Intelligent scheduling control method and system based on multi-objective optimization

Abstract

The invention relates to the technical field of intelligent scheduling, and discloses an intelligent scheduling control method and system based on multi-objective optimization, and the method comprises the steps: obtaining an attribute set of different tasks, obtaining a disturbed task attribute initial value, and carrying out the calculation to obtain an importance score of the tasks during initialization; based on the task attribute initial value and the attribute weight, calculating the conflict degree between the tasks through the task scheduling conflict degree; defining an optimization target according to the importance score of the task and the conflict degree between the tasks, and automatically generating an association rule; and after optimization, calculating the task importance score of the final time step based on the task attribute value and the attribute weight of the final time step, and generating a final scheduling scheme according to the final importance score of each task. According to the method, through dynamic conflict evaluation, adaptive weight adjustment and a closed-loop feedback mechanism, the environmental adaptability, global optimization and stability of a scheduling system are remarkably improved, and resource competition and time delay are effectively reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent scheduling, in particular to an intelligent scheduling control method and system based on multi-objective optimization. BACKGROUND

[0002] As a core work of production management, scheduling control involves multiple aspects such as tasks, resources, time and space. Traditional scheduling methods mostly rely on manual or simple heuristic algorithms, which are not competent in dealing with complex and variable scheduling problems. With the development of information technology, especially under the support of big data, cloud computing and Internet of Things, scheduling problems have become more complex and dynamic. Traditional scheduling methods have been difficult to meet the real needs. Therefore, an intelligent scheduling control method based on multi-objective optimization has emerged, which not only can find the optimal solution under various constraints, but also can respond to the changes of system state in real time, ensuring the stability and reliability of the system.

[0003] However, the above-mentioned existing intelligent scheduling control method based on multi-objective optimization still has the problem of ignoring environmental uncertainty and disturbance in task scheduling. It only considers the priority of tasks, but ignores the mutual influence between tasks. It relies on pre-set rules and priority sorting, lacks the ability to dynamically generate and adjust optimization rules, and is difficult to respond to fluctuations in task attributes in real time. SUMMARY

[0004] In view of the above-mentioned existing problems, the present application is proposed.

[0005] Therefore, the present application provides an intelligent scheduling control method based on multi-objective optimization, which can solve the problem that traditional task scheduling algorithms cannot flexibly adapt to the influence of environmental factors in the change of task attributes (such as execution time, urgency, etc.). The external environment will cause uncertainty in task attributes, which will challenge the stability and effectiveness of task scheduling.

[0006] To solve the above technical problems, the present application provides the following technical scheme, an intelligent scheduling control method based on multi-objective optimization, comprising: obtaining the attribute set of different tasks, obtaining the initial value of the disturbed task attribute, weighting the initial value of the disturbed task attribute and the corresponding attribute weight, and calculating the importance score of the task at initialization; based on the initial value of the task attribute and the attribute weight, the conflict degree between tasks is calculated by task scheduling conflict degree; define the optimization target according to the importance score of the task and the conflict degree between tasks, and automatically generate the associated rules; after optimization, the importance score of the final time step is calculated based on the final time step task attribute value and the attribute weight, and the final scheduling scheme is generated according to the final importance score of each task

[0007] As a preferred scheme of the intelligent scheduling control method based on multi-objective optimization, the attribute set of different tasks comprises applying attribute disturbance to attribute values of each task, controlling disturbance amplitude through a strength coefficient, and simulating environmental uncertainty based on a standard normal distribution random variable.

[0008] The importance score of the task at the initialization stage is obtained by weighted summation of the initial values of the disturbed task attributes through the initialized attribute weight, and the sum of the weights of all attributes is 1.

[0009] As a preferred scheme of the intelligent scheduling control method based on multi-objective optimization, the conflict degree between tasks comprises quantifying the influence of task attribute difference on the conflict degree through an exponential decay function, wherein the task attribute difference is inversely proportional to the conflict degree.

[0010] The spatial adjustment coefficient and the overlap index between tasks are introduced to adjust the additional influence of spatial position difference on the conflict degree.

[0011] As a preferred scheme of the intelligent scheduling control method based on multi-objective optimization, the definition of the optimization target comprises taking the weighted sum of the task importance score and the conflict degree between tasks as the objective function, and balancing the weights between the two through an adjustment parameter.

[0012] The associated rules are generated in real time based on the changes of the objective function, and the task attribute values are dynamically adjusted.

