An aps scheduling method based on marine engineering equipment manufacturing

Abstract

The application is applied to the field of advanced scheduling of ocean engineering equipment manufacturing, in particular to an APS scheduling method based on ocean engineering equipment manufacturing. The method comprises the following steps: 1) constructing a multi-objective scheduling model based on the ocean engineering equipment manufacturing process; 2) performing scheduling calculation on the multi-objective scheduling model by using an improved ant colony algorithm to obtain an optimal scheduling scheme. The application can play an efficient application in production plan making, production personnel arrangement, transportation equipment scheduling application, emergency project emergency processing, leader decision processing and the like, improve production efficiency and avoid production waste.

Description

Technical Field

[0001] This invention is applied to the field of advanced scheduling in marine engineering equipment manufacturing, specifically an APS scheduling method based on marine engineering equipment manufacturing. Background Technology

[0002] Currently, the production optimization models used in the production of large equipment often have the following problems:

[0003] Single-objective processing: Most scheduling results output only the planned start time and planned end time, which cannot achieve the multi-objective requirements of actual production.

[0004] Slow response speed: Traditional scheduling uses an iterative approach to find the optimal value, which consumes a lot of machine resources, has a slow response speed, and results in a poor user experience;

[0005] Insufficient flexibility: Most scheduling models are geared towards customization and cannot be adjusted to meet user needs in terms of multi-objective results, resulting in insufficient user participation.

[0006] While it closely resembles existing scheduling models, it suffers from limitations such as single-objective processing, slow response speed, and poor flexibility, resulting in unsatisfactory scheduling results and difficulty in achieving practical purposes. Summary of the Invention

[0007] To achieve the above-mentioned goal of solving the existing problems of APS scheduling technology in marine engineering equipment manufacturing, this invention adopts a comprehensive technical solution of "ant colony algorithm driving + multi-dimensional parameter configuration + dynamic bottleneck identification". Specifically, it unfolds from three levels: scheduling model construction, core algorithm optimization, and user configuration interaction, forming a scheduling system that can cover the entire production process of marine engineering equipment manufacturing.

[0008] The technical solution adopted by the present invention to achieve the above objectives is as follows:

[0009] An APS scheduling method based on marine engineering equipment manufacturing includes the following steps:

[0010] 1) Based on the marine engineering equipment manufacturing process, a multi-objective scheduling model is constructed;

[0011] 2) Use the improved ant colony algorithm to perform scheduling calculations on the multi-objective scheduling model to obtain the optimal scheduling scheme.

[0012] The multi-objective scheduling model comprises a data layer, a constraint layer, and an objective layer, wherein:

[0013] The data layer is used to collect and store data from the entire process of marine engineering equipment manufacturing in real time, and to interface with the PLC and MES systems on the production site through the OPC UA protocol, while ensuring that the data collection frequency meets the threshold.

[0014] The constraint layer is used to define the boundary conditions for scheduling calculation, which are used as constraints during calculation. It is associated with the user configuration module through a parameterized configuration interface and supports real-time updates of constraint weights.

[0015] The target layer is used to set targets, including production efficiency targets, resource utilization targets, and cost control targets. The weights of each target are dynamically adjusted through the user configuration module and used as optimization targets for improving the fitness function of the ant colony algorithm.

[0016] The data for the entire process of marine engineering equipment manufacturing includes static data and dynamic data. The static data includes equipment parameters, personnel skill matrix, and production process standards, while the dynamic data includes production progress, equipment operating status, and personnel on-duty status.

[0017] The constraints include hard constraints and soft constraints. Hard constraints are production rules that must be met, while soft constraints are rules whose priority can be adjusted according to user needs.

[0018] The improved ant colony algorithm is specifically as follows:

[0019] Initialize the ant colony and pheromone matrix: Treat the scheduled tasks as "ant colony foraging targets" and the combination of equipment and personnel as "path nodes". Set the initial number of ant colonies and set the initial value of the pheromone matrix based on the historical utilization rate of equipment and the matching degree of personnel skills.

