Manufacturing industry advanced plan scheduling system and method based on artificial intelligence

Abstract

The invention belongs to the technical field of scheduling optimization, and discloses a manufacturing industry advanced plan scheduling system and method based on artificial intelligence, and the method comprises the steps: generating an initial scheduling scheme according to production order data and manufacturing resource constraint data; analyzing the initial scheduling scheme to establish a basic time sequence graph, and performing man-hour probability distribution fitting on each task node based on historical production data to obtain a probability time sequence graph; and traversing each task node, calculating and correcting the material ready time based on the man-hour probability distribution function of the previous task node and a preset on-time completion confidence coefficient, updating the plan time of each task node, and obtaining a final production schedule. Adjacent processes form a parallel overlapping relation by starting a time delay mechanism, so that the production cycle is effectively shortened; and meanwhile, the planned time is dynamically adjusted based on the man-hour probability distribution, the time margin can be reserved according to the difference of the actual fluctuation characteristics of each process, and the robustness of the scheduling scheme is improved.

Description

Technical Field

[0001] This invention relates to the field of scheduling optimization technology, and more specifically, to an advanced planning and scheduling system and method for the manufacturing industry based on artificial intelligence. Background Technology

[0002] As the manufacturing industry moves towards intelligence and flexibility, production scheduling, as a core link connecting production planning and shop floor execution, is becoming increasingly important. Advanced planning and scheduling systems, by comprehensively considering multi-dimensional constraints such as order demand, material supply, and equipment capacity, generate executable production operation plans, and have become a key technical means for modern manufacturing enterprises to improve production efficiency and shorten delivery cycles.

[0003] Traditional production scheduling methods typically employ a strict sequential waiting model when dealing with the timing relationships between processes. That is, the next process can only begin production after the previous process has completed all its work and produced all its materials. While this method is logically simple and easy to implement, it suffers from significant efficiency losses in actual production. Taking the machining industry as an example, when the output rate of an upstream process is lower than the consumption rate of a downstream process, the downstream process can start earlier after the upstream process has produced some materials, creating a parallel overlapping relationship between processes and effectively compressing the overall production cycle. However, existing scheduling methods rarely consider this parallel overlapping mechanism based on the matching relationship between material output and consumption rates, resulting in low time utilization of the scheduling schemes.

[0004] Furthermore, various uncertainties are prevalent in the production process, including fluctuations in equipment condition, differences in personnel operation, and changes in material quality. These factors often cause deviations between the actual processing time and the planned time for each step. To address these fluctuations, existing scheduling methods typically employ a fixed buffer time strategy, which involves adding a pre-set time margin between adjacent steps. For example, a fixed 30-minute buffer time is added after each step. However, this fixed buffer strategy has significant limitations: on the one hand, the fixed buffer time may be insufficient to absorb actual fluctuations, leading to delays in subsequent steps due to material supply interruptions; on the other hand, the fixed buffer time may result in unnecessary time waste, reducing equipment utilization and overall capacity. Summary of the Invention

[0005] To overcome the aforementioned problems of the prior art, this invention proposes an advanced planning and scheduling system and method for the manufacturing industry based on artificial intelligence, which is used to solve the above problems.

[0006] This invention provides the following technical solution: Advanced planning and scheduling methods for the manufacturing industry based on artificial intelligence include: An initial scheduling plan is generated based on production order data and manufacturing resource constraint data; Analyze the initial scheduling scheme and establish a basic time series graph connecting the task sequence and resource allocation results; Based on historical production data, the probability distribution function of working hours for each task node in the basic time series graph is calculated to obtain the probability time series graph. Traverse each task node in the probability time sequence graph in order, and calculate the corrected material ready time of the task node based on the time probability distribution function of each preceding task node. The planned time for the current task node is updated based on the corrected material readiness time. After updating the planned time for all task nodes in sequence, the final production schedule is obtained.

[0007] Preferably, the step of generating an initial scheduling plan based on production order data and manufacturing resource constraint data includes: Starting from the product corresponding to the production order, the entire process task and its material dependencies required to produce the product are identified in reverse recursion, and a process dependency chain is generated. Based on the process dependency chain, manufacturing resources are allocated to each process task in sequence, and the planned time is determined to obtain the initial scheduling plan.

