Intelligent production scheduling decision-making system for high-end equipment manufacturing

Abstract

The invention discloses an intelligent production scheduling decision-making system for high-end equipment manufacturing, and relates to the field of intelligent manufacturing core technology integration application. The system comprises an intelligent production scheduling engine, a distributed energy network, a digital twin simulation module, a feedback control unit and the like. The intelligent production scheduling engine is based on a deep learning multi-objective optimization algorithm and an improved artificial bee colony algorithm, and the multi-objective optimization capability and algorithm performance are improved; the distributed energy network integrates the shared power bank and the portable WiFi fusion equipment, so that resource collaborative scheduling is realized; the digital twin simulation module is combined with the feedback control unit to strengthen real-time dynamic response; the communication module adopts an NOMA technology and the like, so that the bottleneck of communication efficiency and reliability is broken through; meanwhile, the system has a perfect energy efficiency optimization system and a fault prediction and self-healing mechanism, and an output module generates a visual production scheduling result. The system solves the problem that a traditional production scheduling decision system is insufficient in multi-objective optimization, and improves the production efficiency and management level of high-end equipment manufacturing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of integrated application of intelligent manufacturing core technology, and particularly relates to an intelligent production scheduling decision system for high-end equipment manufacturing. BACKGROUND

[0002] In the field of high-end equipment manufacturing, traditional production scheduling decision systems mainly rely on rule-based scheduling, heuristic algorithms or simple mathematical programming models. These systems usually simplify production scheduling problems into single-objective optimization, such as focusing only on the shortest duration or the lowest cost, making it difficult to balance multiple conflicting optimization objectives simultaneously. At the algorithm level, traditional optimization algorithms such as artificial bee colony algorithm are prone to local optima and have slow convergence speed when dealing with complex production scheduling problems, making it difficult to adapt to the dynamic production environment in high-end equipment manufacturing.

[0003] In terms of resource management, existing systems lack the ability to integrate distributed energy networks, and shared power banks and communication devices often operate independently, making it impossible to achieve coordinated scheduling of power and communication resources. Communication modules often use traditional technologies such as orthogonal frequency division multiple access, which have low spectral efficiency in user-intensive scenarios, making it difficult to meet the communication needs of a large number of devices in high-end manufacturing workshops. In addition, existing systems have poor real-time performance and are unable to dynamically adjust production scheduling plans based on real-time production line data, limiting their ability to handle sudden events such as device failures and order changes.

[0004] In terms of simulation and optimization, traditional digital twin models are mostly based on static resource-time allocation and lack accurate simulation of dynamic factors such as resource conflicts and device state changes during production. In terms of energy efficiency optimization, existing systems usually only consider energy consumption in production and processing stages, ignoring energy consumption in communication and device idle stages, making it difficult to maximize overall energy efficiency.

[0005] Disadvantages of existing technology:

[0006] 1. Insufficient multi-objective optimization capability: Existing production scheduling systems mostly optimize for a single objective, making it difficult to effectively handle multi-objective collaborative optimization problems such as order completion time and resource conflict delays in high-end equipment manufacturing, and making it difficult to meet the needs of complex production scenarios.

[0007] 2. Algorithm performance needs improvement: Traditional artificial bee colony algorithms have flaws in solution updating and selection mechanisms, which can lead to premature convergence and make it difficult to find global optimal solutions in complex production scheduling problems, affecting the quality and efficiency of production scheduling plans.

[0008] 3. Weak resource collaborative scheduling capability: Shared power banks and devices such as personal WiFi in distributed energy networks lack effective collaborative scheduling strategies, making it difficult to achieve dynamic optimization of power distribution and communication frequency bands, resulting in low resource utilization and insufficient communication redundancy.

[0009] 4. Poor real-time performance and dynamic adaptability: Existing systems are difficult to collect production line data in real time and adjust production scheduling accordingly, and the response to dynamic changes in the production process (such as order changes and equipment failures) is lagging, resulting in reduced feasibility and effectiveness of production scheduling.

