An APS-based dynamic scheduling system for ship production

Abstract

This invention relates to the field of shipbuilding technology and proposes an APS-based dynamic scheduling system for ship production, comprising: a production site perception module, which is user-defined with production processes, production plans, and a site perception layer; a data processing module, which identifies the production status of each work segment according to the production logic and generates a structured production status event flow for each work segment based on the production logic; an APS optimization scheduling module, which solves various constraints in the production process and generates guiding scheduling instructions; a MES execution feedback module, which issues guiding scheduling instructions and monitors production progress according to the guiding scheduling instructions, and senses the progress of the project in real time through the production site perception module; and a performance analysis module, which periodically analyzes the changes in the progress of the project. This invention, by integrating real-time site perception, intelligent data processing, and APS dynamic optimization, effectively improves the accuracy, response speed, and overall operational efficiency of ship production scheduling.

Description

Technical Field

[0001] This invention relates to the field of intelligent ship manufacturing technology, and more specifically, to an APS-based dynamic scheduling system for ship production. Background Technology

[0002] Shipbuilding is a typical engineer-to-order (ETO) and project-driven complex discrete manufacturing industry. Its production process is characterized by extremely complex product structures, long production cycles, multiple parallel projects, highly shared resources, and intertwined constraints. A modern large ship consists of tens of thousands of parts, and its construction process spans multiple stages, including detailed design, procurement, hull section prefabrication, painting, section assembly, equipment installation, dock outfitting, and commissioning, with the entire production cycle lasting several years. During this period, large shipyards typically need to manage multiple ship projects of different models and at different schedules simultaneously. These projects compete for critical and limited scarce resources such as gantry cranes, dry docks / slipways, storage yards, sandblasting rooms, large transport vehicles, and highly skilled labor, leading to a sharp increase in resource conflicts and scheduling difficulties.

[0003] Shipbuilding production is also subject to strict physical space constraints (such as the storage locations of sections in the yard cannot obstruct each other) and complex process logic constraints (such as subsequent sections cannot start if the preceding sections are not completed). The intertwining of these "space-time-logic" triple constraints constitutes a large-scale, nonlinear NP-hard level dynamic scheduling problem.

[0004] For a long time, shipyards have primarily relied on Enterprise Resource Planning (ERP) systems for master production scheduling or manual scheduling using Gantt charts. However, the core logic of ERP systems—Material Requirements Planning (MRP)—is often based on the idealized assumption of unlimited capacity, severely neglecting the actual bottlenecks in key production equipment, manpower, and space, resulting in plans that are unfeasible in reality. Manual scheduling, on the other hand, is inefficient, struggles to handle complex scheduling across multiple projects and large scales, and is prone to problems such as a disconnect between plans and actual production, frequent production bottlenecks, delivery delays, and cost overruns.

[0005] To address these challenges, Advanced Planning and Scheduling (APS) systems are considered a key technological path to achieving digital transformation and intelligent production in shipyards. Unlike traditional ERP systems, APS is an intelligent decision-making tool based on Theory of Constraints (TOC) and advanced optimization algorithms (such as mixed-integer programming, constrained programming, and heuristic algorithms). Its core lies in "finite capacity scheduling," which, considering material supply and multiple capacity constraints, generates feasible and optimal detailed production schedules through mathematical modeling and simulation.

[0006] In general manufacturing, APS has proven effective in improving on-time delivery rates, shortening production cycles, and reducing inventory levels. However, in the shipbuilding sector, despite its production complexity and scheduling challenges far exceeding those of general manufacturing, the application of APS is still in the exploratory and initial development stages, facing the challenge of urgently needing targeted technological breakthroughs.

[0007] Therefore, there is an urgent need for an APS-based dynamic scheduling system for ship production to solve the technical problem that existing APS systems are difficult to implement in the shipbuilding field. Summary of the Invention

[0008] In view of this, the present invention proposes a dynamic scheduling system for ship production based on APS, which aims to solve the technical problem that APS systems are difficult to implement in the shipbuilding field.

[0009] This invention discloses an APS-based dynamic scheduling system for ship production, comprising:

[0010] The production site perception module is defined by the user, including production process, production plan and site perception layer. The site perception layer is equipped with sensor groups in each section, and the production logic of the sensor groups is preset to coordinate and associate according to the production process.

[0011] The data processing module identifies the production status of each work segment according to the production logic and generates a structured production status event stream for each work segment according to the production logic.

[0012] The APS optimization scheduling module integrates a constraint manager and a solver. The constraint manager is connected to a constraint knowledge base, which stores production rules written in a domain-specific language (DSL). The APS optimization scheduling module also presets basic production factor weights and generates guiding scheduling instructions by combining the production status event flow with the rules in the constraint knowledge base through a hybrid solver engine.

[0013] The MES execution feedback module issues guiding scheduling instructions, monitors production progress according to the guiding scheduling instructions, and senses the progress of the project in real time through the production site sensing module.

[0014] The performance analysis module periodically analyzes changes in project progress, generates on-time delivery rate for each work segment, and adjusts the weights of basic production factors for the next cycle based on the on-time delivery rate.

[0015] Preferably, in the production site sensing module, the site sensing layer is equipped with a sensor group in each section, and each section group includes a first sensor group and a second sensor group, wherein:

[0016] The first sensor group is a human-machine interaction terminal that receives information from personnel's active reports or behaviors, including electronic signatures, selection options, and input boxes, and is used to subjectively perceive the production progress of personnel in the production process.

[0017] The second sensor includes an equipment IoT module, an environmental sensor, a tide level gauge, an RFID / BLE positioning system, a face / fingerprint attendance system for workers, and a vision sensor, used to objectively monitor the spatial scheduling and resource occupancy of equipment in the production environment.

[0018] Preferably, the production site sensing module has pre-set collaborative production logic for each sensor group according to the production process, including:

[0019] Each shipbuilding process X is decomposed into several stages, each stage comprising multiple standardizable and repeatable work units Xi. Each work unit Xi is configured with unit constraints, and each work unit Xi is defined by worker group NR, process time GS, and equipment code group SB.

[0020] According to the production process, there is a corresponding task node for each traversal to the i-th work unit Xi, and each task node is configured with node constraints.

