Workflow intelligent management method and system based on artificial intelligence

Abstract

The invention discloses a workflow intelligent management method and system based on artificial intelligence, and relates to the technical field of workflow intelligent management, and the method comprises the steps: employing an incremental learning model to analyze workflow historical data, constructing a probability distribution model of task completion time, and calculating a risk coefficient of a task based on the probability distribution model; according to the risk coefficient, the buffer time and the resource reservation strategy of the task are dynamically adjusted in combination with an expert rule base, and an early warning mechanism is triggered when the risk coefficient exceeds a preset threshold value; and performing classification processing on the early warning tasks through the incremental decision tree, automatically generating a risk response scheme, and feeding back a processing result to the expert rule base for rule updating. According to the invention, the task delay and resource conflict risk can be accurately predicted, the adaptive optimization of the system is realized, and a more intelligent and adaptive workflow management solution is provided for enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field of workflow management, and in particular to an intelligent workflow management method and system based on artificial intelligence. Background Art

[0002] With the deepening of digital transformation of enterprises, workflow management system (WfMS) has become a key tool to improve organizational operational efficiency. Traditional workflow management mainly relies on static process definition and fixed resource allocation strategy, and coordinates task execution through predefined business rules and scheduling algorithms. In recent years, with the rapid development of artificial intelligence technology, especially the widespread application of machine learning and deep learning in the field of process mining and predictive analysis, intelligent workflow management has gradually become a research hotspot. Existing technologies have made significant progress in workflow modeling, task scheduling optimization, resource allocation, etc., such as automatic discovery of workflow models based on process mining, task scheduling optimization based on heuristic algorithms, and resource collaborative allocation based on multi-agent systems.

[0003] However, existing workflow management methods still have many limitations when facing complex and changing business environments. First, traditional static thresholds and fixed rules are difficult to adapt to dynamically changing business needs, resulting in insufficient accuracy in the system's prediction of task delay risks and resource conflicts. Second, existing risk assessment models often use a single assessment indicator, lack comprehensive consideration of multi-dimensional risk factors, and cannot accurately reflect the actual risk status during task execution. Third, existing technologies generally adopt a passive response strategy in resource scheduling, lack active early warning and dynamic adjustment mechanisms, resulting in a significant decrease in system performance under high load or emergency conditions. Finally, existing methods have obvious deficiencies in knowledge accumulation and experience inheritance, making it difficult to achieve continuous optimization and improvement of system performance.

[0004] The existing workflow management technology mainly has technical problems such as low accuracy of task risk prediction, low resource scheduling efficiency, poor system adaptability, etc. This invention effectively solves the above technical problems by introducing incremental learning models, dynamic risk assessment mechanisms and intelligent early warning responses, and provides enterprises with a more intelligent and adaptive workflow management solution. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the present invention provides an artificial intelligence-based workflow intelligent management method and system, which can solve the problems mentioned in the background technology.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: an artificial intelligence-based workflow intelligent management method, comprising: using an incremental learning model to analyze workflow historical data, constructing a probability distribution model of task completion time, and calculating the risk coefficient of the task based on the probability distribution model; the risk coefficient includes the probability of task delay and the probability of resource conflict; according to the risk coefficient, the buffer time and resource reservation strategy of the task are dynamically adjusted in combination with an expert rule library, and an early warning mechanism is triggered when the risk coefficient exceeds a preset threshold; the early warning mechanism sets different early warning levels based on the criticality of the task; the early warning tasks are classified and processed through an incremental decision tree, a risk response plan is automatically generated, and the processing results are fed back to the expert rule library for rule updating.

[0008] As a preferred solution of the workflow intelligent management method based on artificial intelligence described in the present invention, the process of calculating the risk coefficient includes the following steps: using an incremental learning model to extract features of task start and end times, resource occupancy information, and task dependencies in workflow historical data to obtain feature vectors; using the feature extraction results to construct a probability distribution model of task completion time; calculating the task delay probability based on the probability distribution model; calculating the resource conflict probability between tasks; and calculating the risk coefficient of the task through the FMEA method based on the delay probability and the resource conflict probability.

[0009] As a preferred solution of the workflow intelligent management method based on artificial intelligence described in the present invention, wherein: the probability distribution model is expressed by the following formula:

[0010]

[0011] Among them, μ t is the average completion time of historical tasks; is the variance of the completion time; P(t) is the output of the probability distribution model.

[0012] The delay probability is expressed by the following formula: deadline , the probability of delay is:

[0013]

[0014] Through standardization transformation, we get:

[0015]

[0016] Where Φ(x) is the cumulative distribution function of the standard normal distribution; P d is the probability of delay.

[0017] The calculation of the resource conflict probability comprises the following steps:

[0018] The resource-time occupancy set O of the construction task task :

[0019] O task =(r m , t s )|r m ∈R set , t s ∈[T start , T end ];

[0020] Among them, r m is the mth required resource; R set is the set of resources required for the task; s is the time point; T start is the task start time; T end The estimated end time of the task.

