A chemical reagent inventory scheduling method based on experimental demand prediction

Abstract

The present application relates to a kind of chemical reagent inventory scheduling method based on experimental demand prediction, comprising: generating the future consumption prediction of each laboratory reagent based on historical consumption and experimental plan;Obtain each warehouse batch inventory and remaining effective time length, mark as near-expiration batch and locate if it is below early warning threshold;The predicted consumption amount is compared with the corresponding area inventory to determine the reagent gap amount;With gap amount as demand, near-expiration inventory of other areas as priority supply source, a deployment network is constructed;Under the constraint of preferentially consuming near-expiration batch, the deployment scheme with the shortest total transportation distance is solved and instructions are generated.The present application minimizes reagent waste and optimizes logistics efficiency while ensuring the continuity of experimental supply by predicting consumption and forcing the preferential consumption of near-expiration inventory.

Description

Technical Field

[0001] This invention relates to the field of chemical reagent scheduling and management technology, and in particular to a chemical reagent inventory scheduling method based on experimental demand prediction. Background Technology

[0002] In a context where scientific research and experimental activities heavily rely on chemical reagents, the management efficiency of reagents, as core materials, directly impacts the progress and cost control of research projects. Currently, research institutions typically have central warehouses and regional warehouses distributed across different laboratory floors. Reagents are stored in batches, with each batch recorded with its production date and shelf life. However, with the expansion of experimental scale and the increasing frequency of interdisciplinary collaboration, the complexity of reagent management has significantly increased. A prominent contradiction within the industry is that, on the one hand, some laboratories experience reagent stockpiles due to project cycle fluctuations or inadequate planning, or even expired reagents due to forgetfulness; on the other hand, other laboratories may face shortages of such reagents due to sudden experimental needs or accelerated progress, forcing them to make ad-hoc purchases, which increases the logistics costs of emergency procurement and affects the continuity of experiments. This dual imbalance in resource allocation across spatial and temporal dimensions reveals significant shortcomings in the existing management model regarding overall scheduling and proactive intervention.

[0003] Existing reagent management methods largely rely on laboratory information management systems or inventory management software, whose main functions are concentrated on static recording, such as recording the time of receipt, the quantity issued, and simple low-inventory alarms. To address the aforementioned uneven inventory levels, current solutions typically involve managers periodically checking inventory lists and communicating manually within departments via email or instant messaging to allocate surplus or near-expiration reagents from one laboratory to those in need. While this approach can alleviate some immediate crises at a micro level, it suffers from significant lag and inconsistency. First, allocation decisions heavily rely on managers' personal experience and memory, failing to systematically assess the actual consumption trends of each laboratory over a future period. Often, requisitioning only begins passively when a laboratory explicitly signals a "stockout," by which time the experiment may be on the verge of failure. Second, even when near-expiration reagents are identified, the lack of foresight regarding the future experimental plans of all laboratories often results in near-expiration reagents being allocated to laboratories with lower current demand, causing the reagents to remain idle in new storage locations until they expire completely, failing to fundamentally solve the problem of resource waste. Furthermore, manual dispatching lacks overall optimization of transportation routes. In large research parks, frequent and fragmented transportation across buildings is not only inefficient but also increases the risk of damage during handling. Summary of the Invention

[0004] Therefore, the technical problem to be solved by this invention is to overcome the shortcomings of existing chemical reagent management, which relies on static records and passive manual allocation, and cannot combine dynamic changes in shelf life and future experimental consumption trends for global prediction. This leads to waste of near-expiry reagents due to delayed allocation and shortage of urgently needed reagents due to blind scheduling. The invention provides a chemical reagent inventory scheduling method based on experimental demand prediction. It can predict the future consumption of each laboratory and compare it with local inventory to lock in the gap. Under the constraint of mandatory priority consumption of near-expiry batches, it solves the cross-regional allocation scheme with the shortest total transportation distance, thereby minimizing resource waste and optimizing logistics efficiency throughout the entire life cycle of reagents while ensuring the continuity of experimental supply.