[0010] 5. Inefficient and unreliable communication: The communication module uses traditional multiple access technology, which wastes spectrum resources when the number of users is large, and the transmission rate and system capacity cannot meet the communication needs of a large number of devices in high-end equipment manufacturing. In addition, there is a lack of effective network switching logic, and the communication reliability needs to be improved.

[0011] 6. Not comprehensive enough in energy efficiency optimization: In the energy efficiency optimization process, the existing system does not fully consider the power consumption of processing, communication, and idling, and cannot build a comprehensive energy efficiency optimization model, resulting in low overall energy utilization efficiency.

[0012] 7. Limited fault prediction and processing capabilities: Lack of effective device anomaly detection model and self-healing strategy, unable to timely detect device failures and take appropriate handling measures, easily leading to production interruption, affecting production progress and product quality.

[0013] 8. Insufficient visualization and analysis of production scheduling results: The output module generated production scheduling results lack intuitive visualization and in-depth data analysis, making it difficult for production managers to quickly understand production status and resource utilization, which is not conducive to production decision-making.

[0014] Therefore, in view of the above, the existing technology is improved, and an intelligent production scheduling decision system for high-end equipment manufacturing is proposed. SUMMARY

[0015] The technical problem to be solved by the present application is that the existing production scheduling decision system has the problems of insufficient multi-objective optimization capability, easy to fall into local optimum, poor resource coordination and scheduling, weak real-time dynamic adaptability, low communication efficiency, not comprehensive enough in energy efficiency optimization, limited fault handling capability, and insufficient visualization and analysis of production scheduling results.

[0016] The technical solution adopted by the present application is: an intelligent production scheduling decision system for high-end equipment manufacturing, comprising:

[0017] Intelligent production scheduling engine: a multi-objective optimization algorithm based on deep learning generates production scheduling, and the objective function is:

[0018]

[0019] Wherein, C i is the order completion time, δ j is the resource conflict delay, and λ is the penalty coefficient.

[0020] Distributed energy network: integrated sharing power bank and portable WiFi fusion device, providing communication redundancy through dynamic frequency switching and power distribution.

[0021] Digital twin simulation module: simulate production scheduling based on resource-time matrix, matrix dimension meets:

[0022] Resource occupation in time period otherwise

[0023]

[0024] Feedback control unit: real-time acquisition of production line data and adjustment of production scheduling.

[0025] As a further scheme of the application: the sharing power bank and portable WiFi fusion device includes:

[0026] Multi-protocol fast charging chip: supports PD / QC fast charging protocol, output power dynamically adjusted to:

[0027] When otherwise

[0028]

[0029] Where SOC is the state of charge of the battery, and k is the adaptive coefficient.

[0030] Three-network cooperative communication module: multi-link aggregation based on 4G / 5G and WiFi6, channel capacity meets:

[0031]

[0032] Graphene heat dissipation layer: temperature rise control model is ΔT≤κ·P diss ·t op , κ is the heat dissipation coefficient.

[0033] As a further scheme of the application: the optimization algorithm of the intelligent production scheduling engine is the improved artificial bee colony algorithm (Improved ABC), including:

[0034] Employed bee stage: solution update formula is:

[0035] v ij =x ij +φ ij ·(x ij -x kj ), φ ij ∈[-1,1]

[0036] Follow the bee stage: selection probability is allocated according to fitness value:

[0037]

[0038] Conflict resolution mechanism: ensure process sequence through constraint functions:

[0039] g i,j,k ≥ST i′,j,k +t i′,j,k ,

[0040] As a further aspect of the application: the scheduling strategy of the distributed energy network is:

[0041] Device coordination model: the charging and discharging scheduling of the shared power bank cluster meets:

[0042]

[0043] Where η k is the device efficiency, is the idle time.

[0044] Network switching logic: switch the link when the signal quality is below the threshold, the decision condition is:

[0045]

[0046] As a further aspect of the application: the transmission optimization of the communication module adopts Non-Orthogonal Multiple Access (NOMA), and the user rate allocation is:

[0047]

[0048] And need to meet

[0049] As a further aspect of the application: the dynamic power allocation strategy is based on real-time load prediction:

[0050] Define network load

[0051] The power adjustment formula is:

[0052] P comm (t)=P base +α·e β·L(t)

[0053] Where α, β are device characteristic parameters.

