Unified construction and extension method of multi-layer capability system oriented to APS

Abstract

The invention provides a unified construction and extension method of a multi-layer capability system oriented to advanced planning and scheduling (APS). The capability system is composed of a data layer, a target layer, a collaboration layer, a response layer, an intelligent layer, a model / operator layer, an algorithm layer, a template layer and an acceleration layer; and an expansion management device is matched to support plug-in expansion. Modularization, platformization, expandability and acceleratability of the APS are achieved through a unified data model, operator target expression, collaborative modeling and multi-level response, intelligent prediction and risk-driven intervention, multiplexing of a model / operator layer and irrelevant design of solution and acceleration layers. The system serves as a top blueprint of a product and also serves as a general outline of sub-patents and development specifications, and a patent matrix and an ecological business model are conveniently formed.

Description

Technical Field

[0001] This invention belongs to the technical fields of intelligent manufacturing, production planning optimization and industrial software platform. Specifically, it relates to a multi-layer capability system for advanced planning and scheduling (APS) and its unified construction method and extension mechanism, which is applicable to the capability construction, capability reuse, capability extension and platform evolution of APS systems. Background Technology

[0002] Existing APS systems generally adopt a single-layer or weakly layered architecture, lacking a unified capability system, resulting in:

[0003] 1. Data sources are heterogeneous, formats are inconsistent, and reuse is difficult;

[0004] 2. Inconsistent goal setting methods and poor scalability;

[0005] 3. Insufficient collaborative optimization capabilities make it difficult to support collaboration among multiple entities, resources, and regions;

[0006] 4. Lacks multi-level response capabilities, unable to support rolling plans and disturbance responses;

[0007] 5. Lacks intelligent predictive capabilities, making it impossible to identify risks in advance and generate intervention strategies;

[0008] 6. Lacks operatorization and model layer capabilities, making it difficult to achieve model reuse and model compression;

[0009] 7. Lacks solution strategy layer capability, unable to dynamically select solution strategy according to the scenario;

[0010] 8. Lack of industry template capabilities, making it difficult to quickly adapt to industry scenarios;

[0011] 9. Lacking acceleration layer capabilities, it is unable to achieve efficient solutions for large-scale models. Summary of the Invention

[0012] Therefore, a unified multi-layered capability system for APS is needed to construct the underlying capabilities, expressive capabilities, intelligent capabilities, solution capabilities, and extension capabilities of APS. Based on the applicant's previous research on APS data processing, target construction, collaborative optimization, and plan response, this invention proposes an overall framework and unified construction approach for a multi-layered capability system.

[0013] The capability system of this invention can be expanded at different levels of abstraction, including:

[0014] 1. System architecture layer, used to define the overall capability framework and extension mechanisms;

[0015] 2. Functional module layer, used to describe reusable capability units;

[0016] 3. Technical point layer, used to implement specific algorithms, operators or optimization strategies.

[0017] This multi-level structure helps the system achieve capability reuse, flexible expansion, and engineering implementation in different scenarios.

[0018] Based on this, the present invention further provides a multi-layer capability architecture covering the data layer, target layer, collaboration layer, response layer, intelligence layer, model / operator layer, algorithm layer, template layer and acceleration layer, and provides corresponding construction methods and extension mechanisms to realize the modularization, systematization and platformization of the capabilities of the APS system.

[0019] Based on the above abstract hierarchy and capability structure, this invention proposes a multi-layered capability system for APS and its unified construction method, including a data layer, target layer, collaboration layer, response layer, intelligence layer, model / operator layer, algorithm layer, template layer, and acceleration layer. It achieves platformization, modularization, plug-in functionality, and scalability through an extended management mechanism. The main technical solutions include, but are not limited to:

[0020] 1. Data Layer: Unified data acquisition, field mapping, validation, transformation, standardization, modeling and versioning, outputting a unified APS data model.

[0021] 2. Target Layer: The target expression is constructed based on decision factors and operator links, supporting weight governance and soft constraint generation.

[0022] 3. Collaboration Layer: Supports collaborative modeling, collaborative goal construction, conflict detection, and distributed / decomposition solution for multiple subjects, resources, and regions.