[0005] To address the aforementioned technical problems, this invention provides a chemical reagent inventory scheduling method based on experimental demand forecasting, comprising the following steps:

[0006] Acquire historical reagent consumption data and future experimental plan data for each laboratory, and generate future cycle consumption forecasts for different reagents for each laboratory based on the historical consumption data and future experimental plan data.

[0007] Obtain the batch information and current inventory of reagents in the central warehouse and regional warehouse nodes. The batch information includes the production date and shelf life, and calculate the remaining validity period of each reagent batch.

[0008] Based on the preset shelf-life warning threshold, the remaining effective time information of each reagent batch is judged. Reagent batches with a remaining effective time less than the shelf-life warning threshold are marked as near-expiry batches, and the warehouse node location and near-expiry inventory quantity of the near-expiry batches are extracted.

[0009] For each reagent, the predicted future consumption of each laboratory is compared with the current inventory of the reagent in the corresponding regional warehouse node of each laboratory to determine the reagent shortage that the local inventory cannot meet the predicted consumption.

[0010] In response to the existence of reagent shortage, the reagent shortage is taken as the demand, and the near-expiry inventory of batches located in other regional warehouse nodes is taken as the priority supply source, and an allocation network including supply nodes and demand nodes is constructed.

[0011] In the allocation network, allocation schemes are calculated based on a preset allocation optimization model. The allocation optimization model takes the priority consumption of near-expiry batches of inventory as a constraint, and solves the allocation instruction with the shortest total logistics allocation path under the constraint condition. The allocation instruction includes the warehouse to be transferred out, the laboratory to be transferred in, the reagent batch, and the allocation quantity.

[0012] In one embodiment of the present invention, historical reagent consumption data includes a project identifier field, a reagent name field, a reagent actual consumption quantity field, a project execution cycle field, and a consumption record timestamp field; future experimental plan data includes a planned execution project identifier field, a planned execution laboratory field, a planned reagent name field, and a planned execution cycle duration field.

[0013] In one embodiment of the present invention, the predicted future cycle consumption of different reagents for each laboratory is generated based on historical consumption data and future experimental plan data, including:

[0014] Retrieve the project identifier field, reagent name field, actual reagent consumption quantity field, and project execution cycle field from the historical consumption data;

[0015] Historical consumption data is grouped according to the project identifier field. For each project identifier, the total actual consumption of reagents for each reagent name within the project execution cycle corresponding to the project identifier is calculated to form a project-level historical consumption record.

[0016] The planned execution project identifier in the future experimental plan data is matched with the project identifier field in the project-level historical consumption record. If the match is successful, the sum of the actual consumption quantity of reagents in the successfully matched project-level historical consumption record is extracted as the future cycle consumption prediction quantity of the corresponding reagent.

[0017] In one embodiment of the present invention, if the planned execution project identifier contained in the future experimental plan data does not match the same project identifier field in the project-level historical consumption record, then the project type tag corresponding to the planned execution project identifier is obtained.

[0018] Match the project type label with the project type label corresponding to each project identifier in the project-level historical consumption record, and extract the average value of the actual consumption quantity of reagents in all project-level historical consumption records with the same project type label, as the future cycle consumption prediction quantity of the corresponding reagent.

[0019] In one embodiment of the present invention, it further includes:

[0020] Retrieve the historical project execution cycle duration corresponding to the successfully matched project-level historical consumption records;

[0021] Obtain the planned execution period duration corresponding to the planned execution item identifier in the future experiment plan data;

[0022] If the planned execution period is inconsistent with the historical project execution period, the ratio of the planned execution period to the historical project execution period is calculated. The sum of the actual reagent consumption in the successfully matched project-level historical consumption records is multiplied by the ratio to obtain the corrected sum of actual reagent consumption. The corrected sum of actual reagent consumption is then used as the future cycle consumption forecast for the corresponding reagent.

[0023] In one embodiment of the present invention, setting a shelf-life warning threshold based on reagent type includes:

[0024] For room temperature stable reagents, the shelf life warning threshold is set to a remaining effective time of less than or equal to 90 days.

[0025] For refrigeration-sensitive reagents, the shelf-life warning threshold is set to a remaining effective time of less than or equal to 30 days.

[0026] For reagents that require cryopreservation, the shelf life warning threshold is set to a remaining effective time of less than or equal to 60 days.