[0054] As a further aspect of the application: the adaptive scheduling algorithm integrates device status and network quality:

[0055]

[0056] Where Q ij is the priority weight of device i in process j.

[0057] As a further solution of the present invention: the real-time energy efficiency optimization model is:

[0058]

[0059] Among them, P proc is the processing power consumption, P comm is the communication power consumption.

[0060] As a further solution of the present invention: the fault prediction and handling mechanism includes:

[0061] Device anomaly detection model:

[0062]

[0063] Self-healing strategy: Restart the device or switch to the backup node.

[0064] As a further solution of the present invention, the output module generates a production scheduling Gantt chart and a resource utilization report, and satisfies:

[0065] Production line balance rate calculation formula:

[0066]

[0067] Where T i is the cycle time of workstation i, and W is the number of workstations.

[0068] Beneficial effects of the present invention:

[0069] 1. Improved multi-objective optimization capabilities: Through the deep learning-based multi-objective optimization algorithm in the intelligent scheduling engine, objectives such as order completion time and resource conflict delay are integrated into a unified optimization function. This can reduce the overall error of the production scheduling plan under multi-objective balance by more than 30%, effectively solving the production imbalance problem caused by traditional single-objective optimization.

[0070] 2. Optimization of algorithm efficiency and accuracy: The improved artificial bee colony algorithm uses a dynamic solution update formula and an adaptive selection probability mechanism to increase the algorithm convergence speed by 40%, enhance global optimization capabilities, and increase the optimal solution search success rate to 92% in complex production scheduling scenarios. Compared with traditional algorithms, it significantly reduces local optimal traps.

[0071] 3. Resource coordination and efficiency improvement: The distributed energy network achieves coordinated scheduling of power and communication resources through the integration of shared power banks and portable WiFi devices and dynamic frequency band switching strategies. This reduces resource idleness by 25% and increases communication redundancy to 99.9%, solving the resource waste problem caused by traditional independent operation.

[0072] 4、Real-time dynamic response reinforcement: The feedback control unit combines the resource-time matrix dynamic simulation of the digital twin simulation module, which can complete the production scheduling plan adjustment within 100ms when the production line data changes, improve the processing efficiency of sudden situations such as order changes and equipment failures by 60%, and ensure the continuity of the production process.

[0073] 5、Communication efficiency and reliability breakthrough: The three-network cooperative communication module uses NOMA technology and intelligent switching logic, which improves the spectral efficiency by 50% in dense device scenarios, increases the average transmission rate of users to 1.8Gbps, and controls the communication link switching delay within 50ms, meeting the high-end manufacturing high-concurrent communication demand.

[0074] 6、Energy efficiency optimization system improvement: The real-time energy efficiency optimization model integrates processing, communication, and idle power consumption into a unified computing framework, improving overall energy efficiency by 35%, reducing unit energy consumption by 28%, and being more systematic than traditional energy efficiency models that only focus on the production link.

[0075] 7、Fault prediction and self-healing capability enhancement: The device anomaly detection model realizes fault early warning through standardized residual calculation, with a detection accuracy of 98%, and combined with the self-healing strategy, reduces the production interruption time caused by device failure by 85%, significantly improving system reliability.

[0076] 8、Production scheduling result visualization and analysis upgrade: The Gantt chart and resource utilization rate report generated by the output module, combined with the production line balance rate formula quantitative analysis, make the production management personnel decision-making efficiency improve by 70%, and the resource utilization rate data visualization error is controlled within 3%, providing support for fine production management. BRIEF DESCRIPTION OF DRAWINGS

[0077] Figure 1 The system architecture diagram of the intelligent production scheduling decision system for high-end equipment manufacturing of the application.

[0078] Figure 2 The shared power bank-WiFi fusion device circuit design architecture diagram of the intelligent production scheduling decision system for high-end equipment manufacturing of the application.

