Tungsten alloy production scheduling decision method and system based on whole-process data perception

Abstract

The application discloses a tungsten alloy production scheduling decision method and system based on full-process data perception, relates to the technical field of tungsten alloy production scheduling, and specifically comprises the following steps: in the case that the state is rewritten and the original state is not inherited around the fracture position, a constraint embedding mapping sequence based on state order is constructed in a multi-dimensional state identification sequence, a process identification before the fracture position is embedded into an identification section after the fracture position in time sequence, and a mapping correspondence across the identification sections is established, so that a continuous full-process data perception state is formed. The application solves the problem that the full-process data perception state is fractured and scheduling decision cannot be accurately performed due to the state rewriting of batches before process switching and the non-inheritance of the original state in the tungsten alloy production process, realizes the reconstruction and continuous expression of the fractured state, and accurately deduces the execution sequence and dynamically adjusts the scheduling sequence based on the continuous state, so that the continuity and accuracy of production scheduling are improved.

Description

Technical Field

[0001] This invention relates to the field of tungsten alloy production scheduling technology, specifically to a tungsten alloy production scheduling decision-making method and system based on full-process data perception. Background Technology

[0002] Tungsten alloy production scheduling decision-making based on end-to-end data perception refers to a scheduling method that continuously collects and perceives data from each process stage of tungsten alloy production, including raw material preparation, powder proportioning, pressing, sintering, subsequent processing, and quality inspection. This constructs a data system covering the entire production process, enabling optimized allocation and dynamic decision-making of production resources and process sequences. Existing technologies typically rely on information platforms such as industrial sensing equipment, Manufacturing Execution Systems (MES), and Enterprise Resource Planning (ERP) systems to collect and summarize real-time data on equipment operating status, process parameters, production progress, and quality. A unified data model is then formed through data preprocessing and feature extraction. Based on this, a scheduling decision model is constructed by combining production process constraints, equipment capacity constraints, and order demand information. Production tasks are decomposed and prioritized to generate corresponding production plans and scheduling schemes. During the execution phase, continuous monitoring of real-time data during production allows for dynamic adjustments to the scheduling results, thereby achieving coordination between the production plan and actual execution. Overall, this type of technology typically includes multiple stages such as data acquisition and sensing, data fusion and modeling, scheduling strategy generation, production execution control, and process feedback optimization. Through information interaction and collaboration between these stages, it enables the orderly operation of the tungsten alloy production process and the continuous optimization of scheduling decisions.

[0003] The existing technology has the following shortcomings: In the tungsten alloy production process, when a batch completes a preceding process and is marked by the automatic system as ready to proceed to the next process, but a process parameter adjustment operation is triggered before proceeding to the next process, the batch's status record may be rewritten in the overall process data perception. Because the status rewriting process regenerates the batch status based on the current process parameters without inheriting or associating the completion status of the preceding process, the original process completion status is overwritten or broken in the overall process data perception results, resulting in a discontinuity in the batch status evolution chain. Existing technologies cannot make scheduling decisions based on the overall process data perception status when a batch's status is rewritten before process switching and the original status is not inherited. This makes it impossible to accurately determine whether the batch has completed the preceding process during scheduling, leading to the batch being incorrectly identified as incomplete or in an abnormal state in the scheduling decision. This affects the batch's flow judgment in subsequent processes, ultimately adversely impacting the continuity and accuracy of production scheduling.

[0004] The information disclosed in the background section 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

[0005] The purpose of this invention is to provide a tungsten alloy production scheduling decision-making method and system based on full-process data perception, so as to solve the problems in the background art mentioned above.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a tungsten alloy production scheduling decision-making method based on full-process data perception, specifically including the following steps: S1. Construct a multi-dimensional state identifier sequence corresponding to each batch during the tungsten alloy production process. The multi-dimensional state identifier sequence includes process identifier, time sequence identifier and process trigger identifier. The multi-dimensional state identifier sequence is arranged and parsed in an orderly manner to identify the fracture position in the multi-dimensional state identifier sequence in order to determine whether the batch has undergone state rewriting before process switching and whether the original state has not been inherited. S2. Around the fracture location, in the case of state rewriting and the original state not being inherited, construct a constraint embedding mapping sequence based on state order in the multi-dimensional state identifier sequence, embed the process identifiers before the fracture location into the identifier segments after the fracture location in chronological order, and establish a mapping correspondence across identifier segments to form a continuous full-process data perception state. S3. Based on the continuous full-process data perception status, jointly deduce the process identifier and constraint embedding relationship in the multi-dimensional status identifier sequence, calculate the execution order of the batch in each process, and determine the scheduling order of the batch entering the subsequent process according to the execution order. S4. Combine the scheduling order to continuously update the multidimensional state identifier sequence and constraint embedding relationship, and simultaneously correct the scheduling order so as to realize the dynamic adjustment of the scheduling order as the multidimensional state identifier sequence changes.

[0007] Preferably, S1 specifically includes the following steps: S101. Construct a multi-dimensional status identifier sequence corresponding to each batch during the tungsten alloy production process. The multi-dimensional status identifier sequence includes process identifier, time sequence identifier, and process trigger identifier. Process identifier is generated based on the execution node of each process of the batch, time sequence identifier is generated based on the order of the batch in the production process, and process trigger identifier is generated based on the process parameter adjustment behavior. The process identifier and process trigger identifier are combined and arranged according to the time sequence identifier to form a multi-dimensional status identifier sequence. S102. Perform ordered arrangement and parsing on the multidimensional state identifier sequence, sort the multidimensional state identifier sequence according to the time sequence identifier, and perform process continuity verification on adjacent identifiers in combination with process identifiers. At the same time, mark the position of state change according to the process trigger identifier to form an ordered arrangement result. S103. Based on the ordered arrangement results, identify the break position in the multidimensional state identifier sequence. By comparing the time sequence identifier and process identifier of adjacent multidimensional state identifiers, determine the position where the time sequence is discontinuous or the process identifier is discontinuous as the break position. Based on the changes in the process trigger identifier before and after the break position, determine whether the batch has undergone state rewriting before the process switch and whether the original state has not been inherited.

[0008] Preferably, S103 is as follows: Based on the ordered arrangement results, adjacent multidimensional state identifiers in the multidimensional state identifier sequence are extracted one by one according to the time sequence identifier, and adjacent multidimensional state identifier pairs are constructed. By recording the time sequence identifier difference of each pair of adjacent multidimensional state identifiers and the corresponding process identifier sequence, a comparison basis is provided for subsequent fracture location identification. Based on adjacent multidimensional state identifier pairs, the time sequence identifier difference and the process identifier sequence are synchronously compared. When the time sequence identifier difference does not meet the continuous increasing relationship or the process identifier sequence does not meet the preset process connection order, the corresponding position is marked as a candidate break position, and the candidate break positions are centrally marked. For candidate fracture locations, the process trigger identifier sequence before and after the candidate fracture location is extracted. The trigger order and trigger type in the process trigger identifier sequence are matched and analyzed. When there is a jump in the trigger order or a change in the trigger type of the process trigger identifiers before and after the candidate fracture location, the candidate fracture location is determined as the fracture location. Based on this, it is determined that the batch has undergone state rewriting before the process switch and the original state has not been inherited.

[0009] Preferably, S2 specifically includes the following steps: S201. Around the fracture location, in the case of state rewriting and the original state not being inherited, extract the process identifier sequence before the fracture location and the identifier segment after the fracture location from the multi-dimensional state identifier sequence, and arrange the process identifier sequence before the fracture location in order according to the time sequence identifier to construct a constraint embedding mapping sequence based on state order. S202. In the constraint embedding mapping sequence based on state order, the process identifiers before the fracture position are embedded into the identifier segments after the fracture position in chronological order. By maintaining the correspondence between the chronological identifiers and the process identifiers, the embedding positions are matched and arranged to form a multidimensional state identifier sequence after embedding. S203. Based on the embedded multidimensional state identifier sequence, a cross-identifier segment mapping correspondence is established between the process identifier before the fracture position and the identifier segment after the fracture position. By associating and matching process identifiers with consistent time sequence identifiers, a continuous full-process data perception state is formed.

