Titanium plate inventory and production plan collaborative intelligent management system

Abstract

The present application relates to the technical field of intelligent manufacturing and supply chain management, in particular to an intelligent management system for titanium plate inventory and production plan coordination. It comprises: a multi-source data fusion center for the supply chain management system to obtain operation data and generate a collaborative context original data set; a context quantitative analysis unit for receiving the collaborative context original data set and generating market agility demand points and inventory liquidity status points; a collaborative conflict evaluation unit for calculating a collaborative conflict index and determining a collaborative conflict context level; an evolutionary trend prediction unit for generating an evolution sequence of the future collaborative conflict index; and a dynamic decision management unit for generating an immediate conversion strategy in response to the collaborative conflict context level and generating a predictive inventory structure pre-configuration decision. The present application significantly improves the accuracy and objectivity of the decision, enabling enterprises to gain a deeper insight into the core contradictions of operations.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing and supply chain management technology, specifically to an intelligent management system that coordinates titanium plate inventory and production planning. Background Technology

[0002] In the high-end titanium plate manufacturing industry, there is an inherent contradiction between the rigidity and long cycle of production planning and the high uncertainty of market order demand. Traditional production and inventory management models often rely on their own isolated information systems and static strategies, lacking real-time and quantitative assessment of the degree of conflict between the urgency of market demand and the flexibility of internal inventory assets, and also unable to effectively predict the future trend of conflict evolution.

[0003] This management approach leads to widespread decision-making delays, making it difficult to achieve a dynamic balance between ensuring order delivery agility and maintaining high liquidity of inventory assets, thereby increasing the operational risks of delayed order delivery and ineffective capital occupation.

[0004] Therefore, how to build an intelligent collaborative management system that can integrate multi-source heterogeneous data, accurately quantify production and sales conflicts, and make forward-looking predictions and closed-loop dynamic decisions to achieve proactive optimization of production agility and inventory liquidity has become a technical problem that urgently needs to be solved in this field.

[0005] The information disclosed in the background section above is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention discloses an intelligent management system for coordinating titanium plate inventory and production planning. Specifically, the technical solution of this invention includes:

[0007] The multi-source data fusion center is used to acquire operational data from manufacturing execution systems, enterprise resource planning systems, customer relationship management systems, commodity markets, and supply chain management systems, and to fuse and process the acquired operational data to generate a raw dataset for collaborative scenarios.

[0008] The scenario quantitative analysis unit is used to receive the original dataset of collaborative scenarios and perform quantitative calculations to generate market agility demand score and inventory liquidity status score.

[0009] The collaborative conflict assessment unit is used to combine market agility demand score and inventory liquidity status score to calculate collaborative conflict index, and to judge collaborative conflict index according to preset conflict threshold to determine collaborative conflict situation level.

[0010] The evolution trend prediction unit is used to construct a scenario historical sequence based on historical collaborative conflict index, market agility demand score and inventory liquidity status score, and combine it with external macroeconomic data to make time series predictions in order to generate the evolution sequence of future collaborative conflict index.

[0011] The dynamic decision management unit is used to generate immediate transformation strategies in response to the level of collaborative conflict situations, and to generate predictive inventory structure pre-configuration decisions based on the evolution sequence of future collaborative conflict indices.

[0012] Optionally, the process for generating the market agility requirement segment is as follows:

[0013] Obtain the order weighting coefficient, expected profit margin, order completion probability, and expected delivery cycle of potential orders;

[0014] The agility contribution value of each order is obtained by multiplying the order weight coefficient, the expected profit margin of the order, and the order completion probability, and then dividing by the expected delivery cycle of the order.

[0015] Sum the agility contribution values ​​of all potential orders to generate the raw agility score;

[0016] The original agility score is normalized using a pre-defined logic function to generate a market agility demand score.

[0017] Optionally, the process for generating the inventory liquidity status score is as follows:

[0018] Obtain the total material value of each form of material and the universality coefficient calibrated by the process engineer;

[0019] Obtain the total value of all inventory materials;

[0020] Multiply the total value of each form of material by the corresponding universality coefficient to obtain the liquidity value of each form of material.

[0021] The total liquidity value is obtained by summing the liquidity values ​​of all forms of materials.

[0022] The total liquidity value is divided by the total value of all inventory items to generate the current inventory liquidity score.

[0023] Optionally, the calculation process for the collaborative conflict index is as follows:

[0024] Subtracting the inventory liquidity score from the value of 1 yields the degree of inventory illiquidity.

[0025] Multiplying market agility demand by the degree of inventory illiquidity yields the conflict baseline value;

[0026] The conflict baseline value is subjected to a power operation on a preset conflict sensitivity coefficient to generate a collaborative conflict index.

[0027] Optionally, the process for determining the level of collaborative conflict situations is as follows:

[0028] The collaborative conflict index is compared with the first preset threshold and the second preset threshold;

[0029] If the collaborative conflict index is less than the first preset threshold, the collaborative conflict situation is determined to be a low conflict situation.

[0030] If the collaborative conflict index is greater than or equal to the first preset threshold and less than or equal to the second preset threshold, then the collaborative conflict situation level is determined to be a medium conflict situation.

[0031] If the collaborative conflict index is greater than the second preset threshold, the collaborative conflict situation is determined to be a high-conflict situation.