[0020] Dynamic bottleneck identification and path pruning: During the ant colony optimization process, the load rate of each node is calculated in real time. When the load rate of a node is greater than the threshold, it is identified as a bottleneck node and path pruning is immediately triggered to prevent subsequent ants from choosing unnecessary paths to that node. At the same time, the bottleneck node information is fed back to the data layer to update the dynamic data of equipment / personnel.

[0021] Memory optimization and pheromone updating: A "tacit list + elite ant" mechanism is introduced. The tacit list is used to record the nodes that the ants have visited. The 5%-10% of ants with the best objective function value in each iteration are selected as "elite ants". The paths they have visited have an additional 2 times the pheromone concentration. At the same time, the 3 paths with the smallest objective function value in the previous iterations are retained as the initial guiding paths for the next iteration. The pheromone updating adopts a combination of "local update + global update". The local update is executed after the ant completes the single-step path selection, and the global update is executed after all ants have completed one iteration.

[0022] Multi-objective solution output and adjustment: The algorithm terminates when the fluctuation of the objective function value is less than a threshold or the number of iterations reaches a preset upper limit. After the iteration ends, the output is a set of multi-objective Pareto optimal solutions containing scheduling schemes under different objective weights.

[0023] The load rate is specifically:

[0024] Load rate = Node assigned task duration / Node available time.

[0025] It also includes human-computer interaction steps, specifically:

[0026] Get the optimization target weights set by the user in real time, and display the target priority ranking after the weight adjustment in real time;

[0027] The system acquires user-added, modified, and deleted constraint rules in real time, automatically verifies the legality of the rules (i.e., whether they conflict with hard constraints), and updates the constraint layer parameters accordingly.

[0028] To address the needs of sudden emergency projects in marine engineering equipment manufacturing, users can upload emergency task information through the emergency task insertion interface and select one of three insertion strategies: "prioritize insertion during idle periods", "adjust existing task order", or "increase personnel / equipment input". Based on the current scheduling plan, the system quickly calculates the optimal schedule after insertion using the ant colony algorithm, and also indicates the impact of insertion on existing orders.

[0029] An APS scheduling system based on marine engineering equipment manufacturing includes:

[0030] The model building module is used to build multi-objective scheduling models based on the marine engineering equipment manufacturing process.

[0031] The scheduling calculation module is used to perform scheduling calculations on a multi-objective scheduling model using an improved ant colony algorithm to obtain the optimal scheduling scheme.

[0032] It also includes a user configuration module, specifically including:

[0033] The target weight configuration interface is used to dynamically adjust and optimize the target weight using a slider-style interactive method.

[0034] The constraint rule configuration interface is used to add, modify, and delete constraint rules, and automatically validates the legality of the rules.

[0035] The emergency task insertion interface is used to receive emergency task information uploaded by users using three insertion strategies. It quickly calculates the optimal schedule after insertion using the ant colony algorithm, and also indicates the impact on existing orders. The three insertion strategies are: prioritizing insertion during idle periods, adjusting the order of existing tasks, and increasing personnel / equipment input.

[0036] The present invention has the following beneficial effects and advantages:

[0037] Compared with existing APS scheduling technology for marine engineering equipment manufacturing, the present invention, through the above technical solution, achieves significant improvements in three core dimensions: multi-objective processing capability, response speed, and flexibility. The specific beneficial effects are as follows:

[0038] 1. Enhanced multi-target processing capabilities to meet actual production needs.

[0039] Existing technologies often focus solely on the single objective of "order delivery time," failing to consider factors such as personnel, equipment, and cost. This invention constructs a multi-objective model encompassing "production efficiency, resource utilization, and cost control," incorporating configurable objective weights to output scheduling solutions tailored to different production scenarios. Experimental verification shows that, in the production scheduling of a certain component for marine engineering equipment, adopting this solution yields the following results:

[0040] On-time delivery rate has increased from 90% with existing technology to 95%, and for urgent orders, the on-time delivery rate can reach 99%.

[0041] Overall equipment efficiency (OEE) increased from 90% to 95%, with idle time for key equipment such as welding robots and large cranes reduced by 60%.