[0021] The node constraints and unit constraints are used to semantically align physical perception with the progress status in the digital production plan, generating a structured production status event stream, where each event is represented by... This indicates, including:

[0022] ,in ;

[0023] in, It is the i-th work unit; It is the ship's registration number; This refers to the i-th task node in stage m; Di is the work unit identifier; Di is the task node; Si is the timestamp; Si is the confidence level.

[0024] Preferably, the APS optimization scheduling module includes sensor group perception verification rules configured for each unit constraint and node constraint:

[0025] The node constraints are used to manage the lifecycle events of the task, including triggering and completion, wherein triggering and completion are achieved through a joint mapping with data from the first sensor group and the second sensor group;

[0026] The unit constraints are used to ensure compliance and resource accuracy during the execution process, including process monitoring and resource binding, wherein process monitoring and resource binding are achieved through data mapping with the second sensor group.

[0027] Preferably, the node constraints are used to manage task lifecycle events, including triggering and completion for determining task nodes. Whether it can be triggered or completed, i.e., the lifecycle events of the management task, include:

[0028] ;

[0029] in, It represents node constraints, where N is the node identifier; It is the triggering condition. It is the manual operation readiness status collected by the first sensor group; It is the completion status of the preceding task node detected by the second sensor group; It is a condition for completion. The status of the manual operation is collected by the first sensor group. This is the state where all equipment actions have ended, as detected by the second sensor group.

[0030] Preferably, the unit constraints are used to ensure compliance and resource accuracy during execution, including process monitoring, determining whether the current operation is within the normal range, and whether the data collected in real time by the second sensor group meets the process parameter standards, i.e., the operation unit. All associated sensors at the event timestamp All collected measurements must fall within their corresponding preset process window, whereby process monitoring is referred to as:

[0031] ;

[0032] in, It refers to process monitoring within unit constraints; Uk is the rule type identifier. Work unit The start time, Work unit End time, This indicates any time within the work unit; For the first One sensor in Physical quantities collected within the period; This is the preset process window corresponding to the j-th second sensor; This indicates a logical AND operation, meaning that all conditions must be true simultaneously.

[0033] Preferably, the unit constraints are used to ensure compliance and resource accuracy during execution, including resource binding to ensure that the resources used to execute the job are consistent with the plan.

[0034] ;

[0035] in, This refers to resource binding within a unit constraint; Ub is the resource binding identifier. It is the actual worker attendance data obtained from the second sensor. yes A work unit allows for the group of workers to perform that process. This is the actual equipment usage information obtained from the second sensor. yes The set of device code groups that a work unit is allowed to use;

[0036] If the resource binding conditions are not met, it is determined that resources are missing. A hierarchical priority scheduling mechanism is then initiated based on the weights of the basic production factors. Resources are scheduled hierarchically according to the unit's priority to ensure the priority execution of higher-level work units.

[0037] Preferably, the APS optimization scheduling module further presets basic production factor weights, and through a hybrid solution engine, combines the production status event flow with the rules in the constraint knowledge base to generate guiding scheduling instructions, including:

[0038] The APS optimization scheduling module also presets basic production factor weights. Through a hybrid solution engine, it combines the production status event flow with the rules in the constraint knowledge base to generate guiding scheduling instructions, including:

[0039] The weights of the basic production factors include assigning a value to the urgency of each project; assigning a value to the scarcity of materials based on the type, precision, and processing cost of equipment involved in each production segment; and assigning a value to the scarcity of labor based on the type of work, skill level, and training cost involved in each production segment.

[0040] The combination of the production status event flow and the rules in the constraint knowledge base includes: if the event flow determines that the resource binding condition is not met, it is determined that the resource is missing. When allocating resources to any work unit, the optimal boundary value of the candidate resources is generated from the available resource pool that meets the process constraints, prioritizing the nearest idle resource or prioritizing the resource with the lowest priority. The optimal boundary value of the candidate resources is then sorted and filtered sequentially.

[0041] Preferably, the APS optimization scheduling module prioritizes retrieving the nearest idle resource or prioritizes retrieving the resource with the lowest priority to generate the optimal boundary value of the candidate resources, and then sorts and filters the candidate resources according to their optimal boundary values, including:

[0042] Resource gaps will be quantified based on the material scarcity assignment and rules, and allocated preferentially according to the matching degree between the resource gap and the available resource pool. The first priority ranking rule is to remove resources from all tasks of the same importance level; the second priority ranking rule is to match the production resource with the highest matching degree in the idle resource pool; the third priority ranking rule is to match the production resource with the highest matching degree at a lower level than the production task; the fourth priority ranking rule is to further quantify the resource gaps generated after scheduling according to the third priority ranking rule, and allocate production resources based on this second quantification; the fifth priority ranking rule is to further quantify the resource gaps generated after scheduling according to the fourth priority ranking rule, and allocate production resources based on this third quantification; the sixth priority ranking rule is to further quantify the resource gaps generated after scheduling according to the fifth priority ranking rule, determine the degree of impact of the fourth quantification on the cycle of the task at this level, and generate a guiding scheduling instruction permission report for approval.

[0043] Preferably, the performance analysis module adjusts the weights of basic production factors for the next cycle based on the on-time delivery rate, including:

[0044] Obtain the resource gap deferred projects caused by the sixth priority ranking rule. For the on-time delivery rate of the deferred projects, determine the degree of their impact on the project's deferred status in actual construction based on the on-time delivery rate. In the next cycle, adjust the urgency assignment of the deferred projects based on the degree of deferred impact and prohibit the deferred projects from appearing in the candidate pool of the sixth priority ranking rule.

[0045] This invention deploys two types of sensor groups in each work section and configures collaborative production logic for each sensor group based on a preset production process, transforming multi-source heterogeneous field data into a structured production status event stream, thereby significantly improving the real-time performance and confidence of task status identification.

[0046] Furthermore, this invention manages the lifecycle of tasks through node constraints, using manual operation status and equipment completion signals to jointly verify the triggering and completion conditions of tasks; at the same time, it ensures the compliance of the execution process through unit constraints, wherein process monitoring ensures that all sensors in the work unit continuously meet their respective preset process windows during the execution cycle, and resource binding verifies whether the actual participating worker groups and equipment code groups belong to the set allowed by the unit.

