer perceptron and graph neural network model, and process the scheduling order relationship of each task in the task scheduling information in the task processing result corresponding to the t-th round of the genetic algorithm, and output the intermediate features of each task scheduling information.

[0158] In one example, the first determination unit 42 determines the genetic operator of the (t + 1)-th round of the genetic algorithm according to the intermediate features of the task scheduling information, specifically for:

[0159] Process the intermediate features of the task scheduling information based on the reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of the genetic algorithm.

[0160] In one example, when the first determination unit 42 processes the intermediate features of the task scheduling information based on the reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of the genetic algorithm, it is specifically for:

[0161] Establish an objective function of the genetic algorithm according to each task scheduling information; based on the objective function, process the task scheduling information with the highest fitness obtained in the t-th round of the genetic algorithm and the set of task scheduling information obtained in the (t - 1)-th round of the genetic algorithm to determine the reward information; the reward information is used to indicate the selection of the genetic operator; based on the reinforcement learning model, determine the genetic operator corresponding to the intermediate features when the value of the reward information is the largest as the genetic operator of the (t + 1)-th round of the genetic algorithm.

[0162] In one example, each task scheduling information has a fitness, where the fitness characterizes the quality of the task scheduling information.

[0163] In one example, each task in the task scheduling information has task information, and the task information includes the start execution time and task duration of the task; the device provided in this embodiment further includes:

[0164] The processing unit 44 selects the task scheduling information with the highest fitness in the set of task scheduling information, and performs topological sorting on the tasks in the task scheduling information to obtain the processed task scheduling information.

[0165] The third determination unit 45 determines the start time of each task according to the topological relationship between the tasks in the processed task scheduling information and the task information of each task.

[0166] In one example, the device provided in this embodiment further includes:

[0167] An execution unit 46, configured to, after a third determination unit 45 determines the start time of each task according to the topological relationship between tasks in the processed task scheduling information and the task information of each task; execute each task according to the topological relationship between tasks in the processed task set and the start time of each task.

[0168] The device of this embodiment can execute the technical solutions in the above method. The specific implementation process and technical principle are the same, and will not be elaborated here.

[0169] Figure 5 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As Figure 5 shown, the electronic device includes: a memory 51, a processor 52; the memory 51; a memory for storing executable instructions of the processor 52.

[0170] Among them, the processor 52 is configured to execute the method provided in any of the above embodiments.

[0171] The terminal device further includes a receiver 53 and a transmitter 54. The receiver 53 is used to receive instructions and data sent by other devices, and the transmitter 54 is used to send instructions and data to external devices.

[0172] Figure 6 It is a block diagram of an electronic device provided by an embodiment of the present application. The device can be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0173] The device 600 may include one or more of the following components: a processing component 602, a memory 604, a power component 606, a multimedia component 608, an audio component 610, an input / output (I / O) interface 612, a sensor component 614, and a communication component 616.

[0174] The processing component 602 generally controls the overall operation of the device 600, such as operations associated with display, telephone call, data communication, camera operation, and recording operation. The processing component 602 may include one or more processors 620 to execute instructions to complete all or part of the steps of the above method. In addition, the processing component 602 may include one or more modules to facilitate the interaction between the processing component 602 and other components. For example, the processing component 602 may include a multimedia module to facilitate the interaction between the multimedia component 608 and the processing component 602.

[0175] The memory 604 is configured to store various types of data to support the operation of the device 600. Examples of such data include instructions for any application or method operating on the device 600, contact data, phone book data, messages, pictures, videos, and the like. The memory 604 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0176] The power supply component 606 provides power to various components of the device 600. The power supply component 606 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 600.

[0177] The multimedia component 608 includes a screen that provides an output interface between the device 600 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can not only sense the boundaries of touch or swipe actions but also detect the duration and pressure associated with the touch or swipe operation. In some embodiments, the multimedia component 608 includes a front camera and / or a rear camera. When the device 600 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera can receive external multimedia data. Each of the front camera and the rear camera can be a fixed optical lens system or have focal length and optical zoom capabilities.

[0178] The audio component 610 is configured to output and / or input audio signals. For example, the audio component 610 includes a microphone (MIC) that is configured to receive external audio signals when the device 600 is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals can be further stored in the memory 604 or transmitted via the communication component 616. In some embodiments, the audio component 610 further includes a speaker for outputting audio signals.

[0179] The I / O interface 612 provides an interface between the processing component 602 and a peripheral interface module, which can be a keyboard, a click wheel, buttons, etc. These buttons can include, but are not limited to: a home button, a volume button, a power-on button, and a lock button.

