Scheduling optimization method and system for production project management

Abstract

The invention discloses a scheduling optimization method and system for production project management, and relates to the technical field of production scheduling. A scheduling optimization system for production project management comprises a production project management module and a production project scheduling module. According to the invention, through the construction mode of the main project triad and the sub project triad, the complex production project information is standardized and semantized, and the analytic property of the scheduling model to the business logic is effectively enhanced; through scheduling perception coding and clustering processing of sub-project feature vectors, similar production unit tasks which can be merged can be automatically identified, the resource utilization rate is improved, and the redundancy scheduling cost is reduced; historical progress fusion and idle time window information are introduced, a scheduling constraint graph is constructed in combination with a graph attention network model, various resource, personnel and equipment state restrictions can be comprehensively considered, and a globally optimal solution of a task scheduling path is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of production scheduling, and particularly relates to a scheduling optimization method and system for production project management. Background Art

[0002] In modern manufacturing enterprises, production projects usually consist of multiple main projects and several sub-projects. Their scheduling process involves complex resource coordination, task dependencies, and real-time change management. Traditional scheduling methods often struggle to achieve fine-grained scheduling and global optimal allocation in the face of a highly concurrent, multi-constrained, and dynamically changing production environment, resulting in low production efficiency, serious resource waste, and difficulty in adapting to the conflict problems of shared resources among multiple projects. Especially in the scenario of cross-operation of multiple sub-projects, existing methods lack accurate modeling of the topological relationship and scheduling priority between tasks, making it difficult to dynamically optimize the scheduling path and resource allocation.

[0003] Therefore, there is a need for a scheduling optimization method that realizes scheduling optimization at the sub-project level and global rescheduling optimization at the main project level, effectively improving the utilization rate of production resources and scheduling flexibility. Summary of the Invention

[0004] The present invention aims to provide a scheduling optimization method and system for production project management, effectively improving the utilization rate of production resources and scheduling flexibility.

[0005] A scheduling optimization method for production project management includes the following steps: Obtain the production operation project list of the current day; Based on the production operation project list, obtain the project information of several production projects; extract features from the project information to obtain the main project triple set and the sub-project triple set; among them, the main project triple includes the project number, production task, and initial project priority; the sub-project triple includes the project number, production task sub-project, and sub-project production information. Extract production sub-projects from all the main project triple sets and sub-project triple sets to obtain several production sub-project queues; each production sub-project queue represents several production project scheduling units corresponding to a single sub-project; based on any production sub-project queue, obtain the corresponding unfinished production sub-project queue. Use the scheduling constraint model to perform single sub-project scheduling on the production sub-project queue and the unfinished production sub-project queue to obtain an optimized production sub-project queue; re-schedule several optimized production sub-project queues based on the main project triple set to obtain the main project scheduling queue. Complete the optimized scheduling of the production operation project list of the current day based on the main project scheduling queue.

[0006] As a preferred technical solution of the present invention, the specific steps for extracting production sub-items from all main-item triple sets and sub-item triple sets include: Generating sub-item feature vectors based on the sub-item triples in the sub-item triple set; Generating main-item guiding factors with item numbers as identifiers based on the main-item triples in the main-item triple set; Performing scheduling awareness encoding based on the sub-item feature vectors to obtain sub-item scheduling awareness encoding; Clustering based on all sub-item scheduling awareness encodings using a clustering model to obtain several sub-item groups; Generating several production sub-item queues based on several sub-item groups. Each production sub-item queue contains several production item scheduling units for subsequent single sub-item scheduling; the production item scheduling units contain the affiliated item numbers, production item scheduling unit numbers, production information, and progress information.

[0007] As a preferred technical solution of the present invention, the specific steps for performing single sub-item scheduling on the production sub-item queue and the unfinished production sub-item queue using a scheduling constraint model include: The unfinished production sub-item queue contains several production item scheduling units with non-empty progress information; Merging the production sub-item queue and the unfinished production sub-item queue to obtain an integrated production sub-item queue; mapping the integrated production sub-item queue according to the topological structure of the sub-item production information to obtain an integrated production sub-item graph; Introducing a production sub-item idle time window; optimizing the integrated production sub-item graph and the production sub-item idle time window using a scheduling constraint model to obtain an optimized production sub-item queue.

[0008] As a preferred technical solution of the present invention, the specific steps for obtaining an optimized production sub-item queue through optimization in the scheduling constraint model include: Identifying the main-item guiding factors of the production item scheduling units based on the affiliated item numbers of the production item scheduling units in the integrated production sub-item queue; Identifying the required production scheduling features of the production item scheduling units based on the production information of the production item scheduling units in the integrated production sub-item queue; Allocating production resources according to the production scheduling features according to the scheduling constraint set to obtain production scheduling constraint factors; The scheduling constraint set specifically includes: production personnel fatigue constraint, equipment switching cooling constraint, production heat accumulation constraint, equipment production pre-maintenance constraint, production preparation precondition constraint; Match virtual production constraint time windows for production scheduling features based on production scheduling constraint factors; in the idle time windows of production sub-projects, fill in according to the virtual production constraint time windows of production project scheduling units to obtain production sub-project time windows; fuse the features of production sub-project time windows and the integrated production sub-project diagram to obtain a fused production sub-project diagram; Solve the path based on the fused production sub-project diagram and the main project guiding factor pair to obtain an optimized production sub-project queue.