[0008] Preferably, the process-dependent chain generation step includes: Starting with the final process task corresponding to the product, add it to the queue to be expanded; Retrieve the current process task from the queue to be expanded and check the existing inventory of the required input materials; If the existing inventory meets the demand, the input material is marked as inventory supply, and the current process task is completed. If the existing inventory does not meet the demand, identify the source process task that produces the input material, establish the material dependency relationship between the source process task and the current process task, and add the source process task to the queue to be expanded. Repeat the above steps until the queue to be expanded is empty, and you will get a process dependency chain containing all process tasks and their material dependencies.

[0009] Preferably, the step of allocating manufacturing resources and determining planned times for each process task according to the process dependency chain to obtain an initial scheduling scheme includes: The task is traversed sequentially through the process dependency chain. The planned time of the previous process task is obtained based on the material dependency relationship. The start-up delay is calculated by combining the difference between the material output rate of the previous process and the material consumption rate of the current process, and the earliest allowable start time of the material is determined. Search for the earliest idle time of the equipment on the manufacturing resources specified in the current process task, take the later of the earliest allowable start time of the material and the earliest idle time of the equipment as the planned start time, determine the planned end time based on the output demand and the material output rate, and obtain the initial scheduling plan after traversal.

[0010] Preferably, the steps for establishing the basic time series map include: Generate task nodes for each process task in the initial scheduling plan, and generate resource nodes for each manufacturing resource; Based on the material dependency relationship, establish a material rate coupling edge between the source process task node and the current process task node, and record the material output rate, material consumption rate and start-up delay on the material rate coupling edge; Based on the resource allocation results, establish a time-series occupancy edge between the task node and the corresponding resource node, and record the planned start time and planned end time on the time-series occupancy edge.

[0011] Preferably, the step of calculating the probability distribution function of working hours for each task node based on historical production data includes: Determine the process category and associated manufacturing resource object of each task node, retrieve the historical completed records of the process category on the manufacturing resource object in the historical production database, extract the actual processing time and actual output quantity of each record, calculate the actual processing time per unit material, and construct a sample set of unit processing time. The probability distribution of the sample set of unit time consumption is fitted to generate the probability distribution function of unit material time consumption; the probability distribution function of unit material time consumption is scaled according to the output demand of the current task node to generate the time probability distribution function of the task node.

[0012] Preferably, the calculation step for the corrected material readiness time includes: Get all preceding task nodes of the current task node; For each preceding task node, based on its time probability distribution function and the preset on-time completion confidence level, Based on the planned start time of the preceding task node, calculate the confidence material ready time of the current task node relative to the preceding task node; Compare the ready times of all confidence materials and take the latest one as the corrected ready time of the current task node.

[0013] Preferably, the calculation step for the confidence material readiness time includes: Obtain the time probability distribution function of the preceding task node; take the preset on-time completion confidence level as the target probability value, perform the inverse operation on the cumulative distribution function corresponding to the time probability distribution function, solve for the time value that makes the cumulative probability equal to the target probability value, and obtain the confidence completion time of the preceding task node; Calculate the startup delay by replacing the planned time of the preceding task node in the initial scheduling scheme with the confidence completion time; Add the start delay to the planned start time of the preceding task node to obtain the confidence material ready time of the current task node relative to the preceding task node.

[0014] Preferably, the step of updating the planned time of the current task node based on the corrected material readiness time includes: The revised material readiness time is used as the updated planned start time for the current task node; the confidence completion time for the current task node is calculated based on the time probability distribution function of the current task node and the preset on-time completion confidence level; the updated planned end time is obtained by adding the confidence completion time to the updated planned start time.

[0015] This invention also provides an AI-based advanced planning and scheduling system for the manufacturing industry, used to implement an AI-based advanced planning and scheduling method for the manufacturing industry, including: The initial scheduling generation module is used to generate an initial scheduling plan based on production order data and manufacturing resource constraint data; The time sequence graph construction module is used to parse the initial scheduling scheme and establish a basic time sequence graph connecting the task sequence and the resource allocation results; The probability distribution fitting module is used to calculate the time probability distribution function of each task node in the basic time series map based on historical production data, and obtain the probability time series map. The readiness time correction module is used to sequentially traverse each task node in the probability time series graph and calculate the corrected material readiness time of the task node based on the time probability distribution function of each preceding task node. The planned time update module is used to update the planned time of the current task node based on the corrected material readiness time. After updating the planned time of all task nodes in sequence, the final production schedule is obtained.