[0013] As a preferred scheme of the intelligent scheduling control method based on multi-objective optimization, the dynamic adjustment of the task attribute values comprises controlling the adjustment amplitude of the attribute weight through an adaptive factor based on the feedback of the attribute weight and the optimization target at the previous time step.

[0014] The update formula of the attribute weight is:

[0015]

[0016] wherein, represents the attribute weight of the jth attribute at time t; represents the attribute weight of the jth attribute at time t-1; η represents the adaptive factor, ΔU represents the change of the optimization target function value, and ΔS i,j represents the change of the jth attribute value of the ith task;

[0017] The adaptive step is calculated according to the partial derivative of the task attribute to the optimization target, and the update amplitude of the task attribute value is adjusted.

[0018] As a preferred scheme of the intelligent scheduling control method based on multi-objective optimization, the calculation of the adaptive step length comprises: summing the partial derivatives of all attributes through a normalization term, limiting the step length update amplitude of a single attribute, and enabling the update of the task attribute value to be within a reasonable range.

[0019] The calculation formula of the update step length is:

[0020]

[0021] wherein, denotes the update step length of the jth attribute value of the ith task at time t, γ denotes a parameter for adjusting the step length update amplitude, U denotes an optimization objective function, S i,j denotes the jth attribute value of the ith task, and m denotes the total number of task attributes.

[0022] The task attribute value is updated by adjusting the step length and the attribute weight, in combination with the task attribute value at the previous time step, and the update formula of the task attribute value is:

[0023]

[0024] wherein, denotes the jth attribute value of the ith task at time t+1, denotes the jth attribute value of the ith task at time t, denotes the update step length of the jth attribute value of the ith task at time t.

[0025] As a preferred scheme of the intelligent scheduling control method based on multi-objective optimization, the task importance score at the final time step comprises: obtaining the task attribute value of each task at the final time step T after the optimization process is completed The importance score of the task at the final time step T is calculated according to the task attribute value at the final time step T and the attribute weight.

[0026] The specific calculation formula is:

[0027]

[0028] wherein, denotes the importance score of the ith task at time T, denotes the attribute weight of the jth attribute at time T, denotes the jth attribute value of the ith task at time T, and T denotes the time at which the optimization objective function reaches convergence.

[0029] The tasks are sorted according to the final importance scores of the tasks, and a final scheduling scheme is generated.

[0030] Another object of the present application is to provide an intelligent scheduling control system based on multi-objective optimization.

[0031] As a preferred scheme of the intelligent scheduling control system based on multi-objective optimization, the system comprises an attribute processing module, a conflict evaluation module, a rule generation module, and a scheduling execution module.

[0032] The attribute processing module acquires original attributes of tasks, applies random disturbance to the attributes, simulates environmental uncertainty, and calculates initial priority scores of the tasks through weighted calculation.

[0033] The conflict evaluation module quantifies attribute differences between tasks based on task attributes and weights, dynamically calculates the degree of task conflict in combination with spatial positions and overlap indexes.

[0034] The rule generation module constructs a multi-objective optimization function by comprehensively considering task priority and the degree of conflict, dynamically adjusts attribute weights and step lengths, and generates updated rule optimization attribute values.

[0035] The scheduling execution module recalculates task priority based on optimized attributes and weights, generates a final scheduling sequence, and outputs the sequence to an execution unit.

[0036] The present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of an intelligent scheduling control method based on multi-objective optimization when executing the computer program.

[0037] The present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of an intelligent scheduling control method based on multi-objective optimization when executed by a processor.

[0038] The present application has the following advantages: by introducing attribute disturbance, the present application simulates the influence of external environment on task attributes, effectively considers random changes in the environment, provides a more realistic and robust initial priority ranking for task scheduling, and enhances the adaptability and reliability of task scheduling.

[0039] By calculating the conflict degree between tasks at different time steps, considering the influence of task attribute differences and spatial position differences, the present application can accurately evaluate the potential conflict between tasks, optimize the task scheduling process, reduce resource competition, avoid performance bottlenecks, and ensure efficient execution of tasks.

[0040] By comprehensively considering the importance score of tasks and the conflict degree between tasks, the present application balances the importance and conflict degree of tasks, and optimizes the priority and resource allocation between tasks.

[0041] By adaptively adjusting the step length and weight, the instability in the scheduling process can be avoided, and the tasks can continuously obtain reasonable scheduling and resource allocation in a changing environment. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of these drawings.