[0021] Calculating resource conflict probability P based on resource-time intersection c :

[0022]

[0023] Among them, O i is the resource-time occupancy set of task i; O j is the resource-time occupancy set of task i; i ∩O j | represents the intersection cardinality of the occupied sets; |O i | represents the cardinality of the occupancy set; Γ is a set of task pairs that satisfy the following conditions: the execution time periods of task i and task j overlap; the resource requirement sets of task i and task j intersect.

[0024] The risk factor is expressed by the following formula:

[0025] R=O risk ×S risk ×D risk ;

[0026] Among them, O risk is the risk rate: O risk =max(P d , P c )×10;

[0027] S risk S is the risk severity: S risk =-ln(1-P d )×5;

[0028] D risk Risk detection difficulty: Drisk =(1+P c ) 2 ×5.

[0029] As a preferred solution of the workflow intelligent management method based on artificial intelligence described in the present invention, the expert rule base includes a risk coefficient threshold matrix M for different task levels and different operating states. threshold and the task criticality matrix M critical ; The task levels include high-level tasks, medium-level tasks, and low-level tasks, and the operating states include normal state, bottleneck state, and emergency state.

[0030] As a preferred solution of the workflow intelligent management method based on artificial intelligence described in the present invention, the risk coefficient threshold matrix M threshold The expression is:

[0031]

[0032] The task criticality matrix M critical The expression is:

[0033]

[0034] Among them, R tij is the risk threshold of the i-th level task in the j-th state; K ij is the criticality coefficient of the i-th level task in the j-th state; i∈[1,3], represents the task level, i=1: low-level task, i=2: medium-level task, i=3: high-level task; j∈[1,3], represents the running state, j=1: normal state, j=2: bottleneck state, j=3: emergency state; R t21 is the risk threshold of the intermediate task under normal conditions, and is used as the intermediate threshold for the first-level warning judgment; R t31 is the risk threshold of the high-level task under normal conditions, and is used as the high-level threshold for the second-level and third-level warning judgments; K 21 is the criticality coefficient of the intermediate task under normal conditions, and is used as the intermediate threshold for the first-level warning judgment; K 31 It is the criticality coefficient of the high-level task under normal conditions, and is used as the high-level threshold for second-level and third-level warning judgments.

[0035] As a preferred solution of the workflow intelligent management method based on artificial intelligence described in the present invention, wherein: according to the risk coefficient, the buffer time and resource reservation strategy of the task are dynamically adjusted in combination with the expert rule base, and the early warning mechanism is triggered when the risk coefficient exceeds the preset threshold, including the following steps: establishing an expert rule base;

[0036] Calculate the buffer time T of the task buffer:

[0037] T buffer =T base ×(1+α t R task +β t K task );

[0038] Among them, T base is the benchmark buffer time; R task is the risk factor of the current task; K task is the criticality coefficient of the current task; α t is the risk weight coefficient; β t is the criticality weight coefficient.

[0039] Compute the resource reservation ratio P of the task reserve :

[0040]

[0041] Among them, P min is the minimum resource reservation ratio; P max R is the maximum resource reservation ratio; task is the risk factor of the current task; K task is the criticality coefficient of the current task; R threshold is the upper limit of the risk factor threshold; K max is the upper limit of the criticality coefficient.

[0042] The risk factor and the risk factor threshold matrix M threshold Advanced Thresholding in R t31 and the intermediate threshold R t21 Compare the task criticality with the task criticality matrix M critical The advanced threshold K in 31 and the intermediate threshold K 21 Make a comparison.

[0043] According to the comparison results, the warning level is divided into zero-level warning, first-level warning, second-level warning, and third-level warning; when the risk coefficient exceeds R t31 , and the mission criticality exceeds K 31 When the risk factor exceeds R t31 Or the mission criticality exceeds K 31 When the risk factor exceeds R t21 , and the mission criticality exceeds K 21 It is considered as a Level 1 warning when the situation is serious, and it is considered as a Level 0 warning in other cases.

[0044] A corresponding relationship is established between the warning level and the warning response; among which, the third-level warning triggers three warning responses, namely, warning notification, emergency resource pool activation and task re-planning; the second-level warning triggers two warning responses, namely, warning notification and emergency resource pool activation; the first-level warning triggers one warning response, namely, warning notification; and the zero-level warning does not trigger any warning response.

[0045] The buffer time T buffer , the resource reservation ratio P reserve And the warning response Action warn Perform collaborative scheduling and output the scheduling plan matrix.

[0046] As a preferred solution of the workflow intelligent management method based on artificial intelligence described in the present invention, wherein: warning tasks are classified and processed through incremental decision trees, risk response plans are automatically generated, and the processing results are fed back to the expert rule library for rule updating, including the following steps: constructing a feature space of an incremental decision tree based on risk coefficient, task criticality, buffer time, resource reservation ratio, warning level and warning response, and using information gain as a node splitting criterion to construct an incremental decision tree model; setting processing effect evaluation indicators including warning accuracy, response time and resource consumption, pruning and optimizing the decision tree when the processing effect of the subtree is less than or equal to the product of the parent node processing effect and the pruning threshold coefficient, and updating the decision tree leaf node path that meets the rule update threshold to the expert rule library.