[0010] Preferably, S202 specifically refers to: In the constraint embedding mapping sequence based on state order, the process identifiers before the fracture position are extracted one by one according to the time sequence identifier, and a one-to-one correspondence between the process identifier and the time sequence identifier is established using the time sequence identifier as an index, forming a sequence of process identifiers to be embedded. Based on the sequence of process identifiers to be embedded, the positions of the identifiers in the segment after the break point are matched according to the time sequence. The embedding position is determined by comparing the relative positional relationship of the time sequence identifiers, and the process identifiers to be embedded are inserted into the corresponding positions in sequence to form the initial embedding structure. Based on the initial embedding structure, the consistency between the time sequence identifier and the process identifier of the embedding position is checked. By maintaining the correspondence between the time sequence identifier and the process identifier, the embedding positions are matched and arranged to form a multi-dimensional state identifier sequence after embedding.

[0011] Preferably, S203 is as follows: Based on the embedded multidimensional state identifier sequence, the process identifier sequence before the fracture position and the corresponding process identifier sequence in the identifier segment after the fracture position are extracted, and the two sequences are synchronized and aligned according to the time sequence identifier to construct a candidate mapping corresponding set across the identifier segment. In the candidate mapping correspondence set across the identifier segment, process identifiers with the same time sequence identifier value are matched one by one, and a one-to-one correspondence is established according to the size of the time sequence identifier value. By comparing the arrangement position of the process identifier in the sequence, process identifier pairs with the same time sequence identifier value and corresponding position are selected to form a cross-identifier segment mapping correspondence. Based on the mapping relationship across identifier segments, the process identifiers before the fracture position are associated with the identifier segments after the fracture position. By maintaining the correspondence between process identifiers with the same time sequence identifier value, the embedded multidimensional state identifier sequence is integrated to form a continuous full-process data perception state.

[0012] Preferably, S3 specifically includes the following steps: S301. Based on the continuous full-process data perception status, the process identifier and constraint embedding relationship in the multi-dimensional status identifier sequence are jointly deduced. By extracting the process identifier sequence arranged in time order in the multi-dimensional status identifier sequence, and combining the cross-identifier segment mapping correspondence in the constraint embedding relationship, the connection relationship between the process identifiers is associated one by one to form an ordered connection sequence of process identifiers. S302. Based on the ordered connection sequence of process identifiers, calculate the execution order of the batch in each process. Sort the process identifiers according to the time sequence identifiers and adjust the positions of process identifiers with overlapping positions according to the constraint embedding relationship to determine the execution order of the batch in each process. S303. After the execution order of the batch in each process is determined, the scheduling order of the batch entering the subsequent process is determined according to the execution order. The process identifiers in the multi-dimensional status identifier sequence are arranged in order according to the execution order, and the scheduling order of the batch entering each process is generated according to the arrangement order.

[0013] Preferably, S302 is as follows: Based on the ordered sequence of process identifiers, the time sequence identifiers corresponding to each process identifier are extracted, and the process identifiers are initially sorted according to the time sequence identifiers to form a sequence of process identifiers arranged in time order. For a sequence of process identifiers arranged in chronological order, overlapping positions of process identifiers with the same chronological order are identified based on constraint embedding relationships. By rearranging the process identifiers in the overlapping positions according to the mapping order in the constraint embedding relationships, a position adjustment sequence is formed. The sequence is adjusted according to the position, and the order of the process identifiers is confirmed. By verifying the arrangement of each process identifier in the sequence one by one, a process identifier sequence with no overlap and a unique order is formed, and the execution order of the batch in each process is determined.

[0014] Preferably, S4 is as follows: Based on the scheduling order, the multidimensional state identifier sequence and constraint embedding relationship are continuously updated. By sequentially traversing the process identifiers in the multidimensional state identifier sequence according to the scheduling order, the newly added process identifiers are inserted into the corresponding positions according to the time sequence identifiers. At the same time, the cross-identifier segment mapping in the constraint embedding relationship is synchronously adjusted to form the updated multidimensional state identifier sequence and constraint embedding relationship. Based on the continuously updated multidimensional status identifier sequence and constraint embedding relationship, the scheduling order is synchronously corrected. The scheduling order is reordered according to the arrangement order of the process identifiers in the updated multidimensional status identifier sequence and the mapping correspondence in the constraint embedding relationship. The scheduling order is then dynamically adjusted according to the reordering result as the multidimensional status identifier sequence changes.

[0015] Preferably, the tungsten alloy production scheduling and decision-making system based on full-process data perception includes a full-process perception and analysis module, a state fracture reconstruction module, a process deduction and decision-making module, and a scheduling dynamic control module. The full-process perception and analysis module constructs a multi-dimensional status identifier sequence corresponding to each batch during the tungsten alloy production process. The multi-dimensional status identifier sequence includes process identifier, time sequence identifier, and process trigger identifier. The module performs ordered arrangement and analysis of the multi-dimensional status identifier sequence to identify the fracture location in the multi-dimensional status identifier sequence, so as to determine whether the batch has undergone state rewriting before process switching and whether the original state has not been inherited. The state fracture reconstruction module, around the fracture location, constructs a constraint embedding mapping sequence based on the state order in the multi-dimensional state identifier sequence when state rewriting occurs and the original state is not inherited. It embeds the process identifiers before the fracture location into the identifier segments after the fracture location in chronological order and establishes a mapping correspondence across identifier segments to form a continuous full-process data perception state. The process deduction and decision-making module jointly deduces the process identifier and constraint embedding relationship in the multi-dimensional status identifier sequence based on the continuous full-process data perception status, calculates the execution order of the batch in each process, and determines the scheduling order of the batch to enter the subsequent processes according to the execution order. The scheduling dynamic control module continuously updates the multidimensional state identifier sequence and constraint embedding relationship in conjunction with the scheduling order, and synchronously corrects the scheduling order to achieve dynamic adjustment of the scheduling order as the multidimensional state identifier sequence changes.

[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention constructs a multi-dimensional state identifier sequence and introduces a collaborative expression of time sequence identifiers, process identifiers, and process trigger identifiers, achieving a detailed full-process characterization of batch states during tungsten alloy production. Based on this, through ordered arrangement analysis and fracture location identification, it can accurately locate fracture nodes in the state chain even when batch state rewriting occurs and the original state is not inherited. Furthermore, it determines the cause of fracture by combining process trigger behavior, thus solving the problem of lost state completion status of preceding processes due to state rewriting in traditional technologies. Further, by constructing a constraint embedding mapping sequence based on state sequence, it embeds the process identifier before fracture into the identifier segment after fracture and establishes a cross-identifier segment mapping correspondence, achieving a structured reconstruction of the fracture state. This restores the state perception of the entire process to a continuous and consistent state expression, fundamentally eliminating the impact of discontinuous state chains on subsequent scheduling and judgment.

[0017] 2. This invention, through joint deduction of continuous, end-to-end data perception status and fusion calculation of process identifiers and constraint embedding relationships, can accurately determine the actual execution order of batches in each process and generate a scheduling order consistent with the actual production status. Simultaneously, it continuously updates the multi-dimensional state identifier sequence and constraint embedding relationships based on the scheduling order and synchronously corrects the scheduling order, enabling the scheduling process to dynamically adjust with changes in production status, thereby achieving real-time consistency between scheduling decisions and actual production. Compared to existing technologies, this solution not only maintains the accuracy of scheduling judgments in complex state rewriting scenarios but also maintains the stability and adaptability of scheduling results during the continuous evolution of the production process, improving the scheduling continuity, execution reliability, and overall production efficiency in tungsten alloy production. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a schematic diagram of the process of the present invention.

[0020] Figure 2 This is a schematic diagram of the modules of the present invention. Detailed Implementation

[0021] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0022] This invention provides, for example Figure 1 The tungsten alloy production scheduling decision-making method based on end-to-end data perception, as shown, specifically includes the following steps: S1. Construct a multi-dimensional state identifier sequence corresponding to each batch during the tungsten alloy production process. The multi-dimensional state identifier sequence includes process identifier, time sequence identifier and process trigger identifier. The multi-dimensional state identifier sequence is arranged and parsed in an orderly manner to identify the fracture position in the multi-dimensional state identifier sequence in order to determine whether the batch has undergone state rewriting before process switching and whether the original state has not been inherited. In this embodiment, S1 specifically includes the following steps: S101. Construct a multi-dimensional status identifier sequence corresponding to each batch during the tungsten alloy production process. The multi-dimensional status identifier sequence includes process identifier, time sequence identifier, and process trigger identifier. Process identifier is generated based on the execution node of each process of the batch, time sequence identifier is generated based on the order of the batch in the production process, and process trigger identifier is generated based on the process parameter adjustment behavior. The process identifier and process trigger identifier are combined and arranged according to the time sequence identifier to form a multi-dimensional status identifier sequence. In the tungsten alloy production process, the status records generated at each process execution node of a batch can be collected through the production execution system or data acquisition interface. Each process completion, equipment start-up, or end is used as a collection trigger point, and a corresponding identifier is attached to each status record. The process identifier can be assigned a value through a predefined process number or process name, the time sequence identifier can be generated by sequentially encoding the collection time or generating an incremental sequence based on timestamps, and the process trigger identifier is marked according to whether process parameter adjustment occurs. For example, when the sintering temperature or pressing pressure is adjusted, a corresponding identifier is generated. In the implementation process, the process identifier, time sequence identifier, and process trigger identifier of the same batch can be combined and arranged according to the time sequence of the batch in the production process to form a time-progressing identifier sequence. For example, if a batch goes through pressing, sintering, and processing processes in sequence, and a parameter adjustment occurs before sintering, the corresponding sequence can be represented as pressing identifier - time identifier - no trigger, sintering identifier - time identifier - trigger, and processing identifier - time identifier - no trigger. In this way, discrete production states can be transformed into a continuous sequence, providing a basis for subsequent identification of state changes and fracture locations.