[0047] When resource binding conditions are not met, a resource shortage is identified, and APS (Advanced Planning and Scheduling) optimization is initiated to ensure that higher-level tasks have priority in scheduling. During resource allocation, high-value resources are effectively prevented from being occupied by lower-level tasks. For complex resource gaps, a six-level progressive scheduling mechanism is further implemented. In addition, the performance analysis module periodically calculates the on-time delivery rate of each work section. For projects delayed due to resource shortages, their urgency assignment in the next cycle is automatically increased, and they are prohibited from re-entering the candidate pool of the sixth priority ranking rule. This forms a closed-loop optimization mechanism that adaptively adjusts scheduling, execution, performance, and weights, enabling the system to continuously evolve. Effectively addressing industry challenges such as the vast space, high resource mobility, and susceptibility to plan disruptions in shipyards, actual tests show a significant improvement in plan achievement rate and a 10% to 20% reduction in the overall construction cycle, demonstrating outstanding technological advancement and engineering practical value.

[0048] In summary, this invention is highly compatible with the production characteristics of shipbuilding, which involves multiple parallel projects, high resource sharing, and sensitivity to critical paths. In practical applications, shipyards often build several different ship types simultaneously. High-end equipment such as large gantry cranes, laser cutting lines, and scarce talent such as certified high-pressure welders are extremely limited, and docking time and tidal windows are non-renewable. Existing APS's "static binding" or "single-layer scheduling" in such scenarios easily leads to delays in critical tasks due to resources being occupied by ordinary projects. However, the hierarchical scheduling mechanism of this application allows the system to intelligently temporarily allocate resources from low-priority tasks such as civilian ships or outfitting phases, while ensuring the delivery of high-urgency projects such as military-grade projects. Through multiple rounds of quantification, it ensures that the allocation behavior does not trigger a chain reaction of delays, thereby achieving global optimization under complex constraints. Attached Figure Description

[0049] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0050] Figure 1 A functional block diagram of an APS-based ship production dynamic scheduling system provided in an embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram of the execution logic of the data processing module of the APS-based ship production dynamic scheduling system provided in an embodiment of the present invention. Detailed Implementation

[0052] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0053] See Figure 1 As shown, this embodiment discloses a ship production dynamic scheduling system based on APS, including:

[0054] The production site perception module is defined by the user, including production process, production plan and site perception layer. The site perception layer is equipped with sensor groups in each section, and the production logic of collaborative association is preset for each sensor group according to the production process.

[0055] The data processing module identifies the production status of each work segment based on the production logic and generates a structured production status event stream for each work segment based on the production logic.

[0056] The APS optimization scheduling module integrates a constraint manager and a solver. The constraint manager is connected to a constraint knowledge base, which stores production rules written in the domain-specific language DSL. The APS optimization scheduling module also presets basic production factor weights and generates guiding scheduling instructions by combining the production status event flow with the rules in the constraint knowledge base through a hybrid solver engine.

[0057] The MES execution feedback module issues guiding scheduling instructions, monitors production progress according to the guiding schedule, and senses the progress of the project in real time through the production site sensing module.

[0058] The performance analysis module periodically analyzes changes in project progress, generates on-time delivery rates for each work segment, and adjusts the weights of basic production factors for the next cycle based on these on-time delivery rates.

[0059] Specifically, the overall architecture of the APS-based ship production dynamic scheduling system consists of a production site perception module, a data processing module, an APS optimization scheduling module, a MES execution feedback module, and a performance analysis module, forming a closed-loop control system from production data collection to scheduling optimization and execution feedback. The production site perception module is user-defined, with physical devices such as sensor groups, human-machine interaction terminals, and data acquisition nodes deployed in each work section. The sensor groups are configured collaboratively according to process logic to collect real-time information on equipment operation, process completion, and personnel operations. The data processing module, deployed through industrial servers or edge computing nodes, performs logical recognition and formatting processing on the collected signals, generating a structured production status event stream. The APS optimization scheduling module, implemented by an industrial computer or dedicated scheduling server, integrates a constraint manager, solver, and constraint knowledge base. The knowledge base stores various production constraint rules based on a domain-specific language (DSL) and, combined with preset basic production factor weights, generates guiding scheduling instructions in real-time through a hybrid solver engine. The MES execution feedback module issues scheduling instructions and monitors execution progress through production management terminals, industrial control networks, and execution control units, while simultaneously receiving real-time data from the perception module to achieve synchronized adjustments between planning and execution. The performance analysis module utilizes a database and analysis engine to periodically calculate the on-time delivery rate and schedule deviation for each work section and automatically corrects the weight parameters in the APS module to achieve dynamic optimization. Through the above system structure and module coordination, real-time monitoring and adaptive scheduling of the ship production process can be realized, improving resource utilization and the accuracy of production plans, reducing process waiting time and production delays, thereby significantly improving the on-time delivery rate and production efficiency of the overall project.

[0060] Understandably, this APS-based ship production dynamic scheduling system incorporates various types of processors and devices to support real-time computing, data acquisition, and intelligent scheduling. The system core consists of an industrial-grade central processing unit (CPU) server cluster, edge computing nodes, on-site sensing and control units, human-machine interface terminals, data storage and knowledge base servers, and network and communication equipment. The industrial-grade CPU server cluster runs the APS optimized scheduling module, data processing module, and performance analysis module. It employs multi-core x86 or ARM architecture CPUs with a clock speed of 3.0GHz or higher, possessing high I / O throughput and memory bandwidth capabilities. It can be configured as a dual-machine redundant or distributed computing architecture and deployed in enterprise data centers or local server rooms to provide a highly reliable computing environment. Edge computing nodes are deployed in shipyards, section workshops, or work areas to perform preliminary cleaning, formatting, and status identification of sensor data. They utilize low-power industrial computers as hardware, equipped with Intel i7, Ryzen, or ARM Cortex-A series processors, 8 to 32GB of memory and solid-state storage, supporting Ethernet and serial communication, and possessing Docker or lightweight containerization deployment capabilities for real-time data identification logic execution. Field sensing and control units primarily collect sensor signals such as temperature, displacement, pressure, RFID, and images, and execute some process logic. They employ industrial-grade microcontrollers or PLC processing units, such as ARM Cortex-M, STM32, TI DSP, Siemens S7, or Mitsubishi Q series, supporting Modbus-TCP, PROFINET, EtherCAT, or CAN bus communication protocols. Human-machine interface terminals serve as workstation operation devices, used by operators to input production status, sign off, and provide feedback. They employ ARM or Intel embedded processors, with built-in touch display and wireless communication modules, and can run lightweight Linux or Windows IoT systems. The data storage and knowledge base server is used to store structured event streams, DSL constraint rules, scheduling results, and performance data. It is configured with a high-performance CPU, RAID disk array, and SSD caching layer, supporting relational and time-series databases to ensure high-reliability data access and query performance. Network and communication equipment constitute the system's communication infrastructure, employing industrial Ethernet switches, 5G industrial routers, and fiber optic ring network nodes to achieve high-speed and reliable connections from the perception layer to the data center, supporting QoS and link redundancy mechanisms. Through the coordinated deployment of the aforementioned hardware processing units and communication equipment, the system can achieve real-time data acquisition, dynamic scheduling calculation, and closed-loop feedback execution in the ship production process, ensuring the stability of scheduling optimization and the real-time nature of on-site response.