[0009] As a preferred technical solution of the present invention, the specific steps for re-scheduling several optimized production sub-project queues based on the main project triple set include: Extract features based on the project information of the production project to obtain production project requirements; Associate the main project triples of the optimized production sub-project queue; Generate a production project production scheduling diagram based on all main project triples; where the nodes represent production projects; the edges represent the sharing degree of the optimized production sub-project queue shared by two different production projects; Schedule and sort the production project production scheduling diagram based on the main project scheduling priority strength index, and assign the final production priority to each optimized production sub-project queue; the main project scheduling priority strength index includes the main project type urgency, the main project scale, and the sub-project queue complexity; Combine several optimized production sub-project queues with production priorities to obtain a main project scheduling queue.

[0010] As a preferred technical solution of the present invention, the basic model of the scheduling constraint model is a graph attention network model.

[0011] A scheduling optimization system for production project management, including: A production project management module, which includes a feature description unit for obtaining the production operation project list of the day; obtaining the project information of several production projects based on the production operation project list; extracting features from the project information to obtain a main project triple set and a sub-project triple set; where the main project triple includes a project number, a production task, and an initial project priority; the sub-project triple includes a project number, a production task sub-project, and sub-project production information; extracting production sub-projects from all main project triple sets and sub-project triple sets to obtain several production sub-project queues; each production sub-project queue represents several production project scheduling units corresponding to a single sub-project; based on any production sub-project queue, obtain the corresponding unfinished production sub-project queue; The production project scheduling module includes an optimization scheduling unit, which is used to perform single-subproject scheduling on the production subproject queue and the unfinished production subproject queue by using the scheduling constraint model to obtain an optimized production subproject queue; re-schedule several optimized production subproject queues based on the main project triple set to obtain a main project scheduling queue; and complete the optimized scheduling of the production operation project list for the current day based on the main project scheduling queue.

[0012] The present invention has the following advantages: 1. Through the construction method of the main project triple and the subproject triple, the present invention standardizes and semanticizes complex production project information, effectively enhancing the parsability of the scheduling model for business logic; through the scheduling-aware coding and clustering processing of the subproject feature vectors, it can automatically identify similar production unit tasks that can be merged, improving resource utilization rate and reducing redundant scheduling costs; by introducing historical progress fusion and idle time window information, and combining the graph attention network model to construct a scheduling constraint graph, it can comprehensively consider various resource, personnel, and equipment status restrictions, and achieve the global optimal solution of the task scheduling path.

[0013] 2. Through the graph sorting mechanism of the main project scheduling priority, the present invention can uniformly evaluate the scheduling urgency for multiple order projects, accurately determine the execution order between projects, and take into account multiple dimensions such as urgency, resource sharing, and task complexity; by clustering and fusing the same or similar subproject tasks in different production projects, it can maximize the batch processing efficiency, reduce the frequency of process switching and equipment idle time, and achieve a project scheduling method of centralized processing of common tasks and intelligent interspersion of different tasks, greatly improving the production line throughput rate and equipment utilization rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a schematic structural diagram of a scheduling optimization system for production project management adopted in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0016] In this embodiment, the production project is carried out based on the manufacture of power equipment, and the following description will be made in the context of the manufacture of power equipment.

[0017] Embodiment 1. A scheduling optimization method for production project management includes the following steps: Obtain the production operation project list for the current day; In the scenario of power equipment manufacturing, the production operation project list for the day is used to indicate the scope of the production task schedule for the day, representing the list of all production projects entered into the production plan or scheduling system, and is distinguished by the project number as the only identifier.

[0018] Obtain the project information of several production projects based on the production operation project list; perform feature extraction on the project information to obtain the main project triple set and the sub-project triple set; among them, the main project triple includes the project number, production task, and initial project priority; the sub-project triple includes the project number, production task sub-project, and sub-project production information; In this method, the production task in the main project mainly refers to the production goals that the whole project needs to achieve, such as the general assembly of high-voltage switchgear, the complete set of electric energy metering boxes, the integration of control cabinets, etc., and usually appears as the main delivery content in the customer order; while the initial project priority includes the scheduling priority set in the initial stage of the system, which is determined according to factors such as contract delivery date, customer level, project amount, customer credit level, etc., and can be in the form of high, medium, low or numerical priority weight; correspondingly, the production task sub-project in the sub-project refers to the specific process that can be executed as a scheduling unit, such as sheet metal blanking, busbar installation, electrical commissioning, injection molding, etc., which are the technological links that make up the main project; and the sub-project production information includes the detailed execution data of the sub-task, such as process name, affiliated station or workshop, equipment used, person in charge, task status, planned quantity, completed quantity, estimated working hours, resource constraints, etc., and these information constitute the core basis for the scheduling system to allocate personnel, equipment and timing relationships.