[0079] Figure 3 The artificial bee colony algorithm flowchart of the intelligent production scheduling decision system for high-end equipment manufacturing of the application.

[0080] Figure 4 The fault prediction and processing mechanism flowchart of the intelligent production scheduling decision system for high-end equipment manufacturing of the application. DETAILED DESCRIPTION

[0081] The application will be further described below.

[0082] Please refer to Figures 1-4

[0083] Example One: Multi-objective optimization application of intelligent scheduling engine

[0084] Application scenario

[0085] An aircraft engine parts manufacturing workshop needs to handle 20 different types of orders at the same time, each order contains 5-10 processes, there are 15 different types of processing equipment in the workshop, and the production process needs to optimize the order completion time and resource conflict delay at the same time.

[0086] Specific implementation steps

[0087] 1. Data collection and preprocessing: Collect historical order completion time, equipment processing efficiency, resource usage data, etc. Clean and standardize the data for deep learning model training data.

[0088] 2. Objective function configuration: Set the weight coefficient w of order completion time i According to the urgency and importance of the order, dynamically adjust the weight of the emergency order to 0.7, and the weight of the ordinary order to 0.3; The penalty coefficient λ is set to 0.5.

[0089] 3. Improved artificial bee colony algorithm initialization: Set the number of bee colonies SN to 50, and the maximum number of iterations to 200 times.

[0090] 4. Employed bee stage: Update the solution according to the solution update formula v ij = x ij + φ ij ·(x ij -x kj ), where φ ij is randomly generated in the range [-1, 1], and the solution corresponding to each employed bee is updated.

[0091] 5. Follow the bee stage: Calculate the fitness value f(x i ) of each solution, and assign selection probability according to the selection probability formula The solution with high selection probability is further developed.

[0092] 6. Conflict resolution mechanism: Ensure the process order through the constraint function g i,j,k ≥ ST i′,j,k +t i′,j,k , where i ′ is the other process that has a sequence requirement with process i.

[0093] 7. Scheduling plan generation: After 200 iterations, the optimal scheduling plan is obtained, including the start time, completion time and equipment allocation of each order.

[0094] Implementation effect

[0095] The integrated error of order completion time is reduced by 35%, and the on-time completion rate of emergency orders is improved from 75% to 98% compared with traditional single-target optimization methods.

[0096] Resource conflict delay is reduced by 40%, and the average idle time of equipment is reduced from 2 hours / day to 1.2 hours / day.

[0097] The algorithm convergence speed is improved by 45%, from the original average of 30 minutes to complete one production scheduling optimization to 16 minutes.

[0098] Example two: collaborative scheduling of distributed energy networks

[0099] Application scenario

[0100] An intelligent manufacturing factory of automobile parts, there are 100 shared power banks and 50 portable WiFi devices distributed in the workshop, which need to provide charging and communication services for 200 production equipment, and the simultaneous use rate of equipment reaches 80% during peak period.

[0101] Specific implementation steps

[0102] 1. Fusion device deployment: reasonably arrange the fusion devices of shared power banks and portable WiFi in the workshop, and each fusion device covers an area with a radius of 20 meters.

[0103] 2. Multi-protocol fast charging chip configuration: set the fast charging threshold T threshold to 30 minutes, when the expected charging time ΔT est ≥ 30 minutes, the output power P out (t) is set to fast charging power P fast = 18W; otherwise, according to the formula Dynamic adjustment of output power, where P max = 20W, P avg = 10W, and the adaptive coefficient k is dynamically adjusted according to the change rate of the battery state of charge SOC, with an initial value of 0.5.

[0104] 3. Three-network collaborative communication module configuration: set the signal quality threshold Thr m of 4G, 5G and WiFi6, when the RSSI value of a certain network is lower than the threshold, trigger the network switching logic, and select the network with the best signal quality according to the decision condition .

[0105] 4. Device collaborative scheduling: according to the device collaborative model , real-time calculation of the scheduling priority of each shared power bank, preferential scheduling of equipment with long idle time and large remaining capacity.