[0047] To further solve the above technical problems, the present invention provides the following technical solutions: an artificial intelligence-based workflow intelligent management system, comprising: an intelligent analysis module, used to analyze workflow historical data using an incremental learning model, construct a probability distribution model of task completion time, and calculate the risk coefficient of the task based on the probability distribution model; a dynamic adjustment module, used to dynamically adjust the buffer time and resource reservation strategy of the task according to the risk coefficient in combination with an expert rule library, and trigger an early warning mechanism when the risk coefficient exceeds a preset threshold; an intelligent response module, used to classify and process early warning tasks through an incremental decision tree, automatically generate risk response plans, and feed back the processing results to the expert rule library for rule updating.

[0048] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the above-mentioned artificial intelligence-based intelligent workflow management method when executing the computer program.

[0049] 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 the above-mentioned artificial intelligence-based intelligent workflow management method are implemented.

[0050] The beneficial effects of the present invention are as follows: the present invention constructs a probability distribution model of task completion time by adopting an incremental learning model, and calculates the risk coefficient by combining the FMEA method, so as to accurately predict the risk of task delay and resource conflict; dynamically adjusts the task buffer time and resource reservation strategy through the expert rule library, thereby realizing the adaptive optimization of the system; adopts an incremental decision tree to classify and process the warning tasks, and feeds back the processing results to the expert rule library, thus forming a closed-loop continuous optimization mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0052] Figure 1 This is a schematic diagram of the overall process of an intelligent workflow management method based on artificial intelligence proposed by the present invention;

[0053] Figure 2 This is a computer equipment diagram in an artificial intelligence-based workflow intelligent management method proposed by the present invention. DETAILED DESCRIPTION

[0054] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0055] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0056] Example 1, reference Figure 1 , which is an embodiment of the present invention, provides an intelligent workflow management method based on artificial intelligence.

[0057] In the relevant technologies, firstly, traditional static thresholds and fixed rules are difficult to adapt to dynamically changing business needs, resulting in insufficient accuracy in the system's prediction of task delay risks and resource conflicts. Secondly, existing risk assessment models often use a single assessment indicator, lack comprehensive consideration of multi-dimensional risk factors, and cannot accurately reflect the actual risk status during task execution. Thirdly, existing technologies generally adopt a passive response strategy in resource scheduling, lack active early warning and dynamic adjustment mechanisms, resulting in a significant decrease in system performance under high load or emergency conditions. Finally, existing methods have obvious deficiencies in knowledge accumulation and experience inheritance, making it difficult to achieve continuous optimization and improvement of system performance.

[0058] The present application provides an effective solution to the above-mentioned problems. Next, multiple embodiments will be combined to explain in detail how to implement the intelligent workflow management method based on artificial intelligence.

[0059] Figure 1 The overall flow chart of an intelligent workflow management method based on artificial intelligence is shown, including:

[0060] S1: Use the incremental learning model to analyze the workflow historical data, build a probability distribution model of task completion time, and calculate the risk coefficient of the task based on the probability distribution model.

[0061] Specifically, the risk factors include the probability of task delay and the probability of resource conflict.

[0062] S1.1: Use the incremental learning model to learn the task start and end time T in the workflow history data h , resource occupancy information R t , Task Dependency D t Perform feature extraction and obtain the feature vector F = [T t , R t , D t ].

[0063] Among them, the incremental learning model updates the feature weights based on the following function:

[0064] w k+1 =w k +η(y k -f(x k ))x k ;

[0065] Among them, w k is the feature weight of the kth iteration; η is the learning rate; y k is the true value of the kth iteration; f(x k ) is the predicted value of the kth iteration; x k is the input feature of the kth iteration.

[0066] It should be noted that by adopting the incremental learning model for data feature extraction, the online processing capability of historical data is realized. Compared with the traditional batch learning method, this step does not need to load all historical data at one time, which effectively reduces the system storage overhead and achieves the dynamic update effect of the feature extraction process.

[0067] S1.2: Use the feature extraction results to construct the probability distribution model P(t) of task completion time:

[0068]

[0069] Among them, μ t is the average completion time of historical tasks; is the variance of completion time; t i is the completion time of the i-th historical task; N is the total number of historical tasks.

[0070] It should be noted that by constructing a task completion time probability model based on normal distribution, discrete historical data are converted into a continuous probability distribution function. This step provides the necessary mathematical basis for the subsequent delay risk calculation.

[0071] S1.3: Calculate the task delay probability P based on the probability distribution model P(t) in S1.2 d :

[0072] For the specified completion time T deadline , the probability of delay is:

[0073]

[0074] Through standardization transformation, we get:

[0075]

[0076] Where Φ(x) is the cumulative distribution function of the standard normal distribution.