[0032] Optionally, the process for generating an instant conversion strategy is as follows:

[0033] When the collaborative conflict situation is classified as a high-conflict situation, a predictive production strategy is output with the goal of ensuring production agility.

[0034] When the collaborative conflict situation level is medium, output instructions to maintain the current production and inventory strategy;

[0035] When the collaborative conflict situation is classified as low-conflict, the output strategy is to increase upstream reserves with the goal of ensuring inventory liquidity.

[0036] Optionally, the evolution trend prediction unit adopts a long short-term memory network model, taking the historical sequence of the situation and external macroeconomic data as input, to output the evolution sequence of the future collaborative conflict index.

[0037] Optionally, the process for generating predictive inventory structure pre-configuration decisions is as follows:

[0038] By analyzing the evolution sequence of the future collaborative conflict index, we can identify future time points in the sequence where the predicted value exceeds the high conflict threshold, in order to generate an early warning of future conflict peaks.

[0039] In response to early warnings of future conflict peaks, multi-objective optimization calculations are initiated to generate predictive inventory structure pre-configuration decisions.

[0040] Optionally, it also includes a closed-loop feedback execution step, which is as follows:

[0041] Transform real-time conversion strategies or predictive inventory structure pre-configuration decisions into production or procurement instructions, and issue them to the enterprise resource planning system and manufacturing execution system for execution;

[0042] After the instruction is executed, the multi-source data fusion center collects the changed system status in the next calculation cycle, so that all subsequent calculations are updated based on the latest reality, thus forming a closed-loop management process.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. This invention transforms vague market pressures and internal asset flexibility into precise mathematical representations by constructing two core quantitative indicators: market agility demand score and inventory liquidity status score. The system integrates multi-source heterogeneous data and, through a collaborative conflict index model, objectively and in real-time measures the intensity of production-sales contradictions, providing a scientific and unified data benchmark for decision-making. This quantitative assessment method replaces the traditional experience-based management model, significantly improving the accuracy and objectivity of decision-making, enabling enterprises to gain a deeper understanding of core operational contradictions.

[0045] 2. This invention utilizes a Long Short-Term Memory (LSTM) network model, combining internal historical state sequences with external macroeconomic data, to achieve accurate prediction of the future evolution trend of the collaborative conflict index. The system can identify potential peaks in production and sales conflicts in advance and generate early warnings. Subsequently, through a multi-objective optimization algorithm, it proactively generates forward-looking inventory structure pre-configuration decisions. This mode of advance adjustment at a lower cost effectively avoids major operational risks such as delayed order delivery, achieving an upgrade from a passive response to a proactive anticipation management model.

[0046] 3. This invention establishes a dynamic decision-making management mechanism that automatically matches the optimal response strategy based on real-time assessments of high, medium, and low levels of collaborative conflict. In high-conflict situations, the system prioritizes production agility, decisively converting inventory to meet urgent orders; in low-conflict situations, it focuses on ensuring inventory liquidity, increasing upstream general-purpose material reserves. This scenario-driven adaptive decision-making capability ensures that enterprises can achieve a dynamic balance between production agility and inventory liquidity in different market environments, enhancing overall operational resilience.

[0047] 4. This invention achieves globally optimized allocation of production resources and inventory assets through the synergistic effect of real-time conversion strategies and predictive inventory structure pre-configuration decisions. The system accurately guides the conversion and storage of inventory forms, effectively reducing the problems of backlog of specialized materials and shortage of general materials caused by market misjudgments. The closed-loop management process formed by instruction issuance and execution result feedback ensures continuous iteration and optimization of decisions, significantly reducing inventory holding costs, shortening production delivery cycles, and comprehensively improving the company's capital utilization efficiency and market competitiveness. Attached Figure Description

[0048] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0049] Figure 1 This is a system structure diagram of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0051] Example 1:

[0052] Please see Figure 1 An intelligent management system for coordinating titanium plate inventory and production planning includes:

[0053] The multi-source data fusion center is used to acquire operational data from manufacturing execution systems, enterprise resource planning systems, customer relationship management systems, commodity markets, and supply chain management systems, and to fuse and process the acquired operational data to generate a raw dataset for collaborative scenarios.

[0054] The scenario quantitative analysis unit is used to receive the original dataset of collaborative scenarios and perform quantitative calculations to generate market agility demand score and inventory liquidity status score.

[0055] The collaborative conflict assessment unit is used to combine market agility demand score and inventory liquidity status score to calculate collaborative conflict index, and to judge collaborative conflict index according to preset conflict threshold to determine collaborative conflict situation level.

[0056] The evolution trend prediction unit is used to construct a scenario historical sequence based on historical collaborative conflict index, market agility demand score and inventory liquidity status score, and combine it with external macroeconomic data to make time series predictions in order to generate the evolution sequence of future collaborative conflict index.

[0057] The dynamic decision management unit is used to generate immediate transformation strategies in response to the level of collaborative conflict situations, and to generate predictive inventory structure pre-configuration decisions based on the evolution sequence of future collaborative conflict indices.