[0016] This invention provides an advanced planning and scheduling system and method for the manufacturing industry based on artificial intelligence, which has the following beneficial effects: By identifying process tasks and establishing material dependencies through reverse recursion, and calculating start-up delays based on the matching relationship between material output and consumption rates, adjacent processes with material supply and demand relationships can form a parallel overlapping relationship while ensuring material supply, rather than the traditional strict serial waiting mode. This parallel mechanism based on material flow rate effectively compresses the overall production cycle, improves equipment utilization and time efficiency, and is particularly suitable for production scenarios where the output rate of upstream processes differs from the consumption rate of downstream processes.

[0017] By constructing a probability distribution function of work hours for each process based on historical production data, and combining this with a preset on-time completion confidence level to calculate the confidence completion time and correct material readiness time, dynamic adjustment of process planned times is achieved. Compared to traditional scheduling methods based on ideal work hours, this method can adjust planned times differently according to the actual work hour fluctuation characteristics of each process: automatically reserving a longer time margin for processes with large work hour fluctuations, and allocating a shorter time margin for processes with small work hour fluctuations. This dynamic adjustment mechanism based on probability distribution can effectively absorb the impact of work hour fluctuations under a preset confidence level, reducing the risk of overall scheduling failure due to delays in individual processes, while avoiding the time waste caused by a fixed time margin strategy, thus achieving a balanced optimization of scheduling robustness and production efficiency. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the advanced planning and scheduling method for the manufacturing industry based on artificial intelligence, as described in this invention. Figure 2 This is a schematic diagram of the modules of the AI-based advanced planning and scheduling system for the manufacturing industry according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example

[0020] Please see Figure 1 In this embodiment, the advanced planning and scheduling method for the manufacturing industry based on artificial intelligence includes: S1. Generate an initial scheduling plan based on production order data and manufacturing resource constraint data; The steps for generating an initial scheduling plan based on production order data and manufacturing resource constraint data include: Starting from the product corresponding to the production order, the entire process task and its material dependencies required to produce the product are identified in reverse recursion, and a process dependency chain is generated. Based on the process dependency chain, manufacturing resources are allocated to each process task in sequence, and the planned time is determined to obtain the initial scheduling plan.

[0021] The steps for generating the process-dependent chain include: Starting with the final process task corresponding to the product, add it to the queue to be expanded; Retrieve the current process task from the queue to be expanded and check the existing inventory of the required input materials; If the existing inventory meets the demand, the input material is marked as inventory supply, and the current process task is completed. If the existing inventory does not meet the demand, identify the source process task that produces the input material, establish the material dependency relationship between the source process task and the current process task, and add the source process task to the queue to be expanded. Repeat the above steps until the queue to be expanded is empty, and you will get a process dependency chain containing all process tasks and their material dependencies.

[0022] The steps of allocating manufacturing resources and determining planned times for each process task according to the process dependency chain to obtain the initial scheduling scheme include: The task is traversed sequentially through the process dependency chain. The planned time of the previous process task is obtained based on the material dependency relationship. The start-up delay is calculated by combining the difference between the material output rate of the previous process and the material consumption rate of the current process, and the earliest allowable start time of the material is determined. Search for the earliest idle time of the equipment on the manufacturing resources specified in the current process task, take the later of the earliest allowable start time of the material and the earliest idle time of the equipment as the planned start time, determine the planned end time based on the output demand and the material output rate, and obtain the initial scheduling plan after traversal.

[0023] In this embodiment, it should be noted that production order data typically comes from order information in the enterprise's production management system, including key elements such as order number, product type, and required quantity. Manufacturing resource constraint data refers to various resources involved in the production process and their limitations, including but not limited to equipment processing capacity and equipment availability time windows. The initial scheduling plan is a basic production plan generated under ideal conditions, providing a benchmark for subsequent robust optimization.

[0024] It should be noted that the reverse recursion starts from the final product and only expands the actually required process tasks. During the expansion process, whenever a new source process task is identified, the system simultaneously establishes the material dependency relationship between that source process and the current process. Since a process may require multiple input materials, and each material may come from different source processes, the final process dependency chain usually presents as a directed acyclic graph with branching and converging structures. The expansion process continues until all material requirements are met, including conditions such as sufficient existing inventory or tracing back to purchased raw materials not produced by the plant.