[0027] In one embodiment of the present invention, if a reagent batch has been marked as an expiration batch and the sum of the predicted future consumption of this reagent by all laboratories is zero, the expiration batch marking of the reagent batch is revoked, and the reagent batch is marked as a batch to be disposed of and is not included in the construction of the dispensing network.

[0028] In one embodiment of the present invention, constructing a dispatch network including supply nodes and demand nodes includes:

[0029] Iterate through all laboratories marked as having reagent shortages, identify each laboratory with a reagent shortage as a demand node, and assign a corresponding reagent shortage value to each demand node;

[0030] Traverse all warehouse nodes, filter out other warehouse nodes that store near-expiry batches and whose warehouse nodes do not belong to the regional warehouse nodes corresponding to the laboratory that caused the reagent shortage, determine the other warehouse nodes as supply nodes, and assign a corresponding value for the near-expiry inventory to each supply node.

[0031] A directed connection path is established between each supply node and each demand node, and the set of all directed connection paths constitutes the allocation network.

[0032] In one embodiment of the present invention, the allocation optimization model is constrained by prioritizing the consumption of near-expiry batches of inventory, including:

[0033] The total amount of near-expiry inventory held by all supply nodes in the statistical allocation network is taken as the total near-expiry supply.

[0034] The total reagent shortage corresponding to all demand nodes in the statistical allocation network is taken as the total demand.

[0035] The first constraint is to require that the sum of the allocation quantities from all supply nodes to all demand nodes in the allocation network equals the smaller of the total supply and demand for near-expiry batches, so as to force the inventory quantity of near-expiry batches to be consumed first and to the maximum extent during the allocation process.

[0036] In one embodiment of the present invention, solving for the dispatch instruction with the shortest total logistics dispatch path under constraints includes:

[0037] Obtain the path distance between each supply node and each demand node in the allocation network;

[0038] Construct an objective function, which is the sum of the products of the number of dispatches and the corresponding path distance values ​​on all paths in the dispatch network where dispatches occur;

[0039] Under the premise of satisfying the first constraint, the allocation scheme that minimizes the objective function is solved by adjusting the allocation quantity from each supply node to each demand node.

[0040] Based on the allocation plan, an allocation instruction is generated, which includes the warehouse identifier, the receiving laboratory identifier, the reagent batch number, and the allocation quantity.

[0041] The technical solution of the present invention has the following advantages compared with the prior art:

[0042] The chemical reagent inventory scheduling method based on experimental demand forecasting described in this invention proactively identifies the reagent needs and potential shortages of each laboratory by introducing an experimental demand forecasting mechanism. Combined with the identification and priority consumption strategy for reagents nearing their expiration date, it avoids reagent stockpiling and spoilage. Simultaneously, by constructing an allocation network and applying an optimization model, it achieves global scheduling of reagent resources, reducing the frequency and cost of emergency procurement, improving reagent turnover efficiency, and optimizing logistics routes, thereby ensuring the continuity of scientific research experiments and the utilization rate of resources. Attached Figure Description

[0043] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0044] Figure 1 This is a flowchart of the chemical reagent inventory scheduling method based on experimental demand prediction of the present invention;

[0045] Figure 2This is a flowchart illustrating an embodiment of the present invention for generating future cycle consumption predictions for different reagents in various laboratories;

[0046] Figure 3 This is a flowchart illustrating another embodiment of the present invention for generating future cycle consumption predictions for different reagents in various laboratories;

[0047] Figure 4 This is a flowchart illustrating the steps of constructing a dispatch network comprising supply nodes and demand nodes according to the present invention.

[0048] Figure 5 This is a flowchart illustrating the steps of the present invention to solve for the shortest total path of logistics allocation under the condition of satisfying constraints. Detailed Implementation

[0049] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0050] To address the issues of passive response, delayed allocation, and inability to perform global optimization by dynamically incorporating shelf life in the existing technologies mentioned above, refer to Figure 1 As shown, this invention provides a chemical reagent inventory scheduling method based on experimental demand prediction. By automating a series of interrelated data processing and calculation steps through computer programs and systems, it fundamentally changes the decision-making logic of reagent allocation.