[0058] An intelligent management system for coordinating titanium plate inventory and production planning aims to resolve the core contradiction between the high uncertainty of market demand and the rigidity and long cycle of production processes in the high-end titanium plate manufacturing industry. By constructing a complete intelligent management system from data perception, context quantification, conflict assessment, trend prediction to closed-loop decision-making, the system achieves a dynamic balance between production agility and inventory liquidity. Logically, the system consists of multiple functional units deployed on enterprise servers or cloud platforms.

[0059] The purpose of the multi-source data fusion center is to provide a comprehensive, real-time, and standardized data foundation for the intelligent analysis and decision-making of the entire system. The center connects with multiple existing information systems of the enterprise through application programming interfaces (APIs) and direct database connections. It obtains full-process inventory data from the Manufacturing Execution System (MES) and Enterprise Resource Planning (ERP) in real time. This data not only includes physical information such as the quantity and specifications of ingots, slabs, work-in-process, and finished products, but also relates to their financial value information.

[0060] Simultaneously, it retrieves potential order data from the Customer Relationship Management (CRM) system and, in conjunction with an independent sales forecasting system, labels these orders with the probability of completion and expected delivery cycle. Furthermore, it accesses commodity market data via external APIs to obtain price fluctuation information such as titanium raw materials, and acquires capacity and delivery capability data from key suppliers through the Supply Chain Management (SCM) system. All these data, from different sources and with varying structures, are cleaned, aligned, and standardized before being integrated into a unified collaborative contextual raw dataset, serving as the sole and reliable data source for all subsequent analysis units. This deep integration of multi-source data provides the prerequisite for the system to comprehensively and accurately perceive the current business environment.

[0061] The Contextual Quantitative Analysis Unit aims to transform complex, multi-dimensional business scenarios into calculable and comparable standardized core indicators. This unit receives raw datasets of collaborative scenarios generated by a multi-source data fusion center and incorporates two core quantitative models. The first model calculates the market agility demand score, condensing external market demand pressures, such as order urgency, profitability, and certainty, into a single, dimensionless score. The second model calculates the inventory liquidity status score, assessing the flexibility and versatility of converting current inventory assets into downstream products from an internal asset allocation perspective. Through these two key indicators, the system accurately mathematically represents the conflicting market pull and internal asset flexibility in production and inventory management, providing a foundation for subsequent conflict assessment and decision-making.

[0062] The collaborative conflict assessment unit aims to diagnose the degree of production-sales contradiction currently faced by the system in real time. This unit uses market agility demand score and inventory liquidity status score as direct inputs, and calculates a collaborative conflict index through a specially constructed mathematical model. The design logic of this index is that when the market demands high agility but internal inventory liquidity is insufficient, the index will increase significantly, and vice versa, thus accurately measuring the intensity of the conflict between the two. Based on a set of conflict thresholds preset through historical data backtesting analysis, this unit performs real-time discrimination on the calculated collaborative conflict index, automatically classifying the current state of the system into different collaborative conflict situation levels, such as high conflict, medium conflict, or low conflict. This step discretizes continuous conflict states into clear and instructive strategy levels.

[0063] The evolution trend prediction unit aims to endow the system with the ability to anticipate future trends, upgrading it from passive response to proactive prediction. This unit continuously records and stores historical collaborative conflict indices, market agility demand scores, and inventory liquidity status scores calculated by preceding units, constructing a multi-dimensional historical scenario sequence by timestamps. This unit employs a pre-trained time-series prediction model, in this embodiment a Long Short-Term Memory (LSTM) network model, using this historical scenario sequence and externally collected macroeconomic data, such as industry prosperity indices and relevant industrial policies, as input to predict the collaborative conflict index for one or more future production cycles, ultimately generating an evolution sequence of the future collaborative conflict index. This predictive capability is the core of the system's proactive risk avoidance.

[0064] The dynamic decision management unit aims to transform the system's analysis and forecast results into specific, executable business instructions. This unit has two decision-making modes: one is the immediate response mode, which directly responds to the collaborative conflict situation level output by the collaborative conflict assessment unit. Through a preset strategy rule base or a pre-trained reinforcement learning model, it quickly generates immediate transformation strategies, such as issuing instructions for predictive production or increasing upstream reserves. The other is the forward-looking planning mode, which identifies potential future risk peaks based on the evolution sequence of the future collaborative conflict index generated by the evolution trend prediction unit and initiates multi-objective optimization calculations in advance to generate predictive inventory structure pre-configuration decisions. This aims to proactively smooth out peaks and valleys and alleviate future operational pressures through current fine-tuning of the inventory structure.

[0065] Example 2:

[0066] The process of generating market agility requirement scores is as follows:

[0067] Obtain the order weighting coefficient, expected profit margin, order completion probability, and expected delivery cycle of potential orders;

[0068] The agility contribution value of each order is obtained by multiplying the order weight coefficient, the expected profit margin of the order, and the order completion probability, and then dividing by the expected delivery cycle of the order.

[0069] Sum the agility contribution values ​​of all potential orders to generate the raw agility score;

[0070] The original agility score is normalized using a preset logic function to generate a market agility demand score.