[0023] 4. Response layer: Supports rolling plans, event-triggered disturbance responses, policy pushdown, and multi-level linkage.

[0024] 5. Intelligent Layer: Provides disturbance prediction, predictive maintenance, dynamic OEE calculation, risk preference modeling, risk management and optimal intervention strategy generation (functional description, without limiting the specific algorithm implementation).

[0025] 6. Model / Operator Layer: Operator registration and link management; model building, pruning, decomposition, rewriting and caching, supporting model reuse and expression layer abstraction.

[0026] 7. Algorithm layer: solution strategy selection, incremental solution, warm start, parallel solution and strategy switching.

[0027] 8. Template layer: Industry / scenario template library, template parameterization, automatic matching and version management.

[0028] 9. Acceleration layer: Model compression, model decomposition, parallel acceleration, solution acceleration, and solver-independent acceleration strategies.

[0029] 10. Extended Management: The plug-in mechanism supports the registration, management, and runtime loading of operator plug-ins, template plug-ins, prediction plug-ins, solver plug-ins, and acceleration plug-ins.

[0030] Supported by the aforementioned multi-layered capability system and its unified construction and extension mechanism, this invention possesses excellent reusability, scalability, and engineering adaptability in APS scenarios. Specific beneficial effects include:

[0031] 1. Elevate APS from a standalone tool to a capability platform to facilitate modular development and ecosystem expansion;

[0032] 2. Reduce industry adaptation costs through operatorization and template-based methods;

[0033] 3. Enhance planning robustness by enabling proactive intervention driven by prediction through an intelligent layer;

[0034] 4. Improve the solvability and performance of large-scale models through the design of operator / model layers and acceleration layers;

[0035] 5. Enable third-party capability integration and ecosystem building through plug-in extensions. Attached Figure Description

[0036] Figure 1: Schematic diagram of data layer capability structure.

[0037] Figure 2: Schematic diagram of target layer capability structure.

[0038] Figure 3: Schematic diagram of the capability structure of the collaboration layer.

[0039] Figure 4: Schematic diagram of the response layer capability structure.

[0040] Figure 5: Schematic diagram of the capability structure of the intelligent layer.

[0041] Figure 6: Schematic diagram of the capability structure of the model / operator layer.

[0042] Figure 7: Schematic diagram of the algorithm layer capability structure.

[0043] Figure 8: Schematic diagram of template layer capability structure.

[0044] Figure 9: Schematic diagram of the acceleration layer capability structure.

[0045] Figure 10: Schematic diagram of the construction process of APS multi-layer capability system.

[0046] Figure 11: Schematic diagram of the implementation of the APS multi-layer capability system for discrete manufacturing.

[0047] Figure 12: Schematic diagram of patent family tree structure. Detailed Implementation

[0048] The technical solution of the present invention will be described in detail below with reference to specific embodiments, so that those skilled in the art can more clearly understand the concept, technical features and beneficial effects of the present invention. It should be understood that these embodiments are only used to illustrate the basic principles of the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0049] The technical solutions of this invention can be modified and extended in various ways according to actual application needs. As long as they do not depart from the core idea of ​​this invention and the scope defined by the claims, they should be regarded as the objects of protection of this invention.

[0050] Without causing technical conflicts, the technical features or sub-modules in the following embodiments can be combined arbitrarily to adapt to different application scenarios and system architectures.

[0051] The illustrations are for illustrative purposes only. The processes, module divisions, and interaction relationships shown can be adjusted and optimized according to actual application scenarios and should not be construed as the sole limitation of the collaborative decision-making optimization system and method of this invention. Those skilled in the art can adapt and extend the method steps, system structure, or information interaction methods according to specific needs without departing from the core ideas of this invention.

[0052] To enhance the clarity of the technical expression, terms such as "first" and "second" used in the specification are only used to distinguish similar components or steps and do not indicate their execution order or priority; terms such as "comprising" and "having" are open-ended expressions intended to cover other technical features not explicitly listed; terms such as "and / or" and " / " indicate a parallel or selective relationship, and their specific meaning should be understood in conjunction with the context.