[0077] It should be noted that by establishing a calculation method for the probability of task delay and introducing standardized transformation to simplify the calculation process, a quantitative assessment of the risk of task delay is achieved. This step not only provides a quantitative indicator of the risk of delay, but also improves the calculation efficiency through standardized processing, ultimately achieving an accurate measurement effect of the risk of delay.

[0078] S1.4: Calculate the probability of resource conflict between tasks P d :

[0079] First, construct the resource-time occupancy set O of the task task :

[0080] O task=(r m , t s )|r m ∈R set , t s ∈[T start , T end ];

[0081] Among them, r m is the mth required resource; R set is the set of resources required for the task; s is the time point; T start is the task start time; T end The estimated end time of the task.

[0082] Then the conflict probability P is calculated based on the resource-time intersection c :

[0083]

[0084] Among them, O i is the resource-time occupancy set of task i; O j is the resource-time occupancy set of task i; i ∩O j | represents the intersection cardinality of the occupied sets; |O i | represents the cardinality of the occupied set; Γ is the set of task pairs that satisfy the following conditions:

[0085] There is overlap in the execution time periods of task i and task j;

[0086] There is an intersection between the resource requirement sets of task i and task j;

[0087] It should be noted that by proposing a resource conflict assessment scheme based on set operations, unified processing of the time dimension and the resource dimension is achieved. This method transforms the calculation of the resource conflict probability into a set intersection operation, breaking through the limitation of the traditional method of only considering a single dimension, and ultimately achieving an accurate quantification effect of resource conflicts.

[0088] S1.5: Based on the delay probability P d and resource conflict probability P c , the improved FMEA method is used to calculate the risk factor R of the task:

[0089] R=O risk ×S risk ×D risk ;

[0090] The calculation of the three risk factors is as follows:

[0091] Risk Occurrence Rate risk :Orisk =max(P d , P c )×10;

[0092] Risk severity S risk :S risk =-ln(1-P d )×5;

[0093] Risk Detection Difficulty D risk :D risk =(1+P c ) 2 ×5.

[0094] Furthermore, compared with the traditional FMEA method, the present invention mainly makes the following improvements:

[0095] 1. Improve the traditional FMEA based on experience scoring method to a quantitative calculation method based on probability;

[0096] 2. Risk Occurrence Rate risk Instead of using a discrete score of 1-10, we use max(P d , P c ) and multiply by 10 for quantization;

[0097] 3. Risk severity S risk Introducing the logarithmic function - ln(1-P d ) establishes a relationship with the probability of postponement, reflecting the nonlinear loss caused by postponement;

[0098] 4. Risk Detection Difficulty D risk Using the quadratic function (1+P c ) 2 Characterize the increase in detection complexity caused by resource conflicts.

[0099] It should be noted that by constructing an improved FMEA risk assessment scheme, an accurate mapping from probability indicators to risk coefficients is achieved. In this step, the maximum value of the probability of delay and the probability of resource conflict is used as the benchmark for the risk occurrence rate, which reflects the short board theory in the "barrel effect"; the risk severity introduces a logarithmic transformation, which increases rapidly when the probability of delay is close to 1, reflecting the nonlinear loss caused by task delay; the risk detection difficulty strengthens the impact of resource conflicts through quadratic transformation, highlighting the complexity of resource scheduling in multi-task scenarios. This evaluation method based on mathematical transformation not only maintains the versatility of the FMEA framework, but also enhances the discrimination of risk indicators through nonlinear mapping, making the risk assessment results more consistent with the empirical laws in engineering practice, and ultimately providing a reliable quantitative basis for risk management and resource scheduling decisions of workflows. In particular, by unifying the risk of delay and resource conflict under the FMEA framework, a comprehensive evaluation of multi-dimensional risks is achieved, overcoming the problem of the separation of risk assessment indicators in traditional methods, and providing more comprehensive decision support for risk prevention and control of workflow systems.

[0100] S2: According to the risk factor, the buffer time and resource reservation strategy of the task are dynamically adjusted in combination with the expert rule base, and the early warning mechanism is triggered when the risk factor exceeds the preset threshold.

[0101] Specifically, the early warning mechanism sets different warning levels based on the criticality of the mission.

[0102] S2.1: Establish an expert rule base.

[0103] The expert rule base contains the risk factor threshold matrix M for different task levels and different operating states. threshold and the task criticality matrix M critical ,The task levels are divided into three levels: high, medium and low, and the ,operation states include normal, bottleneck and emergency.

[0104] Risk factor threshold matrix M threshold The expression is:

[0105]

[0106] Mission criticality matrix M critical The expression is:

[0107]

[0108] Among them, R tij represents the risk threshold of the i-th level task in the j-th state, in R tij In the above figure, t stands for "threshold", which is used to distinguish that this is a threshold type parameter; K ijIt represents the criticality coefficient of the i-th level task in the j-th state; i∈[1,3] represents the task level, i=1: low-level task, i=2: medium-level task, i=3: high-level task; j∈[1,3] represents the operating state, j=1: normal state, j=2: bottleneck state, j=3: emergency state.