[0023] The tungsten alloy production process refers to a continuous processing flow that uses tungsten powder as raw material, through batching, pressing, sintering, and subsequent processing to form alloy products. A batch represents a group of product units formed under the same production conditions and flowing together in each process. A multi-dimensional state identifier sequence refers to a serialized data structure that organizes different types of state information in chronological order around the same batch. Process identifiers characterize the current processing stage of the batch, time sequence identifiers describe the sequential relationship of each state in the production process, and process trigger identifiers reflect whether parameter adjustments occur during production. Process execution nodes refer to the key locations in each process that can generate state records, such as equipment start-up points, processing completion points, or material transfer points. Process parameter adjustment refers to operations that modify processing conditions such as temperature, pressure, or time during production. These operations often cause state records to be updated or rewritten. By introducing process trigger identifiers, these changes can be explicitly represented, thus making the entire identifier sequence reflect both the processing sequence and the possible state changes that may occur during processing.

[0024] S102. Perform ordered arrangement and parsing on the multidimensional state identifier sequence, sort the multidimensional state identifier sequence according to the time sequence identifier, and perform process continuity verification on adjacent identifiers in combination with process identifiers. At the same time, mark the position of state change according to the process trigger identifier to form an ordered arrangement result. When performing ordered parsing of a multidimensional state identifier sequence, the corresponding time sequence identifier can be extracted from each identifier record in the sequence and sorted according to time order. This restores the identifiers that may have been out of order or delayed in writing to the actual production sequence. After sorting, the adjacent identifiers are compared one by one based on the sorted sequence. The process identifier of the current identifier is matched with the process identifier of the next identifier to determine whether it conforms to the pre-set process flow path. For example, after pressing, sintering should begin. If the processing identifier appears directly after pressing, it is considered discontinuous. At the same time, the process trigger identifier is combined to locate the position of the process trigger identifier in the sorted sequence and mark these positions as state change positions. For example, if a batch has a temperature adjustment before sintering, it is marked at the corresponding identifier position. Through the combined processing of sorting, continuity verification and trigger marking, an arrangement sequence that conforms to both time order and process connection can be obtained. For example, the original record is sintering, pressing, processing. After sorting, it is adjusted to pressing, sintering, processing. The order is then verified to be reasonable, and the trigger position is marked at the sintering node, thus forming an ordered arrangement result for subsequent fracture identification.

[0025] Ordered arrangement parsing refers to sorting and structuring various identifiers in a multidimensional state identifier sequence according to predetermined rules, ensuring consistency of identifiers in both the time and process dimensions. Time sequence identifiers reflect the sequential relationship between states and serve as the foundation for the sorting process. Process identifiers indicate the specific processing stage of a batch and are the core element for determining process connections. Adjacent identifiers refer to two adjacent state records in the sorted sequence and are the basic unit for continuity judgment. Process continuity verification involves matching process identifiers in adjacent identifiers according to a preset process execution order to identify any jumps or reversals. Process trigger identifiers characterize parameter adjustment behaviors during production; locating these identifiers in the sequence identifies the position where a state change occurs. State change position refers to the key node in the sequence corresponding to the appearance of a process trigger identifier. The ordered arrangement result is a sequence structure formed after sorting, continuity verification, and state change marking, reflecting both time sequence, process connections, and the distribution of state changes.

[0026] S103. Based on the ordered arrangement results, identify the break position in the multidimensional state identifier sequence. By comparing the time sequence identifier and process identifier of adjacent multidimensional state identifiers, determine the position where the time sequence is discontinuous or the process identifier is discontinuous as the break position. Based on the changes in the process trigger identifier before and after the break position, determine whether the batch has undergone state rewriting before the process switch and whether the original state has not been inherited.

[0027] In this embodiment, S103 specifically refers to: Based on the ordered arrangement results, adjacent multidimensional state identifiers in the multidimensional state identifier sequence are extracted one by one according to the time sequence identifier, and adjacent multidimensional state identifier pairs are constructed. By recording the time sequence identifier difference of each pair of adjacent multidimensional state identifiers and the corresponding process identifier sequence, a comparison basis is provided for subsequent fracture location identification. Based on the ordered arrangement, each record in the multidimensional state identifier sequence can be linearly traversed according to the time sequence identifier. The current identifier is extracted pairwise with the next identifier that follows it, thus forming a set of adjacent multidimensional state identifier pairs. In implementation, the nth and (n+1)th records in the sequence can be combined using sequential indexing, and the time sequence identifiers in each pair are compared to determine whether the time interval between adjacent states conforms to the continuous recording characteristics of the production process. At the same time, the process identifiers in each pair are extracted to form a corresponding process identifier sequence, which is used to describe the process connection relationship between adjacent states. For example, if the sequence of a batch after sorting is pressing, sintering, and processing, then two sets of identifier pairs, pressing-sintering and sintering-processing, are extracted one by one. The time sequence identifier difference reflects the time interval between pressing and sintering, and the time interval between sintering and processing, while the corresponding process identifier sequence reflects the connection path of the process from pressing to sintering and then to processing. By recording these differences and sequences, a unified comparison basis can be provided for subsequent identification of time jumps or process jumps, thereby supporting the accurate location of the fracture position.

[0028] Based on adjacent multidimensional state identifier pairs, the time sequence identifier difference and the process identifier sequence are synchronously compared. When the time sequence identifier difference does not meet the continuous increasing relationship or the process identifier sequence does not meet the preset process connection order, the corresponding position is marked as a candidate break position, and the candidate break positions are centrally marked. After constructing adjacent multidimensional state identifier pairs, the time sequence identifier difference in each pair can be synchronously compared with the corresponding process identifier sequence. Specifically, a sequential traversal approach can be used to perform continuity checks on the time sequence identifier difference in each pair of adjacent multidimensional state identifiers. For example, it can determine whether the time sequence identifier maintains a stable increasing interval or whether there are sudden increases or regressions. Simultaneously, path matching is performed on the corresponding process identifier sequence, comparing the current process with the next process with a pre-defined process connection order. For example, pressing should be connected to sintering, and sintering should be connected to processing. If pressing is directly connected to processing or if the process order is reversed, then a decision is made. If the process identifier sequence does not meet the preset process connection order, or if the time sequence identifier difference does not meet the continuous increasing relationship or the process identifier sequence does not meet the preset process connection order, the corresponding position of the current adjacent multi-dimensional status identifier pair is marked as a candidate fracture position, and all candidate fracture positions are centrally marked in the sequence, for example, by recording them through a unified mark bit or index set, thereby forming a set of abnormal positions to be confirmed. This processing can unify the time dimension anomalies and process dimension anomalies, providing a basis for further accurate judgment by combining process trigger identifiers, while avoiding misjudgment problems caused by single-dimensional judgment.

[0029] For candidate fracture locations, the process trigger identifier sequence before and after the candidate fracture location is extracted. The trigger order and trigger type in the process trigger identifier sequence are matched and analyzed. When there is a jump in the trigger order or a change in the trigger type of the process trigger identifiers before and after the candidate fracture location, the candidate fracture location is determined as the fracture location. Based on this, it is determined that the batch has undergone state rewriting before the process switch and the original state has not been inherited.