[0180] The sensor assembly 614 includes one or more sensors for providing a status assessment of various aspects of the device 600. For example, the sensor assembly 614 can detect the on / off state of the device 600, the relative positioning of components, such as the display and keypad of the device 600. The sensor assembly 614 can also detect a change in the position of the device 600 or a component of the device 600, the presence or absence of user contact with the device 600, the orientation or acceleration / deceleration of the device 600, and the temperature change of the device 600. The sensor assembly 614 can include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 614 can also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 614 can further include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0181] The communication component 616 is configured to facilitate communication between the device 600 and other devices in a wired or wireless manner. The device 600 can access a wireless network based on communication standards, such as WiFi, 2G, or 3G, or a combination thereof. In an exemplary embodiment, the communication component 616 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 616 further includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0182] In an exemplary embodiment, the device 600 can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for performing the above method.

[0183] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as the memory 604 including instructions, and the above instructions can be executed by the processor 620 of the device 600 to complete the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0184] An embodiment of the present application also provides a non-transitory computer-readable storage medium. When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device can execute the above method.

[0185] The present application also provides a computer program product, which includes: a computer program stored in a readable storage medium. At least one processor of the electronic device can read the computer program from the readable storage medium, and the at least one processor executes the computer program to cause the electronic device to execute the solution provided in any of the above embodiments.

[0186] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any variations, uses, or adaptations of the present application, which follow the general principles of the present application and include common general knowledge or conventional technical means in the technical field not disclosed in the present application. The specification and embodiments are only regarded as exemplary, and the true scope and spirit of the present application are pointed out by the following claims.

[0187] It should be understood that the present application is not limited to the exact structures already described and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. A method for determining task scheduling information based on a genetic algorithm, characterized in that, The method includes: Obtaining a task scheduling information set, where the task scheduling information set includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task. Repeatedly execute the following process until a preset condition is reached, where the initial value of t is 1 and t is a positive integer greater than or equal to 1: Process the task scheduling information set according to the genetic operator of the t-th round of genetic algorithm to obtain the task processing result corresponding to the t-th round of genetic algorithm, where the task processing result represents the optimized set of task scheduling information; Input the task processing result corresponding to the t-th round of genetic algorithm into a preset multi-layer perceptron and graph neural network model, and process the scheduling order relationship of each task in the task scheduling information in the task processing result corresponding to the t-th round of genetic algorithm, and output the intermediate features of each task scheduling information and the fitness of the task scheduling information; where the task scheduling information includes the scheduling order of each task; Determine the genetic operator of the (t + 1)-th round of genetic algorithm according to the intermediate features of each task scheduling information, and determine that the value of t is incremented by 1. Based on the genetic operator obtained when the preset condition is reached, process the task scheduling information set to obtain a set of task scheduling information with higher fitness; where the task scheduling information represents the scheduling order between tasks.

2. The method according to claim 1, wherein Determining the genetic operator of the (t + 1)-th round of genetic algorithm according to the intermediate features of the task scheduling information includes: Processing the intermediate features of the task scheduling information based on a reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of genetic algorithm.

3. The method according to claim 2, wherein Processing the intermediate features of the task scheduling information based on a reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of genetic algorithm includes: Establish an objective function of the genetic algorithm according to each task scheduling information. Based on the objective function, process the task scheduling information with the highest fitness obtained by the t-th round of genetic algorithm and the set of task scheduling information obtained by the (t - 1)-th round of genetic algorithm to determine the reward information; the reward information is used to indicate the selection of the genetic operator. Based on the reinforcement learning model, determine that the genetic operator corresponding to the intermediate features when the value of the reward information is the largest is the genetic operator of the (t + 1)-th round of genetic algorithm.

4. The method according to any one of claims 1 to 3, characterized in that Each task scheduling information has a fitness, where the fitness represents the quality of the task scheduling information.

5. The method according to any one of claims 1-3, characterized in that Each task in the task scheduling information has task information, and the task information includes the start execution time and task duration of the task. The method further includes: Select the task scheduling information with the highest fitness in the task scheduling information set, and perform topological sorting on the tasks in the task scheduling information to obtain the processed task scheduling information. Determine the start time of each task according to the topological relationship between tasks in the processed task scheduling information and the task information of each task.

6. The method according to claim 5, wherein After determining the start time of each task according to the topological relationship between tasks in the processed task scheduling information and the task information of each task, the following is further included: Execute each task according to the topological relationship between tasks in the processed task set and the start time of each task.