[0109] Risk factor threshold matrix M threshold The meaning of each element is:

[0110] R t11 : Risk threshold of low-level tasks under normal conditions;

[0111] R t12 : Risk threshold of low-level tasks in bottleneck state;

[0112] R t13 : Risk threshold of low-level tasks in emergency situations;

[0113] R t21 : The risk threshold of the intermediate task under normal conditions, used as the intermediate threshold for the first-level warning judgment;

[0114] R t22 : Risk threshold of intermediate tasks in bottleneck state;

[0115] R t23 : Risk threshold of medium-level tasks in emergency situations;

[0116] R t31 : The risk threshold of high-level tasks under normal conditions is used as the high-level threshold for second-level and third-level warning judgments;

[0117] R t32 : Risk threshold of high-level tasks in bottleneck state;

[0118] R t33 : Risk threshold of high-level tasks in emergency situations.

[0119] Mission criticality matrix M critical The meaning of each element is:

[0120] K 11 : The criticality coefficient of low-level tasks under normal conditions;

[0121] K 12 : The criticality coefficient of the low-level task in the bottleneck state;

[0122] K 13 : The criticality coefficient of low-level tasks in emergency situations;

[0123] K 21: The criticality coefficient of the intermediate task under normal conditions is used as the intermediate threshold for the first-level warning judgment;

[0124] K 22 : The criticality coefficient of the intermediate task under the bottleneck state;

[0125] K 23 : The criticality coefficient of the intermediate task under emergency conditions;

[0126] K 31 : The criticality coefficient of the high-level task under normal conditions is used as the high-level threshold for the second-level and third-level warning judgments;

[0127] K 32 : The criticality coefficient of the high-level task in the bottleneck state;

[0128] K 33 : The criticality coefficient of high-level tasks under emergency conditions.

[0129] Preferably, the design of the expert rule base adopts a dual matrix (risk coefficient threshold matrix and task criticality matrix) structure, and by taking the task level and the operating state as the two dimensions of the matrix, the thresholds of different task levels under different system states are accurately defined. This dual matrix structure design overcomes the limitation of the prior art that only a single threshold or simple classification is often used, so that the present invention can more accurately characterize the risk and criticality characteristics of the task, and provide a more reliable decision-making basis for subsequent early warning judgment.

[0130] S2.2: Calculate the buffer time T of the task buffer :

[0131] T buffer =T base ×(1+α t R task +β t K task );

[0132] Among them, T base is the benchmark buffer time; R task is the risk factor of the current task; K task is the criticality coefficient of the current task; α t is the risk weight coefficient, ranging from [0, 1]; β t is the criticality weight coefficient, and its value range is [0, 1].

[0133] Preferably, the calculation of buffer time introduces risk weight coefficient and criticality weight coefficient, and quantifies and weights the impact of task risk coefficient and criticality coefficient, so as to break through the problem of overly simple or rigid buffer time setting in traditional methods, realize dynamic adaptive adjustment of buffer time, and significantly improve the flexibility and accuracy of time scheduling.

[0134] S2.3: Calculate the resource reservation ratio P of the task reserve :

[0135]

[0136] Among them, P min is the minimum resource reservation ratio; P max R is the maximum resource reservation ratio; task is the risk factor of the current task; K task is the criticality coefficient of the current task; R threshold is the upper limit of the risk factor threshold; K max is the upper limit of the criticality coefficient.

[0137] Preferably, the calculation of the resource reservation ratio adopts a nonlinear mapping scheme based on the risk coefficient and the criticality coefficient, and establishes a functional relationship between resource reservation and task characteristics. This solves the problem that resource reservation in the existing technology often adopts a fixed ratio or empirical value, realizes the intelligence and precision of resource reservation, and effectively improves resource utilization efficiency.

[0138] S2.4: Compare the risk factor with the advanced threshold R in the risk factor threshold matrix t31 and the intermediate threshold R t21 Compare the task criticality with the high-level threshold K in the task criticality matrix 31 and the intermediate threshold K 21 The comparison is carried out and the warning level is divided into four levels: zero, level 1, level 2 and level 3 according to the comparison results. When the risk factor exceeds R t31 And the mission criticality exceeds K 31 When the risk factor exceeds R t31 Or the mission criticality exceeds K 31 When the risk factor exceeds R t21 And the mission criticality exceeds K 21 It is judged as a level one warning when it occurs, and a level zero warning in other cases.

[0139] Preferably, S2.4 adopts a "double threshold and double indicator" judgment scheme, combining the high and low thresholds of risk coefficient and criticality to build a four-level early warning system to overcome the limitations of existing early warning mechanisms that often rely only on a single indicator or use simple threshold judgments, achieving more comprehensive and accurate risk warnings and significantly improving the system's risk prevention and control capabilities.