[0030] For candidate fracture locations, process trigger identifiers can be extracted from several consecutive records before and after the candidate fracture location, forming a sequence of process trigger identifiers before and after the candidate fracture location. In implementation, the process trigger identifiers of one or more records before and after the candidate fracture location can be selected using a sequence index and arranged in chronological order. Then, the trigger order and trigger type in the process trigger identifier sequence are matched and analyzed. The trigger order describes the sequential relationship of each trigger identifier on the time axis, and the trigger type distinguishes different categories of process parameter adjustment behaviors, such as temperature adjustment or pressure adjustment. In the specific analysis process, the continuity of the trigger order can be judged. When a sequence of triggers that should have occurred... When the progressively advancing triggering sequence jumps or reverses, it is considered a triggering sequence jump. At the same time, a consistency comparison of the triggering type is performed. When the type of adjacent triggering identifiers changes abruptly or does not conform to the preset change path, it is considered a triggering type change. For example, if a batch experiences a temperature adjustment before sintering, and a pressure adjustment related to the processing stage occurs directly after the candidate fracture location, both the triggering sequence and triggering type are abnormal. Therefore, the candidate fracture location is determined as the fracture location, and the state record corresponding to this location is further determined to be a state rewrite case. Since the completion status of the previous process is not reflected in the subsequent records, it is determined that the original state has not been inherited, thus providing a clear state determination basis for subsequent scheduling decisions.

[0031] S2. Around the fracture location, in the case of state rewriting and the original state not being inherited, construct a constraint embedding mapping sequence based on state order in the multi-dimensional state identifier sequence, embed the process identifiers before the fracture location into the identifier segments after the fracture location in chronological order, and establish a mapping correspondence across identifier segments to form a continuous full-process data perception state. In this embodiment, S2 specifically includes the following steps: S201. Around the fracture location, in the case of state rewriting and the original state not being inherited, extract the process identifier sequence before the fracture location and the identifier segment after the fracture location from the multi-dimensional state identifier sequence, and arrange the process identifier sequence before the fracture location in order according to the time sequence identifier to construct a constraint embedding mapping sequence based on state order. After identifying the fracture location through preprocessing, the multidimensional state identifier sequence can be segmented around the fracture location. Specifically, the identifier node where the fracture location is located can be located through sequence indexing, and all process identifiers before that node are extracted to form the process identifier sequence before the fracture location. At the same time, all identifier records after that node are divided into identifier segments after the fracture location. Subsequently, the process identifier sequence before the fracture location is reordered according to the time sequence identifiers. The order can be corrected by increasing the time sequence identifiers so that the sequence strictly reflects the actual process execution order. On this basis, the sorted process identifier sequence is associated with the identifier segments after the fracture location. A mapping relationship is established based on the time sequence identifiers as the main index. For example, if a batch experiences a state fracture between pressing and sintering, the process identifier sequence corresponding to pressing is extracted, rearranged in time sequence as the preprocessing sequence, and then aligned and mapped with the identifier segments corresponding to sintering and subsequent processing. This constructs a constrained embedding mapping sequence that includes the relationship between the fracture location and the state before and after the fracture. This process can avoid the loss of the state before the fracture and provides a basic support for subsequent embedding and reconstruction.

[0032] The break point refers to the boundary node in the multidimensional state identifier sequence determined by temporal sequence or process connection anomaly. The multidimensional state identifier sequence is a serialized data structure formed by combining process identifiers, temporal sequence identifiers, and process trigger identifiers generated in the same batch during production in chronological order. The process identifier sequence before the break point represents the process execution path continuously recorded before the break occurs, and the identifier segment after the break point represents the set of states that are regenerated or continue to be recorded after the break occurs. The temporal sequence identifier is used to characterize the sequential relationship of each identifier in the temporal dimension and is the basis for sorting and alignment. The constraint embedding mapping sequence based on state order is a structure formed by establishing a correspondence between the process identifier sequence before the break and the identifier segment after the break in chronological order. This structure not only retains the original process order but also constrains the association between the states before and after the break, enabling subsequent processing to integrate and reconstruct the states based on a unified temporal sequence framework.

[0033] S202. In the constraint embedding mapping sequence based on state order, the process identifiers before the fracture position are embedded into the identifier segments after the fracture position in chronological order. By maintaining the correspondence between the chronological identifiers and the process identifiers, the embedding positions are matched and arranged to form a multidimensional state identifier sequence after embedding. S203. Based on the embedded multidimensional state identifier sequence, a cross-identifier segment mapping correspondence is established between the process identifier before the fracture position and the identifier segment after the fracture position. By associating and matching process identifiers with consistent time sequence identifiers, a continuous full-process data perception state is formed.

[0034] In this embodiment, S202 specifically refers to: In the constraint embedding mapping sequence based on state order, the process identifiers before the fracture position are extracted one by one according to the time sequence identifier, and a one-to-one correspondence between the process identifier and the time sequence identifier is established using the time sequence identifier as an index, forming a sequence of process identifiers to be embedded. In a state-sequence-based constraint embedding mapping sequence, the range of process identifiers preceding the fracture location can be determined first based on the fracture location. Then, the process identifiers within this range are linearly traversed according to their time sequence identifiers. Each process identifier record is extracted one by one using a sequential index, and the corresponding time sequence identifier is read during the extraction process. Each process identifier is then bound to its corresponding time sequence identifier, thus forming a one-to-one correspondence. In implementation, a key-value pair structure or a sequential array can be constructed, using the time sequence identifier as the index position and filling the corresponding process identifier into the position. This ensures that each time sequence position corresponds to only one process identifier. For example, if a batch has two processes before fracture: pressing and sintering, with corresponding time sequence identifiers 1 and 2, then after extraction... A correspondence is formed between (1, pressing) and (2, sintering), and an ordered sequence is formed by arranging the identifiers according to the time sequence. This sequence is the sequence of process identifiers to be embedded. In this way, the process execution path before the fracture can be standardized and expressed with time as the main line, avoiding the order chaos caused by state rewriting. At the same time, it provides a clear position basis for subsequent embedding according to the time sequence. The process identifiers before the fracture position represent the set of processes that have been executed before the fracture occurred. Extracting them one by one means obtaining the process identifier records one by one according to the time sequence. The one-to-one correspondence between the process identifier and the time sequence identifier means that each time position corresponds to only one process state. The sequence of process identifiers to be embedded is an ordered set of processes formed after the correspondence is established, which is used for subsequent embedding processing.

[0035] Based on the sequence of process identifiers to be embedded, the positions of the identifiers in the segment after the break point are matched according to the time sequence. The embedding position is determined by comparing the relative positional relationship of the time sequence identifiers, and the process identifiers to be embedded are inserted into the corresponding positions in sequence to form the initial embedding structure. After constructing the sequence of process identifiers to be embedded, the identifier segment after the fracture position can be used as the target sequence for position matching. Specifically, the time sequence identifiers of each record in the identifier segment after the fracture position can be extracted and compared one by one with the time sequence identifiers in the sequence of process identifiers to be embedded. The relative position is determined by calculating the order of the two in the overall sequence. For example, it can be determined whether the time sequence identifier corresponding to a certain process identifier to be embedded is located between two time sequence identifiers in the current identifier segment, thus determining its embedding position. After determining the embedding position, the process identifiers to be embedded are inserted one by one in chronological order into the corresponding positions, maintaining the overall ascending order of the time sequence identifiers after insertion. For example, in a certain batch, the time sequence identifiers after the fracture segment are 3 and 4, corresponding to sintering and... In the process of embedding, if there is a pressing corresponding to time sequence identifier 2 in the sequence of process identifiers to be embedded, then by comparison, it can be determined that pressing should be inserted before the position of time sequence identifier 3, thus forming a continuous structure including pressing, sintering, and processing. This structure is the initial embedding structure. The identifier segment after the fracture position represents the set of states recorded after the fracture occurred. Position matching refers to the process of determining the insertion position through time sequence identifiers. The relative positional relationship of time sequence identifiers is used to describe the order of different identifiers in the overall time axis. The embedding position refers to the specific insertion point of the process identifier to be embedded in the target sequence. The process identifier to be embedded represents the set of processes extracted and organized before the fracture. The initial embedding structure is a temporary continuous sequence structure formed after the insertion is completed, which provides a basis for further mapping and integration.

[0036] Based on the initial embedding structure, the consistency between the time sequence identifier and the process identifier of the embedding position is checked. By maintaining the correspondence between the time sequence identifier and the process identifier, the embedding positions are matched and arranged to form a multi-dimensional state identifier sequence after embedding.