[0019] After obtaining the project information of several production projects based on the production operation project list, the specific steps of feature extraction include the following links: First, read each project number from the production operation project list, and call the front-end or database interface to obtain the corresponding complete project information, including customer information, product type, delivery time, project priority, process flow, etc.; Second, divide each production project into a main project and a sub-project level according to the modular structure, where the main project corresponds to the overall production task (such as the assembly of high-voltage switchgear), and the sub-project corresponds to the schedulable process or task unit (such as sheet metal processing, wire harness welding); Then, extract key fields such as project number, project name, customer attribution, initial priority, delivery date, etc. from the main project information to construct a main project triple; Extract sub-task number, affiliated project, process name, equipment resources, person in charge, planned output, status, etc. from the sub-project information to construct a sub-project triple; After that, perform standardization processing on all structured information, including field unification, coding mapping (such as mapping priority to a numerical value), time format normalization, and perform data normalization operations (such as performing Min-Max or Z-score normalization on numerical features such as output, estimated man-hours, and process duration), and finally form a set of main project triples and a set of sub-project triples with clear structure and unified standards, providing a unified data input basis for subsequent scheduling modeling and clustering analysis.

[0020] Extract production sub-projects from all sets of main project triples and sub-project triples to obtain several production sub-project queues; Each production sub-project queue represents several production project scheduling units corresponding to a single sub-project; The specific steps for extracting production sub-projects from all sets of main project triples and sub-project triples include: Generate sub-project feature vectors based on the sub-project triples in the set of sub-project triples; Extract the key attribute information of each sub-project from the set of sub-project triples as input, and the input data structure is a set of structured triple records. Through feature engineering means, these triples are vectorized and transformed. First, categorical variables (such as process name, equipment type) are one-hot encoded or embedded, numerical variables (such as output, man-hours) are normalized, and time variables are time-window vectorized (such as morning and evening shift identification). Finally, a high-dimensional, trainable sub-project feature vector (usually a dense floating-point vector) is constructed to represent the scheduling behavior characteristics; The output of this step is a list of sub-project feature vectors, each vector corresponding to a sub-project, constituting the original input of the subsequent coding and clustering models; In this process, the beneficial effect is to realize the conversion from the original triples to a unified vector space, making the structured task data comparable and measurable, and providing a numerical expression basis for subsequent scheduling tendency analysis and clustering processing.

[0021] Based on the main project triples in the main project triple set, generate main project guiding factors with the project number as the identifier; take the main project triple set as the input, use the project number as the index, and correspondingly integrate multiple main project triples under this project to form a structured main project information block; through feature screening and semantic aggregation operations, perform standardization processing on different fields: for example, map the priority (high / medium / low) to a weight value, convert the customer attribute into a delivery sensitivity score, convert the delivery time into a time urgency level (such as the remaining number of days), and can add context scoring logic for factors such as project scale / importance. The output result is a set of main project guiding factor sets with the project number as the primary key, which is used to propagate and schedule importance information at the sub-project level in the subsequent steps; the beneficial effect of this step is to establish a global tonality constraint for task scheduling, so that subsequent sub-project processing not only clusters based on the technological characteristics of the task itself, but also introduces the business-oriented information of the main project, enhancing the global rationality and responsiveness of the scheduling clustering.

[0022] Perform scheduling-aware encoding based on the sub-project feature vectors to obtain sub-project scheduling-aware encoding; Concatenate (or fuse) each sub-project feature vector with the guiding factor under its corresponding project number, and generate sub-project scheduling-aware encoding through a neural network encoder (such as an MLP or an autoencoder); this encoding retains the technological and resource characteristics of the sub-project, and at the same time integrates the scheduling priority and delivery intention of the main project, so that the final representation vector of each sub-project has the dual characteristics of project structure and project orientation, capturing the relationship between the technological complexity of the task itself and the upper-level scheduling strategy; the beneficial effect of this step is to enable the scheduling model to identify the key processes in high-priority production projects, thus guiding key tasks to form clusters in subsequent clustering.

[0023] Perform clustering on all sub-project scheduling-aware encodings using a clustering model to obtain several sub-project groups; Use a clustering algorithm to group sub-projects. For example, DBSCAN can be selected to automatically determine the number of clusters, or an improved KMeans can be used to set the initial number of clusters according to business-side priors; the input data structure is a set of sub-project vectors, and the output is several sub-project groups, and the sub-projects within each group are highly similar in terms of scheduling tendency, technological processes, equipment requirements, etc.; clustering can combine the main project factor weights and dynamically adjust the distance metric method (such as using scheduling-aware distance) to avoid high-priority sub-projects being diluted and clustered by low-priority ones; the beneficial effect is to classify the sub-projects within the scope of the production operation project list according to scheduling-related attributes, providing a grouping basis for subsequent resource coordination, batch production scheduling, and task merging, and improving the controllability and production efficiency of the scheduling system.