[0071] This embodiment elaborates on the generation process of the market agility demand score in the contextual quantitative analysis unit. Its aim is to accurately quantify abstract market demand pressure into a dimensionless, standardized indicator within the (0,1) interval. ;

[0072] The contextual quantitative analysis unit obtains the order weight coefficient, expected profit margin, order completion probability, and expected delivery cycle of potential orders;

[0073] Order weighting coefficient It refers to the benchmark values ​​that are pre-set by the enterprise's sales strategy department based on the strategic importance of customers or the priority of orders. Its function is to reflect the differentiated importance of different orders at the strategic level and it is an internal sales strategy database of the enterprise.

[0074] Expected profit margin of orders It refers to the profitability of a single potential order calculated based on sales quotes and cost estimates. Its function is to measure the economic value of an order and it is a CRM system.

[0075] Order completion probability This refers to the order conversion probability assessment given by the sales team based on the progress of communication with customers and historical transaction data. Its function is to introduce probability weights to uncertain future demand and it is a CRM system or sales forecasting system.

[0076] Expected delivery time of the order This refers to the time from order confirmation to receiving the finished product that customers expect. Its purpose is to measure the urgency of order demand and it is part of the CRM system.

[0077] To prevent calculation errors, the system calculates the expected delivery time for all orders before performing the calculations. Conduct an inspection, if Less than a preset minimum positive number For example, if it is 0.1 days, then its value is set to... The system processes each potential order. Order weighting coefficient Expected profit margin of orders With order completion probability Multiply and then divide by the expected delivery period of the order. The agility contribution value of each order is obtained. The underlying logic of this calculation is that orders with high strategic importance, high profit margin, high probability of completion and short delivery cycle have higher requirements for the agility of the production system and thus a greater contribution value.

[0078] The system will The agility contribution values ​​of each potential order are summed to generate the raw agility score. The calculation process is represented by the following formula:

[0079] ;

[0080] in, It is the weighted total requirement of the entire market demand on the agility of the enterprise's production system; The dimension of is one part of time, and its value range is not fixed, which is not conducive to the stable calculation of subsequent models;

[0081] To facilitate subsequent calculations, the system uses a preset logical function to normalize the original agility score in order to generate a market agility demand score. In this embodiment, the Sigmoid function is used for this processing:

[0082] ;

[0083] in, The final generated market agility demand score is a dimensionless parameter, calculated in this step.

[0084] is the base of the natural logarithm, and is a constant;

[0085] The gain coefficient refers to a preset parameter used to control the steepness of the normalization curve; its function is to adjust the market agility demand distribution. For the raw score The sensitivity to change is pre-set by experts in the production and sales fields based on historical operating data or strategic objectives;

[0086] The translation parameter refers to a standard or balanced original agility score baseline value in the business scenario. Its function is to define the center point of the agility requirement score, which is also preset by experts based on the average or median of historical data. To ensure the consistency of the formula's dimensions, the parameter... Dimensions and Maintain consistency, representing one-third of the time; parameters The dimension is time, and its value can be set as a parameter. The countdown or history The reciprocal of the average to ensure the product The input is dimensionless and normalized near the business equilibrium point;

[0087] The innovation of this normalization method lies in its utilization of the smooth and bounded properties of logical functions, which not only... It stably maps to the 0,1 interval and effectively solves the interval out-of-bounds problem of traditional linear normalization methods when the system is cold-started or when extreme order data occurs, ensuring the market agility requirements are met. Stability and effectiveness under any circumstances;

[0088] The linear multiplication and division model used in this embodiment is a first-order approximation of the dose-effect relationship of each influencing factor. Its advantages lie in its simple calculation and intuitive business logic. In scenarios where higher model accuracy is required, nonlinear functions can be introduced to transform the variables. For example, a logarithmic function can be used to handle the delivery cycle. To reflect its urgency through nonlinear growth effects, thereby further improving the physical fidelity of the model.

[0089] Example 3:

[0090] The process of generating the inventory liquidity status score is as follows:

[0091] Obtain the total material value of each form of material and the universality coefficient calibrated by the process engineer;

[0092] Obtain the total value of all inventory materials;

[0093] Multiply the total value of each form of material by the corresponding universality coefficient to obtain the liquidity value of each form of material.

[0094] The total liquidity value is obtained by summing the liquidity values ​​of all forms of materials.

[0095] Divide the total liquidity value by the total value of all inventory items to generate the current inventory liquidity score;

[0096] This embodiment elaborates on the generation process of the inventory liquidity status score in the contextual quantitative analysis unit. Its aim is to quantify the overall flexibility and versatility of an enterprise's internal inventory assets into a dimensionless, standardized indicator within the range [0,1]. ;

[0097] The contextual quantitative analysis unit obtains the total material value of each form of material and the universality coefficient calibrated by the process engineer;

[0098] Materials in various forms refer to materials with different degrees of processing that exist in the inventory, such as ingots, slabs of specific specifications, hot-rolled coils, etc.

[0099] Total value of materials This refers to the form of inventory. The total value of all materials is used to measure the economic proportion of this type of material in the total inventory assets. It is the inventory asset module of the ERP system.

[0100] Universality coefficient This refers to a dimensionless parameter between [0,1], which quantifies the ability of a material in a certain form to be transformed into different downstream finished products. The higher the coefficient, the stronger the process flexibility of the material in that form, and the closer it is to a general semi-finished product. This calibration is performed professionally by experienced process engineers based on the production process roadmap. For example, a slab that can be used to produce ten different specifications of finished products will have a significantly higher versatility coefficient than a slab that can only be used to produce two specific specifications of finished products. A specific calibration method is to define the number of downstream finished product (SKU) types that the material j can be directly converted into as... And define the total number of all finished product types that a company can produce as Then the universality coefficient can be initially quantified as Then, the process engineer makes fine adjustments and confirms the results.