[0025] It's important to note that start-up lag refers to the time delay required for the current process to wait after the preceding process has started production. If the output rate of the preceding process is less than the consumption rate of the current process, and both start simultaneously, the current process will be interrupted due to insufficient material supply. Therefore, it's necessary for the preceding process to produce for a period to accumulate materials, ensuring that the accumulated amount from the preceding process, plus the continuous output from the preceding process during the current process's operation, can meet the total consumption requirements of the current process. If the materials required by the current process exist in existing inventory, this inventory can be used as part of the initial accumulation, correspondingly shortening the start-up lag. This mechanism allows for parallel and overlapping relationships between processes, effectively compressing the overall production cycle.

[0026] It should be noted that the earliest permissible start time for materials refers to the earliest moment when the current process can obtain a stable supply of materials. It is calculated by adding the start-up delay to the planned start time of the preceding process. In cases with multiple preceding processes, the earliest permissible start time for materials relative to each preceding process needs to be calculated separately, and then the latest time is taken to ensure a stable supply of all input materials. The planned start time of the current process is the later of the earliest permissible start time for materials and the earliest idle time of the equipment, ensuring that there is both sufficient material supply and available equipment resources when the process starts.

[0027] S2. Analyze the initial scheduling scheme and establish a basic time-series graph connecting the task sequence and resource allocation results; The steps for establishing the basic time series map include: Generate task nodes for each process task in the initial scheduling plan, and generate resource nodes for each manufacturing resource; Based on the material dependency relationship, establish a material rate coupling edge between the source process task node and the current process task node, and record the material output rate, material consumption rate and start-up delay on the material rate coupling edge; Based on the resource allocation results, establish a time-series occupancy edge between the task node and the corresponding resource node, and record the planned start time and planned end time on the time-series occupancy edge.

[0028] In this embodiment, it should be noted that the basic time series graph is a graph-structured data representation used to structurally express the task sequence, resource allocation, and material flow relationship in the initial scheduling scheme in the form of nodes and edges, which facilitates subsequent probability analysis and optimization calculations.

[0029] It should be noted that task nodes correspond to each process task in the initial scheduling plan, and node attributes typically include information such as process type, output requirement, and planned working hours. Resource nodes correspond to each manufacturing resource in the production process, such as processing equipment and production lines, and node attributes typically include information such as resource type and processing capacity.

[0030] It should be noted that the material rate coupling edge is used to characterize the association between process tasks with material supply and demand relationships. This edge connects the source process task node and the current process task node, and the material output rate, material consumption rate, and start-up delay recorded on the edge reflect the material flow characteristics and temporal coupling relationship between the two processes. Through these attributes, the degree of parallel overlap between processes can be characterized.

[0031] It should be noted that the time-series occupancy edge is used to represent the occupancy relationship between task nodes and resource nodes. This edge connects a task node with its assigned resource node, and the planned start time and planned end time recorded on the edge represent the time occupancy interval of the task on the corresponding resource. Through the time-series occupancy edge, the task scheduling status and idle time window of each resource can be intuitively queried.

[0032] S3. Calculate the working time probability distribution function of each task node in the basic time series graph based on historical production data to obtain the probability time series graph; The steps for calculating the probability distribution function of working hours for each task node based on historical production data include: Determine the process category and associated manufacturing resource object of each task node, retrieve the historical completed records of the process category on the manufacturing resource object in the historical production database, extract the actual processing time and actual output quantity of each record, calculate the actual processing time per unit material, and construct a sample set of unit processing time. The probability distribution of the sample set of unit time consumption is fitted to generate the probability distribution function of unit material time consumption; the probability distribution function of unit material time consumption is scaled according to the output demand of the current task node to generate the time probability distribution function of the task node.

[0033] In this embodiment, it should be noted that the time probability distribution function is used to describe the uncertainty of the actual processing time of a process task. Due to factors such as equipment status fluctuations, differences in personnel operation, and changes in material quality during the production process, the actual processing time of the same process often varies in different batches. Quantifying this uncertainty using the probability distribution function provides a mathematical basis for subsequent robustness optimization.