[0024] Generate a number of production sub - project queues based on several sub - project groups. Each production sub - project queue contains a number of production project scheduling units for subsequent single - sub - project scheduling. The production project scheduling unit contains the project number it belongs to, the production project scheduling unit number, production information, and progress information. Restructure the structure of each group, and parse each sub - project into the smallest scheduling unit, that is, the production project scheduling unit. The data structure of this unit includes the project number it belongs to (associated with the upper - level project), the production project scheduling unit number (uniquely identifying this scheduling task), production information (such as processes, equipment, material requirements), current progress (completed / not started / paused), man - hour prediction, etc. Each sub - project group is constructed into a production sub - project queue to maintain the consistency of project - allocated resources and resource commonality for subsequent targeted scheduling optimization. The output result is a number of structured scheduling queues, and each queue represents a set of tasks that can be independently optimized. The beneficial effect is to achieve the integration of sub - tasks between main - project tasks (normalizing similar processes), breaking down project barriers for resource rearrangement scheduling, thereby improving production capacity utilization and avoiding scheduling fragmentation.

[0025] Based on any production sub - project queue, obtain the corresponding unfinished production sub - project queue; obtain the corresponding unfinished production sub - project queue by feature matching based on the sub - project name. Use the scheduling constraint model to perform single - sub - project scheduling on the production sub - project queue and the unfinished production sub - project queue to obtain an optimized production sub - project queue. The basic model of the scheduling constraint model is the graph attention network model. Utilize the ability of the graph attention network model to model the dependency relationship and resource constraint relationship between tasks to achieve intelligent optimization of complex scheduling graphs.

[0026] The specific steps of using the scheduling constraint model to perform single - sub - project scheduling on the production sub - project queue and the unfinished production sub - project queue include: The unfinished production sub - project queue contains a number of production project scheduling units with non - empty progress information. Extract the list of unfinished production tasks, and filter out the task units that are still in the not-started, in-progress, or paused state. The input data includes the production project scheduling unit records of all unfinished sub-projects. Each record contains the project number it belongs to, the production project scheduling unit number, the completed man-hours / quantity, the remaining workload, the task status, etc.; Filter out the task units that still need to participate in subsequent scheduling optimization according to the criterion that the progress information is not empty (such as completed man-hours > 0 or status ≠ completed); The output is the queue of unfinished production sub-projects, where each item is a structured production project scheduling unit record; The beneficial effect of this step is to integrate historical unfinished tasks with today's planned scheduling, making the scheduling model have continuity and traceability, avoiding resource idling and repeated scheduling, and enhancing the system's ability to respond to the actual production status.

[0027] Merge the production sub-project queue and the unfinished production sub-project queue to obtain the comprehensive production sub-project queue; Map the comprehensive production sub-project queue according to the topological structure of the sub-project production information to obtain the comprehensive production sub-project graph (directed acyclic graph); Unify and integrate the current production sub-project queue with the production project scheduling units of the unfinished sub-projects left over from history to form a complete set of tasks to be scheduled; The data input includes: the newly generated production sub-project queue (from the clustering output), the unfinished production sub-project queue obtained in the previous step; Each record is a production project scheduling unit, and the structure should at least include: project number, process name, task status, required equipment, estimated man-hours, resource requirements, etc.; The merging method usually adopts the de-duplication and status update mechanism at the task level. For example, the status of units with the same task number is overwritten, and the remaining quantities of overlapping tasks are merged. The output is the comprehensive production sub-project queue, a unified set of production project scheduling units containing all sub-projects to be processed; The beneficial effect of this step is to uniformly integrate and process the tasks unfinished yesterday and the scheduling tasks newly issued today, construct a real and complete task input space, and lay an accurate data foundation for the subsequent construction of the scheduling graph; The topological structure of the sub-project production information refers to the sequential dependence relationship, resource flow path, and execution logic connection method existing among the various processes or production tasks within the sub-project. Taking the integrated production sub - project queue as the input, the production information of each production project scheduling unit records the process name, process sequence number (or predecessor process field), the affiliated equipment, and resource occupation information; by identifying the relationship between the predecessor and successor processes of each production project scheduling unit, a directed edge is established between scheduling tasks, where nodes represent production project scheduling units and edges represent process dependencies; a DAG (Directed Acyclic Graph) is generated using a process flow template, a product BOM structure, or a predefined process sequence library; the output is an integrated production sub - project diagram, where each node has task execution information and each edge represents the scheduling sequence logic; this graph structure is the basis for subsequent path planning and resource scheduling, used to clarify the task execution order, construct a process constraint graph, prevent parallel conflicts and scheduling reverse order, and ensure the physical feasibility and safety of the scheduling plan.

[0028] Introduce the idle time window of the production sub - project; use the scheduling constraint model to optimize the integrated production sub - project diagram and the idle time window of the production sub - project to obtain an optimized production sub - project queue. The specific steps to obtain an optimized production sub - project queue through optimization in the scheduling constraint model include: Based on the affiliated project number of the production project scheduling unit in the integrated production sub - project queue, identify the main project guiding factor of the production project scheduling unit. Based on the production information of the production project scheduling unit in the integrated production sub - project queue, identify the production scheduling characteristics required by the production project scheduling unit. For each production project scheduling unit, extract its structured production information fields, including process type, resource requirements, equipment model, required process duration, bill of materials, etc. The input data is the record of the production project scheduling unit; the output is the scheduling feature vector of each unit; the data structure is a dense feature vector, encoding task load, resource preference, execution characteristics, etc. The feature extraction method uses rule mapping (such as mapping the assembly of injection - molded parts to equipment type = injection molding machine, process load = high) or an embedding - based encoding model to convert task information into a standard expression for scheduling calculations, providing an input feature basis for subsequent time - window matching and path optimization.