[0042] The total production cost was reduced by 10%, mainly due to the reduction in overtime hours (an average reduction of 80 hours per person per month) and the reduction in equipment energy consumption (due to load balancing, the energy consumption of equipment during idling was reduced by 25%), effectively solving the problem of "single-objective scheduling being out of sync with actual production needs" in existing technologies.

[0043] 2. Significantly faster response time and reduced resource consumption.

[0044] Existing technologies employ traditional iterative algorithms, which often take over 2 hours to compute when handling multi-task, multi-constraint scheduling of marine engineering equipment, and consume significant server resources. This invention optimizes this process through "dynamic bottleneck identification and pruning" and "memory optimization."

[0045] The calculation time is reduced to less than 10 minutes, which is 75% better than the existing technology, meeting the need for "real-time adjustment" of production plans;

[0046] The server CPU utilization rate dropped from 85% to 40%, and memory usage decreased by 50%, avoiding production system lag caused by algorithm calculations, improving the user experience, and solving the problems of "slow response and high resource consumption" in existing technologies.

[0047] 3. Increased flexibility and user engagement, adapting to personalized needs.

[0048] Existing technologies are mostly customized scheduling models, which users cannot adjust target priorities or constraint rules, making them difficult to adapt to the characteristics of "variable order types and dynamic adjustments to production needs" in marine engineering equipment manufacturing. This invention addresses this through a user-configurable interaction module:

[0049] It allows users to adjust target weights and constraint rules within 10 minutes without modifying the underlying algorithm code, and is suitable for different scenarios such as "prioritizing urgent orders" and "low load during equipment maintenance".

[0050] Emergency task insertion response time is ≤5 minutes, which is 95% faster than existing technology (which requires manual rescheduling and takes more than 2 hours). It can also automatically indicate the impact on existing orders, assist users in decision-making, and solve the problems of "poor flexibility and low user participation" of existing technology. Attached Figure Description

[0051] Figure 1 Flowchart of the improved ant colony algorithm in this invention. Detailed Implementation

[0052] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0053] An APS scheduling method based on marine engineering equipment manufacturing includes the following steps:

[0054] 1. Construction of Multi-Objective Scheduling Model

[0055] To address the need for multi-element coordination among "personnel, equipment, and plans" in marine engineering equipment manufacturing, a three-layer scheduling model is constructed, comprising an objective layer, a constraint layer, and a data layer. The functions and relationships of each layer are as follows:

[0056] Data Layer: Serving as the foundation of the model, this layer integrates data from the entire manufacturing process of marine engineering equipment, including both static and dynamic data. Static data encompasses equipment parameters (such as the rated lifting capacity of cranes and the working efficiency of welding robots), personnel skill matrices (such as welder certification levels and assembler proficiency), and production process standards (such as component welding time and coating curing cycle). Dynamic data is collected in real-time, including production progress (such as component processing completion rate), equipment operating status (such as whether there is a malfunction and idle time), and personnel on-duty status (such as attendance and job transfer information). It interfaces with the PLC and MES systems on the production site via the OPC UA protocol, ensuring a data collection frequency of at least once per minute, providing real-time and accurate data support for scheduling calculations.

[0057] Constraint Layer: Defines the boundary conditions for scheduling calculations, adapting to the specific constraints of marine engineering equipment manufacturing, including hard and soft constraints. Hard constraints are production rules that must be met, such as the lifting of large components must be matched with cranes whose rated lifting capacity is greater than or equal to the weight of the component, and high-pressure pipeline welding must be performed by personnel holding special equipment operation certificates. Soft constraints are rules whose priority can be adjusted according to user needs, such as prioritizing the delivery time of urgent orders, maximizing the utilization rate of key equipment, and minimizing the overtime hours of personnel. The constraint layer is associated with the user configuration module through a parameterized configuration interface, supporting real-time updates of constraint weights.

[0058] The target layer clarifies the multi-objective optimization direction, covering three core objectives: production efficiency, resource utilization, and cost control. Production efficiency objectives include on-time order delivery rate and production cycle reduction rate; resource utilization objectives include Overall Equipment Effectiveness (OEE) and the percentage of effective working hours per employee; cost control objectives include equipment energy consumption reduction rate and overtime cost savings rate. The target layer is directly linked to the fitness function of the ant colony algorithm, and the weights of each objective are dynamically adjusted through a user configuration module, enabling flexible adaptation of multi-objective priorities.

