Task scheduling method and device, electronic equipment, storage medium and computer product

Abstract

The invention relates to the technical field of computers, and provides a task scheduling method and device, electronic equipment, a storage medium and a computer product, and the method comprises the steps: determining a state transition probability matrix, an observation probability matrix and state distribution information according to the historical emotion data of a to-be-scheduled employee; inputting the state transition probability matrix, the observation probability matrix and the state distribution information of the to-be-scheduled employees into an emotional state prediction model to obtain a forward probability, a backward probability and a posterior probability output by the emotional state prediction model; wherein the emotional state prediction model is obtained by training a dynamic Bayesian network model according to a state transition probability matrix, an observation probability matrix and state distribution information of the sample employees; and based on the forward probability, the backward probability and the posterior probability, generating a task scheduling scheme in combination with to-be-scheduled task information and a tiger-whale population algorithm. The task execution efficiency in the mobile office environment can be improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular, to a task scheduling method, apparatus, electronic device, storage medium, and computer product. Background Art

[0002] With the rapid development of information technology and the acceleration of the globalization process, mobile office has become the norm in modern enterprise operations. Through mobile devices such as smartphones, tablets, and laptops, employees can complete work tasks at any time and place. This flexible working method not only improves work efficiency and employee satisfaction, but also brings new challenges, especially in task scheduling and work quality management.

[0003] In a mobile office environment, employees face situations such as multitasking, high dispersion, and frequent interruptions, which easily lead to a decline in work quality. In addition, due to the lack of face-to-face communication, it is difficult to monitor and manage employees' emotional states and work pressures in a timely manner. These problems not only affect the overall work efficiency, but also may lead to task delays and an increase in errors. Therefore, how to balance efficiency and quality in a mobile office environment has become an urgent problem to be solved.

[0004] However, when performing task scheduling in a mobile office environment currently, it is usually fixed to schedule based on the priority of tasks and the skill levels of employees. As a result, the efficiency of task execution in a mobile office environment is low. Summary of the Invention

[0005] This application aims to at least solve one of the technical problems existing in the related art. For this purpose, this application provides a task scheduling method, apparatus, electronic device, storage medium, and computer product, to solve the problem of scheduling based on the priority of tasks and the skill levels of employees when performing task scheduling in a mobile office environment currently, and to achieve improving the efficiency of task execution in a mobile office environment.

[0006] According to the task scheduling method of the first aspect embodiment of this application, it includes: Respectively determine a state transition probability matrix, an observation probability matrix, and state distribution information according to the historical emotion data of the employee to be scheduled; Input the state transition probability matrix, the observation probability matrix, and the state distribution information of the employee to be scheduled into an emotion state prediction model, and obtain the forward probability, backward probability, and posterior probability output by the emotion state prediction model; wherein, the emotion state prediction model is obtained by training a dynamic Bayesian network model according to the state transition probability matrix, the observation probability matrix, and the state distribution information of sample employees; Based on the forward probability, the backward probability, and the posterior probability, combine the task information to be scheduled with the killer whale population algorithm to generate a task scheduling plan.

[0007] According to an embodiment of the present application, generating a task scheduling scheme by combining the forward probability, the backward probability, and the posterior probability with the to-be-scheduled task information and the killer whale population algorithm includes: Generating at least one pre-scheduling scheme based on the to-be-scheduled task information; According to the forward probability, the backward probability, and the posterior probability, adjusting the search behavior of the killer whale population algorithm initialized based on each pre-scheduling scheme; wherein, the search behavior of the killer whale population algorithm includes encirclement behavior, spiral update, and random search; the encirclement behavior dynamically adjusts the parameters of the encirclement behavior by using the forward probability and the backward probability; the spiral update dynamically adjusts the step size and frequency of the spiral update by using the posterior probability; the random search adjusts the frequency of the random search by using the forward probability and the backward probability; After the search behavior adjustment is completed, determining the pre-scheduling scheme with the highest population fitness in the killer whale population algorithm as the task scheduling scheme.

[0008] According to an embodiment of the present application, when adjusting the search behavior of the killer whale population algorithm initialized based on each pre-scheduling scheme according to the forward probability, the backward probability, and the posterior probability, for each adjustment process, the following operations are performed: Updating the position according to the adjusted encirclement behavior, spiral update, and random search, and determining the number of iterations or population fitness after the position update; If the number of iterations reaches the preset iteration threshold or the population fitness converges, it is determined that the search behavior adjustment is completed; otherwise, the next search behavior adjustment is performed.

[0009] According to an embodiment of the present application, the emotion state prediction model is determined in the following manner: Respectively determining the state transition probability matrix, the observation probability matrix, and the state distribution information according to the historical emotion data of the sample employees; According to the state transition probability matrix, the observation probability matrix, and the state distribution information of the sample employees, iteratively training a dynamic Bayesian network model by using the expectation maximization algorithm, and obtaining the emotion state prediction model after the iteration is completed.

[0010] According to an embodiment of the present application, when iteratively training a dynamic Bayesian network model by using the expectation maximization algorithm according to the state transition probability matrix, the observation probability matrix, and the state distribution information of the sample employees, for each iterative training process, the following operations are respectively performed: Determining the forward probability according to the state transition probability matrix, the observation probability matrix, and the state distribution information of the sample employees; Determine the backward probability based on the state transition probability matrix, observation probability matrix, state distribution information of the sample employee, and the forward probability; Determine the posterior probability based on the forward probability and the backward probability; Based on the posterior probability, update the observation probability matrix and the state distribution information of the sample employee respectively; Based on the forward probability, the backward probability, and the posterior probability, update the state transition probability matrix of the sample employee; If the state transition probability matrix, observation probability matrix, and state distribution information of the sample employee converge after update or the current iteration number reaches the preset iteration threshold, determine that the iteration is completed; otherwise, perform the next iteration.