[0051] At the outset of the implementation, the system doesn't simply check current inventory levels. Instead, it first acquires detailed historical reagent consumption data from each laboratory, along with pre-entered future experimental plans. Based on the coupled analysis of these two types of data, it generates a predicted consumption volume for each laboratory within a specific future period for each reagent. The core of this step lies in introducing a time-dimensional prediction mechanism, ensuring that subsequent inventory assessments focus on future needs rather than past events. Simultaneously, the system delves into batch management granularity, retrieving detailed identification information for each bottle or batch of reagent from the central warehouse and all regional warehouse databases, including its production date and shelf life, and calculating its precise remaining effective duration in real time.

[0052] After obtaining the two core dynamic data points—future demand and actual inventory—the system automatically screens batches with remaining shelf life below a preset warning threshold, clearly marking them as near-expiration batches and pinpointing their specific warehouse node locations and available quantities. This step proactively identifies high-waste-risk assets. Next, the system enters the core supply-demand matching stage. For each reagent, it compares the previously calculated future consumption forecasts for each laboratory with the current inventory levels of regional warehouse nodes within its corresponding service area, accurately calculating the reagent shortage that local inventory cannot meet. If a shortage exists, the system determines that an inter-regional allocation demand exists. It is worth noting that when constructing the allocation network, this method prioritizes near-expiration batch inventory located in other regional warehouse nodes as the primary source of supply to address the shortage, rather than simply searching for the nearest or largest warehouse. Finally, the system uses a preset allocation optimization model for calculation. The model's operation logic takes the priority consumption of near-expiry batches of inventory as a rigid constraint. Under this mandatory premise, the system uses graph theory or operations research algorithms to solve for the allocation path combination that minimizes the total logistics transportation distance, and finally generates an executable allocation instruction that includes the specific warehouse to be transferred out, the laboratory to be transferred in, the specified near-expiry reagent batches, and the accurate allocation quantity.

[0053] Through the orderly combination of the above-mentioned technical features, the beneficial effects of the technical solution proposed in this invention include: First, by introducing a prediction mechanism based on historical consumption and future plans, the system can detect potential demand gaps in advance, transforming the traditional passive response mode into a proactive supply matching mode. This effectively avoids experimental delays caused by reagent shortages and ensures the smooth progress of research projects. Second, by setting the priority consumption of near-expiry batches as a strong constraint in the optimization model, it ensures that reagents on the verge of expiration can be consumed by laboratories with genuine and urgent needs before they become obsolete. This mechanism directly cuts off the path of reagents naturally becoming obsolete due to prolonged storage, reducing the waste rate of research materials from the source and improving the efficiency of fund utilization. Furthermore, based on prioritizing the consumption of near-expiry reagents, the system further seeks the solution with the shortest total transportation distance. This means that while solving the two core problems of waste and shortage, the system can also automatically merge transportation tasks and plan optimal routes, thereby reducing the logistics manpower and time costs associated with cross-regional allocation and achieving multi-dimensional synergistic optimization of efficiency in the inventory scheduling process. Ultimately, the entire solution integrates scattered inventory information, dynamic shelf-life status, and uncertain experimental needs into a closed-loop automated scheduling process, significantly improving the level of refined management in the field of chemical reagent management for large research institutions.

[0054] In this embodiment, it is necessary to generate the future cycle consumption forecast of different reagents for each laboratory by analyzing historical reagent consumption data and future experimental plan data. Therefore, this application further clarifies the specific composition of the historical reagent consumption data and future experimental plan data, wherein: the historical reagent consumption data includes a project identifier field, a reagent name field, a reagent actual consumption quantity field, a project execution cycle field, and a consumption record timestamp field; the future experimental plan data includes a planned execution project identifier field, a planned execution laboratory field, a planned reagent name field, and a planned execution cycle duration field.