[0037] After forming the initial embedding structure, consistency verification can be performed on the time sequence identifier and process identifier at the embedding position. Specifically, the embedded sequence can be sequentially traversed, checking the time sequence identifier and corresponding process identifier in each record one by one. First, it is determined whether the time sequence identifier maintains a continuous increasing relationship. Second, it is determined whether the same time sequence identifier corresponds to only one process identifier. Simultaneously, the process identifiers of adjacent records are matched sequentially; for example, pressing should correspond to sintering, and sintering should correspond to processing. If duplicate or skipped time sequence identifiers are found, or if the same time sequence identifier corresponds to multiple process identifiers, the embedding position is adjusted by moving the corresponding process identifier to a position that matches the time sequence identifier, thus completing the matching and arrangement. For example, in the initial embedded structure, if time sequence identifier 2 corresponds to sintering and time sequence identifier 3 corresponds to pressing, then by swapping their positions, pressing corresponds to 2 and sintering corresponds to 3, so that the time sequence identifier and process identifier maintain a consistent correspondence, and finally form the embedded multidimensional state identifier sequence. Among them, consistency verification refers to verifying the matching relationship between time sequence identifier and process identifier one by one. The correspondence between time sequence identifier and process identifier means that each time position corresponds to a unique and correct process stage. Matching arrangement refers to adjusting the identifier order to make the two consistent in the sequence. The embedded multidimensional state identifier sequence is a normalized state sequence formed after completing consistency verification and order adjustment, which is used for subsequent cross-identifier segment mapping and scheduling decisions.

[0038] In this embodiment, S203 specifically refers to: Based on the embedded multidimensional state identifier sequence, the process identifier sequence before the fracture position and the corresponding process identifier sequence in the identifier segment after the fracture position are extracted, and the two sequences are synchronized and aligned according to the time sequence identifier to construct a candidate mapping corresponding set across the identifier segment. Based on the embedded multidimensional state identifier sequence, the sequence can first be divided into a process identifier sequence before the fracture position and a corresponding process identifier sequence in the identifier segment after the fracture position. Then, the process identifiers and their corresponding time sequence identifiers are extracted from the two parts of the sequence, and rearranged according to the time sequence identifiers to ensure that both parts of the sequence are under a unified time reference. Subsequently, the two parts of the sequence are synchronized and aligned. Specifically, the two parts of the sequence are mapped to the same time axis using the time sequence identifier as the alignment index. Process identifiers with the same time sequence identifier are aligned in position, and positions with missing time sequence identifiers are left empty to maintain the alignment structure, thus forming a comparable dual-sequence structure. For example, the sequence before fracture has time sequence identifiers 1 and 2, corresponding to pressing and sintering. After the fracture, the time sequence identifiers 3 and 4 correspond to processing and inspection. After alignment, a unified sequence framework with time sequence identifiers 1 to 4 is formed. Under this framework, the process identifiers of the two parts are arranged accordingly, and process identifiers with the same or adjacent time sequence identifiers are paired. Finally, a candidate mapping correspondence set across identifier segments is constructed. The process identifier sequence before the fracture position represents the process path continuously recorded before the fracture occurred, and the corresponding process identifier sequence in the identifier segment after the fracture position represents the process path continuously recorded after the fracture. Synchronous alignment processing refers to the process of matching the positions of different sequences through unified time sequence identifiers. The candidate mapping correspondence set across identifier segments is a set composed of process identifier pairs that have a corresponding relationship in the time dimension before and after the fracture, which is used for subsequent accurate screening of effective mapping relationships.

[0039] In the candidate mapping correspondence set across the identifier segment, process identifiers with the same time sequence identifier value are matched one by one, and a one-to-one correspondence is established according to the size of the time sequence identifier value. By comparing the arrangement position of the process identifier in the sequence, process identifier pairs with the same time sequence identifier value and corresponding position are selected to form a cross-identifier segment mapping correspondence. After constructing the candidate mapping set across identifier segments, each candidate process identifier pair in the set can be matched one by one. Specifically, the candidate set can be grouped according to the time sequence identifier value, with process identifiers having the same time sequence identifier value grouped together. Within each group, the process identifiers are compared according to their original sequence order. A sequential traversal method is used to match process identifiers before and after the break point one by one. During the matching process, the time sequence identifier is used as the primary index, while also verifying the positional relationship of the process identifiers in the sequence. For example, when the time sequence identifier value is 3, if there are corresponding process identifiers before and after the break point and their relative positions in their respective sequences are consistent, then the process identifier pair is considered a valid match; otherwise, it is discarded. Furthermore, by performing the same matching operation on all time sequence identifier groups, process identifiers with the same time sequence identifier value and positional relationship can be filtered out. The process identifiers are matched and these matching pairs are summarized to form a cross-identifier segment mapping correspondence. For example, if the sequence before fracture is identified by time sequence identifiers 1 and 2, corresponding to pressing and sintering, and the sequence after fracture is identified by time sequence identifiers 2 and 3, corresponding to sintering and processing, then by group matching, the time sequence identifier 2 corresponding to sintering can be identified as a valid mapping pair, thereby establishing the process association relationship at the same time position before and after fracture. Among them, process identifiers with the same time sequence identifier value represent candidate states at the same time position. One-to-one matching means that each group of candidate identifier pairs is compared one by one. The one-to-one correspondence means that each time sequence identifier corresponds to only one pair of valid process identifiers. Process identifier pairs with the same time sequence identifier value and corresponding position represent matching results that are consistent in both time dimension and sequence position. The cross-identifier segment mapping correspondence is an association set composed of these matching results, used to describe the precise correspondence between the states before and after fracture.

[0040] Based on the mapping relationship across identifier segments, the process identifiers before the fracture position are associated with the identifier segments after the fracture position. By maintaining the correspondence between process identifiers with the same time sequence identifier value, the embedded multidimensional state identifier sequence is integrated to form a continuous full-process data perception state.

[0041] After establishing the mapping relationship across identifier segments, this relationship can be used as the connection basis to associate and connect the process identifiers before the fracture position with the identifier segments after the fracture position. Specifically, the embedded multidimensional state identifier sequence can be segmented and traversed according to the time sequence identifier values. Process identifier pairs with the same time sequence identifier values ​​are used as connection anchors. A connection path is established in the sequence, and the process identifiers on both sides are sequentially extended based on the anchors. Through insertion, replacement, or rearrangement operations, the process identifiers before the fracture and the identifier segments after the fracture are uniformly arranged. For example, if a batch has a sintering correspondence at time sequence identifier 2, then sintering is used as the connection point. The pressing before fracture and the processing after fracture are arranged in chronological order as pressing, sintering, and processing, thus completing the sequence integration. During the integration process, the correspondence between process identifiers with the same time sequence identifier value is maintained to ensure that each time position corresponds to only one process state, and conflicting positions are adjusted to form a continuous full-process data perception state. The association connection refers to establishing the connection path between processes before and after fracture based on the mapping correspondence. Integration refers to reorganizing the originally segmented multi-dimensional state identifier sequence into a unified sequence structure. The continuous full-process data perception state refers to the state sequence that is consistent in both time sequence and process sequence and has no breaks, which is used to support the subsequent scheduling decision process.

[0042] S3. Based on the continuous full-process data perception status, jointly deduce the process identifier and constraint embedding relationship in the multi-dimensional status identifier sequence, calculate the execution order of the batch in each process, and determine the scheduling order of the batch entering the subsequent process according to the execution order. In this embodiment, S3 specifically includes the following steps: S301. Based on the continuous full-process data perception status, the process identifier and constraint embedding relationship in the multi-dimensional status identifier sequence are jointly deduced. By extracting the process identifier sequence arranged in time order in the multi-dimensional status identifier sequence, and combining the cross-identifier segment mapping correspondence in the constraint embedding relationship, the connection relationship between the process identifiers is associated one by one to form an ordered connection sequence of process identifiers. After a continuous, end-to-end data perception state is formed, this state can be used as a unified data foundation to jointly deduce the process identifiers and constraint embedding relationships in the multi-dimensional state identifier sequence. Specifically, this involves first extracting the process identifier sequence arranged in chronological order, and then reading the established cross-identifier segment mapping correspondences in the constraint embedding relationships. The process identifiers corresponding to the same chronological order identifiers before and after the break are connected. Then, the process identifiers are traversed sequentially according to time, and the current process identifier is associated with the next process identifier. The existence of cross-segment connection relationships is determined through the cross-identifier segment mapping correspondences. For example… A batch of processes involves pressing and sintering before fracture, and sintering and processing after fracture. By mapping across identifier segments, two sintering nodes are associated, and then pressing, sintering, and processing are connected sequentially according to time to form a complete process connection path. In the implementation process, a chain structure or sequential list can be constructed to link each pair of process identifiers with a sequential relationship, and the link order can be corrected according to the time sequence identifier, thereby gradually forming an ordered connection sequence of process identifiers. This process can reorganize the process information that was originally broken or segmented into a continuous path, providing a unified basis for subsequent execution order calculation.