[0011] According to an embodiment of the present application, the updating the state transition probability matrix of the sample employee based on the forward probability, the backward probability, and the posterior probability includes: Determine the joint probability according to the forward probability and the backward probability; Update the state transition probability matrix of the sample employee according to the joint probability and the posterior probability.

[0012] According to the task scheduling device of the second aspect embodiment of the present application, it includes: A determination module, configured to determine a state transition probability matrix, an observation probability matrix, and state distribution information respectively according to the historical emotion data of the employee to be scheduled; A prediction module, configured to input the state transition probability matrix, observation probability matrix, and state distribution information of the employee to be scheduled into an emotion state prediction model, and obtain the forward probability, backward probability, and posterior probability output by the emotion state prediction model; wherein, the emotion state prediction model is trained based on the state transition probability matrix, observation probability matrix, and state distribution information of the sample employee for a dynamic Bayesian network model; A generation module, configured to generate a task scheduling plan based on the forward probability, the backward probability, and the posterior probability, in combination with the information of the task to be scheduled and the killer whale population algorithm.

[0013] According to the electronic device of the third aspect embodiment of the present application, it includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, it implements the task scheduling method as described in any one of the above.

[0014] According to the storage medium of the fourth aspect embodiment of the present application, the storage medium is a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the task scheduling method as described in any one of the above.

[0015] A computer program product according to an embodiment of the fifth aspect of the present application includes a computer program, and when the computer program is executed by a processor, it implements the task scheduling method described in any one of the above.

[0016] One or more of the above technical solutions in the embodiments of the present application have at least the following technical effects: By pre-training a dynamic Bayesian network model according to the state transition probability matrix, observation probability matrix, and state distribution information of sample employees, an emotion state prediction model can be obtained. Thus, the state transition probability matrix, observation probability matrix, and state distribution information can be determined respectively according to the historical emotion data of the employees to be scheduled. Furthermore, by inputting the state transition probability matrix, observation probability matrix, and state distribution information of the employees to be scheduled into the emotion state prediction model, the forward probability, backward probability, and posterior probability output by the emotion state prediction model can be obtained. Further, based on the forward probability, backward probability, and posterior probability, combined with the information of the tasks to be scheduled and the orca population algorithm, a task scheduling plan can be generated. By training the dynamic Bayesian network model to obtain the emotion state prediction model, it is possible to monitor and predict the emotion state of employees in real time according to the emotion state prediction model combined with the state transition probability matrix, observation probability matrix, and state distribution information determined respectively from the historical emotion data of the employees to be scheduled. Then, based on the forward probability, backward probability, and posterior probability, combined with the information of the tasks to be scheduled and the orca population algorithm, a task scheduling plan can be generated, thereby realizing personalized intelligent task scheduling according to the real-time emotion state of employees, which helps to improve the efficiency of task execution in the mobile office environment.

[0017] Additional aspects and advantages of the present application will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0019] Figure 1 is a flowchart of the task scheduling method provided by the embodiments of the present application.

[0020] Figure 2 is a structural schematic diagram of the electronic device provided by the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] The following further describes the implementation manners of the present application in conjunction with the accompanying drawings and embodiments. The following embodiments are used to illustrate the present application, but cannot be used to limit the scope of the present application.

[0022] In the description of the embodiments of the present application, it should be noted that the orientation or positional relationship indicated by the terms "center", "longitudinal", "transverse", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the embodiments of the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation on the embodiments of the present application. In addition, the terms "first", "second", and "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0023] In the description of the embodiments of the present application, it should be noted that unless otherwise clearly specified and limited, the terms "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to specific situations.

[0024] In the embodiments of the present application, unless otherwise clearly specified and limited, the first feature being "on" or "under" the second feature can be that the first and second features are in direct contact, or the first and second features are indirectly in contact through an intermediate medium. Moreover, the first feature being "above", "over" and "on" the second feature can be that the first feature is directly above or obliquely above the second feature, or simply means that the first feature has a higher horizontal height than the second feature. The first feature being "under", "beneath" and "under" the second feature can be that the first feature is directly below or obliquely below the second feature, or simply means that the first feature has a lower horizontal height than the second feature.

[0025] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the embodiments of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0026] It should be noted that currently, when performing task allocation and task scheduling in a mobile office environment, it is usually fixedly based on the priority of tasks and the skill levels of employees for allocation and scheduling, and cannot be automatically and dynamically adjusted according to the real-time emotional state of employees, resulting in poor real-time performance. When employees are in a low mood or fatigued, they are assigned complex tasks, leading to a decline in work efficiency. In addition, when performing task scheduling when employees are in a low mood or fatigued, it may lead to unstable task completion quality, increasing the error rate and rework rate, and wasting enterprise resources.

[0027] Based on this, the present application proposes a task scheduling method, device, electronic device, storage medium and computer product.

[0028] Figure 1 is a schematic flowchart of the task scheduling method provided by the embodiment of the present application, as Figure 1 shown, the task scheduling method includes: Step 110, respectively determine a state transition probability matrix, an observation probability matrix and state distribution information according to the historical emotion data of the employee to be scheduled.

[0029] Step 120, input the state transition probability matrix, the observation probability matrix and the state distribution information of the employee to be scheduled into an emotion state prediction model, and obtain the forward probability, the backward probability and the posterior probability output by the emotion state prediction model; wherein, the emotion state prediction model is trained on a dynamic Bayesian network model according to the state transition probability matrix, the observation probability matrix and the state distribution information of sample employees.