[0055] Based on the above historical reagent consumption data and future experimental plan data, referring to Figure 2 As shown, this application further proposes a specific method for generating future cycle consumption forecasts for different reagents in various laboratories, including: First, obtaining the project identifier field, reagent name field, actual reagent consumption quantity field, and project execution cycle field from historical consumption data. This step aims to extract key information related to the project from the original historical reagent consumption data. The project identifier field is used to uniquely identify an experimental project, ensuring the accuracy of subsequent data processing; the reagent name field clarifies the specific type of reagent consumed; the actual reagent consumption quantity field records the actual amount of the reagent used in a specific project, which is the basis for quantifying consumption; the project execution cycle field provides the time range in which the project occurred, helping to correlate consumption with the time dimension. Obtaining these fields is a prerequisite for conducting refined, project-level consumption analysis, and can be obtained from a Laboratory Information Management System (LIMS) or Enterprise Resource Planning (ERP) system through data parsing, database queries, or API calls.

[0056] Secondly, historical consumption data is grouped according to the project identifier field. For each project identifier, the total actual consumption quantity of each reagent name within the project execution cycle corresponding to that project identifier is calculated, forming a project-level historical consumption record. The core of this step is to structure and aggregate the scattered historical consumption data. By grouping based on the project identifier field, all reagent consumption records belonging to the same experimental project can be aggregated together. Subsequently, within each project group, the total consumption of each reagent throughout the entire project execution cycle is further calculated. This processing method transforms the original, potentially fragmented consumption records into project-level historical consumption records centered on "project-reagent-total consumption," thus clearly showing the overall demand pattern of each project for various reagents, providing a more macroscopic and meaningful reference benchmark for subsequent future demand forecasting.

[0057] Finally, the planned project identifier in the future experimental plan data is matched with the project identifier field in the project-level historical consumption records. If a match is found, the sum of the actual reagent consumption quantities in the successfully matched project-level historical consumption records is extracted as the predicted future cycle consumption of the corresponding reagent. This step is crucial for achieving prediction based on project similarity. The system obtains the unique identifier (planned project identifier) ​​of the project to be executed in the future experimental plan data and attempts to find historical projects with the same project identifier field in the existing project-level historical consumption records. If a match is successful, it is considered that the future project and the historical project have a high degree of consistency in reagent consumption patterns. Therefore, the sum of the actual reagent consumption quantities corresponding to the successfully matched historical project is directly extracted as the predicted future cycle consumption of the corresponding reagent for the planned project. This direct mapping method can maximize the use of historical experience, ensure the accuracy and relevance of the prediction results, and is especially suitable for experimental projects with high repetition and standardization.

[0058] However, in practice, new experimental projects may not always have identical historical projects as references, or historical data may lack project identifiers that perfectly match the new project. In such cases, relying solely on direct matching will lead to an inability to accurately predict reagent consumption for new experimental projects, thus affecting the accuracy and efficiency of inventory scheduling.

[0059] In response, this application further proposes that if the planned execution project identifier contained in the future experimental plan data does not match the same project identifier field in the project-level historical consumption record, then the project type label corresponding to the planned execution project identifier is obtained; the project type label is matched with the project type label corresponding to each project identifier in the project-level historical consumption record, and the average value of the actual consumption quantity of reagents in all project-level historical consumption records with the same project type label is extracted as the future cycle consumption prediction quantity of the corresponding reagent.

[0060] Specifically, when the system finds that the project identifier in a future experimental plan cannot be completely matched in the historical records, it will further retrieve the "project type tag" associated with the project to be executed in that plan. The project type tag is metadata that categorizes experimental projects, such as "cell culture experiment," "gene sequencing experiment," and "protein purification experiment." These tags are usually assigned by the user or automatically by the system when the experimental plan is created, and are used to describe the nature, purpose, or main technical approach of the project. This tag can be obtained by querying the project attribute database in the project management system, or by extracting keywords from the project description text using natural language processing techniques for categorization.

[0061] After obtaining the project type tags for future experiment plans, the system iterates through existing project-level historical consumption records. For each historical record, the system searches for the project type tag associated with its corresponding project identifier. Then, it compares the project type tags of future experiment plans with these historical project type tags to identify all historical project records with the same project type tag. This matching can be exact matching, requiring complete tag consistency, or fuzzy matching, such as using semantic similarity algorithms to identify highly relevant tags.