[0053] To clearly present the overall structure of the multi-layered capability system of this invention, based on the patent family tree structure diagram shown in Figure 12, the overall capability is divided into a data layer, target layer, collaboration layer, response layer, intelligence layer, operator / model layer, algorithm layer, template layer, and acceleration layer. An extension management mechanism is used to achieve plug-in-like expansion of the system. Figure 12 illustrates the hierarchical relationship and expansion logic of the above capability system. The typical implementation methods of each capability layer will be described in detail below according to the hierarchical order defined by this family tree structure.

[0054] Example 1: Data Layer Capabilities

[0055] As shown in Figure 1, this embodiment provides an implementation method for APS data layer capabilities to support unified access, standardized processing, and model-based representation of multi-source data.

[0056] The data layer includes a field mapping module, a data validation module, a data transformation module, a data standardization module, a data modeling module, and a data versioning module.

[0057] In a typical embodiment, the system collects data such as orders, processes, resources, equipment, and inventory from multiple sources, including MES, ERP, and WMS. A field mapping module maps field names from different sources to a unified field system. A data validation module performs integrity, type, and constraint checks. A data conversion module performs unit, time format, and structure conversions. A data standardization module converts the data into a unified APS data model. A data modeling module constructs order, process, resource, and constraint models. A data versioning module records data snapshots to support rolling planning and disturbance response.

[0058] The final output is a unified APS data model for use by the target layer, collaboration layer, and algorithm layer.

[0059] Example 2: Target Layer Capabilities

[0060] As shown in Figure 2, this embodiment provides an implementation method for APS target layer capabilities, which is used to construct operatorized target representations based on a unified APS data model.

[0061] The target layer includes a decision factor generation module, an operatorized target construction module, a weight governance module, a soft constraint generation module, and a target expression output module.

[0062] In a typical embodiment, the system generates decision factors such as delivery deviation, equipment load, and number of changeovers based on order, process, and resource data; combines the above factors into a target expression based on an operatorization mechanism; normalizes and prioritizes the target weights through a weight governance module; constructs penalty terms through a soft constraint generation module; and finally outputs the operatorized target expression for use by the collaboration layer and algorithm layer.

[0063] Example 3: Collaboration Layer Capabilities

[0064] As shown in Figure 3, this embodiment provides an implementation method for APS collaboration layer capabilities to support collaborative modeling and collaborative solving of multiple subjects, multiple resources, and multiple regions.

[0065] The collaboration layer includes a collaboration modeling module, a collaboration goal construction module, a collaboration conflict detection module, and a collaboration solution module.

[0066] In a typical embodiment, the system constructs a collaborative model based on multiple factories, multiple production lines, or multiple resource pools; generates cross-agent objectives through a collaborative objective construction module; identifies resource conflicts, capacity conflicts, and delivery date conflicts through a collaborative conflict detection module; and performs cross-agent solving through a collaborative solution module and outputs collaborative decision results.

[0067] Example 4: Response Layer Capabilities

[0068] As shown in Figure 4, this embodiment provides an implementation method for APS response layer capabilities to support rolling planning, disturbance response, and policy pushdown.

[0069] The response layer includes a rolling planning module, a disturbance detection module, a disturbance response module, and a policy pushdown module.

[0070] In a typical embodiment, the system detects disturbance events such as equipment failure, order changes, or material shortages based on real-time data; performs local adjustments or global reordering through the disturbance response module; synchronizes the adjusted plan to the execution system through the strategy push-down module; and updates the plan periodically through the rolling plan module to maintain the real-time nature of the plan.

[0071] Example 5: Smart Layer Capabilities

[0072] As shown in Figure 5, this embodiment provides an implementation method for APS intelligent layer capabilities to support disturbance prediction, predictive maintenance, and risk management.

[0073] The intelligent layer includes a disturbance prediction module, an equipment health assessment module, a dynamic OEE calculation module, a risk preference modeling module, and an intervention strategy generation module.

[0074] In a typical embodiment, the system predicts the probability of equipment failure, order delay risk, or production capacity bottleneck based on historical and real-time data; generates a risk score through a risk preference modeling module; and outputs adjustment suggestions through an intervention strategy generation module for use by the target layer and response layer.

[0075] Example 6: Model / Operator Layer Capabilities

[0076] As shown in Figure 6, this embodiment provides an implementation method for APS model / operator layer capabilities to support operator registration, model building, model pruning, model decomposition, and model caching.