[0043] Figure 1 A flowchart of an intelligent scheduling control method based on multi-objective optimization is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0044] In order to make the above-mentioned objects, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should be within the scope of protection of the present application.

[0045] Embodiment 1, refer to Figure 1 For the first embodiment of the present application, the embodiment provides an intelligent scheduling control method based on multi-objective optimization, comprising:

[0046] S1: Obtain the attribute set of different tasks, simulate random changes by introducing attribute disturbance, obtain the initial value of the disturbed task attribute, and calculate the importance score of the task at initialization according to the weighted initial value of the disturbed task attribute and the corresponding attribute weight.

[0047] Further, in the task scheduling process, different tasks have different attributes, the attribute set of different tasks is obtained, and the attribute set of the i-th task is represented as follows:

[0048] x i ={x i,1 ,x i,2 ,…,x i,m},j∈{1,2,…,m}

[0049] Wherein, x i,j represents the j-th attribute original value of the i-th task, such as execution time, urgency; m represents the number of attributes;

[0050] Since the attributes of tasks are susceptible to external environment, such as system load, hardware failure, the attribute perturbation is introduced to simulate random changes, adjust the attribute values of tasks, reflect the influence of randomness and uncertainty in the environment, and adjust the amplitude of attribute perturbation through the strength coefficient of perturbation, so that the attribute perturbation has an impact on the task attributes within a reasonable range. Finally, the initial value of the perturbed task attribute is obtained, and the calculation formula is:

[0051]

[0052] wherein, represents the initial value of the jth attribute of the ith task after perturbation; ∈ represents the strength coefficient of perturbation, ∈>0, which is used to control the size of the perturbation and can be set according to the specific implementation scene, which is not limited here; represents a standard normal distribution random variable, that is, a random number sampled from a standard normal distribution. The mean of the standard normal distribution is 0, and the standard deviation is 1, reflecting the random changes of the task attributes.

[0053] In an optional embodiment, adjusting the task attribute value to generate the initial importance score can be achieved by directional perturbation based on scene classification. Specifically, according to the historical data of the task execution environment (such as system load, failure frequency), the environment is divided into steady state mode (low perturbation) and dynamic mode (high perturbation). According to the real-time environment state, the scene mode is matched, the corresponding perturbation amplitude is selected, and the directional perturbation (such as adjusting only the execution time or resource demand) is applied to the task attribute value. In dynamic mode, the weight compensation (such as 10% weight increase) is added to the urgency attribute to ensure that the emergency task is scheduled first.

[0054] In another optional embodiment, adjusting the task attribute value to generate the initial importance score can also be achieved by adaptive perturbation based on reinforcement learning. Specifically, a reinforcement learning model is constructed to take the success rate of scheduling (such as the rate of completing tasks on time) as the reward signal, and the model is trained through historical scheduling data to learn the optimal perturbation amplitude in different environments; Real-time acquisition of environmental parameters (such as CPU load, network delay), input model to predict the current optimal perturbation amplitude, and perturbation (such as execution time ± predicted value) is applied to the task attribute value according to the prediction result. Record the actual effect of each scheduling, dynamically update the model parameters, and continuously optimize the perturbation strategy.

[0055] In order to prioritize the tasks according to their importance in the scheduling process, the relative importance of each task needs to be evaluated, specifically, each task attribute has a corresponding initial attribute weight, which reflects the relative importance of each attribute in evaluating the importance of the task, for example, a task with longer execution time may require a higher weight to prioritize, while a task with higher urgency may be assigned a higher weight to ensure timely completion, and the sum of the weights of all task attributes must be 1 to ensure the rationality of the weighted sum, the importance score of the task at initialization is calculated by weighting the initial value of the task attribute after disturbance and the corresponding attribute weight, which reflects the relative priority of the task in the scheduling process, the calculation formula is:

[0056]

[0057] wherein, represents the importance score of the i-th task at initialization, which reflects the relative priority of the task in the scheduling process; represents the initial attribute weight of the j-th attribute, and satisfies ensuring that the sum of the weights of all attributes of each task is 1, so that the total weight distribution of the task on all attributes is reasonable, which can be set according to the specific implementation scenario, and is not limited here;

[0058] The initial importance score can provide a preliminary and reasonable priority ranking basis for task scheduling, laying a foundation for subsequent task scheduling decisions.