[0029] Allocate production resources according to the production scheduling characteristics according to the scheduling constraint set to obtain production scheduling constraint factors. Use a pre - trained production resource allocation model to identify production scheduling characteristics and output production scheduling constraint factors. The pre-trained production resource allocation model can be constructed and trained by a BP neural network model. The specific steps are as follows: Extract the production project scheduling unit records of the past year from the historical scheduling logs as the training set. These records include input features such as task processes, equipment types, product specifications, batch sizes, man-hours, main project priorities, and personnel shifts, as well as the constraint factor labels actually assigned by manual or expert systems at that time; for example, rest interval durations, equipment cooling buffers, allowable peak heat loads, pre-maintenance insertion marks, and preparation lead times, etc.; all numerical features are normalized, and categorical features are encoded by one-hot or embedding; subsequently, the samples are randomly divided into an 80% training set and a 20% validation set, and the data is input into a BP network with two hidden layers. The loss function selects multi-output mean squared error, and the optimizer uses Adam (initial learning rate 0.001); during the training process, monitor the loss and R² metrics on the validation set. If the validation loss does not decrease significantly within multiple training rounds, trigger early stopping to save the weights and export the model; the final output of the model is the production scheduling constraint factors, corresponding to five types of constraint parameters for fatigue, cooling, heat, maintenance, and preparation lead times that each input production project scheduling unit needs to comply with, which can be directly called by subsequent scheduling algorithms to achieve fast, data-driven resource allocation decisions.

[0030] The scheduling constraint set specifically includes: production personnel fatigue constraint, equipment switching cooling constraint, production heat accumulation constraint, equipment production pre-maintenance constraint, production preparation lead constraint; The production personnel fatigue constraint means that during the production scheduling process, it is necessary to consider the continuous working hours and rest times of production personnel to avoid efficiency decline, operation errors, or safety hazards caused by overwork. It is particularly crucial in key processes (such as high-voltage wiring and precision welding); rules such as a single person's continuous working time not exceeding X hours and daily total working time limit can be set; in the scheduling constraint model, forced rest windows will be automatically inserted or tasks will be excluded from being scheduled during this time period; The equipment switching cooling constraint indicates that when switching equipment between different process tasks (for example, from injection molding to compression molding), physical processing processes such as cleaning, cooling, and preheating may be required to avoid material residues or temperature abnormalities affecting product quality. For example: it is necessary to judge whether the process type changes for adjacent tasks and automatically insert a cooling time buffer (such as 30 minutes or dynamically predicted according to equipment type); otherwise, scheduling will result in equipment contamination or process failure; The cumulative heat production constraint means that when certain hot - processing tasks (such as welding, drying, and molding) are arranged continuously, the internal heat load in the equipment or workshop continues to increase. Exceeding the threshold will cause the equipment to overheat and the production environment to become unstable. It is necessary to monitor the density or frequency of the same type of high - heat tasks per unit time; insert low - heat tasks or adjustment sections for waiting to cool in the scheduling path to form a heat load distribution map and balance the production scheduling rhythm; the equipment production pre - maintenance constraint refers to predictive maintenance for key production equipment (such as injection molding machines, CNC machine tools, etc.) based on their usage frequency, historical health data, or warning indicators. Arrange maintenance or calibration before a failure occurs to ensure the stable operation of the equipment; predict the future maintenance window based on the equipment operation data, actively block the production scheduling line at this time point, and only allow re - arranging tasks after the maintenance is completed; the scheduling engine needs to have the ability to embed maintenance nodes; The pre - production preparation constraint means that before some processes start, it is necessary to prepare resources such as special tooling, fixtures, molds, gauges, and materials in advance. If the pre - production preparation tasks are not completed, the main tasks cannot start on time; it is necessary to analyze the pre - production dependencies of each production project scheduling unit to ensure that its pre - production preparation tasks are arranged before the main tasks and have sufficient time to complete, and model the constraints by defining the dependency edges from tooling preparation to the main process.