[0077] The operator layer includes an operator registration module, an operator link management module, and an operator combination verification module; the model layer includes a model building module, a model pruning module, a model decomposition module, a model rewriting module, and a model caching module.

[0078] In a typical embodiment, the system constructs an APS optimization model based on operator links; performs redundant constraint and redundant variable pruning through a model pruning module; performs structured decomposition through a model decomposition module; and caches the model structure through a model caching module to improve solution performance.

[0079] Example 7: Algorithm Layer Capabilities

[0080] As shown in Figure 7, this embodiment provides an implementation method for the APS algorithm layer capability, which supports solution strategy selection, incremental solution, hot start and parallel solution.

[0081] The algorithm layer includes a solution strategy selection module, an incremental solution module, a warm start module, and a parallel solution module.

[0082] In a typical embodiment, the system automatically selects a solution strategy based on model size, time window, and resource load; performs local solution when disturbances occur through an incremental solution module; reuses the previous solution results through a warm-start module; and improves solution efficiency through a parallel solution module.

[0083] Example 8: Template Layer Capabilities

[0084] As shown in Figure 8, this embodiment provides an implementation method for APS template layer capabilities to support the reuse of industry templates and scene templates.

[0085] The template layer includes an industry template library, a scenario template library, a template parameterization module, a template auto-matching module, and a template versioning module.

[0086] In a typical embodiment, the system automatically selects templates based on industry characteristics; fills in industry parameters through the template parameterization module; and manages template evolution through the template versioning module.

[0087] Example 9: Acceleration Layer Capabilities

[0088] As shown in Figure 9, this embodiment provides an implementation method for the APS acceleration layer capability, which supports model compression, model decomposition, parallel acceleration, and solver-independent acceleration.

[0089] The acceleration layer includes a model compression module, a model decomposition module, a parallel acceleration module, and a solver-independent acceleration module.

[0090] In a typical embodiment, the system performs variable aggregation, constraint aggregation, and time granularity compression based on the model structure; performs structured decomposition through the model decomposition module; improves the model construction and solution speed through the parallel acceleration module; and provides a unified acceleration interface for different solvers through the solver-independent acceleration module.

[0091] Example 10: A unified construction method for APS multi-layer capability system

[0092] As shown in Figure 10, this embodiment provides a unified method for constructing a multi-layered capability system for APS. This method constructs APS's data capabilities, target capabilities, collaborative capabilities, response capabilities, intelligent capabilities, operator / model capabilities, algorithm capabilities, template capabilities, and acceleration capabilities layer by layer according to predefined abstraction levels and capability structures, ultimately forming an scalable capability system. This method is applicable to the initial construction, capability expansion, and engineering implementation of APS systems.

[0093] The method includes the following steps:

[0094] S1: Data Layer Construction

[0095] Data on orders, processes, resources, equipment, and inventory is collected from multiple systems (such as MES, ERP, and WMS). Through mechanisms such as field mapping, data validation, data transformation, data standardization, and data modeling, a unified APS data model is generated to provide standardized input for subsequent capability layers.

[0096] S2: Target Layer Construction

[0097] Based on the unified APS data model, decision factors such as delivery deviation, equipment load, and switching costs are generated. The target expression is constructed through an operatorization mechanism, and weight governance and soft constraint generation are performed to form an operatorized target expression.

[0098] S3: Collaboration Layer Construction

[0099] A collaborative model is built based on multiple production lines, factories, or resource pools. Resource conflicts, capacity conflicts, and delivery conflicts are identified through collaborative conflict detection, and cross-entity collaborative objectives are constructed to provide a collaborative structure for subsequent solutions.

[0100] S4: Response Layer Construction

[0101] Based on real-time data, a rolling planning mechanism, a disturbance detection mechanism, and a disturbance response mechanism are built to support local adjustments or global reordering of disturbance events such as equipment failure, order changes, and material shortages.

[0102] S5: Building the Smart Layer

[0103] Based on historical and real-time data, disturbance prediction models, equipment health assessment models, dynamic OEE calculation models, and risk preference models are constructed, and intervention strategies are generated to provide intelligent assistance capabilities for the target and response layers.