[0030] Step 130, based on the forward probability, the backward probability and the posterior probability, combine the task information to be scheduled and the orca population algorithm to generate a task scheduling plan.

[0031] The execution subject of the task scheduling method provided by the embodiment of the present application can be a computer device, and the computer device can be, for example, a mobile phone, a tablet computer, a notebook computer, a palm computer, an in-vehicle electronic device, a wearable device, a super mobile personal computer (Ultra-mobile Personal Computer, UMPC), a netbook or a personal digital assistant (Personal Digital Assistant, PDA), etc. It should be noted that all the data that needs to be obtained in the present application is obtained through formal channels after being authorized by relevant users.

[0032] A task scheduling device can be set or connected in the computer device of the present application, whereby the task scheduling device can be controlled to execute the task scheduling method of the present application.

[0033] It should be noted that the present application can be applied in at least the following technical fields: remote work; project management; intelligent manufacturing. In a specific scenario, the employees to be scheduled can be one or more employees who need to be assigned and scheduled tasks. The sample employees can be the same as or different from the employees to be scheduled.

[0034] In the present application, the historical emotion data of employees can include the emotion state data of employees and the observable data of employees in different emotion states.

[0035] As shown in Table 1 below, the emotion state data of employees in the present application is the emotion state of users at different time points, including the initial emotion state at the beginning of each working day or task. The emotion state can include five types: relaxed, focused, fatigued, anxious, and stressed.

[0036] Table 1 Emotion State Table

[0037] As shown in Table 2 below, the observable data of employees in different emotion states includes facial expressions (such as smiling, frowning, calm), voice intonations (such as high-pitched, low-pitched, peaceful), typing speed (such as the number of words typed per minute), etc.

[0038] Table 2 Observable Data Table

[0039] The above data can be collected through mobile devices used by employees, such as smartphones, laptops, etc. It is collected through cameras, microphones, typing software, etc. on the mobile devices.

[0040] After collecting the above data, the data can be cleaned and stored to facilitate the subsequent construction of an emotion state prediction model and emotion state prediction. When cleaning the data, first remove noise and outliers to ensure the accuracy of the data. Then, normalize the data from different sources to ensure the consistency and comparability of the data. After cleaning the data, store the data in a relational database or a NoSQL database. A NoSQL (Not Only SQL) database is a non-relational database that does not rely on traditional table forms to store data but uses multiple data models to meet different application requirements.

[0041] It should be noted that based on the historical emotion data of employees, the present application can determine the state transition probability matrix, observation probability matrix, and state distribution information of the corresponding employees (such as employees to be scheduled or sample employees).

[0042] Specifically, the present application can, based on the emotional state data, count the frequency of each emotional state transitioning to other emotional states, and calculate the state transition probability matrix A. Then, based on the emotional state and observable data, count the frequency of observing specific data in each emotional state, and calculate the observation probability matrix B. Finally, based on the initial state data in the historical emotional data, count the frequency of each emotional state occurring at the initial time point, and calculate the initial state distribution information h.

[0043] Thus, based on the state transition probability matrix, observation probability matrix, and state distribution information of the sample employees, the dynamic Bayesian network model can be trained, and after the training is completed, an emotional state prediction model can be obtained that can output the forward probability, backward probability, and posterior probability based on the input data. It should be noted that the forward probability, backward probability, and posterior probability of the employees in the present application can be used to reflect the likelihood distribution of the employees in different emotional states, so as to predict the emotional state of the employees.

[0044] Furthermore, after obtaining the state transition probability matrix, observation probability matrix, and state distribution information of the employee to be scheduled, input the state transition probability matrix, observation probability matrix, and state distribution information of the employee to be scheduled into the emotional state prediction model, and the forward probability, backward probability, and posterior probability output by the emotional state prediction model can be obtained.

[0045] Among them, the forward probability can predict the probability of observing employee data at time t (representing the observation data sequence from time 1 to time t, that is, the set of all observation data from the initial time point to the current time point t) and being in state , and can consider the cumulative impact of all observation data from the initial state to the current time point; while the backward probability can predict the probability of observing the employee in state at time t and observing employee data from time t + 1 to the final time T . The backward probability can consider the impact of all observation data from the current time point to the future. Therefore, the combination of the forward probability and the backward probability can comprehensively consider the emotional state information over the entire time series, rather than just the emotional state at a certain moment. This comprehensive consideration enables more informed decisions to be made when adjusting the search parameters of the orca population algorithm, thereby improving the accuracy of task scheduling.

[0046] And the posterior probability represents the probability of observing all data at time t and the employee being in state The probability, the posterior probability combines the information of the forward probability and the backward probability, providing a comprehensive probability estimate for determining the likelihood of an employee being in a certain state at a specific time point. Therefore, the posterior probability can provide a global perspective, considering all data impacts from the initial time to the current time and from the current time to the future time, which enables the posterior probability to more accurately predict the state of the system at the current time point.

[0047] This application can initialize the orca population algorithm according to the information of the task to be scheduled. Further, according to the forward probability, backward probability, and posterior probability of the employee to be scheduled, the parameters of the orca population algorithm combined with the information of the task to be scheduled are dynamically adjusted to achieve intelligent mobile office task scheduling by combining the mental state of the employee, and then an optimal task scheduling plan can be obtained.

[0048] Furthermore, the task can be assigned to the employee to be scheduled according to the task scheduling plan, so that the task matches the best state of the employee and the work efficiency is improved.