[0031] Match the virtual production constraint time window for the production scheduling characteristics based on the production scheduling constraint factors; in the idle time window of the production sub - project, fill it according to the virtual production constraint time window of the production project scheduling unit to obtain the production sub - project time window; fuse the features of the production sub - project time window and the comprehensive production sub - project diagram to obtain the fused production sub - project diagram; Solve the path based on the fused production sub - project diagram and the main project guiding factor pair to obtain the optimized production sub - project queue; After extracting the production scheduling constraint factors, combine the scheduling characteristics of each production project scheduling unit with its corresponding scheduling constraint factors to predict the theoretically schedulable time period for this task, that is, generate the virtual production constraint time window; this process combines multiple methods: for example, hard - rule reasoning can directly exclude clearly unavailable time periods such as maintenance windows and equipment cooling periods; the prediction model based on historical resource load can evaluate the future idle possibility of resources; and the resource schedulable weight model is used to identify the matching priority between tasks and resources; ultimately, each production project scheduling unit will obtain a set of candidate virtual time periods, providing constraint boundaries for subsequent path planning; Based on the idle time window of the existing production sub-projects, the above virtual time window and the actual resource gaps are matched by intersection. The input includes the available time periods of each device, workstation, and worker currently obtained from the MES or prediction system, as well as the task time window inferred in the previous step. The sliding window algorithm or interval tree structure is used to quickly identify which tasks are truly executable in which time periods, thereby forming an accurate time window for the production sub-projects, which is recorded as the legal available time range of each production project scheduling unit on the scheduling resources. The time window information of the production sub-projects is integrated with the comprehensive production sub-project diagram to form a composite scheduling diagram with task dependencies and resource time constraints. This integration process includes embedding the adjustable time window information of each graph node (i.e., the production project scheduling unit) and adjusting the edge weights according to resource characteristics and time characteristics. For example, a penalty is set for task connections across shift segments, or the edge cost is increased for paths that require equipment replacement. This structured time-constrained scheduling diagram not only retains the technological order of the tasks but also superimposes the schedulable time dimension, thus laying the foundation for multi-objective path optimization. Path solving is jointly carried out based on the integrated scheduling diagram and the leading factors of the main projects to which each production project scheduling unit belongs. The path optimization method can choose heuristic search based on graph structure (such as topological sorting with priority, A*), or introduce deep learning scheduling strategies based on graph neural networks, or combine constraint satisfaction and greedy algorithms for multi-objective optimization. The leading factors of the main projects play a strategic role at this stage, which is used to dynamically adjust the scheduling priorities of task nodes. For example, tasks approaching the delivery date are given priority, and tasks with a higher customer level receive preferential resource matching. The final output is an optimized scheduling sequence that meets multiple conditions such as scheduling order, resource conflicts, time windows, and policy orientation, that is, an optimized production sub-project queue, in which each task clearly defines its execution order, scheduling time period, and the resources used.

[0032] Based on the main project triple set, several optimized production sub-project queues are rescheduled to obtain the main project scheduling queue. Based on the main project scheduling queue, the optimization scheduling of the production operation project list for the current day is completed. The specific steps for rescheduling several optimized production sub-project queues based on the main project triple set include: Feature extraction is performed based on the project information of the production project to obtain the production project requirements. Taking each production item included in the production operation item list as input, by accessing item information (such as customer type, product model, planned delivery time, order type, urgency, etc.), data features representing project scheduling tendencies and scheduling priorities are extracted. These features are mapped into a production item demand vector through structured fields. For example, the vector dimension represents the urgency of the delivery date, the project volume (measured by the number of processes and man-hours), the customer level, whether it is a special order, etc.; the data structure is a key-value mapping table, with the input being a dictionary of item information and the output being the demand vector corresponding to each item. The beneficial effect of this step is to convert unstructured project attributes into a data form that can be used for scheduling modeling, providing a decision basis for subsequent sorting and graph construction.

[0033] Associate the main project triples that optimize the production sub-item queue; associate each optimized production sub-item queue with the main project triples from which it originated, mainly through the mapping of the main project number. Generate a production item production scheduling graph based on all main project triples; among them, the nodes represent production items; the edges represent the sharing degree of the optimized production sub-item queues shared by two different production items; by analyzing whether there are multiple items sharing the same sub-items in the optimized sub-item queue (i.e., the same process tasks can be combined for scheduling), the sharing strength between two items is constructed (such as the number of shared tasks, shared weight); it can be represented in the form of an adjacency matrix or an edge-weighted graph, and the edge weight value is the number of shared sub-item queues or the normalized weight. Perform scheduling sorting on the production item production scheduling graph based on the main project scheduling priority strength index, and assign the final production priority to each optimized production sub-item queue; the main project scheduling priority strength index includes the urgency of the main project type, the scale of the main project, and the complexity of the sub-item queue. In the key steps of production scheduling and ranking based on the scheduling priority intensity index of the main project, it is first necessary to comprehensively evaluate the scheduling urgency of each main project; this evaluation consists of three core dimensions: the first is the type urgency of the main project, which mainly reflects whether the project delivery deadline is approaching, whether the customer level is relatively high, or whether the project belongs to an extremely urgent order, etc.; the second is the scale of the main project, which is usually measured by the number of sub-projects included in the project, the overall estimated working hours or the total resource consumption; the third is the complexity of the sub-project queue, specifically including the topological depth of the process path, the dependency density between tasks, the frequency of resource conflicts, etc. After quantification and normalization, these three dimensions are integrated into a scoring vector of the main project, which is used to express the priority intensity of the project in the scheduling system. A sorting algorithm is used for scheduling priority inference: a topological sorting algorithm with weight awareness can be selected to comprehensively consider the urgency of a single project and the coupling impact in the overall task network; each main project will obtain a scheduling priority value, and based on this, a global execution order is assigned to the associated optimized sub-project queue; this process not only ensures the rationality of task scheduling, but also effectively introduces the resource competition relationship and scheduling coordination between projects, making the entire scheduling system tend to be globally optimized.