[0101] The system simultaneously retrieves the total value of all inventory materials. This value is from the ERP system and represents the total value of all materials in all forms. The sum of all additions;

[0102] The system for each material form The total value of materials in all forms With the corresponding universality coefficient Multiplying these together yields the flow value of materials in each form; this product... The innovative significance lies in the fact that it is not the original financial value of the material, but its value that can be flexibly allocated in the production process; the more versatile the material, the greater its contribution to improving the overall system's ability to cope with variables, and the higher its liquidity value.

[0103] The system will The total liquidity value is obtained by summing the liquidity values ​​of the various forms of materials.

[0104] During the calculation, the system first determines the total value of all inventory materials. Is it greater than zero? If If the value is zero, then the current inventory liquidity status will be directly divided. If the value is 0, the following division operation is performed: The system divides the total liquidity value by the total value of all inventory items to generate the current inventory liquidity score. The entire calculation process is represented by the following formula:

[0105] ;

[0106] in, The current inventory liquidity status is represented by a dimensionless parameter, which is calculated in this step.

[0107] This represents the total number of different types of inventory materials, which is an integer defined by the material master data in the ERP system.

[0108] Total liquidity value;

[0109] The value of is between [0,1]; when A higher value indicates that the company's inventory contains a higher proportion of versatile upstream or intermediate products, demonstrating good asset flexibility and the ability to quickly respond to diverse downstream orders; conversely, if... A lower value indicates that the inventory consists mostly of specialized finished products or work-in-progress, which are difficult to switch to other products and have poor inventory liquidity.

[0110] Example 4:

[0111] The calculation process for the collaborative conflict index is as follows:

[0112] Subtracting the inventory liquidity score from the value of 1 yields the degree of inventory illiquidity.

[0113] Multiplying market agility demand by the degree of inventory illiquidity yields the conflict baseline value;

[0114] The conflict baseline value is subjected to a power operation on a preset conflict sensitivity coefficient to generate a collaborative conflict index.

[0115] This embodiment elaborates on the calculation process of the collaborative conflict index in the collaborative conflict assessment unit; the purpose of this index is to divide the market agility demand, which represents the external market pull, into... The current state of inventory liquidity, which represents the flexibility of internal assets, is divided into... These two core opposing indicators are merged into a single indicator that can directly measure the system's decision-making pressure. ;

[0116] The system will subtract the current inventory liquidity score from the value of 1. To obtain the degree of inventory illiquidity ;

[0117] Inventory illiquidity It is a derived indicator that serves to inversely measure the degree of rigidity or specialization in an inventory system; when At a very high level, Close to 0; when When very low, Approaching 1; this shift makes the indicator more intuitively positively correlated with conflict or risk;

[0118] The system will segment market agility requirements With the degree of inventory illiquidity Multiply to obtain the conflict baseline value This step is the core logic of conflict modeling: only when market agility is required... High, and at the same time, the degree of inventory illiquidity is high. When the values ​​of the two are both very high, the contradiction between them will become prominent and their product will increase significantly; if either of the two values ​​is very low, the product will be very small, indicating that the system is in a good cooperative state and there is no significant conflict.

[0119] The system performs a power operation on the conflict baseline value using a preset conflict sensitivity coefficient to generate a collaborative conflict index. ;

[0120] Conflict sensitivity coefficient This refers to a preset, adjustable parameter that can be optimized through backtesting on historical data; its function is to adjust the collaborative conflict index. right and The degree of responsiveness to change; to calibrate this parameter, historical operational data can be collected to construct a model based on historical market agility requirements. Historical inventory liquidity status and the corresponding historical order delay delivery rate The dataset is composed of data; through regression analysis, an optimal selection is made. The value makes the conflict index calculated based on historical data... Actual order delay delivery rate The correlation between them reaches its maximum.

[0121] The complete calculation formula is as follows:

[0122] ;

[0123] in, , where is the cooperative conflict index, is a dimensionless parameter, and is calculated in this step.

[0124] Example 5:

[0125] The process for determining the level of collaborative conflict situations is as follows:

[0126] The collaborative conflict index is compared with the first preset threshold and the second preset threshold;

[0127] If the collaborative conflict index is less than the first preset threshold, the collaborative conflict situation is determined to be a low conflict situation.

[0128] If the collaborative conflict index is greater than or equal to the first preset threshold and less than or equal to the second preset threshold, then the collaborative conflict situation level is determined to be a medium conflict situation.

[0129] If the collaborative conflict index is greater than the second preset threshold, the collaborative conflict situation is determined to be a high conflict situation.

[0130] This embodiment is based on the calculated collaborative conflict index. This further elaborates on the process of determining the level of collaborative conflict situations; the purpose of this process is to transform continuous index values ​​into discrete strategy levels with clear business implications, thereby directly driving the subsequent decision-making system.