[0061] In some embodiments of this application, the production site sensing module, the site sensing layer, is equipped with a sensor group in each section, and each section group includes a first sensor group and a second sensor group, wherein:

[0062] The first sensor group is a human-machine interaction terminal that receives information from personnel's active reports or behaviors, including electronic signatures, selection options, and input boxes, and is used to subjectively perceive the production progress of personnel in the production process.

[0063] The second sensor includes an equipment IoT module, an environmental sensor, a tide level gauge, an RFID / BLE positioning system, a face / fingerprint attendance system for workers, and a vision sensor, used to objectively monitor the spatial scheduling and resource occupancy of equipment in the production environment.

[0064] Specifically, the production site sensing module achieves multi-dimensional perception of the shipbuilding production process by deploying sensor groups in each work section. Each work section is equipped with a first sensor group and a second sensor group to form a subjective and objective perception system. The first sensor group mainly consists of a human-machine interface terminal, including an electronic signature interface, operation selection options, and input boxes, used to collect information on personnel's proactive behavior and status feedback during the production process. Operators can use the terminal to input process completion status, workstation changes, quality confirmations, or anomaly reports, thereby achieving both subjective perception and manual confirmation of production progress. The second sensor group is used to acquire objective on-site operational status data. This includes IoT modules deployed on key equipment and tooling to collect data on equipment status, power, vibration, and location during operation; temperature, humidity, and air pressure sensors installed in the workshop or dock environment to monitor construction environmental conditions; tide gauges installed at the dock or dock location to obtain real-time water level changes to assist production scheduling; RFID or BLE positioning devices distributed in material, component, and work areas to determine the spatial location and movement path of workpieces, equipment, and personnel; facial recognition terminals or fingerprint attendance devices configured in worker work groups to record attendance status and work group distribution; and industrial cameras and depth vision sensors for visual inspection to identify the actual progress and resource occupancy of welding, assembly, or hoisting operations. Through the coordinated deployment of the above sensor groups, the system can simultaneously acquire subjective input from personnel and objective monitoring data from equipment, forming a multi-source fusion on-site perception network. This provides structured production status information input for subsequent data processing modules, enabling real-time accurate identification of production status and providing a data foundation for dynamic scheduling.

[0065] In this embodiment of the application, the production site sensing module is constructed using industrial-grade equipment. The human-machine interface terminal is either Advantech TPC-1571H or Siemens Simatic IPC377E industrial touch screen; equipment monitoring uses a Siemens IoT2040 gateway, Beckhoff CX5130 controller with Schneider PM5350 power monitor and IFM VSA004 vibration sensor; environmental monitoring uses a Vaisala PTB330 pressure transmitter, Sensirion SHT35 temperature and humidity sensor and TSI DustTrakDRX industrial dust monitor; tidal level monitoring uses a Valeport TideMaster or OTT RLS radar level gauge; the positioning system includes a Zebra FX9600 RFID reader, Impinj R420 antenna and Ubisense UWB positioning system; the personnel identification terminal uses Hikvision DS-K1T341A or Suprema FaceStation 2; and the visual inspection equipment includes a Baslerace U industrial camera and a Basler... The system utilizes a blaze-101 depth camera and an NVIDIA Jetson Xavier NX edge computing unit. For data and networking, it employs a Siemens SCALANCE X204-2 switch, a Huawei AR502H 5G industrial router, an Advantech UNO-1372G edge node, and a Dell PowerEdge R750 central server. All equipment features industrial-grade protection and anti-interference capabilities, adapting to the high humidity, high dust, and strong electromagnetic environments of shipbuilding production sites, enabling stable acquisition and dynamic scheduling of multi-source data.

[0066] See Figure 2 As shown, in some embodiments of this application, the production site sensing module has pre-set collaborative production logic for each sensor group according to the production process, including:

[0067] Each shipbuilding process X is decomposed into several stages, each stage comprising multiple standardizable and repeatable work units Xi. Each work unit Xi is configured with unit constraints, and each work unit Xi is defined by worker group NR, process time GS, and equipment code group SB.

[0068] According to the production process, there is a corresponding task node for each traversal to the i-th work unit Xi, and each task node is configured with node constraints.

[0069] Node constraints and cell constraints are used to semantically align physical perception with progress status in digital production planning, generating a structured flow of production status events, where each event is represented by... This indicates, including:

[0070] ,in ;

[0071] in, It is the i-th work unit; It is the ship's registration number; This refers to the i-th task node in stage m; Di is the work unit identifier; Di is the task node; Si is the timestamp; Si is the confidence level.

[0072] In this embodiment of the application, when the work unit When it belongs to a tide-sensitive process, confidence level Generated in the following way:

[0073] in:

[0074] This represents the basic confidence level obtained based on the fusion of multi-source sensor data;

[0075] This is the tidal compliance factor, calculated based on whether the current moment falls within a preset legal operating window;

[0076] like If it is not a tide-sensitive process, then .

[0077] The legal operating window is predefined by the shipbuilding process. For example, section hoisting must be carried out within 30 minutes before and after high tide, and underwater welding must be carried out within 60 minutes before and after low tide.

[0078] This mechanism allows for consideration of environmental constraints during the event generation phase, effectively filtering out failure start-up events caused by unmet tidal conditions, and improving the accuracy and reliability of production status identification.