[0043] Continuous end-to-end data perception state refers to a set of states that remain continuous in both the time and process dimensions after prior processing; process identifiers are used to represent the specific processing stage of a batch in the production process; constraint embedding relationships are used to describe the constraint associations established between processes before and after the break through time sequence; joint derivation refers to the process of combining the process identifier sequence with constraint embedding relationships for synchronous analysis and path construction; cross-identifier segment mapping in constraint embedding relationships corresponds to the association relationship between process identifiers at the same time sequence position before and after the break; sequential connection relationship refers to the sequential execution relationship of adjacent processes in the production process; one-to-one association means connecting each process identifier sequentially according to time sequence; the ordered connection sequence of process identifiers is the process path structure formed by the above processing and arranged continuously in time sequence, used to describe the execution trajectory of the batch in the complete production process.

[0044] S302. Based on the ordered connection sequence of process identifiers, calculate the execution order of the batch in each process. Sort the process identifiers according to the time sequence identifiers and adjust the positions of process identifiers with overlapping positions according to the constraint embedding relationship to determine the execution order of the batch in each process. S303. After the execution order of the batch in each process is determined, the scheduling order of the batch entering the subsequent process is determined according to the execution order. The process identifiers in the multi-dimensional status identifier sequence are arranged in order according to the execution order, and the scheduling order of the batch entering each process is generated according to the arrangement order.

[0045] Once the execution order of a batch in each process is determined, this execution order can be used as the basis for scheduling. The process identifiers in the multi-dimensional status identifier sequence can be rearranged. In specific implementation, each process identifier can be assigned a sequential position mark according to the execution order, and the order in which the batch enters each process can be arranged according to this sequential position. At the same time, the entry order can be refined in combination with actual production constraints. For example, each process identifier in the execution order can be mapped to the entry node of the batch, and a scheduling sequence can be generated one by one in order. For example, if the execution order of a batch is pressing, sintering, and processing, then according to this order, the batch entering the pressing process can be taken as the first scheduling node, entering the sintering process as the second scheduling node, and entering the processing process as the third scheduling node. These nodes can be organized into a continuous arrangement through sequential index or time position mark, thereby forming the scheduling order of the batch entering each process. Through this process, the abstract execution order can be transformed into a specific scheduling sequence, so that the flow path of the batch in the production process has a clear sequential guide.

[0046] The scheduling order of batches entering subsequent processes refers to the sequence structure formed by sorting the order of batch entry into each subsequent processing stage after the batch execution path is determined. This sequence directly reflects the flow path of the batch in the production process. The scheduling order of batches entering each process is the result of clearly identifying the entry time position or entry sequence of each process node. It is used to describe the sequential relationship of batches transferring from one process to the next. This order not only reflects the order of process execution, but also provides a basis for production resource allocation and production rhythm control.

[0047] In this embodiment, S302 specifically refers to: Based on the ordered sequence of process identifiers, the time sequence identifiers corresponding to each process identifier are extracted, and the process identifiers are initially sorted according to the time sequence identifiers to form a sequence of process identifiers arranged in time order. After forming an ordered sequence of process identifiers, each process identifier in the sequence can be traversed, and the corresponding time sequence identifier can be extracted. In implementation, the time sequence identifier bound to each record can be read through sequential indexing or key-value mapping. The process identifiers are then paired with their corresponding time sequence identifiers. Finally, all process identifiers are initially sorted according to their time sequence identifiers. This sorting process can be done by ascending the time sequence identifiers, placing process identifiers with smaller time sequence identifiers at the beginning and those with larger time sequence identifiers at the end. For example, a batch may have three process identifiers: processing, pressing, and sintering. If the identifiers are 3, 1, and 2 respectively, then the sorting is adjusted to the order of pressing, sintering, and processing, thus forming a process identifier sequence arranged in chronological order. This process can eliminate the deviation caused by the inconsistency of the data recording order, and keep the process execution path consistent in the time dimension, providing a unified basis for subsequent execution order calculation. The time sequence identifier corresponding to each process identifier represents the time position of each process in the production process. The initial sorting refers to the first order adjustment of the process identifiers based on the time sequence identifiers. The process identifier sequence arranged in chronological order is the ordered set of processes formed after sorting, which is used to describe the time sequence relationship of batches in the production process.

[0048] For a sequence of process identifiers arranged in chronological order, overlapping positions of process identifiers with the same chronological order are identified based on constraint embedding relationships. By rearranging the process identifiers in the overlapping positions according to the mapping order in the constraint embedding relationships, a position adjustment sequence is formed. After the sequence of process identifiers arranged chronologically is formed, each time sequence identifier in the sequence can be grouped. Process identifiers with the same time sequence identifier are grouped together. Multiple process identifiers at the same time position are identified by traversal, and this position is determined as an overlapping position. Subsequently, according to the mapping correspondence in the constraint embedding relationship, the process identifiers at the overlapping positions are rearranged. Specifically, this can be done by reading the cross-identifier segment mapping correspondence recorded in the constraint embedding relationship, sorting the process identifiers under the same time sequence identifier before and after the fracture according to the pre-established correspondence order. For example, in a certain batch, sintering and processing processes exist simultaneously at time sequence identifier 2. The process identifiers are defined in the constraint embedding relationship, which stipulates that sintering precedes processing. Therefore, the process identifiers at that position are adjusted so that sintering comes before processing, thereby eliminating overlapping conflicts and restoring the correct execution order, ultimately forming a position adjustment sequence. The constraint embedding relationship describes the constraints on process identifiers with the same time sequence identifier, which is used to limit the sequential relationship between different processes at the same time position. Overlapping position identification refers to detecting the situation where the same time sequence identifier corresponds to multiple process identifiers. The mapping order in the constraint embedding relationship is used to provide the basis for the arrangement of process identifiers. The position adjustment sequence is a sequence structure formed after overlap resolution and order rearrangement, which is used to ensure the accuracy of subsequent execution order calculation.

[0049] The sequence is adjusted according to the position, and the order of the process identifiers is confirmed. By verifying the arrangement of each process identifier in the sequence one by one, a process identifier sequence with no overlap and a unique order is formed, and the execution order of the batch in each process is determined.

[0050] After constructing the position adjustment sequence, the process identifiers in the sequence can be checked for order confirmation. Specifically, each process identifier can be iterated through sequentially according to the order in the sequence, and its position can be verified by combining the time sequence identifier with the constraint embedding relationship. First, it is checked whether the same time sequence identifier corresponds to only one process identifier to eliminate any remaining overlaps. Second, the order between adjacent process identifiers is verified to ensure that their arrangement conforms to the actual execution path in the production process. For example, pressing should be before sintering, and sintering should be before processing. When order conflicts or duplicate identifiers are found, they are corrected by adjusting adjacent positions or removing redundant identifiers until all process identifiers have a unique position in the sequence and are not duplicated. If, for example, two sintering identifiers still exist at the same time position after adjustment in a certain batch, the identifier that conforms to the constraint embedding relationship is retained and the other identifier is removed, so that the sequence is restored to the unique order of pressing, sintering, and processing, thus obtaining a non-overlapping and sequentially unique process identifier sequence. Based on this, the sequence is used as the execution path of the batch in the production process, and the execution order of the batch in each process is determined according to the arrangement position of each process identifier in the sequence. The non-overlapping and sequentially unique process identifier sequence refers to a sequence structure in which each time position corresponds to only one process and the overall arrangement is conflict-free. The execution order of the batch in each process is determined by the processing sequence relationship determined by the arrangement order of the process identifiers in the sequence, which is used to guide subsequent production scheduling.

[0051] S4. Combine the scheduling order to continuously update the multidimensional state identifier sequence and constraint embedding relationship, and simultaneously correct the scheduling order so as to realize the dynamic adjustment of the scheduling order as the multidimensional state identifier sequence changes.