[0059] 2. Scheduling calculation based on improved ant colony algorithm

[0060] like Figure 1 As shown, based on the traditional ant colony algorithm, three key optimizations are implemented to address the problems of "slow response speed and low optimization accuracy" in the scheduling of marine engineering equipment manufacturing. The specific process is as follows:

[0061] Step 1: Initialize the ant colony and pheromone matrix: Take the scheduled tasks (such as welding component A and assembling component B) as "ant colony foraging targets" and the combination of equipment and personnel as "path nodes". Initialize the number of ants to [50-100] (which can be configured according to the complexity of the task). The initial value of the pheromone matrix is ​​set according to the historical utilization rate of the equipment and the matching degree of personnel skills. For example, the initial value of the pheromone of the corresponding node of the equipment with high utilization rate is increased by 20% to enhance the directionality of the initial optimization.

[0062] Step 2: Dynamic Bottleneck Identification and Path Pruning: During the ant colony optimization process, a "bottleneck identification module" is embedded to calculate the load rate of each node in real time (load rate = node's assigned task duration / node's available time). When a node's load rate is ≥90%, it is identified as a bottleneck node, and path pruning is immediately triggered, preventing subsequent ants from choosing unnecessary paths (such as non-urgent task paths) to that node. Simultaneously, the bottleneck node information is fed back to the data layer to update the dynamic data of equipment / personnel. This operation can reduce more than 30% of invalid iteration paths and shorten computation time.

[0063] Step 3: Memory Optimization and Pheromone Update: A "tacit list + elite ant" mechanism is introduced. The tabu list records the nodes already visited by the ants, avoiding redundant calculations. The top 5%-10% of ants with the best objective function value in each iteration are selected as "elite ants," and their paths have an additional 2 times the pheromone concentration. Simultaneously, the three best paths from previous iterations are retained as the initial guiding paths for the next iteration. Pheromone updates employ a combination of "local update + global update." Local updates are performed after the ant completes its single-step path selection to prevent premature path fixation; global updates are performed after all ants have completed one iteration, reinforcing the pheromone accumulation of the optimal path.

[0064] Step 4: Multi-objective Solution Output and Adjustment: The algorithm iteration termination condition is set to "the objective function value fluctuation ≤ 1% for 10 consecutive iterations" or "the number of iterations reaches a preset upper limit (e.g., 200 times)". After the iteration ends, a multi-objective Pareto optimal solution set is output, including scheduling schemes under different objective weights (e.g., "on-time delivery rate priority" and "equipment utilization rate priority" schemes). The solution set is displayed through a visual interface, allowing users to view key indicators of each scheme (e.g., order delay days, equipment idle rate), and modify the objective weights through the "parameter fine-tuning" function. The algorithm recalculates in real time and outputs the adjusted scheme, realizing a closed loop of "user configuration - algorithm calculation - scheme optimization".

[0065] 3. User configuration interaction module

[0066] To enhance user engagement and scheduling flexibility, a visual user configuration interaction module was designed, with core functionalities including:

[0067] Target weight configuration: Provides a slider-style interactive interface, allowing users to drag the slider to adjust the weights of targets such as "on-time delivery rate", "equipment utilization rate", and "cost control" (the total weight is 100%). The system displays the target priority ranking after the weight adjustment in real time, such as "on-time delivery rate 60% > equipment utilization rate 30% > cost control 10%".

[0068] Constraint rule configuration: Users can add, modify, and delete constraint rules. For example, they can add a time constraint such as "Monday to Friday 8:00-18:00 is the main production period, and only emergency tasks are arranged on weekends", or modify an equipment constraint such as "the maximum continuous working time of the crane shall not exceed 4 hours". After configuration, the system will automatically verify the legality of the rule (such as whether it conflicts with hard constraints) and update the constraint layer parameters.