[0079] In this embodiment of the application, based on the tidal compliance factor It also has a preset monthly / seasonal correction factor. This leads to a more accurate confidence model;

[0080] Monthly Correction Factor ,include:

[0081] Monthly Correction Factor It can also be set based on historical meteorological data or shipyard experience.

[0082] The confidence formula is:

[0083] in, This refers to the compliance factor based on the tidal window of the day; This refers to the seasonal / weather risk correction factor based on the current calendar month.

[0084] Understandably, by further refining ship production process X into multiple stages and establishing a structure within each stage composed of standardized and repeatable work units Xi, the complex ship production process can be modularized and digitally expressed. Traditional ship production management typically uses "sections" or "processes" as units, with coarse-grained scheduling. The system can only allocate resources based on the overall process progress, failing to reflect the temporal dependencies and resource consumption of different links within the same process. This leads to problems such as planning delays, opaque bottleneck positions, and low resource utilization. In this solution, each work unit Xi is defined as an independently describable production behavior unit, configured with corresponding unit constraints to characterize the sequential dependencies and conflict conditions of the operation in terms of time, resources, space, or process logic.

[0085] Furthermore, each work unit Xi is defined through worker group NR, process time GS, and equipment code group SB, forming a three-element production element mapping of "people-time-equipment". Here, worker group NR represents a set of personnel with specific skill levels or job responsibilities; process time GS reflects the expected duration or allowable fluctuation range of the work unit under standard operating conditions; and equipment code group SB indicates the equipment, tooling, or workstation codes required to perform the operation. The core advantage of this is that it can transform ship production activities into optimizable units driven by standardized constraints in APS scheduling calculations, achieving dynamic rearrangement and accurate prediction based on resource load and process logic. Compared with existing technologies, traditional shipbuilding production planning often relies on experience or static templates, failing to achieve real-time constraint adjustment and inter-process linkage optimization. This invention, through explicit constraint modeling of work unit Xi, enables the system to achieve fine-grained calculation and real-time scheduling of human and machine resources at the data layer, thereby significantly improving production flexibility and plan execution accuracy.

[0086] Understandably, typical implementations of Advanced Planning and Scheduling (APS) systems in manufacturing industries, including shipbuilding, automotive, and heavy equipment, often use "processes" or "work orders" as the basic scheduling unit, resulting in a relatively coarse-grained structure. Traditional APS models often treat a production task, such as "segment closure," as a single operational node, only associating it with general resource types, such as "requiring 1 crane + 2 welders," and estimated working hours, lacking a refined breakdown of the internal structure of the operation. Resource allocation is mostly based on static rules such as earliest available time, fixed binding, or simple priority queues, and resources such as workers and equipment are often abstracted into a homogeneous pool, without distinguishing skill levels, equipment performance, or scarcity differences. This leads to a disconnect between planning and execution, making it difficult to cope with the highly complex, strongly constrained, and highly volatile real-world scenarios in shipbuilding. In contrast, this embodiment systematically decomposes each ship production process X into several logical stages (such as prefabrication, assembly, welding, and inspection), and each stage is further refined into multiple standardized and repeatable work units Xi. Each work unit Xi is precisely defined by worker group NR, process time GS, and equipment code group SB, forming the smallest execution unit integrating process, resources, and time. That is, the work unit Xi of this application can directly correspond to actual workstation operations such as CO2 fillet welding at the third rib on the port side. Its standardized attributes facilitate process reuse, quality traceability, and digital twin mapping. The triple definition provides a structured input basis for subsequent resource binding verification, confidence calculation, and dynamic scheduling, achieving a good balance between complexity and controllability.

[0087] In some embodiments of this application, the APS optimization scheduling module includes sensor group perception verification rules configured for each unit constraint and node constraint:

[0088] Node constraints are used to manage task lifecycle events, including triggering and completion, which are achieved through a joint mapping with data from the first and second sensor groups;

[0089] Unit constraints are used to ensure compliance and resource accuracy during execution, including process monitoring and resource binding, which are achieved through data mapping with the second sensor group.

[0090] In some embodiments of this application, node constraints are used to manage task lifecycle events, including triggering and completion for determining task nodes. Whether it can be triggered or completed, i.e., the lifecycle events of the management task, include:

[0091] ;

[0092] in, It represents node constraints, where N is the node identifier; It is the triggering condition. It is the manual operation readiness status collected by the first sensor group; It is the completion status of the preceding task node detected by the second sensor group; It is a condition for completion. The status of the manual operation is collected by the first sensor group. This is the state where all equipment actions have ended, as detected by the second sensor group.

[0093] In some embodiments of this application, unit constraints are used to ensure compliance and resource accuracy during execution, including process monitoring, determining whether the current operation is within the normal range, and whether the data collected in real time by the second sensor group meets the process parameter standards, i.e., the operation unit. All associated sensors at the event timestamp All collected measurements must fall within their corresponding preset process window, where process monitoring is represented as follows:

[0094] ;

[0095] in, It refers to process monitoring within unit constraints; Uk is the rule type identifier. Work unit The start time, Work unit End time, This indicates any time within the work unit; For the first One sensor in Physical quantities collected within the period; This is the preset process window corresponding to the j-th second sensor; This indicates a logical AND operation, meaning that all conditions must be true simultaneously.

[0096] It is understandable that the real-time acquisition and judgment of process parameters upon which process monitoring relies is achieved through existing mature industrial sensing and data processing technologies. For example, devices such as Schneider Electric power monitors, IFM vibration sensors, and Basler industrial cameras are used to acquire physical quantities such as temperature, current, vibration, and displacement of the work unit during execution, and this data is then connected to edge computing nodes for analysis. The aforementioned sensor deployment and data acquisition methods are well-known technologies in the field, and related hardware has been widely used in complex industrial scenarios such as shipbuilding. The focus of this embodiment is not on how to acquire raw measurement data, but on dynamically comparing the acquired multi-source sensor data with a preset process window to determine whether the current operation is running within the normal range. This determination result serves as the basis for process compliance in unit constraints, thereby affecting the credibility of scheduling decisions and the priority of resource allocation. Therefore, those skilled in the art can select existing mature solutions based on actual working conditions for specific sensor types and data acquisition methods, and there is no need to elaborate on them in this application.