[0104] S6: Operator Layer / Model Layer Construction

[0105] An operator system is constructed through operator registration, operator link management, and operator combination verification; an APS optimization model is built based on the operator link, and model pruning, model decomposition, model rewriting, and model caching are performed to improve the solvability and reusability of the model.

[0106] S7: Algorithm Layer Construction

[0107] The solution strategy is selected based on model size, time window, and resource load. Incremental solution, hot start, parallel solution, and strategy switching mechanisms are constructed to improve solution efficiency and stability.

[0108] S8: Template Layer Construction

[0109] Build industry template libraries and scenario template libraries, realize industry knowledge reuse through template parameterization and automatic template matching mechanisms, and manage template evolution through template versioning mechanisms.

[0110] S9: Accelerated Layer Construction

[0111] Based on the model structure, model compression, model decomposition and parallel acceleration are performed. A unified acceleration interface is provided for different solvers through a solver-independent acceleration mechanism to improve the performance of model building and solving.

[0112] S10: Extended Management Build

[0113] A plugin registration, plugin lifecycle management, and plugin runtime loading mechanism are established to support capability expansion, module replacement, and third-party capability integration, forming a scalable APS multi-layer capability system.

[0114] Example 11: Application Example of APS Multi-Layer Capability System for Discrete Manufacturing

[0115] As shown in Figure 11, this embodiment provides a comprehensive application method of the APS multi-layer capability system in discrete manufacturing scenarios, which is used to illustrate the collaborative working process of the capabilities of each layer of the present invention.

[0116] In a typical discrete manufacturing scenario, enterprises need to plan and schedule orders, processes, equipment, personnel, and materials in a unified manner. The system first collects data from MES, ERP, and equipment through the data layer, and then generates a unified APS data model through field mapping, validation, transformation, and standardization.

[0117] The target layer generates decision factors such as delivery deviation, equipment load, and changeover costs based on a unified data model, and constructs an operator-based target expression. The collaboration layer builds a collaboration model based on multiple production lines, identifies resource conflicts, and performs collaborative problem-solving.

[0118] The response layer detects equipment failures and order changes based on real-time data, and performs local adjustments or global rearrangements through the disturbance response module.

[0119] The intelligent layer predicts the probability of equipment failure and the risk of order delays based on historical data, and generates intervention strategies for the target layer and response layer to invoke.

[0120] The model / operator layer constructs an APS optimization model based on operator links, and improves the solvability of the model through model pruning and model decomposition.

[0121] The algorithm layer selects a solution strategy based on the model size and improves solution efficiency through incremental solution and warm start.

[0122] The template layer automatically selects industry templates based on industry characteristics and performs parametric filling.

[0123] The acceleration layer improves solution performance through model compression, model decomposition, and parallel acceleration.

[0124] Finally, the system outputs executable APS scheduling results and synchronizes them to the execution system through the policy push-down module, realizing the modularization, platformization, and scalability of APS.

Claims

1. A multi-layered capability system for Advanced Planning and Scheduling (APS), characterized in that, include: 1) Data layer capability device, used to perform data acquisition, field mapping, data verification, data transformation, data standardization, data modeling and data versioning, and output a unified APS data model; 2) Target layer capability device, used to generate decision factors, construct operator-based target expressions, govern target weights and generate soft constraints based on the unified APS data model; 3) Collaboration layer capability device, used to perform multi-agent collaborative modeling, collaborative target construction, collaborative conflict detection and collaborative solution. 4) Response layer capability device, used to execute rolling plans, event-triggered disturbance response, policy pushdown and multi-level linkage; 5) Intelligent layer capability device, used to perform disturbance prediction, predictive maintenance, dynamic OEE calculation, risk preference modeling, risk management and generate intervention strategies based on historical and real-time data; 6) Operator / model layer capabilities, used for operator registration and link management, model building, model pruning, model decomposition, model rewriting, and model caching; 7) Algorithm layer capabilities, used for solution strategy selection, incremental solving, warm start, parallel solving, and solution strategy switching; 8) Template layer capabilities, used for industry / scenario template management, template parameterization, automatic template matching, and template versioning; 9) Acceleration layer capabilities, used to provide model compression, model decomposition, parallel acceleration, solution acceleration, and solver-independent acceleration interfaces; 10) An extended management device for registering, managing, and running operator plugins, template plugins, prediction plugins, solver plugins, and acceleration plugins in a plug-in manner; wherein, each capability device works together through capability interfaces to achieve modularity, scalability, and acceleration of APS.