[0059] S2: Based on the task attribute value and the attribute weight, the task scheduling conflict degree calculation algorithm is used to calculate the conflict degree between tasks.

[0060] Further, based on the task attribute value and the attribute weight, the task scheduling conflict degree calculation algorithm is used to calculate the conflict degree between tasks to adapt to the changing task attributes and spatial distance between tasks, and to reduce the performance bottleneck caused by resource competition;

[0061] Since each task has several attributes, such as execution time, resource demand, and urgency, the differences in different attributes between tasks determine whether the tasks will conflict, the task scheduling conflict degree calculation algorithm calculates the conflict degree of each pair of tasks at different time steps, compares the differences in task attributes, and can effectively evaluate the potential conflict between tasks; an exponential decay function is introduced to make the influence of task attribute difference on conflict degree present nonlinear change, and a decay factor is introduced to control the sensitivity of task attribute difference to conflict degree, specifically, the greater the difference between task attributes, the smaller the conflict degree, which can truly reflect the complexity in the scheduling process;

[0062] In addition to the task attribute difference, the calculation of the conflict degree can be further refined by introducing an overlap index between tasks, and a spatial adjustment coefficient is introduced to adjust the influence of overlap on the conflict degree.

[0063] The calculation formula of the conflict degree between tasks is:

[0064]

[0065] wherein, represents the conflict degree between the ith task and the kth task at time t; represents the influence of the task attribute difference on the conflict degree quantified by an exponential decay function, and the conflict degree decreases as the attribute difference between tasks increases; a represents a decay factor for controlling the sensitivity of the task attribute difference to the conflict degree, which can be set according to specific implementation scenarios and is not limited herein; represents the sum of the differences between the ith task and the kth task in all m attribute dimensions; represents the attribute weight of the jth attribute at time t; represents the absolute value of the attribute difference between the ith task and the kth task at time t, reflecting the degree of difference in attribute values; β represents a spatial adjustment coefficient for adjusting the influence of spatial difference on the conflict degree; Overlap(i, k) represents the overlap index of the ith task and the kth task, which can be set according to specific implementation scenarios and is not described herein.

[0066] In an optional embodiment, the conflict degree between tasks can be obtained based on a machine learning conflict prediction model. Specifically, historical scheduling data is collected, including task attributes (execution time, resource demand), inter-task conflict records (such as resource preemption, time overlap), and scheduling results (success / failure), and conflict degree labels (such as low conflict: 0-0.3, medium conflict: 0.3-0.7, high conflict: 0.7-1) are labeled.

[0067] A random forest or neural network model is used, the input task attribute pair (such as the attribute difference between task A and task B), the output predicted conflict probability, the optimization of model parameters, and the consistency of the prediction result with the actual conflict record are ensured; the current task attribute pair is input in real time, the conflict probability is predicted by the model, and the original exponential decay function calculation is replaced. The predicted result is weighted and corrected in combination with the spatial adjustment coefficient (such as the physical distance of the device).

[0068] In another optional embodiment, the conflict degree between tasks is obtained through conflict network analysis of graph theory, specifically, tasks are abstracted as nodes, resource competition or time overlap between tasks is abstracted as edges, and edge weight represents conflict strength (such as resource overlap rate, time overlap duration); a graph clustering algorithm (such as community detection) is used to identify high-conflict task groups, and a node centrality index (such as degree centrality) is used to quantify the influence of a single task on global conflict; the graph structure is updated (such as adding new task nodes, adjusting edge weights) according to the real-time task state, and the centrality index and clustering results are combined to dynamically adjust the task scheduling order.

[0069] S3: defining an optimization target according to the importance score of the task and the conflict degree between tasks, automatically generating a correlation rule according to the current task attribute and the optimization target for guiding the update of the task attribute, and calculating the importance score of the task at the final time step according to the final time step task attribute value and the attribute weight after the optimization process is completed, and generating a final scheduling scheme according to the final importance score of each task.