[0049] According to the task scheduling method of the embodiments of this application, by pre-training the dynamic Bayesian network model according to the state transition probability matrix, observation probability matrix, and state distribution information of the sample employees, an emotion state prediction model can be obtained. Thus, the state transition probability matrix, observation probability matrix, and state distribution information can be determined respectively according to the historical emotion data of the employee to be scheduled. Furthermore, by inputting the state transition probability matrix, observation probability matrix, and state distribution information of the employee to be scheduled into the emotion state prediction model, the forward probability, backward probability, and posterior probability output by the emotion state prediction model can be obtained. Further, based on the forward probability, backward probability, and posterior probability, combined with the information of the task to be scheduled and the orca population algorithm, a task scheduling plan can be generated. By training the dynamic Bayesian network model to obtain the emotion state prediction model, it is possible to monitor and predict the emotion state of employees in real time according to the emotion state prediction model combined with the state transition probability matrix, observation probability matrix, and state distribution information determined respectively from the historical emotion data of the employee to be scheduled. Furthermore, by generating a task scheduling plan based on the forward probability, backward probability, and posterior probability, combined with the information of the task to be scheduled and the orca population algorithm, personalized intelligent task scheduling can be realized according to the real-time emotion state of employees, which helps to improve the efficiency of task execution in a mobile office environment.

[0050] This application realizes intelligent scheduling by dynamically adjusting task allocation through real-time monitoring of the emotion state of employees, thereby ensuring that each employee performs tasks in the best state and improving work efficiency.

[0051] This application can adjust the task difficulty and load when the employee is in a bad mood, reduce the error rate and rework rate, and improve the task completion quality.

[0052] Thus, work efficiency and productivity can be improved.

[0053] This application can also optimize task allocation and resource utilization, improve the utilization rate of equipment, technology, and human resources, avoid resource waste, and achieve efficient resource allocation. Moreover, by reducing the error rate, improving work efficiency, and optimizing resource utilization, the operating cost can be reduced, and the profitability of the enterprise can be improved.

[0054] Based on the above embodiments, the emotional state prediction model is determined in the following manner: Respectively determine the state transition probability matrix, observation probability matrix, and state distribution information according to the historical emotional data of the sample employees; According to the state transition probability matrix, observation probability matrix, and state distribution information of the sample employees, use the expectation maximization algorithm to iteratively train the dynamic Bayesian network model. After the iteration is completed, the emotional state prediction model is obtained.

[0055] It should be noted that this application can construct a dynamic emotional Bayesian network model (which can be abbreviated as a network model later) based on the dynamic Bayesian network (DBN) model combined with historical emotional data, and then train the dynamic emotional Bayesian network model.

[0056] Specifically, state variables and observation variables can be defined first, and then, a state transition probability matrix and an observation probability matrix are constructed, and the initial state distribution is determined.

[0057] Among them, when defining state variables and observation variables, it can be defined that: state variables : The emotional state of employees, such as relaxed, focused, fatigued, anxious, stressed; observation variables : The observable data obtained, such as facial expressions, speech intonation, typing speed.

[0058] Furthermore, when this application statistically calculates the frequency of each emotional state transitioning to other emotional states according to the emotional state data in the historical emotional data and calculates the state transition probability matrix A, it is assumed that there is historical emotional state transition data as shown in Table 3 below (only partial data is listed): Table 3

[0059] Based on these data, calculate the transition probability: (1) The transition probability from the "focused" state: Transition to the "fatigued" state: ; Transition to the "anxious" state: ; Maintain the "focused" state: ; (2) Transition probability from the "fatigued" state: Transition to the "relaxed" state: ; Transition to the "anxious" state: ; (3) Transition probability from the "relaxed" state: Transition to the "focused" state: ; Transition to the "fatigued" state: ; Assume that the transition probability matrix A for all states can be represented as: ; where each represents the probability of transitioning from state i to state j. Based on the above example data, part of the matrix can be: .

[0060] When calculating the observation probability matrix B, assume there is historical observation data as shown in Table 4 below (only partial data is listed): Table 4 .

[0061] Based on this data, calculate the observation probabilities: (1) Observation probability of the "focused" state: Observation of "smile": ; Observation of "frown": ; Observation of "calm": ; (2) Observation probability of the "fatigued" state: Observation of "smile": ; Observation of "frown": ; (3) Observation probability of the "relaxed" state: Observation of "smile": ; Observation of "frown": ; Assume that the transition probability matrix B for all states can be represented as: ; where each represents the probability of observing data in the emotional state . Based on the above example data, part of the matrix can be: 。

[0062] When calculating the initial state distribution information, the occurrence times of each emotional state at the initial time point can be counted according to the above-mentioned historical emotion data obtained. For example, if there are 100 pieces of initial state data, among which 30 are "concentrated", 40 are "relaxed", 20 are "fatigued", and 10 are "anxious".

[0063] Normalize the counted frequencies to obtain the initial state distribution. For example, assume the statistical results of the initial state frequencies are as follows: "concentrated" state: 30 times; "relaxed" state: 40 times; "fatigued" state: 20 times; "anxious" state: 10 times. Then, after normalization, the initial state distribution h is: 。

[0064] Furthermore, using the state transition probability matrix, observation probability matrix, and state distribution information of the sample employees, the parameters of the state transition probability matrix and observation probability matrix in the dynamic emotion Bayesian network model are estimated through the expectation-maximization algorithm to complete model training and obtain the emotion state prediction model.

[0065] This application can monitor and predict the emotion states of employees in real time by constructing a dynamic emotion Bayesian network model, thereby realizing personalized intelligent task allocation and scheduling according to the real-time emotion states of employees.