2. The system according to claim 1, wherein the field mapping of the data layer capability device supports multi-source field alias resolution and priority rules.

3. The system according to claim 1, wherein the operatorized target representation of the target layer capability device supports the construction of composite operators and piecewise penalty terms.

4. The system according to claim 1, wherein the collaborative layer capability device supports distributed solution and decomposition coordination algorithms to achieve cross-subject collaborative optimization.

5. The system according to claim 1, wherein the response layer capability device supports two types of disturbance response strategies: event-triggered local adjustment and global rearrangement.

6. The system according to claim 1, wherein the intelligent layer capability device outputs disturbance vectors, device health indicators, dynamic OEE values ​​and risk scores in the form of interfaces for use by the target layer and the algorithm layer.

7. The system according to claim 1, wherein the operator layer / model layer capability device supports operator metadata registration, operator combination verification, and operator link version management.

8. The system according to claim 1, wherein the model pruning module performs pruning of constraints and variables based on feasibility pre-check and redundant constraint identification, and records the pruning mapping to support the backtracking mapping of the solution.

9. The system according to claim 1, wherein the algorithm layer capability device supports automatic selection of solution strategy based on model size and time window, and supports hot start and incremental update.

10. The system according to claim 1, wherein the template layer capability device includes an industry template library and supports automatic template matching and parameterized filling based on industry characteristics.

11. The system according to claim 1, wherein the acceleration layer capability device provides external "model preprocessing → parallel construction → solution preparation" functional interfaces to support acceleration of any solver, without limiting the implementation details of specific acceleration algorithms.

12. A unified construction method for a multi-layered capability system oriented towards APS, characterized in that, Includes the following steps: S1. Construct data layer capabilities, perform data acquisition, field mapping, data validation, data transformation, data standardization and data versioning, and output a unified APS data model; S2. Based on the unified APS data model, construct target layer capabilities, generate decision factors and construct operator-based target expressions; S3. Construct a collaborative layer capability to achieve multi-agent collaborative modeling, collaborative goal construction, and collaborative solution; S4. Build response layer capabilities to achieve rolling planning, event-triggered disturbance response, and multi-level linkage; S5. Build intelligence layer capabilities to achieve disturbance prediction, predictive maintenance, dynamic OEE calculation, risk preference modeling, risk management, and intervention strategy generation; S6. Build operator / model layer capabilities to achieve operator registration, model building, model pruning, model decomposition, model rewriting, and model caching. S7. Build algorithm layer capabilities to enable solution strategy selection, incremental solution, hot start and parallel solution; S8. Build template layer capabilities to achieve industry / scenario template management and template parameterization; S9. Build acceleration layer capabilities to achieve model compression, model decomposition, parallel acceleration, and solver-independent acceleration interfaces; S10. Build an extended management system to support plug-in registration and operation of operator plug-ins, template plug-ins, prediction plug-ins, solver plug-ins, and acceleration plug-ins.

13. The method according to claim 12, wherein the data acquisition in step S1 includes any one or more methods such as API pull, message queue subscription, file import and direct database connection.

14. The method according to claim 12, wherein the perturbation prediction in step S5 provides prediction results to the target layer and the response layer by defining an input / output interface, without limiting the specific implementation of the prediction model.

15. The method of claim 12, wherein the model pruning in step S6 is triggered by feasibility pre-check and redundant constraint identification, and the pruning map is recorded to support solution backtracking.

16. The method according to claim 12, wherein the acceleration layer in step S9 provides a standardized interface to support the access and hot switching of different solvers.

17. A computer-readable storage medium having stored thereon computer-executable instructions, which, when executed by a processor, cause the processor to perform the method of any one of claims 12 to 16.

18. The "capability device" described in the claims may be implemented in hardware, software, or a combination of hardware and software; the "interface" includes, but is not limited to, API, message queue, file exchange, or database view; without departing from the spirit and substance of the invention, the various functions may be executed in parallel or serially in any order.

Images