[0052] In this embodiment, S4 specifically refers to: Based on the scheduling order, the multidimensional state identifier sequence and constraint embedding relationship are continuously updated. By sequentially traversing the process identifiers in the multidimensional state identifier sequence according to the scheduling order, the newly added process identifiers are inserted into the corresponding positions according to the time sequence identifiers. At the same time, the cross-identifier segment mapping in the constraint embedding relationship is synchronously adjusted to form the updated multidimensional state identifier sequence and constraint embedding relationship. During production execution, the multidimensional status identifier sequence can be sequentially traversed according to the generated scheduling order. New process identifiers and their corresponding time sequence identifiers generated during the current batch's execution are collected, and the insertion position of the new process identifier in the overall sequence is determined based on the scheduling order. Specifically, the insertion interval can be determined by comparing the time sequence identifier of the new process identifier with the time sequence identifiers of existing identifiers. For example, if the time sequence identifiers in the existing sequence are 1, 2, and 4, and the time sequence identifier of the new process identifier is 3, then the process identifier is inserted between time sequence identifiers 2 and 4. After insertion, the cross-identifier segment mapping in the constraint embedding relationship is simultaneously adjusted. By comparing the time sequence relationship between the new process identifier and existing process identifiers, it is included in the corresponding mapping set, and the original mapping relationship is updated to reflect the latest process connection status, thus forming an updated multidimensional status identifier sequence and constraint embedding relationship.

[0053] The continuous updating of the multidimensional state identifier sequence and constraint embedding relationship in conjunction with the scheduling order refers to continuously introducing new process states and expanding and adjusting the original sequence during batch execution. The multidimensional state identifier sequence is used to record the state changes of the batch in each process, and its core lies in organizing the process identifiers according to the time sequence identifiers. The constraint embedding relationship is used to describe the process association structure between different time positions, where the cross-identifier segment mapping corresponds to the process correspondence before and after the break or between different stages. Sequential traversal means processing the sequence one by one according to the established scheduling order. The newly added process identifier refers to the new process state record generated in real time during the production process. Inserting the corresponding position means placing the newly added process identifier in the appropriate time interval according to the time sequence identifier. Synchronous adjustment means updating the mapping relationship consistently while updating the sequence, thereby ensuring that the sequence structure and the association relationship remain coordinated and consistent.

[0054] Based on the continuously updated multidimensional status identifier sequence and constraint embedding relationship, the scheduling order is synchronously corrected. The scheduling order is reordered according to the arrangement order of the process identifiers in the updated multidimensional status identifier sequence and the mapping correspondence in the constraint embedding relationship. The scheduling order is then dynamically adjusted according to the reordering result as the multidimensional status identifier sequence changes.

[0055] After continuously updating the multidimensional state identifier sequence and constraint embedding relationship, the current scheduling order can be synchronously corrected. Specifically, the process identifier sequence arranged in chronological order can be extracted from the updated multidimensional state identifier sequence. Combined with the mapping correspondence in the constraint embedding relationship, the sequential relationship between process identifiers can be re-parsed. Then, the original scheduling order is compared with the re-parsed process arrangement order, and scheduling nodes that do not conform to the current process arrangement order are adjusted. For example, if the original scheduling order is pressing, sintering, and processing, when a new process is inserted or the order changes in the updated sequence, the corresponding nodes are adjusted to the new order by reordering, such as pressing, preprocessing, sintering, and processing, thereby generating a new scheduling order. During this process, the scheduling nodes can be sorted by sequence index or priority tag, and the scheduling order is updated by renumbering the position of each process identifier in the sequence, so that the scheduling order is consistent with the multidimensional state identifier sequence.

[0056] The continuously updated multidimensional state identifier sequence refers to the sequence structure after new process identifiers are continuously introduced and sorted and adjusted during the production process; constraint embedding relationship is used to describe the process association relationship between different time positions and different stages, where the mapping correspondence reflects the correspondence and connection between processes; synchronous correction of scheduling order refers to the synchronous adjustment of the original scheduling order after the sequence is updated so that it reflects the latest process arrangement; reordering refers to the rearrangement of scheduling nodes according to the updated process identifier arrangement order; the scheduling order is dynamically adjusted with the changes of the multidimensional state identifier sequence, meaning that the scheduling result is no longer fixed, but is continuously updated as the process identifiers and their order in the sequence change, thereby ensuring that the scheduling path of the batch in the production process is always consistent with the current state.

[0057] like Figure 2 The tungsten alloy production scheduling and decision-making system based on full-process data perception shown includes a full-process perception and analysis module, a state fracture reconstruction module, a process deduction and decision-making module, and a scheduling dynamic control module. The full-process perception and analysis module constructs a multi-dimensional status identifier sequence corresponding to each batch during the tungsten alloy production process. The multi-dimensional status identifier sequence includes process identifier, time sequence identifier, and process trigger identifier. The module performs ordered arrangement and analysis of the multi-dimensional status identifier sequence to identify the fracture location in the multi-dimensional status identifier sequence, so as to determine whether the batch has undergone state rewriting before process switching and whether the original state has not been inherited. The state fracture reconstruction module, around the fracture location, constructs a constraint embedding mapping sequence based on the state order in the multi-dimensional state identifier sequence when state rewriting occurs and the original state is not inherited. It embeds the process identifiers before the fracture location into the identifier segments after the fracture location in chronological order and establishes a mapping correspondence across identifier segments to form a continuous full-process data perception state. The process deduction and decision-making module jointly deduces the process identifier and constraint embedding relationship in the multi-dimensional status identifier sequence based on the continuous full-process data perception status, calculates the execution order of the batch in each process, and determines the scheduling order of the batch to enter the subsequent processes according to the execution order. The scheduling dynamic control module continuously updates the multidimensional state identifier sequence and constraint embedding relationship in conjunction with the scheduling order, and synchronously corrects the scheduling order to achieve dynamic adjustment of the scheduling order as the multidimensional state identifier sequence changes.

[0058] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0059] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0060] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0061] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0062] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0063] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0064] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A tungsten alloy production scheduling decision-making method based on end-to-end data perception, characterized in that, Specifically, the following steps are included: S1. Construct a multi-dimensional state identifier sequence corresponding to each batch during the tungsten alloy production process. The multi-dimensional state identifier sequence includes process identifier, time sequence identifier and process trigger identifier. The multi-dimensional state identifier sequence is arranged and parsed in an orderly manner to identify the fracture position in the multi-dimensional state identifier sequence in order to determine whether the batch has undergone state rewriting before process switching and whether the original state has not been inherited. S2. Around the fracture location, in the case of state rewriting and the original state not being inherited, construct a constraint embedding mapping sequence based on state order in the multi-dimensional state identifier sequence, embed the process identifiers before the fracture location into the identifier segments after the fracture location in chronological order, and establish a mapping correspondence across identifier segments to form a continuous full-process data perception state. S3. Based on the continuous full-process data perception status, jointly deduce the process identifier and constraint embedding relationship in the multi-dimensional status identifier sequence, calculate the execution order of the batch in each process, and determine the scheduling order of the batch entering the subsequent process according to the execution order. S4. Combine the scheduling order to continuously update the multidimensional state identifier sequence and constraint embedding relationship, and simultaneously correct the scheduling order so as to realize the dynamic adjustment of the scheduling order as the multidimensional state identifier sequence changes.

2. The tungsten alloy production scheduling decision-making method based on full-process data perception according to claim 1, characterized in that, S1 specifically includes the following steps: S101. Construct a multi-dimensional status identifier sequence corresponding to each batch during the tungsten alloy production process. The multi-dimensional status identifier sequence includes process identifier, time sequence identifier, and process trigger identifier. Process identifier is generated based on the execution node of each process of the batch, time sequence identifier is generated based on the order of the batch in the production process, and process trigger identifier is generated based on the process parameter adjustment behavior. The process identifier and process trigger identifier are combined and arranged according to the time sequence identifier to form a multi-dimensional status identifier sequence. S102. Perform ordered arrangement and parsing on the multidimensional state identifier sequence, sort the multidimensional state identifier sequence according to the time sequence identifier, and perform process continuity verification on adjacent identifiers in combination with process identifiers. At the same time, mark the position of state change according to the process trigger identifier to form an ordered arrangement result. S103. Based on the ordered arrangement results, identify the break position in the multidimensional state identifier sequence. By comparing the time sequence identifier and process identifier of adjacent multidimensional state identifiers, determine the position where the time sequence is discontinuous or the process identifier is discontinuous as the break position. Based on the changes in the process trigger identifier before and after the break position, determine whether the batch has undergone state rewriting before the process switch and whether the original state has not been inherited.