[0097] In some embodiments of this application, unit constraints are used to ensure compliance and resource accuracy during execution, including resource binding to ensure that the resources used to execute the job are consistent with the plan:

[0098] ;

[0099] in, This refers to resource binding within a unit constraint; Ub is the resource binding identifier. It is the actual worker attendance data obtained from the second sensor. yes A work unit allows for the group of workers to perform that process. This is the actual equipment usage information obtained from the second sensor. yes The set of device code groups that a work unit is allowed to use;

[0100] If the resource binding conditions are not met, it is determined that resources are missing. A hierarchical priority scheduling mechanism is then activated based on the weight of basic production factors. Resources are scheduled hierarchically according to the unit's priority to ensure the priority execution of higher-level work units.

[0101] It is understandable that unit constraints are a key mechanism to ensure the alignment of physical execution with digital plans; resource binding checks confirm whether the actual worker groups and equipment used conform to the preset plan. The personnel identification and equipment usage status perception upon which resource binding verification relies can be achieved through existing mature industrial sensing and identification technologies, such as using RFID, facial recognition, or UWB positioning to obtain the actual worker groups participating in the operation, and using equipment IoT interfaces or power monitoring to obtain the actual activated equipment codes. The above data collection methods are well-known technologies in the field, and related hardware (such as facial recognition terminals, equipment gateways, sensors, etc.) has been widely deployed in the manufacturing industry. The focus of this embodiment is not on how to collect this raw data, but on logically comparing the collected results with a preset set of allowed resources, and using the comparison results as a key criterion for unit constraints, thereby driving dynamic scheduling decisions. Therefore, those skilled in the art can directly use existing technologies to complete the specific implementation of data collection, and it is unnecessary to elaborate on it in this application.

[0102] In some embodiments of this application, the APS optimization scheduling module also presets basic production factor weights, and generates guiding scheduling instructions by combining the production status event flow and rules in the constraint knowledge base through a hybrid solution engine, including:

[0103] The APS optimization scheduling module also has preset basic production factor weights. Through a hybrid solution engine, it combines the production status event flow with rules from the constraint knowledge base to generate guiding scheduling instructions, including:

[0104] The weights of basic production factors include assigning a value to the urgency of each project; assigning a value to the scarcity of materials based on the type of equipment, precision, and processing cost involved in each production segment; and assigning a value to the scarcity of labor based on the type of work, skill level, and training cost involved in each production segment.

[0105] Combining the rules in the production status event flow and constraint knowledge base, including: if the event flow determines that the resource binding condition is not met, it is determined that the resource is missing. When allocating resources to any work unit, the nearest idle resource or the resource with the lowest priority is selected from the available resource pool that meets the process constraints to generate the optimal boundary value of the candidate resources, and the candidate resources are sorted and filtered in order from the optimal boundary value.

[0106] Specifically, priority is given to retrieving the nearest available resource or the resource with the lowest priority, in order to prevent high-value resources from being occupied by low-level tasks, including:

[0107] From the available resource pool, the priority of the filterable resource pools is divided according to the matching degree. The filtering formula for the resource with the lowest priority and the closest distance from the specific project is as follows:

[0108] ;

[0109] in, It is the optimal boundary value of the candidate resource. :resource To the work unit The distance; The value is assigned based on the scarcity of the resource required; the smaller the value, the more common the resource. A balance coefficient between distance and priority; This is the key logic for ensuring that resource capabilities match task requirements. Indicates the skill type of resource r. Work unit Required skills.

[0110] Specifically, This is a configurable balancing coefficient used to adjust the weight between proximity principle and resource economy (e.g., λ=0.7 in large dry docks, emphasizing distance; λ=0.3 in precision outfitting areas, emphasizing resource matching) and varies according to the actual situation.

[0111] Understandably, for example, when both welding machines (one high-end laser welding machine with p=0.9 and one ordinary welding machine with p=0.3) can complete a non-critical welding task, the system will automatically select the ordinary welding machine, even if it is slightly farther away; however, if the task is a critical weld for hull closure, the required p value will be larger, and the laser welding machine will be selected.

[0112] In some embodiments of this application, the APS optimization scheduling module generates optimal boundary values ​​for candidate resources by prioritizing the nearest idle resource or prioritizing the resource with the lowest priority, and then sorts and filters the candidate resources sequentially from their optimal boundary values, including:

[0113] Resource gaps will be quantified based on material scarcity assignment and the rules governing it. Priority allocation will be based on the matching degree between resource gaps and the available resource pool. The first priority allocation rule removes resources from all tasks of the same importance. The second priority allocation rule matches the production resource with the highest matching degree from the idle resource pool. The third priority allocation rule matches the production resource with the highest matching degree from a lower-level production task. The fourth priority allocation rule further quantifies the resource gaps generated after scheduling according to the third priority allocation rule and allocates production resources based on this second quantification. The fifth priority allocation rule quantifies the resource gaps generated after scheduling according to the fourth priority allocation rule a third time and allocates production resources based on this third quantification. The sixth priority allocation rule quantifies the resource gaps generated after scheduling according to the fifth priority allocation rule a fourth time, assesses the impact of the fourth quantification on the cycle of the task at this level, and generates a guiding scheduling instruction permission report for submission and approval.

[0114] Understandable The optimal boundary value for candidate resources is used to filter specific types of resource needs, while the priority allocation rule is the process of extracting and allocating these resources from projects level by level. Based on project urgency, task level, and resource weight, resources are dynamically extracted and reallocated from low-priority projects or non-critical paths. This mechanism realizes a three-layer collaborative logic: "type matching is the prerequisite, economic optimization is the strategy, and hierarchical scheduling is the execution path," thereby maximizing overall resource utilization efficiency while ensuring the supply of resources for high-level tasks. In existing technologies, Advanced Planning and Scheduling (APS) systems have been applied in discrete manufacturing fields such as shipbuilding and heavy industry, and their typical implementations are mostly based on static rules or simple heuristic algorithms for resource allocation. For example, traditional APS typically employs strategies such as "first-come, first-served," "earliest available time first," or "fixed resource binding," selecting the first available device or personnel that meets the type requirements from the resource pool when a task is triggered. This lacks dynamic consideration of resource scarcity, project urgency, and task hierarchy. While some advanced systems introduce priority queues, they only support single-level scheduling—meaning that in case of resource conflicts, adjustments are made only between tasks of the same level, without the ability to backtrack allocated resources across projects or levels. Furthermore, there is no multi-round quantitative assessment or progressive rescheduling mechanism for resource shortages. In addition, existing APS generally treat resources as "homogeneous units," failing to distinguish the economic value differences between high-end and ordinary welding machines, or between senior technicians and auxiliary workers. This results in high-value resources often being occupied by low-priority tasks, causing overall efficiency losses.