[0062] Once all historical project-level consumption records with the same project type tag as the future experimental plan are identified, the system extracts the actual consumption quantity of a specific reagent from these matching historical records. Then, the average of these actual consumption quantities is calculated. As a statistical measure, the average reflects the typical consumption level of a particular reagent for a certain type of experimental project in the past, thus providing a reasonable consumption prediction benchmark for new projects lacking direct historical project matches. The final calculated average will be designated as the predicted future periodic consumption quantity of the corresponding reagent in the future experimental plan.

[0063] In some existing implementations, when generating future cycle consumption forecasts for different reagents for each laboratory based on historical consumption data and future experimental plan data, the forecast is typically determined by matching the item identifiers in the future experimental plan data with the item identifiers in the historical consumption data. However, this method of directly matching and using historical consumption figures may overlook the potential differences between the execution cycle duration of the future experimental plan and the matched historical item execution cycle duration. Directly using historical consumption data without adjusting for cycle duration may result in inaccurate forecasts of future reagent demand, thereby affecting the rationality of inventory scheduling and causing reagent waste or supply shortages.

[0064] In this regard, refer to Figure 3 As shown, this application further proposes that when generating the future cycle consumption forecast for different reagents in each laboratory, it also includes: determining the future cycle consumption forecast based on different execution durations.

[0065] Specifically, the system retrieves the historical project execution cycle duration corresponding to the successfully matched project-level historical consumption records. It extracts execution cycle information related to specific projects from the matched historical data. This historical project execution cycle duration refers to the length of time a project lasted from start to finish, such as in days, weeks, or months. Obtaining this duration is the basis for subsequent consumption adjustments, ensuring that the actual execution time span of historical projects is considered when making predictions. Simultaneously, the system retrieves the planned execution cycle duration corresponding to the planned project identifiers in the future experimental plan data. This step is used to obtain the planned execution cycle duration of the project to be predicted from the future experimental plan data. This duration refers to the expected duration of the future experimental project, also in days, weeks, or months. Obtaining this duration is crucial for assessing the reagent requirements of future experiments, as it directly reflects the scale and duration of the future experiments.

[0066] If the planned execution period differs from the historical project execution period, the system calculates the ratio of the planned execution period to the historical project execution period. When the planned execution period of a future experiment differs from the execution period of a matched historical project, directly using historical consumption data for prediction will result in bias. Therefore, it is necessary to calculate the ratio of these two period lengths, which serves as an adjustment coefficient to quantify the degree of difference between future and historical projects in the time dimension. For example, if the planned execution period of a future project is twice that of a historical project, the ratio is 2. Based on this, the system multiplies the sum of the actual reagent consumption quantities in the successfully matched project-level historical consumption records by the ratio to obtain the corrected sum of actual reagent consumption quantities. After obtaining the period length ratio, the sum of the actual reagent consumption quantities previously extracted from the project-level historical consumption records is multiplied by this ratio. The purpose of this multiplication operation is to proportionally amplify or reduce the historical consumption data based on the proportional relationship between future and historical projects in the time dimension, thereby obtaining a corrected total reagent consumption that better reflects the actual situation of the future project. Finally, the sum of the corrected actual reagent consumption quantities is used as the predicted future periodic consumption quantity for the corresponding reagent. The sum of the actual reagent consumption quantities after correction for the period length ratio is ultimately determined as the predicted future periodic consumption quantity for the corresponding reagent.

[0067] In this application, the determination of near-expiry batches relies on a preset shelf-life warning threshold. If a single fixed shelf-life warning threshold is used, it may not be able to fully take into account the different storage conditions and stability of different chemical reagents due to their different physicochemical properties. This may result in inaccurate identification of near-expiry reagents, which may trigger warnings too early or too late, affecting the efficiency of inventory management and the effective utilization of reagents.

[0068] To this end, this application further proposes setting shelf-life warning thresholds based on reagent type, aiming to dynamically adjust the triggering conditions for shelf-life warnings according to the inherent properties and storage requirements of chemical reagents. Chemical reagents are diverse, and their stability is affected by various factors such as temperature, humidity, and light, with different reagents exhibiting varying degrees of sensitivity to these factors. By classifying reagents and setting specific warning thresholds for each type, the actual usable lifespan of the reagents can be reflected more accurately.