[0140] S2.5: Establish a corresponding relationship between the warning level and the warning response, where the third-level warning triggers three warning responses: warning notification, emergency resource pool activation and task re-planning; the second-level warning triggers two warning responses: warning notification and emergency resource pool activation; the first-level warning triggers one warning response: warning notification; and the zero-level warning does not trigger a warning response.

[0141] S2.6: Buffer time P buffer , Resource reservation ratio P reserve and warning response action warn To perform collaborative scheduling:

[0142] Schedule = f(T buffer , P reserve , Action warn );

[0143] Among them, f is the scheduling function, and the output scheduling matrix is:

[0144] Schedule = [Task id T adj R adj A exe ];

[0145] Among them, Task id T is the task identification number; adj The adjusted schedule; R adj is the adjusted resource allocation; A exe The type of alert response to execute.

[0146] It should be noted that the scheduling plan matrix includes four dimensions: task identification number, adjusted time schedule, adjusted resource allocation and executed warning response type. The scheduling plan matrix is ​​used to coordinate and dynamically adjust the time dimension, resource dimension and warning dimension of the task.

[0147] Preferably, collaborative scheduling uses a multi-dimensional scheduling matrix structure that includes task identification numbers, time arrangements, resource allocation, and warning response types, and achieves unified coordination of the time dimension, resource dimension, and warning dimension through scheduling functions. This design solves the problem that the adjustments of various dimensions in the prior art are often independent of each other and lack coordination, realizes the systematic and holistic nature of task adjustments, and significantly improves the scientificity and feasibility of the scheduling plan.

[0148] S3: Classify and process early warning tasks through incremental decision trees, automatically generate risk response plans, and feed back processing results to the expert rule base for rule updates.

[0149] S3.1: Construct the feature space of the incremental decision tree based on risk factor, mission criticality, buffer time, resource reservation ratio, warning level and warning response, and use information gain as the node splitting criterion to construct the incremental decision tree model;

[0150] S3.2: Set processing effect evaluation indicators including warning accuracy, response time and resource consumption, and prune the decision tree when the processing effect of the subtree is less than or equal to the product of the parent node processing effect and the pruning threshold coefficient, and update the decision tree tree subnode path that meets the rule update threshold to the expert rule base.

[0151] In summary, the present invention constructs a probability distribution model of task completion time by adopting an incremental learning model, and calculates the risk coefficient in combination with the FMEA method, so as to accurately predict the risk of task delay and resource conflict; dynamically adjusts the task buffer time and resource reservation strategy through the expert rule base, thereby realizing the adaptive optimization of the system; adopts an incremental decision tree to classify and process the warning tasks, and feeds back the processing results to the expert rule base, thus forming a closed-loop continuous optimization mechanism.

[0152] Embodiment 2 is an embodiment of the present invention, which provides an artificial intelligence-based workflow intelligent management system, including: an intelligent analysis module, which is used to analyze workflow historical data using an incremental learning model, construct a probability distribution model of task completion time, and calculate the risk coefficient of the task based on the probability distribution model; a dynamic adjustment module, which is used to dynamically adjust the buffer time and resource reservation strategy of the task according to the risk coefficient in combination with the expert rule library, and trigger the early warning mechanism when the risk coefficient exceeds the preset threshold; an intelligent response module, which is used to classify and process the early warning tasks through an incremental decision tree, automatically generate risk response plans, and feed back the processing results to the expert rule library for rule updating.

[0153] Example 3, reference Figure 2, is an embodiment of the present invention, which is different from the previous embodiment in that: if the function is implemented in the form of a software functional 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 solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0154] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0155] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0156] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0157] Example 4 is an embodiment of the present invention, which provides an intelligent workflow management method based on artificial intelligence. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0158] To verify the effectiveness of the present invention, the present invention selected 100 actual workflow tasks of a certain enterprise as test samples, and processed these tasks using the traditional workflow management method and the intelligent management method of the present invention respectively. In the experiment, the present invention focused on key indicators such as the task completion rate on schedule, resource utilization efficiency, and risk warning accuracy. The traditional method uses a fixed task plan and resource allocation strategy, while the method of the present invention dynamically adjusts task parameters and performs intelligent warning based on an incremental learning model. The experiment lasted for 3 months, and tests were carried out under normal load, high load, and emergency conditions.

[0159] Table 1: Comparative data of different management methods on various indicators

[0160] Performance Indicators Traditional methods The present invention Improvement Task completion rate 68.5％ 92.0％ 23.5％ Resource efficiency 72.3％ 91.0％ 18.7％ System response time (ms) 850 551 35.2％ Risk warning accuracy 60.8％ 92.3％ 31.5％ Timely rate of risk handling 65.2％ 89.8％ 24.6％ Resource conflict rate 25.3％ 8.7％ 65.6％