[0069] Emergency Task Insertion Configuration: To address the needs of sudden emergency projects in marine engineering equipment manufacturing, an emergency task insertion interface is designed. Users can upload emergency task information (such as task name, required equipment, and delivery deadline) and select one of three insertion strategies: "prioritize insertion during idle time periods," "adjust existing task order," or "increase personnel / equipment input." Based on the current scheduling scheme, the system quickly calculates the optimal schedule after insertion using an ant colony algorithm, while also indicating the impact of insertion on existing orders (such as the number of delay days).

[0070] Example

[0071] Taking an emergency task insertion scenario as an example, the implementation of the present invention is illustrated as follows:

[0072] 1. Implementation conditions

[0073] Basic scenario: Production has been running for 12 hours. Component A has been cut (C1 is idle) and is being welded (W1 has been working for 4 hours); Component B has been cut (C2 is idle) and is waiting to be welded; Components C, D, and E have not started.

[0074] Urgent task: Add "Emergency support component for offshore platform (component F)", which needs to be delivered within 24 hours. The process is as follows: cutting (requires 25mm steel plate, 4 hours) → welding (ST50 strength, 6 hours) → painting (fire retardant coating, 3 hours) → assembly (independent assembly, 5 hours). The delivery time must be prioritized.

[0075] User requirements: Select the strategy of "prioritize insertion of idle time slots + adjust the order of non-critical tasks" through the "Emergency Task Insertion" interface, and set the target weights as "delivery on time rate 80%, equipment utilization rate 15%, cost control 5%". Add the constraint rule "component F welding must use W3 (only W3 supports ST50), assembly can occupy Q1 idle time slot".

[0076] 2. Implementation process

[0077] Emergency task data entry: Users upload the process, duration, and equipment requirements of component F, select the insertion strategy and target weight, and the system automatically verifies: W3 is currently idle (originally planned for component C welding, which can be postponed), P2 supports fireproof coating, and Q1 is idle before component A assembly, so the constraints are valid.

[0078] Dynamic bottleneck identification: When the algorithm calculates the initial path, it finds that welding component F in W3 will take 6 hours. If component C is welded as originally planned, the load rate of W3 will reach 95% (6h + 12h for component C = 18h, available time 20h). This is identified as a potential bottleneck, triggering path pruning and adjusting the welding of component C to W2 (W2 was originally planned to weld component B, but component B has a more flexible delivery period and can be postponed).

[0079] The algorithm iterates rapidly: Since only the order of local tasks needs to be adjusted, the number of ants is set to 50. When the 50th iteration is completed, the termination condition is met, and the adjusted solution is output: "Component F → C1 cutting (C1 idle, 4h) → W3 welding (6h) → P2 painting (3h) → Q1+Z1 assembly (5h); Component B welding is adjusted to W2 (delayed by 4h); Component C welding is adjusted to W2 (after component B welding is completed)", and at the same time, it prompts "The delivery time of component B is delayed by 2 hours, which is within the allowable range (the original delivery time has a 5-hour buffer)".

[0080] Solution Confirmation and Execution: After reviewing the "Emergency Task Delivery Period (22 hours, meeting 24-hour requirements)" and "Delay Impact of Component B" on the solution preview interface, the user confirms the execution. The system updates the scheduling plan to the MES system in real time to guide on-site production.

[0081] 3. Implementation Results

[0082] Emergency mission delivery: Component F was delivered within 22 hours, achieving a 100% on-time delivery rate, meeting the emergency needs of offshore platforms, and representing a significant improvement over existing technologies (manual scheduling adjustments take 4 hours, and the delivery period of Component F may exceed 28 hours).

[0083] Impact on existing tasks: Only the delivery date of component B will be delayed by 2 hours (within the allowable buffer range), while the delivery dates of other components will remain unchanged, to avoid overall production disruption;

[0084] Resource utilization rate: The utilization rate of W3 is increased to 90% (6h component F + 12h component C). The idle time of C1, P2 and Q1 is fully utilized, there is no idle equipment, and only one welder is needed to work overtime for component F for 2 hours (overtime cost is controllable). This achieves the goal of "prioritizing urgent tasks + minimizing interference with existing production", which is the best application of this invention in dynamic production scenarios.