[0069] For room-temperature stable reagents, the shelf-life warning threshold is set at 90 days or less of remaining effective time. Room-temperature stable reagents typically refer to those that can maintain their chemical properties and activity for a long time at room temperature, and their degradation rate is relatively slow. Setting the shelf-life warning threshold at 90 days, or three months, aims to provide a relatively lenient yet still effectively manageable time window for these reagents. This allows sufficient time for preparation and use before the reagents approach their expiration date, while avoiding frequent and unnecessary warnings triggered by excessively short thresholds, thus reducing the management burden.

[0070] For refrigeration-sensitive reagents, the shelf-life warning threshold is set at 30 days or less of remaining effective time. Refrigeration-sensitive reagents are those that require low-temperature storage (typically 2-8°C) to maintain their stability and activity. These reagents may degrade significantly at room temperature or near their expiration date. Setting the shelf-life warning threshold to 30 days, or one month, reflects the more time-sensitive nature of these reagents. This shorter threshold ensures that the system can promptly identify and initiate the dispensing process before the reagent's activity begins to decline significantly, maximizing the effective use of the reagent.

[0071] For reagents requiring cryopreservation, the shelf-life warning threshold is set at 60 days or less of remaining effective time. Reagents requiring cryopreservation (typically -20°C or lower) have the highest stability requirements. They can be stored for extended periods under cryopreservation conditions, but their stability can rapidly decrease once thawed or if storage conditions change slightly. Setting the shelf-life warning threshold at 60 days, or two months, aims to balance their longer stability under cryopreservation conditions with the preparation time before actual use. This threshold considers both the long lifespan of these reagents under strict storage conditions and allows for a reasonable buffer time for removal from cryopreservation, thawing, repackaging, and transportation, ensuring they are used within their optimal activity period.

[0072] In real-world applications, if a reagent batch is marked as nearing its expiration date, but the sum of all laboratories' predicted future consumption of that reagent is zero (meaning no laboratory has a need for it), including it in the allocation network would not only increase the computational burden on the allocation model but also fail to effectively address the actual consumption of that batch of reagents. This could ultimately lead to the reagents expiring and being scrapped, resulting in resource waste.

[0073] In this regard, this application further proposes that if a reagent batch has been marked as an expiration batch, and the sum of the predicted future consumption of this reagent by all laboratories is zero, then the expiration batch mark of the reagent batch shall be removed, and the reagent batch shall be marked as a batch to be disposed of and shall not participate in the construction of the allocation network.

[0074] Specifically, when the system determines that the remaining validity period of a reagent batch is less than a preset shelf-life warning threshold and marks it as a near-expiration batch, it further aggregates the predicted consumption of that specific reagent from all laboratories over the future period. If the sum of the predicted consumption for that reagent over the future period is zero, it indicates that no laboratory has a demand for that near-expiration reagent in the foreseeable experimental plan. In this case, the system will perform a reversal operation, that is, remove the near-expiration batch mark from the reagent batch, so that it is no longer considered as near-expiration inventory requiring priority allocation. Subsequently, the reagent batch will be remarked as a pending disposal batch. A pending disposal batch is a special inventory status used to identify reagents that are nearing their expiration date but have been confirmed to have no internal consumption need, prompting management personnel to take other non-allocation measures for disposal, such as scrapping, donating, or transferring to other non-experimental uses. Reagent batches marked as pending disposal will explicitly not participate in the subsequent allocation network construction process.

[0075] In some embodiments described above, a method is proposed to address reagent shortages by treating the shortage as demand and prioritizing the near-expiry inventory of batches located in other regional warehouse nodes as the primary supply source, thus constructing an allocation network that includes supply and demand nodes. However, in practice, clearly and accurately identifying and defining these supply and demand nodes, and establishing their potential allocation relationships, is a key challenge in building an effective allocation network. If the definitions of supply and demand nodes are unclear, or their connections are incomplete, it will directly affect the accuracy and efficiency of subsequent allocation optimization models, potentially leading to allocation schemes that fail to effectively address reagent shortages or fully utilize near-expiry inventory.