[0161] As shown in Table 1, it can be seen from the comparison data of six key performance indicators that the present invention has achieved significant performance improvement in all aspects compared with the traditional method. The most prominent is the improvement of resource conflict rate, which has dropped significantly from 25.3% of the traditional method to 8.7%, a decrease of 65.6%, which fully verifies the effectiveness of the resource conflict prediction mechanism based on the probability distribution model proposed by the present invention. The task completion rate and risk warning accuracy rate increased by 23.5% and 31.5% respectively, reaching a high level of 92.0% and 92.3%, indicating that the incremental learning model of the present invention can accurately grasp the task rules, timely warn and adjust the strategy. The system response time was reduced from 850ms to 551ms, an increase of 35.2%, indicating that the decision tree model of the present invention has good real-time performance. The resource utilization efficiency was increased to 91.0%, an increase of 18.7% over the traditional method, thanks to the dynamic resource reservation strategy proposed by the present invention. The 24.6% increase in the risk handling timeliness rate reflects the advantages of the present invention in risk response. Overall, these data strongly prove the comprehensive advantages of the present invention in all aspects of intelligent workflow management.

[0162] Table 2 System performance data under different operating conditions

[0163]

[0164]

[0165] As shown in Table 2, by comparing the system performance under three different operating states (normal load, high load, and emergency state), the adaptive ability and stability of the present invention are deeply demonstrated. Under normal load conditions, the performance improvement of the present invention is relatively mild compared with the traditional method, and the warning accuracy, resource utilization rate and task completion rate are increased by 18.3%, 17.7% and 19.5% respectively. However, as the system pressure increases, the advantages of the present invention become more and more obvious. Under high load conditions, the improvement of the three indicators has increased significantly, reaching 35.4%, 25.3% and 28.4% respectively. Especially in emergency conditions, the present invention shows excellent performance stability, with a warning accuracy of up to 92.1%, which is 40.7% higher than the traditional method, and the improvement of resource utilization rate and task completion rate has also reached 30.9% and 32.3% respectively. It is worth noting that the present invention maintains a relatively stable high performance under different pressure conditions (the warning accuracy is always maintained between 91% and 93%), while the performance of the traditional method decreases significantly with the increase of pressure (the warning accuracy drops from 75.2% to 51.4%). This comparison fully proves that the management method based on incremental learning and dynamic adjustment proposed in the present invention has excellent pressure adaptability and system stability, and is particularly suitable for handling complex and changeable workflow scenarios.

[0166] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention 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 invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent workflow management method based on artificial intelligence, characterized in that: include: An incremental learning model is used to analyze the workflow historical data, a probability distribution model of task completion time is constructed, and a risk coefficient of the task is calculated based on the probability distribution model; the risk coefficient includes the probability of task delay and the probability of resource conflict; According to the risk coefficient, the buffer time and resource reservation strategy of the task are dynamically adjusted in combination with the expert rule base, and the early warning mechanism is triggered when the risk coefficient exceeds the preset threshold; the early warning mechanism sets different early warning levels based on the criticality of the task; The early warning tasks are classified and processed through incremental decision trees, risk response plans are automatically generated, and the processing results are fed back to the expert rule base for rule updating.

2. The method for intelligent workflow management based on artificial intelligence according to claim 1, characterized in that: The process of calculating the risk factor includes the following steps: An incremental learning model is used to extract features from task start and end times, resource occupancy information, and task dependencies in workflow history data to obtain feature vectors. Use the feature extraction results to build a probability distribution model of task completion time; Calculating the task delay probability based on the probability distribution model; Calculate the probability of resource conflicts between tasks; Based on the delay probability and the resource conflict probability, the risk coefficient of the task is calculated by using the FMEA method.

3. The method for intelligent workflow management based on artificial intelligence according to claim 2, characterized in that: The probability distribution model is expressed by the following formula: Among them, μ t is the average completion time of historical tasks; is the variance of the completion time; P(t) is the output of the probability distribution model; The delay probability is expressed by the following formula: For the specified completion time T deadline , the probability of delay is: Through standardization transformation, we get: Where Φ(x) is the cumulative distribution function of the standard normal distribution; P d is the probability of postponement; The calculation of the resource conflict probability comprises the following steps: The resource-time occupancy set O of the construction task task : O task =(r m ,t s )|r m ∈R set ,t s ∈[T start ,T end ]; Among them, r m is the mth required resource; R set is the set of resources required for the task; s is the time point; T start is the task start time; T end Estimated end time for the task; Calculating resource conflict probability P based on resource-time intersection c : Among them, Q i is the resource-time occupancy set of task i; O j is the resource-time occupancy set of task i; i ∪O j | represents the intersection cardinality of the occupied sets; |O i | represents the cardinality of the occupied set; Γ is the set of task pairs that satisfy the following conditions: There is overlap in the execution time periods of task i and task j; There is an intersection between the resource requirement sets of task i and task j; The risk factor is expressed by the following formula: R=O risk ×S risk ×D risk ; Among them, O risk is the risk rate: O risk =max(P d , P c )×10; S risk S is the risk severity: S risk =-ln(1-P d )×5; D risk Risk detection difficulty: D risk =(1+P c ) 2 ×5.