[0131] The collaboration conflict assessment unit is initialized with two key thresholds: a first threshold and a second threshold. These thresholds are not arbitrarily set, but rather based on in-depth backtesting analysis of the company's historical operational data. Specifically, it analyzes historical collaboration conflict indices. The correlation between the sequence and corresponding negative outcomes such as order delays and supply chain disruptions is analyzed to identify the inflection point where the probability of future operational risks begins to rise sharply when the index exceeds a certain value. In this embodiment, the following is determined through analysis:

[0132] The first preset threshold is set to 0.3;

[0133] The second preset threshold is set to 0.75;

[0134] During system operation, the collaborative conflict assessment unit calculates the collaborative conflict index in real time. The current level of collaborative conflict situation is determined by comparing the data with the first and second preset thresholds and according to the following rules:

[0135] If the collaborative conflict index If the situation is classified as low-conflict, it indicates that the current market demand pressure is relatively small, or the inventory assets have very ample liquidity, the system's production and sales coordination is good, and the operational risk is low.

[0136] If the collaborative conflict index If the situation is classified as a medium-level conflict situation, it indicates that the system has experienced some production and sales pressure, and there is a certain degree of mismatch between market demand and inventory flexibility, which needs attention, but has not yet reached a high-risk level.

[0137] If the collaborative conflict index If the situation is classified as a high-conflict situation, it indicates that the market's demand for agility far exceeds the current flexible supply capacity of inventory, the system's production and sales contradictions are acute, and the risk of future order delays is extremely high, requiring immediate intervention.

[0138] Example 6:

[0139] When the collaborative conflict situation is classified as a high-conflict situation, a predictive production strategy is output with the goal of ensuring production agility.

[0140] When the collaborative conflict situation level is medium, output instructions to maintain the current production and inventory strategy;

[0141] When the collaborative conflict situation is classified as low-conflict situation, the output strategy is to increase upstream reserves with the goal of ensuring inventory liquidity.

[0142] This embodiment follows the determined collaborative conflict situation level and describes the specific generation process of the instant transformation strategy in the dynamic decision management unit. The core of this process is based on a preset decision rule base or a pre-trained reinforcement learning decision model, which maps different situation levels to the optimal instant response strategy, thereby forming a complete perception-evaluation-decision link.

[0143] The specific strategy generation logic is as follows:

[0144] When the collaborative conflict assessment unit determines the collaborative conflict situation to be a high-conflict situation, the dynamic decision management unit will prioritize and overridden the goal of ensuring production agility, and automatically output a predictive production strategy aimed at ensuring production agility. This strategy is then translated into specific execution instructions and issued to the ERP / MES system. For example, the most versatile slabs in the inventory may be pre-processed according to the specifications of potential orders with the highest probability of completion in the sales forecast system, and converted into dedicated work-in-process. Although this will reduce the overall liquidity of the inventory, in the current high-conflict situation, it is a necessary and optimal choice to sacrifice long-term flexibility to meet the urgent market demand.

[0145] When the collaborative conflict situation is classified as a medium-level conflict situation, the dynamic decision management unit determines that the current system pressure is controllable and no radical adjustment measures are needed. At this time, it will output an instruction to maintain the current production and inventory strategy. This means that the system will continue to execute according to the original production plan without any additional inventory transformation. The goal of this strategy is to maintain system stability and avoid unnecessary production costs or changes in inventory structure due to overreaction.

[0146] When the collaborative conflict situation is classified as low-conflict, the dynamic decision-making management unit determines that the current market pressure is low, making it the optimal time to optimize inventory structure and enhance future risk resistance. At this time, it will prioritize ensuring inventory liquidity and automatically output a strategy to increase upstream reserves. This strategy is translated into specific execution instructions, such as increasing the procurement or production of general-purpose ingots, or retroactively or temporarily halting the conversion of some downstream work-in-process to more specialized forms. This aims to improve the current inventory liquidity situation. To prepare for potential peaks in agility requirements in the future, and to stockpile resources in advance.

[0147] Example 7:

[0148] The evolution trend prediction unit uses a long short-term memory network model, taking historical situational sequences and external macroeconomic data as inputs, to output the evolution sequence of the future collaborative conflict index.

[0149] This embodiment provides a detailed explanation of the specific technical implementation of the evolution trend prediction unit; in order to achieve the functional leap from reactive decision-making to predictive decision-making, the core of this unit is the adoption of an advanced time series prediction model;

[0150] The evolution trend prediction unit adopts the Long Short-Term Memory (LSTM) network model. The fundamental technical consideration for choosing the LSTM model is that the evolution process of production and sales coordination conflict has a significant time dependence. That is, the future conflict state is not only related to the current state, but also affected by a variety of long-term factors such as market demand fluctuations and inventory structure adjustments in the past period. Through its unique gating mechanism, the LSTM network can learn and remember long-term dependencies in time series data very effectively. This gives it a significant advantage over traditional statistical models or simple recurrent neural networks (RNNs) when dealing with industrial forecasting problems with complex dynamics and historical memory effects.

[0151] In model applications, the evolutionary trend prediction unit takes historical context sequences and external macroeconomic data as inputs;

[0152] Contextual history sequence This refers to the time continuously recorded by the system up to the current time. The internal state vector sequence, which at each time step contains at least a cooperative conflict index. Market agility requirements And the current status of inventory liquidity This multidimensional sequence is the core basis for the model to learn the internal evolutionary laws of the system.