[0034] It should be noted that historical completed records refer to the actual execution data of past production tasks of the same process category on the same manufacturing resource. The combination of the same process category and the same manufacturing resource is used for retrieval because different equipment may have different processing efficiencies and fluctuation characteristics even when performing the same process. The actual processing time per unit of material is calculated by dividing the actual processing time by the actual output quantity, normalizing production records of different batches into comparable unit processing time data.

[0035] It should be noted that probability distribution fitting is the process of determining the type of probability distribution and its corresponding parameters based on the sample data. Common distribution types include normal distribution, log-normal distribution, and Weibull distribution, etc., and the most suitable distribution type can be selected based on the statistical characteristics of the sample data. The probability distribution function of the unit material consumption time obtained after fitting can describe the probabilistic characteristics of the time required to produce a unit of material.

[0036] It should be noted that scaling refers to amplifying the probability distribution function of unit material consumption time based on the output demand of the current task node. For example, if the unit material consumption time follows a certain distribution, and the current task requires the production of 100 units of material, then the scaling parameter of this distribution is multiplied by 100 to obtain the probability distribution function of the total working time for producing 100 units of material. The transformed working time probability distribution function reflects the probabilistic characteristics of the time required to complete all output demands of the current task node.

[0037] S4. Traverse each task node in the probability time sequence graph in order, and calculate the corrected material ready time of the task node based on the working time probability distribution function of each preceding task node. The calculation steps for the corrected material readiness time include: Get all preceding task nodes of the current task node; For each preceding task node, based on its time probability distribution function and the preset on-time completion confidence level, Based on the planned start time of the preceding task node, calculate the confidence material ready time of the current task node relative to the preceding task node; Compare the ready times of all confidence materials and take the latest one as the corrected ready time of the current task node.

[0038] The calculation steps for the confidence material readiness time include: Obtain the time probability distribution function of the preceding task node; take the preset on-time completion confidence level as the target probability value, perform the inverse operation on the cumulative distribution function corresponding to the time probability distribution function, solve for the time value that makes the cumulative probability equal to the target probability value, and obtain the confidence completion time of the preceding task node; Calculate the startup delay by replacing the planned time of the preceding task node in the initial scheduling scheme with the confidence completion time; Add the start delay to the planned start time of the preceding task node to obtain the confidence material ready time of the current task node relative to the preceding task node.

[0039] In this embodiment, it should be noted that the corrected material readiness time is a correction to the earliest allowed start time of materials in the initial scheduling scheme. The initial scheduling scheme is calculated based on ideal working hours and does not consider the uncertainty of working hours. The corrected material readiness time, on the other hand, is based on the working hour probability distribution function and calculates the time point at which materials can be stably supplied under a preset confidence level, thereby improving the scheduling scheme's tolerance to working hour fluctuations.

[0040] It should be noted that the preset on-time completion confidence level refers to the expected probability of on-time completion, which is usually set according to production management needs. For example, setting it to 95% means that there is a 95% probability of on-time completion. The higher the confidence level, the stricter the tolerance requirement for time fluctuations, and the longer the calculated confidence completion time.

[0041] It's important to note that the cumulative distribution function is the integral form of the probability distribution function, representing the probability that a random variable is less than or equal to a certain value. Performing the inverse operation on the cumulative distribution function, i.e., given a target probability value, calculates the corresponding work time. For example, if the preset confidence level for on-time completion is 95%, the inverse operation calculates the work time value that makes the probability of "the actual work time not exceeding this length" equal to 95%. This work time value is the confidence completion time. The confidence completion time represents the processing time required to complete the process under the preset confidence level.

[0042] It should be noted that the confidence completion time is usually longer than the planned time in the initial scheduling plan. This is because the planned time is generally set based on average or ideal time, while the confidence completion time needs to cover a certain proportion of time fluctuations. Recalculating the start-up lag with the confidence completion time instead of the planned time is equivalent to taking into account possible delays in previous processes, thus making the material readiness time of the current process more reliable.

[0043] It should be noted that when there are multiple preceding task nodes, the confidence material readiness time of the current task node relative to each preceding task node needs to be calculated separately. Since the current task node must wait for all input materials to be stably supplied before it can start, the latest of all confidence material readiness times is taken as the corrected material readiness time to ensure that the material supply of all preceding task nodes can meet the confidence requirements at that time.