[0034] Combine several optimized production sub-project queues with production priorities to obtain the main project scheduling queue. According to the scheduling priorities of the main projects, combine and sort the subordinate optimized sub-project queues to generate the main project scheduling queue. The sub-project queues of each main project are classified and sorted according to the priority blocks to form a hierarchical and ordered task arrangement structure.

[0035] This embodiment is based on the production scenario of power equipment and demonstrates how to apply the described scheduling optimization method to perform collaborative and intelligent scheduling of orders from multiple customers; for example, an electrical manufacturing company receives the following production operation project list on a certain day, including a total of 5 production projects from 3 customers, and the main products are electric energy metering boxes and meter inspection boxes. Each project is parsed into a main project triple, extracting information such as its project number, customer attribution, delivery date, product type, priority, etc.; each project is split into multiple sub-project triples, and each sub-project represents an independently schedulable process task, such as wiring and assembly, shell injection molding, final inspection and packaging, etc.; for all sub-projects, extract process types, resource requirements, task volumes, equipment types, etc. as feature vectors, and use a clustering algorithm (such as dynamic K-means with time window constraints) to merge and classify tasks to obtain multiple production sub-project queues. For example, the injection molding part assembly tasks in multiple projects will be merged and scheduled to form a unified production sub-project queue across projects to maximize batch processing and resource utilization.

[0036] Retrieve the unfinished tasks of the same type from the previous day's production log, merge them with the new tasks to construct a comprehensive task queue, and construct a directed acyclic graph based on the process flow to represent the execution order and dependency relationship between sub-projects; for each sub-project queue, introduce a scheduling constraint model, including: worker fatigue constraint (continuous working hours < 6 hours); equipment switching cooling interval (such as 30 minutes required when an injection molding machine switches); cumulative heat load control (disperse drying / welding tasks); equipment preventive maintenance window (block task arrangement according to the prediction model); tooling preparation lead time (assembly fixtures need to be prepared 1 hour in advance).

[0037] Combined with the idle time window of the actual resources, infer the schedulable window of each task unit, and perform path optimization based on the scheduling graph and resource constraints to output a scheduling sequence that meets all physical and organizational conditions; calculate the scheduling intensity indicators for each main project, including: type urgency (distinguish emergency orders or non-emergency orders); project scale (such as the number of sub-projects, total working hours); sub-project queue complexity (such as the depth of the dependency graph, resource conflict density); construct the main project production scheduling graph based on the sharing degree of the same sub-project queue among multiple projects; determine the global scheduling order of the main project through topological sorting and weight scoring. Finally, generate the main project scheduling queue, issue the corresponding optimized production sub-project queue one by one according to the priority, import it into the MES system for the actual task allocation of the day, and in the result, high-priority tasks will occupy resources first, and low-priority tasks will be delayed or wait to ensure the on-time delivery of key projects.

[0038] Example 2, a scheduling optimization system for production project management, see Figure 1 as shown, including: The production project management module includes a feature description unit for obtaining the list of production operation projects for the day; obtaining the project information of several production projects based on the list of production operation projects; extracting features from the project information to obtain the main project triple set and the sub-project triple set; among them, the main project triple includes the project number, production task, and the initial priority of the project; the sub-project triple includes the project number, production task sub-project, and sub-project production information; extract production sub-projects from all the main project triple sets and sub-project triple sets to obtain several production sub-project queues; each production sub-project queue represents several production project scheduling units corresponding to a single sub-project; based on any production sub-project queue, obtain the corresponding unfinished production sub-project queue. The production project scheduling module includes an optimization scheduling unit, which is used to perform single-subproject scheduling on the production sub-project queue and the unfinished production sub-project queue by using the scheduling constraint model to obtain an optimized production sub-project queue; re-schedule several optimized production sub-project queues based on the main project triple set to obtain a main project scheduling queue; and complete the optimized scheduling of the production operation project list for the current day based on the main project scheduling queue.

[0039] It should be understood that those of ordinary skill in the art can make improvements or transformations according to the above description, and all such improvements and transformations shall fall within the protection scope of the appended claims of the present invention. The parts not described in detail in this specification belong to the prior art well-known to those skilled in the art.

Claims

1. A scheduling optimization method for production project management, characterized in that, It includes the following steps: Obtain the list of production operation items for the current day; Based on the list of production operation items, obtain the project information of several production projects; extract features from the project information to obtain the main project triple set and the sub-project triple set; among them, the main project triple includes the project number, production task, and initial project priority; the sub-project triple includes the project number, production task sub-project, and sub-project production information; Extract production sub-projects from all the main project triple sets and sub-project triple sets to obtain several production sub-project queues; each production sub-project queue represents several production project scheduling units corresponding to a single sub-project; based on any production sub-project queue, obtain the corresponding unfinished production sub-project queue; Use the scheduling constraint model to perform single sub-project scheduling on the production sub-project queue and the unfinished production sub-project queue to obtain an optimized production sub-project queue; based on the main project triple set, re-schedule several optimized production sub-project queues to obtain the main project scheduling queue; Based on the main project scheduling queue, complete the optimized scheduling of the production operation item list for the current day.