[0066] Based on the above embodiments, when iteratively training the dynamic Bayesian network model using the state transition probability matrix, observation probability matrix, and state distribution information of the sample employees with the expectation-maximization algorithm, for each iterative training process, the following operations are respectively performed: Determine the forward probability according to the state transition probability matrix, observation probability matrix, and state distribution information of the sample employees; Determine the backward probability according to the state transition probability matrix, observation probability matrix, state distribution information, and forward probability of the sample employees; Determine the posterior probability based on the forward probability and backward probability; Update the observation probability matrix and state distribution information of the sample employees respectively based on the posterior probability; Update the state transition probability matrix of the sample employees based on the forward probability, backward probability, and posterior probability; If the state transition probability matrix, observation probability matrix, and state distribution information of the sample employees converge after update or the current iteration number reaches the preset iteration threshold, determine that the iteration is completed; otherwise, perform the next iteration.

[0067] Furthermore, updating the state transition probability matrix of the sample employees based on the forward probability, backward probability, and posterior probability includes: Determine the joint probability based on the forward probability and the backward probability; Update the state transition probability matrix of the sample employee according to the joint probability and the posterior probability.

[0068] Specifically, when the present application iteratively trains the dynamic Bayesian network model using the expectation maximization algorithm according to the state transition probability matrix, the observation probability matrix, and the state distribution information of the sample employee, the following steps may be executed: Step (1): E step, calculate the posterior distribution of the hidden state under the given current parameters; 1. Forward algorithm: Calculate the forward probability : ; ; where N is the total number of emotional states; is the probability of the initial state ; is the probability of observing in state ; 2. Backward algorithm: Calculate the backward probability : ; ; where T represents the length of the observation sequence, i.e., the total number of time steps. For example, if data is collected once a day, then it may be the number of data points in a day; if data is collected once an hour, then it is the number of observed data points per hour in a day. Therefore, there are no more observed values after time T; 3. Forward-backward algorithm: Calculate the posterior probability : ; The present application combines the forward probability and the backward probability to calculate the posterior probability, comprehensively considering the influence of all observed data from the initial time to the current time and all observed data from the current time to the future time, so as to achieve a comprehensive prediction of the emotional state and optimize the resource utilization; Step (2): M step, update the network model parameters , A, B; 1. Update the initial state distribution : ; where is the probability of being in state i at time t = 1; 2. Update the transition probability matrix A: ; Among them, represents the joint probability, which is the probability of being in state i at time t and transitioning to state j at time t + 1; the joint probability can be determined according to the forward probability and the backward probability through the following formula: ; 3. Update the observation probability matrix B: ; Among them, is an indicator function that takes the value of 1 when the observed value is j, and 0 otherwise; Step (3): Iteration and convergence: Repeatedly perform the E-step and the M-step until the parameters converge or reach a preset iteration (number) threshold; Step (4): Verification: Use the cross-validation method to divide the data into a training set and a validation set, train the model and verify its accuracy, and adjust the model parameters according to the verification results to ensure the generalization ability of the model.

[0069] Through the above process, an emotional state prediction model can be obtained. By inputting the state transition probability matrix, the observation probability matrix, and the state distribution information of the employee into the emotional state prediction model, the forward probability, the backward probability, and the posterior probability output by the model can be obtained to reflect the probability distribution of the employee in different emotional states, thereby predicting the emotional state of the employee.

[0070] This application can construct a dynamic emotion Bayesian network model to monitor and predict the emotional state of employees in real time, thereby realizing personalized intelligent task allocation and scheduling according to the real-time emotional state of employees.

[0071] On the basis of comprehensively considering historical and future data, optimize resource allocation, improve the accuracy of emotional state prediction, and ensure that the scheduling scheme can be adjusted based on more comprehensive data.

[0072] Based on the above embodiments, based on the forward probability, the backward probability, and the posterior probability, combined with the task information to be scheduled and the orca population algorithm, a task scheduling scheme is generated, including: Generate at least one pre-scheduling scheme based on the task information to be scheduled; According to the forward probability, the backward probability, and the posterior probability, adjust the search behavior of the orca population algorithm initialized based on each pre-scheduling scheme; among them, the search behavior of the orca population algorithm includes the encirclement behavior, the spiral update, and the random search; the encirclement behavior dynamically adjusts the parameters of the encirclement behavior using the forward probability and the backward probability; the spiral update dynamically adjusts the step size and frequency of the spiral update using the posterior probability; the random search adjusts the frequency of the random search using the forward probability and the backward probability; After the search behavior adjustment is completed, the pre-scheduling scheme with the highest population fitness in the killer whale population algorithm is determined as the task scheduling scheme.

[0073] Based on the above embodiments, when adjusting the search behavior of the killer whale population algorithm initialized based on each pre-scheduling scheme according to the forward probability, backward probability, and posterior probability, for each adjustment process, the following operations are performed: Update the position according to the adjusted encirclement behavior, spiral update, and random search, and determine the number of iterations or population fitness after the position update; If the number of iterations reaches the preset iteration threshold or the population fitness converges, it is determined that the search behavior adjustment is completed; otherwise, the next search behavior adjustment is performed.

[0074] Specifically, the present application can generate at least one pre-scheduling scheme by combining the task information to be scheduled with the employee information. It should be noted that this process can be implemented by means of manual generation, generation using traditional schemes, or generation using innovative schemes, which are not limited in the present application.

[0075] Furthermore, the killer whale population initialization of the killer whale population algorithm is performed according to each pre-scheduling scheme.