[0044] S5. Update the planned time of the current task node according to the corrected material readiness time. After updating the planned time of all task nodes in sequence, the final production schedule is obtained.

[0045] The step of updating the planned time of the current task node based on the corrected material readiness time includes: The revised material readiness time is used as the updated planned start time for the current task node; the confidence completion time for the current task node is calculated based on the time probability distribution function of the current task node and the preset on-time completion confidence level; the updated planned end time is obtained by adding the confidence completion time to the updated planned start time.

[0046] In this embodiment, it should be noted that the process of updating the planned time involves directly using the corrected material readiness time as the planned start time of the current task node, while using the confidence completion time to replace the planned time in the initial schedule to calculate the planned end time. In this way, both the planned start and end times are corrected based on confidence levels, ensuring consistency in the time arrangement. The planned times of each task node are updated sequentially, allowing subsequent task nodes to reference the updated planned times of previous task nodes during calculation, thus achieving step-by-step propagation along the process dependency chain. This final production schedule, based on the probability distribution of work hours and a pre-set confidence level, sets reasonable production times for each process. It can absorb the impact of work hour fluctuations under the pre-set confidence level, reducing the risk of overall schedule failure due to delays in individual processes and improving the robustness of the production plan. Example

[0047] Please see Figure 2 This invention provides an advanced planning and scheduling system for the manufacturing industry based on artificial intelligence, used to implement an advanced planning and scheduling method for the manufacturing industry based on artificial intelligence, including: The initial scheduling generation module is used to generate an initial scheduling plan based on production order data and manufacturing resource constraint data; The time sequence graph construction module is used to parse the initial scheduling scheme and establish a basic time sequence graph connecting the task sequence and the resource allocation results; The probability distribution fitting module is used to calculate the time probability distribution function of each task node in the basic time series map based on historical production data, and obtain the probability time series map. The readiness time correction module is used to sequentially traverse each task node in the probability time series graph and calculate the corrected material readiness time of the task node based on the time probability distribution function of each preceding task node. The planned time update module is used to update the planned time of the current task node based on the corrected material readiness time. After updating the planned time of all task nodes in sequence, the final production schedule is obtained.

[0048] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only one method, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0049] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

[0050] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An advanced planning and scheduling method for the manufacturing industry based on artificial intelligence, characterized in that, include: An initial scheduling plan is generated based on production order data and manufacturing resource constraint data; Analyze the initial scheduling scheme and establish a basic time series graph connecting the task sequence and resource allocation results; Based on historical production data, the probability distribution function of working hours for each task node in the basic time series graph is calculated to obtain the probability time series graph. Traverse each task node in the probability time sequence graph in order, and calculate the corrected material ready time of the task node based on the time probability distribution function of each preceding task node. The planned time for the current task node is updated based on the corrected material readiness time. After updating the planned time for all task nodes in sequence, the final production schedule is obtained.

2. The advanced planning and scheduling method for the manufacturing industry based on artificial intelligence according to claim 1, characterized in that, The steps for generating an initial scheduling plan based on production order data and manufacturing resource constraint data include: Starting from the product corresponding to the production order, the entire process task and its material dependencies required to produce the product are identified in reverse recursion, and a process dependency chain is generated. Based on the process dependency chain, manufacturing resources are allocated to each process task in sequence, and the planned time is determined to obtain the initial scheduling plan.

3. The advanced planning and scheduling method for the manufacturing industry based on artificial intelligence according to claim 2, characterized in that, The steps for generating the process-dependent chain include: Starting with the final process task corresponding to the product, add it to the queue to be expanded; Retrieve the current process task from the queue to be expanded and check the existing inventory of the required input materials; If the existing inventory meets the demand, the input material is marked as inventory supply, and the current process task is completed. If the existing inventory does not meet the demand, identify the source process task that produces the input material, establish the material dependency relationship between the source process task and the current process task, and add the source process task to the queue to be expanded. Repeat the above steps until the queue to be expanded is empty, and you will get a process dependency chain containing all process tasks and their material dependencies.