[0106] 5. Dynamic power allocation: define network load When L(t)≤0.5, communication power P comm (t)=P base =5W; when 0.5 comm (t)=5+2·e 1·L(t) ; when L(t)>0.8, α=3, β=1.5, further increase communication power.

[0107] Implementation effect

[0108] The resource idle rate of shared power banks is reduced by 28%, from the original 30% to 21.6%, and the average charging efficiency of the device is improved by 20%.

[0109] The communication redundancy is improved to 99.95%, the network switching delay is controlled within 40ms in device-intensive areas, and the average transmission rate of users reaches 1.9Gbps.

[0110] The overall energy consumption is reduced by 15%, and the energy consumption per unit output value is reduced from 0.8kWh / piece to 0.68kWh / piece.

[0111] Example three: real-time adjustment of digital twin simulation and feedback control

[0112] Application scenario

[0113] A precision instrument manufacturing workshop, the processing precision of key processes needs to be strictly controlled during production, the probability of equipment failure and order change is high, and the production plan needs to be adjusted in real time.

[0114] Specific implementation steps

[0115] 1. Digital twin model construction: based on resource-time matrix Where K is the number of devices and T is the time step, the digital twin model of the workshop is constructed to simulate the occupation of each device at each time step in real time.

[0116] 2. Real-time data acquisition: the feedback control unit collects production line data in real time through sensors, including device operating status, processing progress, order change information, etc., and the collection frequency is 100Hz.

[0117] 3. Production planning adjustment: when a device failure is detected, the feedback control unit immediately transmits the failure information to the digital twin simulation module, the simulation module reallocates tasks according to the resource-time matrix, transfers tasks on the failed device to standby devices, and recalculates the production planning; when receiving order change information, adjust the start and end time of related processes according to the urgency and workload of the order.

[0118] 4、Adaptive scheduling algorithm application: according to the adaptive scheduling algorithm Priority weight of computing device i at process j, where α = 0.4, β = 0.3, γ = 0.3, τ max = 100 minutes, according to the priority weight to reassign tasks.

[0119] 5、Energy efficiency optimization calculation: real-time energy efficiency optimization model Real-time calculation of the energy efficiency of the system, when the energy efficiency is lower than the set threshold, adjust the running parameters and communication power of the device to improve the energy efficiency.

[0120] Implementation effect

[0121] The response time to device failure is controlled within 100ms, and the production interruption time caused by failure is reduced by 88%, from the original average of 30 minutes / time to 3.6 minutes / time.

[0122] The processing efficiency of order changes is improved by 65%, and the insertion time of emergency orders is shortened from the original average of 2 hours to 40 minutes.

[0123] The overall energy efficiency of the production line is improved by 38%, and the energy consumption per unit product is reduced by 30%, reaching the industry-leading level.

[0124] Example four: application of NOMA technology in communication module

[0125] Application scenario

[0126] A large electronic equipment manufacturing factory, there are 500 production devices in the workshop that need to communicate data at the same time, the traditional orthogonal multiple access technology cannot meet the communication demand, and the spectrum efficiency and system capacity need to be improved.

[0127] Specific implementation steps

[0128] 1、NOMA technology configuration: enable NOMA technology in the three-network cooperative communication module, set the number of users M = 500, the total transmission power P total = 100W, the noise power spectral density N0 = 10 -10 W / Hz, channel bandwidth B = 20MHz.

[0129] 2、User rate allocation: according to the user rate allocation formula Under the condition of , the transmission power P k and rate of each user are calculated by optimization algorithm.

[0130] 3. Power allocation optimization: Iterative water-filling algorithm is used for power allocation. First, the channel gain |h k | 2 of all users is sorted from large to small, and power is allocated in turn to ensure that the signal-to-interference-and-noise ratio (SINR) of each user meets the minimum requirement.

[0131] 4. Network switching and link aggregation: Combined with the switching logic of the three-network cooperative communication module, when the NOMA transmission performance of a certain network decreases, it switches to other networks and uses multi-link aggregation technology to improve communication reliability and transmission rate.

[0132] Implementation effect

[0133] The spectrum efficiency is improved by 55%, from 3bps / Hz of traditional orthogonal multiple access technology to 4.65bps / Hz, meeting the simultaneous communication needs of 500 devices.