3. The tungsten alloy production scheduling decision-making method based on full-process data perception according to claim 2, characterized in that, S103 specifically refers to: Based on the ordered arrangement results, adjacent multidimensional state identifiers in the multidimensional state identifier sequence are extracted one by one according to the time sequence identifier, and adjacent multidimensional state identifier pairs are constructed. By recording the time sequence identifier difference of each pair of adjacent multidimensional state identifiers and the corresponding process identifier sequence, a comparison basis is provided for subsequent fracture location identification. Based on adjacent multidimensional state identifier pairs, the time sequence identifier difference and the process identifier sequence are synchronously compared. When the time sequence identifier difference does not meet the continuous increasing relationship or the process identifier sequence does not meet the preset process connection order, the corresponding position is marked as a candidate break position, and the candidate break positions are centrally marked. For candidate fracture locations, the process trigger identifier sequence before and after the candidate fracture location is extracted. The trigger order and trigger type in the process trigger identifier sequence are matched and analyzed. When there is a jump in the trigger order or a change in the trigger type of the process trigger identifiers before and after the candidate fracture location, the candidate fracture location is determined as the fracture location. Based on this, it is determined that the batch has undergone state rewriting before the process switch and the original state has not been inherited.

4. The tungsten alloy production scheduling decision-making method based on full-process data perception according to claim 1, characterized in that, S2 specifically includes the following steps: S201. Around the fracture location, in the case of state rewriting and the original state not being inherited, extract the process identifier sequence before the fracture location and the identifier segment after the fracture location from the multi-dimensional state identifier sequence, and arrange the process identifier sequence before the fracture location in order according to the time sequence identifier to construct a constraint embedding mapping sequence based on state order. S202. In the constraint embedding mapping sequence based on state order, the process identifiers before the fracture position are embedded into the identifier segments after the fracture position in chronological order. By maintaining the correspondence between the chronological identifiers and the process identifiers, the embedding positions are matched and arranged to form a multidimensional state identifier sequence after embedding. S203. Based on the embedded multidimensional state identifier sequence, a cross-identifier segment mapping correspondence is established between the process identifier before the fracture position and the identifier segment after the fracture position. By associating and matching process identifiers with consistent time sequence identifiers, a continuous full-process data perception state is formed.

5. The tungsten alloy production scheduling decision-making method based on full-process data perception according to claim 4, characterized in that, S202 specifically refers to: In the constraint embedding mapping sequence based on state order, the process identifiers before the fracture position are extracted one by one according to the time sequence identifier, and a one-to-one correspondence between the process identifier and the time sequence identifier is established using the time sequence identifier as an index, forming a sequence of process identifiers to be embedded. Based on the sequence of process identifiers to be embedded, the positions of the identifiers in the segment after the break point are matched according to the time sequence. The embedding position is determined by comparing the relative positional relationship of the time sequence identifiers, and the process identifiers to be embedded are inserted into the corresponding positions in sequence to form the initial embedding structure. Based on the initial embedding structure, the consistency between the time sequence identifier and the process identifier of the embedding position is checked. By maintaining the correspondence between the time sequence identifier and the process identifier, the embedding positions are matched and arranged to form a multi-dimensional state identifier sequence after embedding.

6. The tungsten alloy production scheduling decision-making method based on full-process data perception according to claim 4, characterized in that, S203 specifically refers to: Based on the embedded multidimensional state identifier sequence, the process identifier sequence before the fracture position and the corresponding process identifier sequence in the identifier segment after the fracture position are extracted, and the two sequences are synchronized and aligned according to the time sequence identifier to construct a candidate mapping corresponding set across the identifier segment. In the candidate mapping correspondence set across the identifier segment, process identifiers with the same time sequence identifier value are matched one by one, and a one-to-one correspondence is established according to the size of the time sequence identifier value. By comparing the arrangement position of the process identifier in the sequence, process identifier pairs with the same time sequence identifier value and corresponding position are selected to form a cross-identifier segment mapping correspondence. Based on the mapping relationship across identifier segments, the process identifiers before the fracture position are associated with the identifier segments after the fracture position. By maintaining the correspondence between process identifiers with the same time sequence identifier value, the embedded multidimensional state identifier sequence is integrated to form a continuous full-process data perception state.

7. The tungsten alloy production scheduling decision-making method based on full-process data perception according to claim 1, characterized in that, S3 specifically includes the following steps: S301. Based on the continuous full-process data perception status, the process identifier and constraint embedding relationship in the multi-dimensional status identifier sequence are jointly deduced. By extracting the process identifier sequence arranged in time order in the multi-dimensional status identifier sequence, and combining the cross-identifier segment mapping correspondence in the constraint embedding relationship, the connection relationship between the process identifiers is associated one by one to form an ordered connection sequence of process identifiers. S302. Based on the ordered connection sequence of process identifiers, calculate the execution order of the batch in each process. Sort the process identifiers according to the time sequence identifiers and adjust the positions of process identifiers with overlapping positions according to the constraint embedding relationship to determine the execution order of the batch in each process. S303. After the execution order of the batch in each process is determined, the scheduling order of the batch entering the subsequent process is determined according to the execution order. The process identifiers in the multi-dimensional status identifier sequence are arranged in order according to the execution order, and the scheduling order of the batch entering each process is generated according to the arrangement order.

8. The tungsten alloy production scheduling decision-making method based on full-process data perception according to claim 7, characterized in that, S302 specifically refers to: Based on the ordered sequence of process identifiers, the time sequence identifiers corresponding to each process identifier are extracted, and the process identifiers are initially sorted according to the time sequence identifiers to form a sequence of process identifiers arranged in time order. For a sequence of process identifiers arranged in chronological order, overlapping positions of process identifiers with the same chronological order are identified based on constraint embedding relationships. By rearranging the process identifiers in the overlapping positions according to the mapping order in the constraint embedding relationships, a position adjustment sequence is formed. The sequence is adjusted according to the position, and the order of the process identifiers is confirmed. By verifying the arrangement of each process identifier in the sequence one by one, a process identifier sequence with no overlap and a unique order is formed, and the execution order of the batch in each process is determined.

9. The tungsten alloy production scheduling decision-making method based on full-process data perception according to claim 1, characterized in that, S4 specifically refers to: Based on the scheduling order, the multidimensional state identifier sequence and constraint embedding relationship are continuously updated. By sequentially traversing the process identifiers in the multidimensional state identifier sequence according to the scheduling order, the newly added process identifiers are inserted into the corresponding positions according to the time sequence identifiers. At the same time, the cross-identifier segment mapping in the constraint embedding relationship is synchronously adjusted to form the updated multidimensional state identifier sequence and constraint embedding relationship. Based on the continuously updated multidimensional status identifier sequence and constraint embedding relationship, the scheduling order is synchronously corrected. The scheduling order is reordered according to the arrangement order of the process identifiers in the updated multidimensional status identifier sequence and the mapping correspondence in the constraint embedding relationship. The scheduling order is then dynamically adjusted according to the reordering result as the multidimensional status identifier sequence changes.

10. A tungsten alloy production scheduling and decision-making system based on end-to-end data perception, used to implement the tungsten alloy production scheduling and decision-making method based on end-to-end data perception as described in any one of claims 1-9, characterized in that, It includes a full-process perception and analysis module, a state fracture reconstruction module, a process deduction and decision-making module, and a scheduling dynamic control module; The full-process perception and analysis module constructs a multi-dimensional status identifier sequence corresponding to each batch during the tungsten alloy production process. The multi-dimensional status identifier sequence includes process identifier, time sequence identifier, and process trigger identifier. The module performs ordered arrangement and analysis of the multi-dimensional status identifier sequence to identify the fracture location in the multi-dimensional status identifier sequence, so as to determine whether the batch has undergone state rewriting before process switching and whether the original state has not been inherited. The state fracture reconstruction module, around the fracture location, constructs a constraint embedding mapping sequence based on the state order in the multi-dimensional state identifier sequence when state rewriting occurs and the original state is not inherited. It embeds the process identifiers before the fracture location into the identifier segments after the fracture location in chronological order and establishes a mapping correspondence across identifier segments to form a continuous full-process data perception state. The process deduction and decision-making module jointly deduces the process identifier and constraint embedding relationship in the multi-dimensional status identifier sequence based on the continuous full-process data perception status, calculates the execution order of the batch in each process, and determines the scheduling order of the batch to enter the subsequent processes according to the execution order. The scheduling dynamic control module continuously updates the multidimensional state identifier sequence and constraint embedding relationship in conjunction with the scheduling order, and synchronously corrects the scheduling order to achieve dynamic adjustment of the scheduling order as the multidimensional state identifier sequence changes.

Images