[0076] In the present application, the killer whale population represents diverse scheduling schemes. By simulating the predation behavior of killer whales, the task scheduling can be optimized. Combining the emotional state data of employees, the search strategy is dynamically adjusted to gradually approach the global optimal scheduling scheme, thereby improving the efficiency and quality of task allocation and realizing intelligent mobile office task scheduling.

[0077] Each killer whale in the killer whale population represents a possible scheduling scheme. In the initialization stage, each individual (i.e., killer whale) in the population is randomly assigned an initial task scheduling scheme. Each scheduling scheme can be represented as a vector, which contains information such as task allocation, task priority, and task execution time. In the initialization stage, an initial position vector is randomly generated to ensure coverage of the search space.

[0078] After the initialization is completed, the present application can adjust the search behavior of the initialized killer whale population algorithm according to the predicted forward probability, backward probability, and posterior probability, thereby dynamically adjusting (iteratively optimizing) the parameters of the killer whale population algorithm to realize intelligent mobile office task scheduling in combination with the mental state of employees.

[0079] Among them, the search behavior of the killer whale population algorithm includes encirclement behavior, spiral update, and random search; the encirclement behavior dynamically adjusts the parameters of the encirclement behavior using the forward probability and backward probability.

[0080] Encirclement behavior: Dynamically adjust the parameters of the encirclement behavior using the forward probability and backward probability.

[0081] Adjustment factor: ; ; wherein, linearly decreases from 2 to 0; is a random number between [0, 1].

[0082] Dynamically adjust the values of A and C according to the forward probability and the backward probability : ; ; Update the position: ; ; wherein, represents the position of the current orca; represents the optimal position in the current population.

[0083] By utilizing the forward probability and the backward probability characterizing the emotional state of employees, the search parameters can be adjusted more precisely, improving the accuracy of task scheduling.

[0084] Spiral update: Utilize the posterior probability to dynamically adjust the step size and frequency of the spiral update.

[0085] Adjust the step size and frequency: ; ; wherein, is a constant defining the spiral shape; is a random number between [-1, 1].

[0086] Dynamically adjust according to the posterior probability and : ; ; Dynamically adjusting the search step size and frequency according to the posterior probability can improve the flexibility and accuracy of the search.

[0087] Random search: Utilize the forward probability and the backward probability to adjust the frequency of the random search.

[0088] ; wherein, It is a randomly selected position.

[0089] According to the forward probability and the backward probability Adjust the frequency of random search: ; By dynamically adjusting the random search frequency, the diversity and coverage of the search space can be effectively enhanced.

[0090] This application can comprehensively utilize the forward probability, backward probability, and posterior probability to dynamically adjust all search behaviors of the killer whale population algorithm to ensure that the search behaviors are adapted to the emotional state of the employees.

[0091] Among them, the following emotional state mappings are included: Relaxed state: According to the forward probability and the backward probability Decrease and values to enhance local search; Focused state: Adjust according to the posterior probability and values to maintain search balance; Fatigue state: According to the forward probability and the backward probability Increase and values to enhance global search; Anxiety state: Decrease according to the posterior probability value to increase the frequency of spiral update; Stress state: Increase and values according to the forward probability and the backward probability to reallocate tasks.

[0092] Specifically, when this application adjusts the search behavior of the killer whale population algorithm after initialization based on each pre-scheduling scheme, multiple rounds of iteration can be performed, and the population fitness is calculated during each adjustment process. When calculating the population fitness, it is necessary to comprehensively consider the task completion time, resource utilization rate, and the emotional state of the employees: ; Wherein, P is the total task completion time; R is the resource utilization rate, including the human resource utilization rate, the technical resource utilization rate, etc. The human resource rate refers to the work idle and busy rate of employees, and the idle or busy state can be determined according to the usage status of employees' office software; The technical resource utilization rate includes the utilization rates of computing resources, network resources, and hardware resources, etc., such as the utilization rates of remote servers, network bandwidth, cloud storage, and company printers; E is the employee emotional state output by the dynamic emotion Bayesian network model.

[0093] During each adjustment process, the position is also updated according to the encirclement behavior, spiral update, and random search until the preset iteration number threshold is reached or the fitness value no longer changes significantly (i.e., converges), and then the iteration is terminated.

[0094] Furthermore, after the search behavior adjustment is completed, the pre-scheduling scheme with the highest population fitness in the orca population algorithm can be determined as the task scheduling scheme.

[0095] This application can also monitor the output optimal task scheduling scheme, discover problems in a timely manner and make adjustments to ensure the continuous effectiveness and adaptability of the scheme.

[0096] This application combines the dynamic emotion Bayesian network with the orca population algorithm to optimize the task scheduling in the mobile office scenario. By real-time monitoring and predicting the emotional state of employees, the optimization algorithm parameters are dynamically adjusted to achieve intelligent and efficient task allocation combined with the emotional state of employees.

[0097] Specifically, this application uses the forward probability, backward probability, and posterior probability output by the dynamic emotion Bayesian network model to dynamically adjust the search parameters of the orca population algorithm, improve the search efficiency and accuracy, and enhance the adaptability and robustness of the algorithm.

[0098] Moreover, the search parameters are adjusted according to the emotional state of employees, so that when the employees are in a good state, the global search is increased to improve the comprehensiveness of task allocation; when the employees are in a poor state, the local search is increased to improve the accuracy of task allocation; enabling the orca population search algorithm to adapt to changes in different emotional states, avoiding falling into local optimal solutions, and improving the overall efficiency and stability of task scheduling.

[0099] The task scheduling device provided by this application will be described below. The task scheduling device described below can be mutually corresponding and referred to the task scheduling method described above.