2. The scheduling optimization method for production project management according to claim 1, wherein The specific steps for extracting production sub-projects from all the main project triple sets and sub-project triple sets include: Generate sub-project feature vectors based on the sub-project triples in the sub-project triple set; Generate a main project guiding factor with the project number as the identifier based on the main project triples in the main project triple set; Perform scheduling-aware encoding based on the sub-project feature vectors to obtain sub-project scheduling-aware encodings; Cluster all the sub-project scheduling-aware encodings using a clustering model to obtain several sub-project groups; Generate several production sub-project queues based on several sub-project groups; each production sub-project queue contains several production project scheduling units for subsequent single sub-project scheduling; the production project scheduling unit contains the affiliated project number, production project scheduling unit number, production information, and progress information.

3. A scheduling optimization method for production project management according to claim 2, characterized in that, The specific steps for using the scheduling constraint model to perform single sub-project scheduling on the production sub-project queue and the unfinished production sub-project queue include: The unfinished production sub-project queue contains several production project scheduling units with non-empty progress information; Merge the production sub-project queue and the unfinished production sub-project queue to obtain an integrated production sub-project queue; map the integrated production sub-project queue according to the topological structure of the sub-project production information to obtain an integrated production sub-project graph; Introduce the idle time window of the production sub-project; use the scheduling constraint model to optimize the integrated production sub-project graph and the idle time window of the production sub-project to obtain an optimized production sub-project queue.

4. A scheduling optimization method for production project management according to claim 3, characterized in that The specific steps for obtaining the optimized production sub-project queue through optimization in the scheduling constraint model include: Based on the affiliated project number of the production project scheduling unit in the integrated production sub-project queue, identify the main project guiding factor of the production project scheduling unit; Based on the production information of the production project scheduling unit in the integrated production sub-project queue, identify the required production scheduling features of the production project scheduling unit; Allocate production resources according to the production scheduling features according to the scheduling constraint set to obtain the production scheduling constraint factor; The scheduling constraint set specifically includes: production staff fatigue constraint, equipment switching cooling constraint, production heat accumulation constraint, equipment production pre-maintenance constraint, and production preparation precondition constraint; Based on the production scheduling constraint factors, match the virtual production constraint time window for the production scheduling characteristics; in the idle time window of the production sub-project, fill it according to the virtual production constraint time window of the production project scheduling unit to obtain the production sub-project time window; fuse the features of the production sub-project time window and the comprehensive production sub-project diagram to obtain the fused production sub-project diagram; Based on the fused production sub-project diagram and the main project guiding factor pair, solve the path to obtain the optimized production sub-project queue.

5. A scheduling optimization method for production project management according to claim 4, characterized in that The specific steps for re-scheduling several optimized production sub-project queues based on the main project triple set include: Extract features based on the project information of the production project to obtain the production project requirements; Associate the main project triples of the optimized production sub-project queue; Generate a production project production scheduling diagram based on all the main project triples; where the nodes represent production projects; the edges represent the sharing degree of the optimized production sub-project queue shared by two different production projects; Schedule and sort the production project production scheduling diagram based on the main project scheduling priority intensity index, and assign the final production priority to each optimized production sub-project queue; the main project scheduling priority intensity index includes the emergency degree of the main project type, the scale of the main project, and the complexity of the sub-project queue; Combine several optimized production sub-project queues with production priorities to obtain the main project scheduling queue.

6. The scheduling optimization method for production project management according to claim 5, characterized in that, The basic model of the scheduling constraint model is the graph attention network model.

7. A scheduling optimization system for production project management, characterized in that, The system, such as a scheduling optimization method for production project management described in any one of the above claims 1-6, includes: The production project management module includes a feature description unit for obtaining the production operation project list of the day; obtaining the project information of several production projects based on the production operation project list; extracting features from the project information to obtain the main project triple set and the sub-project triple set; where the main project triple includes the project number, production task, and project initial priority; the sub-project triple includes the project number, production task sub-project, and sub-project production information; extract production sub-projects from all the main project triple sets and sub-project triple sets to obtain several production sub-project queues; each production sub-project queue represents several production project scheduling units corresponding to a single sub-project; based on any production sub-project queue, obtain the corresponding unfinished production sub-project queue; The production project scheduling module includes an optimization scheduling unit for using the scheduling constraint model to perform single sub-project scheduling on the production sub-project queue and the unfinished production sub-project queue to obtain the optimized production sub-project queue; re-scheduling several optimized production sub-project queues based on the main project triple set to obtain the main project scheduling queue; completing the optimized scheduling of the production operation project list of the day based on the main project scheduling queue.

Images