[0085] Compared with existing APS scheduling technology for marine engineering equipment manufacturing, the core innovations of this invention are reflected in the following three aspects, and each innovation is achieved through clear structural / algorithm improvements:

[0086] (I) Innovative Hierarchical Architecture of Multi-Objective Scheduling Model

[0087] Existing technologies mostly employ a flat model of "single objective + fixed constraints," which cannot adapt to the multi-element collaborative requirements of marine engineering equipment manufacturing. This invention innovatively constructs a three-layer architecture of "data layer - constraint layer - objective layer," achieving flexible association between multiple objectives and constraints through structural layering.

[0088] The data layer innovatively introduces a "static + dynamic" dual-source data acquisition mechanism, which connects to the production site system in real time through the OPC UA protocol, solving the problems of "data lag and static" in existing technologies;

[0089] The constraint layer features an innovative design with a parameterized configuration interface combining "hard constraints" and "soft constraints," enabling users to dynamically adjust constraint rules and overcoming the limitations of existing technologies where "constraints are fixed and cannot be changed."

[0090] The innovation at the target level covers three categories of objectives: "production efficiency, resource utilization, and cost control." By configuring weights, it enables dynamic switching of priorities among multiple objectives, thus overcoming the shortcomings of existing technologies that rely on "single-objective scheduling."

[0091] (II) Innovation in Efficiency Optimization of Ant Colony Algorithm

[0092] To address the issues of "iterative redundancy and slow response" in existing traditional ant colony algorithms, this invention achieves a breakthrough in efficiency through two algorithmic improvements, both of which are implemented through specific modular structures:

[0093] An innovative "Dynamic Bottleneck Identification and Path Pruning Module" has been added. By calculating the node load rate in real time (load rate = allocated time / available time), it automatically prunes unnecessary paths of bottleneck nodes, reducing invalid iterations by more than 30%. This module ensures the real-time nature of bottleneck identification by dynamically linking with the data layer.

[0094] The innovative "memory optimization + elite ant" mechanism is introduced. By recording historical paths through a taboo table and strengthening the pheromone of the optimal path with elite ants, the historical optimal path is retained as the initial guide, which reduces the number of algorithm iterations by 40% and improves the response speed by 75%. This mechanism is achieved through the collaboration of the "memory module" and "pheromone update module" in the algorithm's computation layer.

[0095] (III) Interactive Innovation in User-Participatory Scheduling

[0096] Existing technologies are mostly "black box" scheduling methods, where users cannot participate in algorithmic decision-making, resulting in poor flexibility. This invention innovatively designs a "user configuration interaction module," enabling deep user participation through a visual interface structure:

[0097] An innovative "target weight slider configuration" interface has been developed, allowing users to intuitively adjust target priorities without requiring professional algorithm knowledge, thus solving the problem of "high user participation threshold" in existing technologies.

[0098] The innovative "emergency task insertion + impact prediction" function enables rapid entry of emergency tasks through a dedicated interactive interface and automatically prompts the impact on existing orders, assisting users in decision-making and breaking through the bottleneck of existing technology where "emergency task processing is time-consuming and the impact is uncontrollable".

Claims

1. An APS scheduling method based on marine engineering equipment manufacturing, characterized in that, Includes the following steps: 1) Based on the marine engineering equipment manufacturing process, a multi-objective scheduling model is constructed; 2) Use the improved ant colony algorithm to perform scheduling calculations on the multi-objective scheduling model to obtain the optimal scheduling scheme.

2. The APS scheduling method based on marine engineering equipment manufacturing according to claim 1, characterized in that, The multi-objective scheduling model comprises a data layer, a constraint layer, and an objective layer, wherein: The data layer is used to collect and store data from the entire process of marine engineering equipment manufacturing in real time, and to interface with the PLC and MES systems on the production site through the OPC UA protocol, while ensuring that the data collection frequency meets the threshold. The constraint layer is used to define the boundary conditions for scheduling calculation, which are used as constraints during calculation. It is associated with the user configuration module through a parameterized configuration interface and supports real-time updates of constraint weights. The target layer is used to set targets, including production efficiency targets, resource utilization targets, and cost control targets. The weights of each target are dynamically adjusted through the user configuration module and used as optimization targets for improving the fitness function of the ant colony algorithm.