4. The method for intelligent workflow management based on artificial intelligence according to claim 3, characterized in that: The expert rule base contains risk coefficient threshold matrices M for different task levels and different operating states. threshold and the task criticality matrix M critical ; The task levels include high-level tasks, medium-level tasks, and low-level tasks, and the operating states include normal state, bottleneck state, and emergency state.

5. The method for intelligent workflow management based on artificial intelligence according to claim 4, characterized in that: The risk coefficient threshold matrix M threshold The expression is: The task criticality matrix M critical The expression is: Among them, R tij is the risk threshold of the i-th level task in the j-th state; K ij is the criticality coefficient of the i-th level task in the j-th state; i∈[1,3], represents the task level, i=1: low-level task, i=2: medium-level task, i=3: high-level task; j∈[1,3], represents the running state, j=1: normal state, j=2: bottleneck state, j=3: emergency state; R t21 is the risk threshold of the intermediate task under normal conditions, and is used as the intermediate threshold for the first-level warning judgment; K t31 is the risk threshold of the high-level task under normal conditions, and is used as the high-level threshold for the second-level and third-level warning judgments; K 21 is the criticality coefficient of the intermediate task under normal conditions, and is used as the intermediate threshold for the first-level warning judgment; K 31 It is the criticality coefficient of the high-level task under normal conditions, and is used as the high-level threshold for second-level and third-level warning judgments.

6. The method for intelligent workflow management based on artificial intelligence according to claim 5, characterized in that: According to the risk coefficient, the buffer time and resource reservation strategy of the task are dynamically adjusted in combination with the expert rule base, and when the risk coefficient exceeds the preset threshold, the early warning mechanism is triggered, including the following steps: Establish an expert rule base; Calculate the buffer time R of the task buffer : T buffer =T base ×(1+a t R task +b t K task ); Among them, T base is the benchmark buffer time; R task is the risk factor of the current task; K task is the criticality coefficient of the current task; α t is the risk weight coefficient; β t is the criticality weight coefficient; Compute the resource reservation ratio P of the task reserve : Among them, P min is the minimum resource reservation ratio; P max R is the maximum resource reservation ratio; task is the risk factor of the current task; K task is the criticality coefficient of the current task; R threshold is the upper limit of the risk factor threshold; K max is the upper limit of the criticality coefficient; The risk factor and the risk factor threshold matrix M threshold Advanced Thresholding in R t31 and the intermediate threshold R t21 Compare the task criticality with the task criticality matrix M critical The advanced threshold K in 31 and the intermediate threshold K 21 Make a comparison; According to the comparison results, the warning level is divided into zero-level warning, first-level warning, second-level warning, and third-level warning; when the risk coefficient exceeds R t31 , and the mission criticality exceeds K 31 When the risk factor exceeds R t31 Or the mission criticality exceeds K 31 When the risk factor exceeds R t21 , and the mission criticality exceeds K 21 When the warning is detected, it is considered as a Level 1 warning, and in other cases, it is considered as a Level 0 warning. Establish a corresponding relationship between the warning level and the warning response; wherein, the third-level warning triggers three warning responses, namely, warning notification, emergency resource pool activation and task re-planning; the second-level warning triggers two warning responses, namely, warning notification and emergency resource pool activation; the first-level warning triggers one warning response, namely, warning notification; and the zero-level warning does not trigger any warning response; The buffer time T buffer , the resource reservation ratio P reserve And the warning response Action warn Perform collaborative scheduling and output the scheduling plan matrix.

7. The method for intelligent workflow management based on artificial intelligence according to claim 6, characterized in that: The early warning tasks are classified and processed through the incremental decision tree, risk response plans are automatically generated, and the processing results are fed back to the expert rule base for rule updating, including the following steps: The feature space of the incremental decision tree is constructed based on risk coefficient, mission criticality, buffer time, resource reservation ratio, warning level and warning response, and the incremental decision tree model is constructed using information gain as the node splitting criterion; The processing effect evaluation index including warning accuracy, response time and resource consumption is set. When the processing effect of the subtree is less than or equal to the product of the processing effect of the parent node and the pruning threshold coefficient, the decision tree is pruned and optimized, and the decision tree tree subnode path that meets the rule update threshold is updated to the expert rule base.

8. A system using the artificial intelligence-based workflow intelligent management method according to any one of claims 1 to 7, characterized in that: include: An intelligent analysis module is used to analyze the workflow historical data using an incremental learning model, construct a probability distribution model for task completion time, and calculate the risk coefficient of the task based on the probability distribution model; A dynamic adjustment module, used to dynamically adjust the buffer time and resource reservation strategy of the task according to the risk coefficient in combination with the expert rule base, and trigger an early warning mechanism when the risk coefficient exceeds a preset threshold; The intelligent response module is used to classify and process early warning tasks through an incremental decision tree, automatically generate risk response plans, and feed back the processing results to the expert rule base for rule updating.

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 steps of the intelligent workflow management method based on artificial intelligence described in any one of claims 1 to 7 are implemented.

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 the intelligent workflow management method based on artificial intelligence described in any one of claims 1 to 7 are implemented.

Images