[0153] External macroeconomic data This refers to external environmental data vectors related to the titanium plate industry, such as commodity price indices, downstream industry prosperity indices, and changes in relevant trade policies; As input, its technical purpose is to enable models to capture future market demand trends driven by changes in the macro environment, thereby improving the accuracy and foresight of forecasts;

[0154] The model's workflow is to stitch or merge the results... and The input is fed into a pre-trained LSTM network, which, through its complex nonlinear transformations, outputs an evolution sequence of the future cooperative conflict index. ; It is a vector that contains information about the future. The collaboration conflict index is calculated over a time step, such as each week within the next quarter. The predicted value.

[0155] Example 8:

[0156] The process of generating predictive inventory structure pre-configuration decisions is as follows:

[0157] By analyzing the evolution sequence of the future collaborative conflict index, we can identify future time points in the sequence where the predicted value exceeds the high conflict threshold, in order to generate an early warning of future conflict peaks.

[0158] In response to early warnings of future conflict peaks, multi-objective optimization calculations are initiated to generate predictive inventory structure pre-configuration decisions;

[0159] This embodiment is based on the predicted evolution sequence of the future collaborative conflict index. This paper further elaborates on the generation process of predictive inventory structure pre-configuration decisions in the dynamic decision management unit. This process aims to use predictive information to shift decision-making behavior from post-event remediation to pre-event avoidance, and proactively optimize the future system state.

[0160] The system analyzes the evolution sequence of the future collaborative conflict index and identifies future time points in the sequence where the predicted value exceeds the high conflict threshold, in order to generate a warning of future conflict peaks.

[0161] The dynamic decision management unit receives prediction sequences from the evolution trend prediction unit. It uses the same high conflict threshold, such as 0.75, as the standard for determining the situation level, to examine each predicted value in the sequence one by one. Once any one or more future points in time are discovered... Predicted value The system will automatically mark these time points and generate a warning of future conflict peaks; this warning information includes not only the time when the risk occurs, but also the predicted severity of the conflict.

[0162] In response to an early warning of future conflict peaks, the system initiates multi-objective optimization calculations to generate predictive inventory structure pre-configuration decisions.

[0163] Once an early warning is received, the dynamic decision-making management unit will not wait until the risk arrives to respond passively, but will immediately activate a multi-objective optimization calculation module. The optimization objective of this module is to minimize the current production adjustment costs and maximize the improvement of the system state at the future early warning time, that is, to minimize the collaborative conflict index at that time. The optimization process seeks the Pareto optimal solution between the MES and ERP systems. This optimization process is also constrained by production constraints such as real-time equipment capacity, material availability, and personnel load obtained from the MES and ERP systems. The decision variables of the optimization algorithm, such as the genetic algorithm or the particle swarm optimization algorithm, are a series of inventory adjustment actions that can be performed at present, such as how many general slabs to produce and how many ingots to purchase.

[0164] The algorithm simulates the execution of these actions to assess future market agility requirements. and inventory liquidity The impact, and recalculate the warning time point. The value is used to arrive at a set of operations that can smooth out future conflict peaks at the lowest cost and to the greatest extent possible in the present; this set of operations is the predictive inventory structure pre-configuration decision.

[0165] For example, if the system predicts a peak in demand in four weeks, the decision might not be to immediately begin large-scale production. Instead, it might recommend moderately increasing upstream general-purpose ingot reserves within the next week and so on. This way, when the high-agility demand orders actually arrive four weeks later, the system will have a more flexible inventory base, effectively preventing issues from arising at that time. The index surged dramatically.

[0166] Example 9:

[0167] It also includes a closed-loop feedback execution step, which is as follows:

[0168] Transform real-time conversion strategies or predictive inventory structure pre-configuration decisions into production or procurement instructions, and issue them to the enterprise resource planning system and manufacturing execution system for execution;

[0169] After the instruction is executed, the multi-source data fusion center collects the changed system status in the next calculation cycle, so that all subsequent calculations are updated based on the latest reality, thus forming a closed-loop management process.

[0170] Based on Example 8, this embodiment further elaborates on the closed-loop feedback execution steps to ensure the continuous and adaptive operation of the system. This step is a key link connecting intelligent decision-making with the physical world and enabling the system to have self-iteration and optimization capabilities, constituting the last mile of the entire management system.

[0171] Whether it's the real-time conversion strategies generated by the dynamic decision management unit or the predictive inventory structure pre-configuration decisions, these decision-making schemes generated at the information level will be automatically converted into production or procurement instructions by the system and issued to the Enterprise Resource Planning (ERP) system and Manufacturing Execution System (MES) for execution. For example, predictive production strategies will be converted into production work orders for specific materials and issued to the MES system; strategies to increase upstream reserves may be converted into purchase orders and issued to the ERP system. This step puts virtual decisions into practice in the physical world.

[0172] After the instruction is executed, the multi-source data fusion center, as the core of the closed loop, collects the changed system status in the next calculation cycle. When the MES reports the completion of the production work order or the ERP confirms the arrival of the purchase order, the actual inventory status of the enterprise, including quantity, form, and value, has undergone real changes. In the next calculation cycle, such as one hour or one day later, the multi-source data fusion center, as the starting point of the system, will collect the latest inventory data from the MES and ERP, which already includes the results of this execution.