[0115] In contrast, the APS optimization scheduling module in this embodiment achieves a fundamental breakthrough in technical architecture and scheduling logic. First, the system not only defines resource type compatibility but also differentiates each resource through material scarcity assignment and manual scarcity assignment. Second, in resource shortage scenarios, instead of simply reporting errors or making local adjustments, it constructs a six-level progressive priority allocation rule: from removing redundant resources from tasks of the same level, to matching the optimal item in the idle pool, and then borrowing resources from lower levels step by step. Furthermore, it performs secondary, tertiary, and even quaternary quantitative assessments on the new gaps created after each borrowing, ultimately generating guiding scheduling instructions requiring approval based on the degree of impact on the task cycle. This mechanism is essentially a flexible resource rebalancing strategy with a feedback loop, its core being "step-by-step backtracking, dynamic relocation, and controllable impact."

[0116] In some embodiments of this application, the performance analysis module adjusts the weights of basic production factors for the next cycle based on the on-time delivery rate, including:

[0117] For projects that have been deferred due to resource gaps caused by the sixth priority allocation rule, the on-time delivery rate of the deferred projects is used to determine the degree of impact on the project's deferred status during actual construction. Based on the degree of impact, the urgency assignment of the deferred projects is adjusted in the next cycle, and the deferred projects are prohibited from appearing in the candidate pool of the sixth priority allocation rule.

[0118] Specifically, On-Time Delivery Rate (OTDR) is defined as the number of vessels (or sections / nodes) actually completed and delivered on the contractually agreed or planned completion dates within a statistical period. The general formula for its calculation is:

[0119] In practical applications, "on-time completion" allows for a reasonable window of adjustment; exceeding this window is considered a delay. For example, if a shipyard plans to deliver four ships per quarter, with three of them due by the contractually agreed date... If a vessel completes undocking or sea trials within one day, and one vessel experiences a significant delay, then its on-time delivery rate is [percentage missing]. This indicator is widely used to measure a shipyard's planning execution capabilities, supply chain coordination efficiency, and production stability, and is one of the core performance parameters in the international shipbuilding industry's KPI system.

[0120] In summary, this embodiment provides an APS-based dynamic scheduling system for ship production. It generates a structured event flow through multi-source sensor fusion and pre-set production logic, and utilizes node and unit constraints to achieve task lifecycle management and execution compliance verification. When resources are scarce, it constructs task priorities by assigning values ​​based on material and labor scarcity, and prioritizes the use of the nearest available resource with the lowest priority during scheduling. This is supplemented by a six-level progressive gap response mechanism and a weighted self-optimization closed loop based on on-time delivery rate, effectively ensuring the priority execution of high-level tasks, improving the utilization efficiency of high-value resources, and significantly enhancing the planning stability and scheduling flexibility of the shipbuilding process.

[0121] Furthermore, the intelligent scheduling scheme for ship production proposed in this embodiment has the following advantages compared to existing technologies:

[0122] First, the solution innovatively incorporates external environmental factors such as tidal levels and seasonal weather into the production event confidence calculation model. Through dynamic correction factors (such as monthly risk coefficients and real-time tidal compliance), it effectively filters out "failure to start" events caused by environmental infeasibility, significantly improving the realism of status perception and the reliability of scheduling decisions. Second, the system constructs a two-layer verification mechanism of "unit constraints" and "node constraints": unit constraints verify the resource binding of the actual participating worker groups and equipment codes at the physical execution level, ensuring that operations comply with process specifications; node constraints control the triggering and completion conditions of tasks at the planning logic level, achieving a two-way closed-loop alignment between digital planning and physical execution. Based on this, the solution proposes a dynamic scheduling strategy based on "resource economy priority." By minimizing the objective function weighted by distance and scarcity, it prioritizes the use of ordinary or nearby resources while satisfying process compatibility, avoiding the occupation of high-value equipment and scarce personnel by low-level tasks, thereby optimizing overall resource utilization efficiency. When resource shortages occur, the system further activates a "hierarchical priority scheduling mechanism." Based on the urgency of the project and the task level, resources are elastically backtracked and reallocated from non-critical paths to ensure the priority execution of high-level work units. This capability relies on the deep fusion of multi-source heterogeneous sensors (including UWB positioning, RFID tracking, power monitoring, industrial vision, and tide gauges), generating high-fidelity, confidence-based structured event streams through edge computing. This enables the entire scheduling system to possess closed-loop intelligent characteristics of strong perception, accurate judgment, and rapid response, fundamentally solving core pain points in traditional shipbuilding such as the disconnect between planning and execution, severe resource misallocation, and weak anti-disturbance capabilities.