[0100] Furthermore, this application also provides a task scheduling device.

[0101] The task scheduling device includes: A determination module, configured to respectively determine the state transition probability matrix, the observation probability matrix, and the state distribution information according to the historical emotion data of the employees to be scheduled; A prediction module, configured to input the state transition probability matrix, the observation probability matrix, and the state distribution information of the employee to be scheduled into an emotional state prediction model, and obtain the forward probability, the backward probability, and the posterior probability output by the emotional state prediction model; wherein, the emotional state prediction model is obtained by training a dynamic Bayesian network model according to the state transition probability matrix, the observation probability matrix, and the state distribution information of sample employees; A generation module, configured to generate a task scheduling plan based on the forward probability, the backward probability, and the posterior probability, in combination with the task information to be scheduled and the killer whale population algorithm.

[0102] The task scheduling device of the present application trains a dynamic Bayesian network model in advance according to the state transition probability matrix, the observation probability matrix, and the state distribution information of sample employees to obtain an emotional state prediction model. Thus, the state transition probability matrix, the observation probability matrix, and the state distribution information can be respectively determined according to the historical emotional data of the employee to be scheduled. Furthermore, by inputting the state transition probability matrix, the observation probability matrix, and the state distribution information of the employee to be scheduled into the emotional state prediction model, the forward probability, the backward probability, and the posterior probability output by the emotional state prediction model can be obtained. Further, based on the forward probability, the backward probability, and the posterior probability, in combination with the task information to be scheduled and the killer whale population algorithm, a task scheduling plan can be generated. By training a dynamic Bayesian network model to obtain an emotional state prediction model, the emotional state of employees can be monitored and predicted in real time according to the emotional state prediction model in combination with the state transition probability matrix, the observation probability matrix, and the state distribution information respectively determined from the historical emotional data of the employee to be scheduled. Furthermore, based on the forward probability, the backward probability, and the posterior probability, in combination with the task information to be scheduled and the killer whale population algorithm, a task scheduling plan can be generated, thereby realizing personalized intelligent task scheduling according to the real-time emotional state of employees, which helps to improve the efficiency of task execution in a mobile office environment.

[0103] In one embodiment, the generation module is specifically configured to: Generate at least one pre-scheduling plan based on the task information to be scheduled; According to the forward probability, the backward probability, and the posterior probability, adjust the search behavior of the killer whale population algorithm initialized based on each pre-scheduling plan; wherein, the search behavior of the killer whale population algorithm includes encirclement behavior, spiral update, and random search; the encirclement behavior dynamically adjusts the parameters of the encirclement behavior by using the forward probability and the backward probability; the spiral update dynamically adjusts the step size and frequency of the spiral update by using the posterior probability; the random search adjusts the frequency of the random search by using the forward probability and the backward probability; Determine the pre-scheduling plan with the highest population fitness in the killer whale population algorithm after the search behavior adjustment as the task scheduling plan.

[0104] In one embodiment, when the generating module is further configured to adjust the search behavior of the killer whale population algorithm after initialization based on each pre-scheduling scheme according to the forward probability, the backward probability, and the posterior probability, for each adjustment process, the following operations are performed: Update the position according to the adjusted encirclement behavior, spiral update, and random search, and determine the number of iterations or population fitness after the position update; If the number of iterations reaches the preset iteration threshold or the population fitness converges, it is determined that the search behavior adjustment is completed; otherwise, the next search behavior adjustment is performed.

[0105] Figure 2 An example of the physical structure diagram of an electronic device is shown as Figure 2 shown. The electronic device may include: a processor 210, a communications interface 220, a memory 230, and a communication bus 240. Among them, the processor 210, the communications interface 220, and the memory 230 complete mutual communication through the communication bus 240. The processor 210 may call the logical instructions in the memory 230 to execute the following method: respectively determine the state transition probability matrix, the observation probability matrix, and the state distribution information according to the historical emotion data of the employee to be scheduled; Input the state transition probability matrix, the observation probability matrix, and the state distribution information of the employee to be scheduled into the emotion state prediction model, and obtain the forward probability, the backward probability, and the posterior probability output by the emotion state prediction model; wherein, the emotion state prediction model is obtained by training a dynamic Bayesian network model according to the state transition probability matrix, the observation probability matrix, and the state distribution information of the sample employee; Based on the forward probability, the backward probability, and the posterior probability, combine the information of the task to be scheduled and the killer whale population algorithm to generate a task scheduling scheme.

[0106] In addition, when the logical instructions in the above-mentioned memory 230 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0107] In another aspect, an embodiment of this application further provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is configured to execute the methods provided in the above-mentioned various embodiments. For example, it includes: respectively determining a state transition probability matrix, an observation probability matrix, and state distribution information according to the historical emotion data of the employee to be scheduled; Inputting the state transition probability matrix, the observation probability matrix, and the state distribution information of the employee to be scheduled into an emotion state prediction model to obtain the forward probability, backward probability, and posterior probability output by the emotion state prediction model; wherein, the emotion state prediction model is obtained by training a dynamic Bayesian network model according to the state transition probability matrix, the observation probability matrix, and the state distribution information of sample employees; Based on the forward probability, the backward probability, and the posterior probability, combining the information of the task to be scheduled and the orca population algorithm to generate a task scheduling plan.