[0173] Based on the updated data, all subsequent calculations are updated according to the latest reality to form a closed-loop management process; the newly collected data will generate a new original dataset of collaborative scenarios, which will lead the scenario quantitative analysis unit to calculate new market agility requirements. And the current status of inventory liquidity Then, the collaborative conflict assessment unit will calculate a new collaborative conflict index. And determine the new situation level; the entire perception-evaluation-decision chain will begin a new cycle based on this updated reality.

[0174] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An intelligent management system for coordinating titanium plate inventory and production planning, characterized in that, include: The multi-source data fusion center is used to acquire operational data from manufacturing execution systems, enterprise resource planning systems, customer relationship management systems, commodity markets, and supply chain management systems, and to fuse and process the acquired operational data to generate a raw dataset for collaborative scenarios. The scenario quantitative analysis unit is used to receive the original dataset of collaborative scenarios and perform quantitative calculations to generate market agility demand score and inventory liquidity status score. The collaborative conflict assessment unit is used to combine market agility demand score and inventory liquidity status score to calculate collaborative conflict index, and to judge collaborative conflict index according to preset conflict threshold to determine collaborative conflict situation level. The evolution trend prediction unit is used to construct a scenario historical sequence based on historical collaborative conflict index, market agility demand score and inventory liquidity status score, and combine it with external macroeconomic data to make time series predictions in order to generate the evolution sequence of future collaborative conflict index. The dynamic decision management unit is used to generate immediate transformation strategies in response to the level of collaborative conflict situations, and to generate predictive inventory structure pre-configuration decisions based on the evolution sequence of future collaborative conflict indices. The process of generating market agility requirement scores is as follows: Obtain the order weighting coefficient, expected profit margin, order completion probability, and expected delivery cycle of potential orders; The agility contribution value of each order is obtained by multiplying the order weight coefficient, the expected profit margin of the order, and the order completion probability, and then dividing by the expected delivery cycle of the order. Sum the agility contribution values ​​of all potential orders to generate the raw agility score; The original agility score is normalized using a preset logic function to generate a market agility demand score. The process of generating the inventory liquidity status score is as follows: Obtain the total material value of each form of material and the universality coefficient calibrated by the process engineer; Obtain the total value of all inventory materials; Multiply the total value of each form of material by the corresponding universality coefficient to obtain the liquidity value of each form of material. The total liquidity value is obtained by summing the liquidity values ​​of all forms of materials. Divide the total liquidity value by the total value of all inventory items to generate the current inventory liquidity score; The calculation process for the collaborative conflict index is as follows: Subtracting the inventory liquidity score from the value of 1 yields the degree of inventory illiquidity. Multiplying market agility demand by the degree of inventory illiquidity yields the conflict baseline value; The conflict baseline value is subjected to a power operation on a preset conflict sensitivity coefficient to generate a collaborative conflict index.

2. The intelligent management system for coordinating titanium plate inventory and production planning according to claim 1, characterized in that, The process for determining the level of collaborative conflict situations is as follows: The collaborative conflict index is compared with the first preset threshold and the second preset threshold; If the collaborative conflict index is less than the first preset threshold, the collaborative conflict situation is determined to be a low conflict situation. If the collaborative conflict index is greater than or equal to the first preset threshold and less than or equal to the second preset threshold, then the collaborative conflict situation level is determined to be a medium conflict situation. If the collaborative conflict index is greater than the second preset threshold, the collaborative conflict situation is determined to be a high-conflict situation.

3. The intelligent management system for coordinating titanium plate inventory and production planning according to claim 2, characterized in that, The process of generating an instant conversion strategy is as follows: When the collaborative conflict situation is classified as a high-conflict situation, a predictive production strategy is output with the goal of ensuring production agility. When the collaborative conflict situation level is medium, output instructions to maintain the current production and inventory strategy; When the collaborative conflict situation is classified as low-conflict, the output strategy is to increase upstream reserves with the goal of ensuring inventory liquidity.

4. The intelligent management system for coordinating titanium plate inventory and production planning according to claim 1, characterized in that, The evolution trend prediction unit uses a long short-term memory network model, taking historical situational sequences and external macroeconomic data as inputs, to output the evolution sequence of the future collaborative conflict index.

5. The intelligent management system for coordinating titanium plate inventory and production planning according to claim 1, characterized in that, The process of generating predictive inventory structure pre-configuration decisions is as follows: By analyzing the evolution sequence of the future collaborative conflict index, we can identify future time points in the sequence where the predicted value exceeds the high conflict threshold, in order to generate an early warning of future conflict peaks. In response to early warnings of future conflict peaks, multi-objective optimization calculations are initiated to generate predictive inventory structure pre-configuration decisions.

6. The intelligent management system for coordinating titanium plate inventory and production planning according to claim 5, characterized in that, It also includes a closed-loop feedback execution step, which is as follows: Transform real-time conversion strategies or predictive inventory structure pre-configuration decisions into production or procurement instructions, and issue them to the enterprise resource planning system and manufacturing execution system for execution; After the instruction is executed, the multi-source data fusion center collects the changed system status in the next calculation cycle, so that all subsequent calculations are updated based on the latest reality, thus forming a closed-loop management process.

Images