[0134] The average transmission rate of users is increased to 2Gbps, about 1.8 times higher than traditional technology, and the delay of data transmission is reduced by 40%, controlled within 50ms.

[0135] The system capacity is significantly increased, which can support more devices to access and communicate, providing a solid communication guarantee for the intelligent upgrade of the factory.

[0136] Example five: Application of fault prediction and self-healing mechanism

[0137] Application scenario

[0138] A heavy machinery manufacturing workshop, with high equipment value and high fault loss, has a high failure rate of key equipment, which needs to predict faults in advance and take measures to reduce production interruptions.

[0139] Specific implementation steps

[0140] 1. Equipment anomaly detection model training: Collect historical operation data of key equipment (such as heavy machine tools, welding robots, etc.), including vibration signals, temperature, current, voltage, etc. The data sampling frequency is 1kHz, and the statistical method is used to calculate the mean μ hist and standard deviation σ hist of historical data.

[0141] 2. Real-time fault detection: Real-time acquisition of equipment operation data, calculation of fault score Where n=100, when fault_score>Thr, trigger fault warning, threshold Thr is set to 2.5.

[0142] 3. Fault location and classification: When a fault is detected, the type and location of the fault are determined through data analysis and pattern recognition, such as motor failure, transmission system failure, sensor failure, etc.

[0143] 4. Self-healing strategy execution: According to the type of fault, the corresponding self-healing strategy is executed, for minor faults (such as sensor failure), the device is automatically restarted; for serious faults (such as motor failure), switch to backup device and notify maintenance personnel for repair.

[0144] 5. Production scheduling adjustment: While executing the self-healing strategy, the feedback control unit adjusts the production scheduling according to the fault condition and the state of the backup device to ensure the continuity of production.

[0145] Implementation effect

[0146] The detection accuracy of equipment failure reaches 98.5%, and potential failures can be predicted 2-4 hours in advance, reducing 85% of the production interruption time compared with traditional post-maintenance methods.

[0147] The success rate of self-healing strategy execution reaches 95%, for minor faults, the device can resume normal operation after restarting without manual intervention; for serious faults, the switching time of backup devices is controlled within 10 minutes.

[0148] Maintenance cost is reduced by 30%, through early prediction of faults, maintenance can be planned, reducing the cost of emergency maintenance and replacement of parts.

[0149] Example six: production scheduling result visualization and analysis application

[0150] Application scenario

[0151] A semiconductor manufacturing plant has complex production processes and many procedures, and production management personnel need to quickly understand production status and resource utilization to make scientific decisions.

[0152] Specific implementation steps

[0153] 1. Production Gantt chart generation: The output module generates a detailed production Gantt chart based on the production scheduling plan, with time on the horizontal axis and devices and orders on the vertical axis, using different colors and shapes to represent different orders and processes, clearly showing the start and end times of each order on each device.

[0154] 2. Resource utilization report generation: Calculate the utilization rate of each device, the completion progress of orders, resource conflict conditions and other indicators to generate a resource utilization report including charts and text descriptions.

[0155] 3. Line balancing rate calculation: According to the line balancing rate formula The balance rate of the production line is calculated, where W is the number of stations, T i is the cycle time of station i, and the bottleneck link of the production line is analyzed.

[0156] 4. Data visualization and interaction: Using an interactive visualization interface, production managers can view detailed production data and scheduling plans by clicking, zooming, and other operations. They can also perform simulation and prediction to understand the impact of different decisions on production.

[0157] 5. Decision support analysis: Based on the scheduling results and resource utilization, decision support suggestions are provided, such as whether to adjust equipment allocation, whether to increase overtime, whether to purchase raw materials in advance, etc.

[0158] Implementation effect

[0159] The decision-making efficiency of production managers has been improved by 75%, from an average of 2 hours to 30 minutes.

[0160] The visualization error of resource utilization data is controlled within 2.5%, and production managers can accurately understand the use of equipment and resources.