[0070] Further, in order to solve the balance problem between the importance of the task and the conflict between the tasks, an optimization target is introduced, which consists of two parts: the importance score of the task and the conflict degree between the tasks; the importance score of the task is a quantification of the current priority of each task, and the conflict degree reflects the adverse effects that may be caused by resource or time conflict between the tasks; by setting an adjustment parameter, the relative importance between the importance score of the task and the conflict between the tasks can be adjusted at each time step, so as to achieve the purpose of optimizing scheduling;

[0071] The definition of the optimization target function is as follows:

[0072]

[0073] Wherein, U represents the optimization target function; n represents the number of tasks; represents the importance score of the i-th task at time t; λ represents an adjustment parameter for balancing the importance score of the task and the conflict between the tasks; represents the conflict degree between the i-th task and the k-th task at time t; represents the conflict degree between all tasks through double summation;

[0074] At each time step, a correlation rule is automatically generated according to the current task attribute and the optimization target for guiding the update of the task attribute; the construction process of the correlation rule is automatic, based on the real-time feedback of the interaction between the task attribute and the optimization target, when the task attribute changes, a rule is automatically formed to indicate whether the task attribute should be increased or decreased, so as to meet the overall scheduling optimization target;

[0075] The attribute weight is dynamically adjusted based on the attribute weight of the previous time step and the feedback of the optimization objective, and the sensitivity of the attribute weight adjustment is controlled through an adaptive factor to ensure that the task can be properly valued in a changing environment;

[0076] The updating formula of the attribute weight is:

[0077]

[0078] wherein, represents the attribute weight of the jth attribute at time t; represents the attribute weight of the jth attribute at time t-1; η represents an adaptive factor for controlling the sensitivity of the attribute weight adjustment, which can be set according to the specific implementation scenario and is not limited herein; ΔU represents the change of the optimization objective function value; ΔS i,j represents the change of the jth attribute value of the ith task.

[0079] In order to ensure the stability in the updating process, the updating amplitude of the step is adjusted, and the calculation of the step is not only based on the current value of the task attribute, but also considers the influence of the task on the optimization objective. If the task attribute value plays a greater role in the change of the optimization objective, the step will be increased accordingly, and vice versa. Through the adaptive adjustment of the step, the instability in the scheduling process can be avoided, and the updating of each task can be ensured within a reasonable range;

[0080] The calculation formula of the updating step is:

[0081]

[0082] wherein, represents the updating step of the jth attribute value of the ith task at time t, which determines the updating amplitude of the task attribute at each time step; γ represents a parameter for adjusting the updating amplitude of the step, which can be set according to the specific implementation scenario and is not limited herein; represents the influence degree of the jth attribute value of the ith task on the optimization objective, which is calculated through the partial derivative of the objective function with respect to the task attribute; represents the overall influence of the changes of all attributes on the optimization objective function, which is used as a normalization term to adjust the step of each attribute, so as to avoid the step of part of the attributes being too large or too small;

[0083] Through the adjustment of the step and the attribute weight, the task attribute value is updated in combination with the task attribute value of the previous time step, and the updating formula of the task attribute value is:

[0084]

[0085] wherein, denotes the value of the i-th task's j-th attribute at time t+1; denotes the value of the i-th task's j-th attribute at time t; denotes the update step of the i-th task's j-th attribute at time t, which determines the update amplitude of the task attribute at each time step; denotes the attribute weight of the j-th attribute at time t;

[0086] After the optimization process is completed, the task attribute value of each task at the final time step T is obtained According to the final time step T task attribute value and attribute weight, the importance score of the task at the final time step T is calculated, and the formula is as follows:

[0087]

[0088] wherein, denotes the importance score of the i-th task at time T; denotes the attribute weight of the j-th attribute at time T; denotes the value of the i-th task's j-th attribute at time T; T denotes the time at which the optimization objective function reaches convergence, which can be set according to specific implementation scenarios, and is not limited herein;

[0089] According to the final importance score of each task, the tasks are sorted, and the final scheduling scheme is generated according to the importance score of the task.

[0090] Embodiment 2, which is an embodiment of the present application, provides an intelligent scheduling control method based on multi-objective optimization. In order to verify the beneficial effects of the present application, scientific demonstration is carried out through experiments.

[0091] On the assembly line of the electric energy meter production factory, three batches of electric energy meters are calibrated and detected, and the assembly line has limited equipment resources (such as calibration instruments and detection tables). The task order is reasonably arranged to avoid resource competition and delay. The three tasks are:

[0092] Task 1: Calibrate 100 ordinary household electric energy meters (small batch, ordinary priority);

[0093] Task 2: Detect 50 high-precision electric energy meters (urgent order, need to be completed as soon as possible);

[0094] Task 3: Calibrate 200 industrial electric energy meters (large batch, but not urgent).