4. The advanced planning and scheduling method for the manufacturing industry based on artificial intelligence according to claim 3, characterized in that, The steps of allocating manufacturing resources and determining planned times for each process task according to the process dependency chain to obtain the initial scheduling scheme include: The task is traversed sequentially through the process dependency chain. The planned time of the previous process task is obtained based on the material dependency relationship. The start-up delay is calculated by combining the difference between the material output rate of the previous process and the material consumption rate of the current process, and the earliest allowable start time of the material is determined. Search for the earliest idle time of the equipment on the manufacturing resources specified in the current process task, take the later of the earliest allowed start time of the material and the earliest idle time of the equipment as the planned start time, determine the planned end time based on the output demand and the output speed of the material, and obtain the initial scheduling plan after the traversal is completed.

5. The advanced planning and scheduling method for the manufacturing industry based on artificial intelligence according to claim 4, characterized in that, The steps for establishing the basic time series map include: Generate task nodes for each process task in the initial scheduling plan, and generate resource nodes for each manufacturing resource; Based on the material dependency relationship, a material rate coupling edge is established between the source process task node and the current process task node. The material output rate, material consumption rate and start-up delay are recorded on the material rate coupling edge. Based on the resource allocation results, establish a time-series occupancy edge between the task node and the corresponding resource node, and record the planned start time and planned end time on the time-series occupancy edge.

6. The advanced planning and scheduling method for the manufacturing industry based on artificial intelligence according to claim 5, characterized in that, The steps for calculating the probability distribution function of working hours for each task node based on historical production data include: Determine the process category and associated manufacturing resource object of each task node, retrieve the historical completed records of the process category on the manufacturing resource object in the historical production database, extract the actual processing time and actual output quantity of each record, calculate the actual processing time per unit material, and construct a sample set of unit processing time. The probability distribution of the sample set of unit time consumption is fitted to generate the probability distribution function of unit material time consumption; the probability distribution function of unit material time consumption is scaled according to the output demand of the current task node to generate the time probability distribution function of the task node.

7. The advanced planning and scheduling method for the manufacturing industry based on artificial intelligence according to claim 6, characterized in that, The calculation steps for the corrected material readiness time include: Get all preceding task nodes of the current task node; For each preceding task node, based on its time probability distribution function and the preset on-time completion confidence level, Based on the planned start time of the preceding task node, calculate the confidence material ready time of the current task node relative to the preceding task node; Compare the ready times of all confidence materials and take the latest one as the corrected ready time of the current task node.

8. The advanced planning and scheduling method for the manufacturing industry based on artificial intelligence according to claim 7, characterized in that, The calculation steps for the confidence material readiness time include: Obtain the time probability distribution function of the preceding task node; take the preset on-time completion confidence level as the target probability value, perform the inverse operation on the cumulative distribution function corresponding to the time probability distribution function, solve for the time value that makes the cumulative probability equal to the target probability value, and obtain the confidence completion time of the preceding task node; Calculate the startup delay by replacing the planned time of the preceding task node in the initial scheduling scheme with the confidence completion time; Add the start delay to the planned start time of the preceding task node to obtain the confidence material ready time of the current task node relative to the preceding task node.

9. The advanced planning and scheduling method for the manufacturing industry based on artificial intelligence according to claim 8, characterized in that, The step of updating the planned time of the current task node based on the corrected material readiness time includes: The revised material readiness time is used as the updated planned start time for the current task node; the confidence completion time for the current task node is calculated based on the time probability distribution function of the current task node and the preset on-time completion confidence level; the updated planned end time is obtained by adding the confidence completion time to the updated planned start time.

10. An AI-based advanced planning and scheduling system for the manufacturing industry, used to implement the AI-based advanced planning and scheduling method for the manufacturing industry as described in any one of claims 1-9, characterized in that, include: The initial scheduling generation module is used to generate an initial scheduling plan based on production order data and manufacturing resource constraint data; The time sequence graph construction module is used to parse the initial scheduling scheme and establish a basic time sequence graph connecting the task sequence and the resource allocation results; The probability distribution fitting module is used to calculate the time probability distribution function of each task node in the basic time series map based on historical production data, and obtain the probability time series map. The readiness time correction module is used to sequentially traverse each task node in the probability time series graph and calculate the corrected material readiness time of the task node based on the time probability distribution function of each preceding task node. The planned time update module is used to update the planned time of the current task node based on the corrected material readiness time. After updating the planned time of all task nodes in sequence, the final production schedule is obtained.

Images