3. The APS scheduling method based on marine engineering equipment manufacturing according to claim 2, characterized in that, The data for the entire process of marine engineering equipment manufacturing includes static data and dynamic data. The static data includes equipment parameters, personnel skill matrix, and production process standards, while the dynamic data includes production progress, equipment operating status, and personnel on-duty status.

4. The APS scheduling method based on marine engineering equipment manufacturing according to claim 2, characterized in that, The constraints include hard constraints and soft constraints. Hard constraints are production rules that must be met, while soft constraints are rules whose priority can be adjusted according to user needs.

5. The APS scheduling method based on marine engineering equipment manufacturing according to claim 2, characterized in that, The improved ant colony algorithm is specifically as follows: Initialize the ant colony and pheromone matrix: Treat the scheduled tasks as "ant colony foraging targets" and the combination of equipment and personnel as "path nodes". Set the initial number of ant colonies and set the initial value of the pheromone matrix based on the historical utilization rate of equipment and the matching degree of personnel skills. Dynamic bottleneck identification and path pruning: During the ant colony optimization process, the load rate of each node is calculated in real time. When the load rate of a node is greater than the threshold, it is identified as a bottleneck node and path pruning is immediately triggered to prevent subsequent ants from choosing unnecessary paths to that node. At the same time, the bottleneck node information is fed back to the data layer to update the dynamic data of equipment / personnel. Memory optimization and pheromone updating: Introduce a "taboo list + elite ant" mechanism, whereby the taboo list is used to record the nodes that the ants have visited; The 5%-10% of ants with the best objective function value in each iteration are selected as "elite ants". The paths they have walked are given an additional 2 times the pheromone concentration. At the same time, the three paths with the smallest objective function value in the previous iterations are retained as the initial guiding paths for the next iteration. The pheromone update adopts a combination of "local update + global update". The local update is executed after the ant completes the single-step path selection, and the global update is executed after all ants have completed one iteration. Multi-objective solution output and adjustment: The algorithm terminates when the fluctuation of the objective function value is less than a threshold or the number of iterations reaches a preset upper limit. After the iteration ends, the output is a set of multi-objective Pareto optimal solutions containing scheduling schemes under different objective weights.

6. The APS scheduling method based on marine engineering equipment manufacturing according to claim 5, characterized in that, The load rate is specifically: Load rate = Node assigned task duration / Node available time.

7. The APS scheduling method based on marine engineering equipment manufacturing according to claim 1, characterized in that, It also includes human-computer interaction steps, specifically: Get the optimization target weights set by the user in real time, and display the target priority ranking after the weight adjustment in real time; The system acquires user-added, modified, and deleted constraint rules in real time, automatically verifies the legality of the rules (i.e., whether they conflict with hard constraints), and updates the constraint layer parameters accordingly. To address the needs of sudden emergency projects in marine engineering equipment manufacturing, users can upload emergency task information through the emergency task insertion interface and select one of three insertion strategies: "prioritize insertion during idle periods", "adjust existing task order", or "increase personnel / equipment input". Based on the current scheduling scheme, the system quickly calculates the optimal schedule after insertion using the ant colony algorithm, and also indicates the impact of insertion on existing orders.

8. An APS scheduling system based on marine engineering equipment manufacturing, characterized in that, include: The model building module is used to build multi-objective scheduling models based on the marine engineering equipment manufacturing process. The scheduling calculation module is used to perform scheduling calculations on multi-objective scheduling models using an improved ant colony algorithm to obtain the optimal scheduling scheme.

9. An APS scheduling system based on marine engineering equipment manufacturing according to claim 8, characterized in that, It also includes a user configuration module, specifically including: The target weight configuration interface is used to dynamically adjust and optimize the target weight using a slider-style interactive method. The constraint rule configuration interface is used to add, modify, and delete constraint rules, and automatically validates the legality of the rules. The emergency task insertion interface is used to receive emergency task information uploaded by users using three insertion strategies. It quickly calculates the optimal schedule after insertion using the ant colony algorithm, and also indicates the impact on existing orders. The three insertion strategies are: prioritizing insertion during idle periods, adjusting the order of existing tasks, and increasing personnel / equipment input.

Images