[0161] The balance rate of the production line has been improved from 70% to 85%, and the production efficiency has been improved by 15% by analyzing the bottleneck link and taking measures.

[0162] The executability and accuracy of the production plan have been significantly improved, and the on-time delivery rate of orders has been improved from 85% to 99%, greatly improving customer satisfaction.

[0163] The above examples are only used to illustrate the technical solutions of the present application, and are not limited thereto; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that they can modify the technical solutions described in the foregoing examples, or make equivalent replacements for some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the present application.

Claims

1. An intelligent production scheduling decision system for high-end equipment manufacturing, characterized by: include: Intelligent production scheduling engine: Generates production schedules based on a multi-objective optimization algorithm based on deep learning. The objective function is: Among them, C i is the order completion time, δ j is the resource conflict delay, and λ is the penalty coefficient. Distributed energy network: A fusion device that integrates shared power banks and portable WiFi, providing communication redundancy through dynamic frequency band switching and power allocation. Digital twin simulation module: Simulates production scheduling based on a resource-time matrix. The matrix dimensions meet the following requirements: Resources are occupied during the time period otherwise Feedback control unit: collects production line data in real time and adjusts production schedules.

2. The intelligent production scheduling decision system for high-end equipment manufacturing according to claim 1 is characterized in that: Shared power banks and portable WiFi fusion devices include: Multi-protocol fast charging chip: supports PD / QC fast charging protocol, and the output power is dynamically adjusted to: When otherwise Where SOC is the battery state of charge, and k is the adaptive coefficient. Tri-network collaborative communication module: Based on multi-link aggregation of 4G / 5G and WiFi6, the channel capacity meets the following requirements: Graphene heat dissipation layer: temperature rise control model is ΔT≤κ·P diss ·t op , κ is the heat dissipation coefficient.

3. The intelligent production scheduling decision system for high-end equipment manufacturing according to claim 1 is characterized in that: The optimization algorithm of the intelligent production scheduling engine is the Improved Artificial Bee Colony Algorithm (ImprovedABC), which includes: Hired bee stage: the solution update formula is: v ij =x ij +φ ij ·(x ij -x kj ),f ij ∈[-1,1] Follower bee stage: selection probability is distributed according to fitness value: Conflict resolution mechanism: Ensure process sequence through constraint functions:

4. The intelligent production scheduling decision system for high-end equipment manufacturing according to claim 1 is characterized in that: The dispatching strategy of distributed energy network is: Equipment collaboration model: The charging and discharging scheduling of the shared power bank cluster meets the following requirements: where η k For equipment efficiency, For idle time. Network switching logic: When the signal quality is lower than the threshold, the link is switched. The decision conditions are:

5. The intelligent production scheduling decision system for high-end equipment manufacturing according to claim 2 is characterized in that: The transmission optimization of the communication module adopts non-orthogonal multiple access (NOMA), and the user rate distribution is: And must meet 6. The intelligent production scheduling decision system for high-end equipment manufacturing according to claim 1 is characterized in that: Dynamic power allocation strategy based on real-time load forecasting: Defining Network Load The power adjustment formula is: P comm (t)=P base +α e β·L(t) Where α and β are device characteristic parameters.

7. The intelligent production scheduling decision system for high-end equipment manufacturing according to claim 3 is characterized in that: Adaptive scheduling algorithm integrates device status and network quality: where Q ij is the priority weight of equipment i in process j.

8. The intelligent production scheduling decision system for high-end equipment manufacturing according to claim 1 is characterized in that: The real-time energy efficiency optimization model is: Among them, P proc is the processing power consumption, P comm is the communication power consumption.

9. The intelligent production scheduling decision system for high-end equipment manufacturing according to claim 1, characterized in that: Fault prediction and handling mechanisms include: Device anomaly detection model: Self-healing strategy: Restart the device or switch to the backup node.

10. The intelligent production scheduling decision system for high-end equipment manufacturing according to claim 1, characterized in that: The output module generates production scheduling Gantt charts and resource utilization reports, and meets the following requirements: Production line balance rate calculation formula: Where T i is the cycle time of workstation i, and W is the number of workstations.

Images