[0123] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0124] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0125] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0126] These computer program instructions can also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. An APS-based dynamic scheduling system for ship production, characterized in that, include: The production site perception module is defined by the user, including production process, production plan and site perception layer. The site perception layer is equipped with sensor groups in each section, and the production logic of the sensor groups is preset to coordinate and associate according to the production process. The data processing module identifies the production status of each work segment according to the production logic and generates a structured production status event stream for each work segment according to the production logic. The APS optimization scheduling module integrates a constraint manager and a solver. The constraint manager is connected to a constraint knowledge base, which stores production rules written in a domain-specific language (DSL). The APS optimization scheduling module also presets basic production factor weights and generates guiding scheduling instructions by combining the production status event flow with the rules in the constraint knowledge base through a hybrid solver engine. The MES execution feedback module issues guiding scheduling instructions, monitors production progress according to the guiding scheduling instructions, and senses project progress in real time through the production site sensing module. The performance analysis module periodically analyzes changes in project progress, generates on-time delivery rate for each work segment, and adjusts the weights of basic production factors for the next cycle based on the on-time delivery rate. The production site sensing module, the site sensing layer, has a sensor group set in each section, and each section group includes a first sensor group and a second sensor group, wherein: The first sensor group is a human-machine interaction terminal that receives information from personnel's active reports or behaviors, including electronic signatures, selection options, and input boxes, and is used to subjectively perceive the production progress of personnel in the production process. The second sensor includes an equipment IoT module, an environmental sensor, a tide level gauge, an RFID / BLE positioning system, a face / fingerprint attendance system for workers, and a vision sensor, which are used to objectively perceive the spatial scheduling and resource occupancy of the equipment in the production environment. The production site sensing module has pre-set collaborative production logic for each sensor group based on the production process, including: Each ship production process X is decomposed into several stages. Each stage includes multiple standardized and repeatable work units Xi. Each work unit Xi is configured with unit constraints, and each work unit Xi is defined by worker group NR, process time GS, and equipment code group SB. According to the production process, there is a corresponding task node for each traversal to the i-th work unit Xi, and each task node is configured with node constraints. The node constraints and unit constraints are used to semantically align physical perception with the progress status in the digital production plan, generating a structured production status event stream, where each event is represented by... This indicates, including: wherein ; in, It is the i-th work unit; It is the ship's registration number; This refers to the i-th task node in stage m; Di is the work unit identifier; Di is the task node; It is a timestamp; Si is the confidence level; The APS optimization scheduling module includes sensor group perception verification rules configured for each unit constraint and node constraint: The node constraints are used to manage the lifecycle events of the task, including triggering and completion, wherein triggering and completion are achieved through a joint mapping with data from the first sensor group and the second sensor group; The unit constraints are used to ensure compliance and resource accuracy during the execution process, including process monitoring and resource binding, wherein process monitoring and resource binding are achieved through data mapping with the second sensor group; The node constraints are used to manage task lifecycle events, including triggering and completion events used to determine task nodes. Whether it can be triggered or completed, i.e., the lifecycle events of the management task, include: ； in, It represents node constraints, where N is the node identifier; It is the triggering condition. It is the manual operation readiness status collected by the first sensor group; It is the completion status of the preceding task node detected by the second sensor group; It is a condition for completion. The status of the manual operation is collected by the first sensor group. This is the state where all equipment actions have ended, as detected by the second sensor group. The unit constraints are used to ensure compliance and resource accuracy during execution, including process monitoring to determine whether the current operation is within the normal range, and whether the data collected in real time by the second sensor group meets the process parameter standards, i.e., the operation unit. All associated sensors at the event timestamp All collected measurements must fall within their corresponding preset process window, whereby process monitoring is referred to as: ； in, It refers to process monitoring within unit constraints; Uk is the rule type identifier. Work unit The start time, Work unit End time, This indicates any time within the work unit; For the first One sensor in Physical quantities collected within the period; This is the preset process window corresponding to the j-th second sensor; This indicates a logical AND operation, meaning that all conditions must be true simultaneously. The unit constraints are used to ensure compliance and resource accuracy during execution, including resource binding, to ensure that the resources used to perform the job are consistent with the plan: ； in, This refers to resource binding within a unit constraint; Ub is the resource binding identifier. It is the actual worker attendance data obtained from the second sensor. yes A work unit allows for the group of workers to perform that process. This is the actual equipment usage information obtained from the second sensor. yes The set of device code groups that a work unit is allowed to use; If the resource binding conditions are not met, it is determined that resources are missing. A hierarchical priority scheduling mechanism is then initiated based on the weights of the basic production factors. Resources are scheduled hierarchically according to the unit's priority to ensure the priority execution of higher-level work units.

2. The ship production dynamic scheduling system based on APS according to claim 1, characterized in that, The APS optimization scheduling module also presets basic production factor weights. Through a hybrid solution engine, it combines the production status event flow with the rules in the constraint knowledge base to generate guiding scheduling instructions, including: The weights of the basic production factors include assigning a value to the urgency of each project; assigning a value to the scarcity of materials based on the type, precision, and processing cost of equipment involved in each production segment; and assigning a value to the scarcity of labor based on the type of work, skill level, and training cost involved in each production segment. The combination of the production status event flow and the rules in the constraint knowledge base includes: if the event flow determines that the resource binding condition is not met, it is determined that the resource is missing. When allocating resources to any work unit, the optimal boundary value of the candidate resources is generated from the available resource pool that meets the process constraints, prioritizing the nearest idle resource or prioritizing the resource with the lowest priority. The optimal boundary value of the candidate resources is then sorted and filtered sequentially.

3. The ship production dynamic scheduling system based on APS according to claim 2, characterized in that, The APS optimization scheduling module prioritizes retrieving the nearest idle resource or the resource with the lowest priority to generate optimal boundary values ​​for candidate resources, and then sorts and filters these candidate resources sequentially based on their optimal boundary values, including: Resource gaps will be quantified based on the material scarcity assignment and rules, and sorted according to the matching degree between resource gaps and available resources in the pool. The first priority ranking rule is to remove resources from all tasks of the same importance level; the second priority ranking rule is to match the production resource with the highest matching degree in the idle resource pool; the third priority ranking rule is to match the production resource with the highest matching degree at a lower level than the task; the fourth priority ranking rule is to further quantify the resource gaps generated after scheduling according to the third priority ranking rule, and allocate production resources based on this second quantification; the fifth priority ranking rule is to further quantify the resource gaps generated after scheduling according to the fourth priority ranking rule, and allocate production resources based on this third quantification; the sixth priority ranking rule is to further quantify the resource gaps generated after scheduling according to the fifth priority ranking rule, determine the degree of impact of these four quantifications on the cycle of the task at this level, and generate a guiding scheduling instruction permission report for submission and approval.

4. The ship production dynamic scheduling system based on APS according to claim 3, characterized in that, The performance analysis module adjusts the weights of basic production factors for the next cycle based on the on-time delivery rate, including: Obtain the resource gap deferral projects caused by the sixth priority ranking rule. For the on-time delivery rate of the deferral projects, determine the degree of their impact on the project's deferral in actual construction based on the on-time delivery rate. In the next cycle, adjust the urgency assignment of the deferral projects based on the degree of deferral impact and prohibit the deferral projects from appearing in the candidate pool of the sixth priority ranking rule.

Images