[0095] Each task has three key attributes:

[0096] Execution time (unit: hour): time required for task completion;

[0097] Urgency (range 0-10): Reflects the priority of the task, 10 indicates the most urgent;

[0098] Resource demand (range 0-10): The degree of demand for pipeline equipment by the task, 10 indicates the highest demand;

[0099] The scheduling target is to generate a reasonable execution order through intelligent algorithms, considering the importance of tasks (priority) and conflicts between tasks (resource competition), to improve pipeline efficiency and meet customer needs.

[0100] In actual production, task attributes are determined based on historical data and order demand, assuming the initial attributes are as follows:

[0101] Task 1 (ordinary electric energy meter calibration): x1 = {2, 4, 3}: execution time 2 hours (100 units, each 1.2 minutes, total about 120 minutes); Urgency 4 (ordinary order, delivery time is relatively loose); Resource demand 3 (only one calibration instrument is needed, demand is low).

[0102] Task 2 (high-precision electric energy meter detection): x2 = {1, 8, 6}: execution time 1 hour (50 units, each 1.2 minutes, total about 60 minutes); Urgency 8 (customer needs, needs to be delivered first); Resource demand 6 (high-precision detection table is needed, resource occupation is more).

[0103] Task 3 (industrial electric energy meter calibration): x3 = {3, 2, 4}: execution time 3 hours (200 units, each 0.9 minutes, total about 180 minutes); Urgency 2 (industrial order, long delivery cycle); Resource demand 4 (two calibration instruments are needed, medium demand).

[0104] During pipeline operation, slight fluctuations in task attributes may occur due to factors such as equipment load, worker fatigue, or temperature changes. To simulate uncertainty, attribute perturbations are introduced, with the formula:

[0105]

[0106] ∈ = 0.1: Perturbation intensity coefficient, controls the fluctuation amplitude (set to within 10%, consistent with reality).

[0107] Standard normal distribution random number, mean 0, standard deviation 1, simulates randomness.

[0108] Assume random number sampling results (generated by statistical software, consistent with normal distribution):

[0109] Task 1: {0.5, -0.3, 0.8}

[0110] Task 2: {-0.2, 0.7, -0.4}

[0111] Task 3: {0.9, -0.6, 0.2}

[0112] Calculate attribute values after perturbation:

[0113] Task 1:

[0114] Execution time:

[0115] Urgency:

[0116] Resource requirement:

[0117] Result:

[0118] By analogy:

[0119]

[0120] The values after perturbation have slight changes, such as the execution time of Task 1 increasing from 2 hours to 2.05 hours, possibly due to slower device preheating, and the urgency of Task 2 rising from 8 to 8.07, possibly due to customer phone urging. These fluctuations can reflect real situations.

[0121] Initialize importance score: determine task priority

[0122] Importance score formula: Let the initial weight w (0) = {0.3, 0.5, 0.2},

[0123] Execution time weight 0.3: tasks with long execution times need to be reasonably arranged; urgency weight 0.5: higher priority for urgent tasks; resource requirement weight 0.2: resource occupation has less impact, and the sum 0.3 + 0.5 + 0.2 = 1 ensures reasonable scoring;

[0124] Calculate:

[0125]

[0126] Task 2 has the highest score (5.521) due to its high urgency (8.07), which meets the actual demand of "urgent orders first". Tasks 1 (3.216) and 3 (2.701) have lower scores, reflecting their ordinary priority, which provides a basis for preliminary scheduling.

[0127] Tasks may conflict due to resource competition (such as limited number of calibration instruments) or spatial location (device distribution), such as Task 2 and Task 3 needing calibration instruments at the same time, which may cause waiting;

[0128] Parameter settings:

[0129] α=0.5: attenuation factor, controlling the impact of attribute differences on conflict;

[0130] β = 0.1: spatial adjustment coefficient, reflecting the impact of operation overlap;

[0131] When t=0, the weight is still w (0) ={0.3,0.5,0.2};

[0132] Assume that the operating area overlap index is:

[0133] Task 1 and Task 2: Overlap(1,2)=5;

[0134] Task 1 and Task 3: Overlap(1,3)=8;

[0135] Task 2 and Task 3: Overlap(2,3)=3.