[0108] In another aspect, an embodiment of this application further provides a computer program product, on which a computer program is stored. When the computer program is executed by a processor, it is configured to execute the methods provided in the above-mentioned various embodiments. For example, it includes: respectively determining a state transition probability matrix, an observation probability matrix, and state distribution information according to the historical emotion data of the employee to be scheduled; Inputting the state transition probability matrix, the observation probability matrix, and the state distribution information of the employee to be scheduled into an emotion state prediction model to obtain the forward probability, backward probability, and posterior probability output by the emotion state prediction model; wherein, the emotion state prediction model is obtained by training a dynamic Bayesian network model according to the state transition probability matrix, the observation probability matrix, and the state distribution information of sample employees; Based on the forward probability, the backward probability, and the posterior probability, a task scheduling scheme is generated by combining the information of the task to be scheduled with the orca population algorithm.

[0109] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative effort.

[0110] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution or the part that contributes to the related technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the present application, rather than to limit the present application. Although the present application has been described in detail with reference to the embodiments, those of ordinary skill in the art should understand that various combinations, modifications, or equivalent replacements of the technical solutions of the present application do not depart from the spirit and scope of the technical solutions of the present application.

Claims

1. A task scheduling method, characterized in that, Including: Respectively determine the state transition probability matrix, the observation probability matrix and the state distribution information according to the historical emotion data of the employees to be scheduled; Input the state transition probability matrix, the observation probability matrix and the state distribution information of the employees to be scheduled into the emotion state prediction model, and obtain the forward probability, the backward probability and the posterior probability output by the emotion state prediction model; wherein, the emotion state prediction model is trained based on the state transition probability matrix, the observation probability matrix and the state distribution information of the sample employees for the dynamic Bayesian network model; Based on the forward probability, the backward probability and the posterior probability, combine the information of the tasks to be scheduled and the killer whale population algorithm to generate a task scheduling plan.

2. The task scheduling method according to claim 1, wherein The generating a task scheduling plan by combining the information of the tasks to be scheduled and the killer whale population algorithm based on the forward probability, the backward probability and the posterior probability includes: Generate at least one pre-scheduling plan based on the information of the tasks to be scheduled; According to the forward probability, the backward probability and the posterior probability, adjust the search behavior of the killer whale population algorithm initialized based on each pre-scheduling plan; wherein, the search behavior of the killer whale population algorithm includes the encirclement behavior, the spiral update and the random search; the encirclement behavior dynamically adjusts the parameters of the encirclement behavior by using the forward probability and the backward probability; the spiral update dynamically adjusts the step size and frequency of the spiral update by using the posterior probability; the random search adjusts the frequency of the random search by using the forward probability and the backward probability; Determine the pre-scheduling plan with the highest population fitness in the killer whale population algorithm after the search behavior adjustment as the task scheduling plan.

3. The task scheduling method according to claim 2, wherein When adjusting the search behavior of the killer whale population algorithm initialized based on each pre-scheduling plan according to the forward probability, the backward probability and the posterior probability, for each adjustment process, perform the following operations: Update the position according to the adjusted encirclement behavior, spiral update and random search, and determine the number of iterations or the population fitness after the position update; If the number of iterations reaches the preset iteration number threshold or the population fitness converges, determine that the search behavior adjustment is completed; otherwise, perform the next search behavior adjustment.

4. The task scheduling method according to claim 1, wherein The emotion state prediction model is determined in the following way: Respectively determine the state transition probability matrix, the observation probability matrix and the state distribution information according to the historical emotion data of the sample employees; According to the state transition probability matrix, the observation probability matrix and the state distribution information of the sample employees, use the expectation maximization algorithm to iteratively train the dynamic Bayesian network model, and obtain the emotion state prediction model after the iteration is completed.

5. The task scheduling method according to claim 4, wherein, When using the expectation maximization algorithm to iteratively train the dynamic Bayesian network model according to the state transition probability matrix, the observation probability matrix and the state distribution information of the sample employees, for each iterative training process, perform the following operations respectively: Determine the forward probability according to the state transition probability matrix, the observation probability matrix and the state distribution information of the sample employees; Determine the backward probability according to the state transition probability matrix, the observation probability matrix, the state distribution information of the sample employees and the forward probability; Determine the posterior probability based on the forward probability and the backward probability; Update the observation probability matrix and the state distribution information of the sample employee respectively based on the posterior probability; Update the state transition probability matrix of the sample employee based on the forward probability, the backward probability and the posterior probability; If the state transition probability matrix, the observation probability matrix and the state distribution information of the sample employee converge after the update or the current iteration number reaches the preset iteration threshold, determine that the iteration is completed; Otherwise, perform the next iteration.

6. The task scheduling method according to claim 5, characterized in that The step of updating the state transition probability matrix of the sample employee based on the forward probability, the backward probability and the posterior probability includes: Determine the joint probability according to the forward probability and the backward probability; Update the state transition probability matrix of the sample employee according to the joint probability and the posterior probability.

7. A task scheduling device, characterized in that, including: A determination module, configured to determine a state transition probability matrix, an observation probability matrix and state distribution information respectively according to the historical emotion data of the employee to be scheduled; A prediction module, configured to input the state transition probability matrix, the observation probability matrix and the state distribution information of the employee to be scheduled into an emotion state prediction model, and obtain the forward probability, the backward probability and the posterior probability output by the emotion state prediction model; wherein, the emotion state prediction model is obtained by training a dynamic Bayesian network model according to the state transition probability matrix, the observation probability matrix and the state distribution information of the sample employee; A generation module, configured to generate a task scheduling plan based on the forward probability, the backward probability and the posterior probability, in combination with the information of the task to be scheduled and the orca population algorithm.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the task scheduling method according to any one of claims 1-6.

9. A storage medium, the storage medium being a non-transitory computer-readable storage medium, having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the task scheduling method according to any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the task scheduling method according to any one of claims 1-6.

Images