7. An apparatus for determining task scheduling information based on a genetic algorithm, characterized in that, The device includes: An acquisition unit that acquires a task scheduling information set, where the task scheduling information set includes at least one task scheduling information; the task scheduling information is the scheduling order of at least one task. A first determination unit that repeatedly executes the following process until a preset condition is reached, where the initial value of t is 1 and t is a positive integer greater than or equal to 1: Process the task scheduling information set according to the genetic operator of the t-th round of genetic algorithm to obtain the task processing result corresponding to the t-th round of genetic algorithm, where the task processing result represents the optimized set of task scheduling information; Input the task processing result corresponding to the t-th round of genetic algorithm into the preset multi-layer perceptron and graph neural network model, and process the scheduling order relationship of each task in the task scheduling information in the task processing result corresponding to the t-th round of genetic algorithm, and output the intermediate feature of each task scheduling information and the fitness of the task scheduling information; where the task scheduling information includes the scheduling order of each task; Determine the genetic operator of the (t + 1)-th round of genetic algorithm according to the intermediate features of each task scheduling information, and determine that the value of t is incremented by 1. A second determination unit that processes the task scheduling information set based on the genetic operator obtained when the preset condition is reached to obtain a set of task scheduling information with higher fitness; where the task scheduling information represents the scheduling order between tasks.

8. The device according to claim 7, characterized in that, The first determination unit processes the task processing result corresponding to the t-th round of genetic algorithm to obtain the intermediate feature of the task scheduling information, specifically for: Input the task processing result corresponding to the t-th round of genetic algorithm into the preset multi-layer perceptron and graph neural network model for processing to obtain the intermediate feature of the task scheduling information.

9. The device according to claim 8, characterized in that, When the first determination unit inputs the task processing result corresponding to the t-th round of genetic algorithm into the preset multi-layer perceptron and graph neural network model for processing to obtain the intermediate feature of the task scheduling information, it includes: Input the task processing result corresponding to the t-th round of genetic algorithm into the preset multi-layer perceptron and graph neural network model, and process the scheduling order relationship of each task in the task scheduling information in the task processing result corresponding to the t-th round of genetic algorithm, and output the intermediate feature of each task scheduling information.

10. The device according to claim 7, characterized in that, When the first determination unit determines the genetic operator of the (t + 1)-th round of genetic algorithm according to the intermediate feature of the task scheduling information, it is specifically for: Process the intermediate feature of the task scheduling information based on the reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of genetic algorithm.

11. The device according to claim 10, wherein, The first determination unit processes the intermediate features of the task scheduling information based on the reinforcement learning model to obtain the genetic operator of the (t + 1)-th round of the genetic algorithm, specifically for: establishing an objective function of the genetic algorithm according to each piece of the task scheduling information; processing the task scheduling information with the highest fitness obtained by the t-th round of the genetic algorithm and the set of task scheduling information obtained by the (t - 1)-th round of the genetic algorithm based on the objective function to determine reward information; the reward information is used to indicate the selection of the genetic operator; determining, based on the reinforcement learning model, the genetic operator corresponding to the intermediate features when the value of the reward information is maximized as the genetic operator of the (t + 1)-th round of the genetic algorithm.

12. The device according to any one of claims 7-11, characterized in that, Each piece of the task scheduling information has a fitness, where the fitness characterizes the quality of the task scheduling information.

13. The device according to any one of claims 7-11, characterized in that, Each task in the task scheduling information has task information, and the task information includes the start execution time and the task duration of the task. The device further includes: a processing unit, which selects the task scheduling information with the highest fitness in the set of task scheduling information and performs topological sorting on the tasks in the task scheduling information to obtain the processed task scheduling information; a third determination unit, which determines the start time of each task according to the topological relationship between the tasks in the processed task scheduling information and the task information of each task.

14. The device according to claim 13, characterized in that, The device further includes: an execution unit, which is configured to, after the third determination unit determines the start time of each task according to the topological relationship between the tasks in the processed task scheduling information and the task information of each task; execute each task according to the topological relationship between the tasks in the processed task set and the start time of each task.

15. An electronic device, characterized in that, including: a processor and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1-6.

16. A computer-readable storage medium, characterized in that, Computer-executable instructions are stored in the computer-readable storage medium, and when the computer-executable instructions are executed by the processor, they are used to implement the method according to any one of claims 1-6.

17. A computer program product, characterized in that, including a computer program, which when executed by the processor implements the method according to any one of claims 1-6.

Images