[0136] Calculate the degree of conflict:

[0137] Attribute difference: |2.05-0.98|=1.07 (time difference)

[0138] ∣3.97-8.07∣=4.1 (poor urgency)

[0139] ∣3.08-5.96∣=2.88 (poor resource demand)

[0140] Weighted sum:

[0141] 0.3 1.07 + 0.5 4.1 + 0.2 2.88 = 0.321 + 2.05 + 0.576 = 2.947 Index term:

[0142] exp(-0.5·2.947)=exp(-1.4735)≈0.229

[0143] Space item:

[0144] 1+0.1·5=1.5

[0145] Degree of conflict:

[0146]

[0147] Similarly, we can get:

[0148] The degree of conflict between Task 1 and Task 3: The degree of conflict between Task 2 and Task 3: Optimizing the objective function: Balancing importance and conflict

[0149] n = 3 (number of tasks)

[0150] λ = 0.5 (tuning parameter, balancing importance and impact of conflicts)

[0151] Time t = 0 (initial state)

[0152] Importance score sum: 3.216 + 5.521 + 2.701 = 11.438

[0153] Conflict degree sum (all task pairs, computed bidirectionally):

[0154] 0.3435 + 0.3435 + 0.8442 + 0.8442 + 0.169 + 0.169 = 2.7134

[0155] η = 0.1

[0156] Assume ΔU = 0.5, ΔS 1,1 = 0.1, ΔS 2,2 = 0.2, ΔS 3,3 = 0.05

[0157] Compute:

[0158]

[0159] Normalize (sum 1.475): w (1) = {0.305, 0.424, 0.271}

[0160] Update step and attribute values

[0161] Step update: γ = 0.2,

[0162]

[0163] Attribute update:

[0164]

[0165] Final importance scores

[0166] T = 1, w (1) = {0.305, 0.424, 0.271}

[0167] Task one:

[0168] Task two:

[0169] Task three:

[0170] Scheduling scheme:

[0171] Ranking: Task 2 (5.3357) -> Task 1 (3.1432) -> Task 3 (2.8544)

[0172] Task 2 is prioritized due to high urgency and low conflict; Task 1 and Task 3 are executed in turn, reducing the risk of mechanical hand operation overlap.

[0173] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalent replaced without departing from the spirit and scope of the present application, and all of them should be covered in the scope of the claims of the present application.

[0174] Embodiment 3 is the third embodiment of the present application, which is different from the first two embodiments:

[0175] If the function is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the part of the present application which essentially contributes to the prior art or the part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0176] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered a list of executable instructions for implementing logical functions, and can be specifically embodied in any computer-readable medium for use by an instruction execution system, apparatus or device, such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, apparatus or device and execute the instructions, or in conjunction with these instructions execution system, apparatus or device. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transport programs for use by an instruction execution system, apparatus or device, or in conjunction with these instruction execution system, apparatus or device.

[0177] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Additionally, the computer readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example, via optical scanning of the paper or other suitable medium, then compiled, interpreted, or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.

[0178] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the embodiments described above, various steps or methods can be implemented, for example, by software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following technologies, known in the art, or a combination thereof, can be used: discrete logic circuitry having logic gates for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), etc.

[0179] Embodiment 4, as an embodiment of the present application, provides an intelligent scheduling control system based on multi-objective optimization, comprising an attribute processing module, a conflict evaluation module, a rule generation module, and a scheduling execution module.

[0180] The attribute processing module acquires the original attributes of the tasks, applies random disturbance to the attributes, simulates environmental uncertainty, and calculates the initial priority score of the tasks by weighting.

[0181] The conflict evaluation module quantifies the attribute difference between tasks based on the task attributes and weights, dynamically calculates the degree of task conflict in combination with the spatial position and overlap index.

[0182] The rule generation module constructs a multi-objective optimization function by synthesizing the task priority and the degree of conflict, dynamically adjusts the attribute weight and step size, and generates updated rules to optimize the attribute value.

[0183] The scheduling execution module recalculates the task priority according to the optimized attributes and weights, generates the final scheduling sequence, and outputs it to the execution unit.

[0184] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced, without departing from the spirit and scope of the technical solutions of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. An intelligent scheduling control method based on multi-objective optimization, characterized by: include, Obtain the attribute sets of different tasks, obtain the initial values ​​of the task attributes after perturbation, and weight them according to the initial values ​​of the task attributes after perturbation and the corresponding attribute weights to calculate the importance score of the task at the time of initialization; Based on the initial value and attribute weight of the task attributes, the conflict degree between tasks is calculated by task scheduling conflict degree; Define optimization goals based on task importance scores and the degree of conflict between tasks, and automatically generate association rules; After optimization, the task importance score of the final time step is calculated based on the task attribute value and attribute weight of the final time step, and the final scheduling plan is generated according to the final importance score of each task.

2. The intelligent scheduling control method based on multi-objective optimization according to claim 1, characterized in that: Acquiring attribute sets of different tasks includes applying attribute perturbations to attribute values ​​of each task, controlling the perturbation amplitude through an intensity coefficient, and simulating environmental uncertainty based on standard normal distribution random variables; The initial values ​​of the perturbed task attributes are weighted and summed by initializing the attribute weights to obtain the importance score of the task at initialization, where the sum of the weights of all attributes is 1.

3. The intelligent scheduling control method based on multi-objective optimization according to claim 2, characterized in that: The degree of conflict between tasks includes quantifying the effect of task attribute differences on the degree of conflict through an exponential decay function, wherein the task attribute differences are inversely proportional to the degree of conflict; The spatial adjustment coefficient and the overlap index between tasks were introduced to regulate the additional impact of spatial position differences on the conflict level.

4. The intelligent scheduling control method based on multi-objective optimization according to claim 3, characterized in that: Defining the optimization objective includes taking the weighted sum of the task importance score and the conflict degree between tasks as the objective function, and balancing the weights between the two by adjusting parameters; Based on the changes in the objective function, association rules are generated in real time to dynamically adjust task attribute values.

5. The intelligent scheduling control method based on multi-objective optimization according to claim 4, characterized in that: The dynamic adjustment of the task attribute value includes controlling the adjustment range of the attribute weight through an adaptive factor based on the attribute weight of the previous time step and the feedback of the optimization target; The updating formula of attribute weight is: in, represents the attribute weight of the j-th attribute at time t; represents the attribute weight of the jth attribute at time t-1; η represents the adaptive factor, ΔU represents the change in the optimization objective function value, and ΔS i,j Indicates the change in the jth attribute value of the i-th task; The adaptive step size is calculated based on the partial derivatives of the task attributes with respect to the optimization objective, and the update amplitude of the task attribute value is adjusted.

6. The intelligent scheduling control method based on multi-objective optimization according to claim 5, characterized in that: The adaptive step size calculation includes summing the partial derivatives of all attributes through a normalization term, limiting the step size update amplitude of a single attribute, and making the update of the task attribute value within a reasonable range; The calculation formula for the update step size is: in, represents the update step size of the jth attribute value of the i-th task at time t, γ represents the parameter used to adjust the step size update amplitude, U represents the optimization objective function, S i,j represents the jth attribute value of the i-th task, and m represents the total number of task attributes; By adjusting the step size and attribute weight, the task attribute value is updated in combination with the task attribute value of the previous time step. The update formula of the task attribute value is: in, represents the jth attribute value of the i-th task at time t+1, represents the jth attribute value of the i-th task at time t, represents the update step of the jth attribute value of the i-th task at time t.

7. The intelligent scheduling control method based on multi-objective optimization according to claim 6, characterized in that: The task importance score of the final time step includes: after the optimization process is completed, the task attribute value of each task at the final time step T is obtained. Calculate the importance score of the task at the final time step T based on the task attribute value and attribute weight at the final time step T; The specific calculation formula is: in, represents the importance score of the i-th task at time T, represents the attribute weight of the j-th attribute at time T, represents the jth attribute value of the i-th task at time T, where T represents the time when the optimization objective function reaches convergence; Sort tasks according to their final importance score and generate a final scheduling plan.

8. A system using the intelligent scheduling control method based on multi-objective optimization according to any one of claims 1 to 7, characterized in that: It includes attribute processing module, conflict assessment module, rule generation module and scheduling execution module; The attribute processing module obtains the original attributes of the task, applies random perturbations to the attributes, simulates environmental uncertainty, and calculates the initial priority score of the task through weighted calculation; The conflict assessment module quantifies the attribute differences between tasks based on task attributes and weights, and dynamically calculates the degree of task conflict by combining spatial position and overlap index; The rule generation module integrates task priorities and conflict levels to construct a multi-objective optimization function; dynamically adjusts attribute weights and step sizes to generate updated rule-optimized attribute values; The scheduling execution module recalculates the task priorities according to the optimized attributes and weights, generates a final scheduling order, and outputs it to the execution unit.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the processor implements the steps of an intelligent scheduling control method based on multi-objective optimization according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of an intelligent scheduling control method based on multi-objective optimization according to any one of claims 1 to 7 are implemented.

Images