[0076] In this regard, refer to Figure 4As shown, this application further proposes constructing an allocation network containing supply nodes and demand nodes, including: First, traversing all laboratories marked as having reagent shortages, identifying each laboratory with a reagent shortage as a demand node, and assigning a corresponding reagent shortage value to each demand node. This step aims to clarify the demand side in the allocation network. The system will identify these laboratories one by one based on the list of laboratories whose local inventory cannot meet the predicted future consumption, as determined in the previous steps. For each identified laboratory, its shortage of a specific reagent is taken as its demand, and it is defined as an independent demand node in the allocation network. For example, this can be achieved by querying the database of laboratory records marked with reagent shortages, or by traversing a list containing all laboratories and their reagent shortage information. Each demand node not only represents a specific laboratory but also carries the specific quantity of a specific reagent required by that laboratory, providing a quantitative basis for subsequent allocation decisions.

[0077] Then, all warehouse nodes are traversed to filter out other warehouse nodes that store near-expiry batches of reagents and whose warehouse nodes do not belong to the regional warehouse nodes corresponding to the laboratory causing the reagent shortage. These other warehouse nodes are identified as supply nodes, and each supply node is assigned a corresponding near-expiry batch inventory value. This step is used to identify suppliers in the allocation network, with a particular emphasis on prioritizing the use of near-expiry batch inventory. The system checks all central warehouses and regional warehouse nodes, filtering out warehouses that currently store near-expiry batches of reagents. Furthermore, to achieve cross-regional allocation and optimize resource allocation, the filtering criteria are further limited to these warehouse nodes not belonging to the regional warehouse nodes corresponding to the laboratory causing the reagent shortage. This means that if a laboratory A has a reagent shortage, the regional warehouse node to which laboratory A belongs will not be considered a supply node for that shortage; instead, warehouses in other regions will be sought to provide near-expiry inventory. For each eligible warehouse node, its near-expiry inventory quantity is used as its supply quantity, and it is defined as an independent supply node in the allocation network. For example, this can be done through database queries, combined with warehouse location information, near-expiry batch markings, and the association with the laboratory region. Each supply node represents a warehouse with the capacity to supply near-expiry reagents, and specifies the quantity of near-expiry reagents it can supply.

[0078] Finally, a directed connection path is established between each supply node and each demand node. The set of all directed connections constitutes the allocation network. This step aims to construct the potential allocation relationships between suppliers and demanders, forming a complete allocation network. Once the supply and demand nodes are explicitly identified and assigned corresponding supply and demand quantities, the system establishes logical connections between all possible supply and demand nodes. These connections are "directed," indicating that reagents can flow from supply nodes to demand nodes. For example, this can be represented using a graph structure from graph theory, where nodes represent suppliers and demanders, and edges represent potential allocation paths. This network structure clearly shows all possible allocation routes, providing a basic framework for subsequent allocation optimization models, enabling the model to find the optimal allocation scheme among all potential supply-demand pairs.

[0079] To achieve the above objectives, the system first accurately calculates the total amount of near-expiry inventory held by all supply nodes in the allocation network, using this as the total near-expiry supply. This step aims to comprehensively understand the total amount of near-expiry reagent resources currently available for priority allocation, providing an accurate data foundation for subsequent optimization decisions. Specifically, the system iterates through all identified supply nodes, which are warehouse nodes that store near-expiry batches and do not belong to the regional warehouse nodes corresponding to the laboratories causing reagent shortages. For each supply node, the system extracts its near-expiry inventory quantity and sums these quantities to obtain a value representing the total near-expiry reagent supply capacity of the entire network.

[0080] Simultaneously, the system also calculates the sum of reagent shortages corresponding to all demand nodes in the allocation network, using this as the total demand. This is to accurately assess the actual scale of demand for a specific reagent across the entire network. Specifically, the system iterates through all laboratories marked as having reagent shortages; these laboratories are identified as demand nodes. For each demand node, the system obtains its corresponding reagent shortage and sums these shortages to obtain a value representing the total demand of all laboratories.

[0081] Based on this, this application establishes a first constraint requiring that the sum of the quantities allocated from all supply nodes to all demand nodes in the allocation network must equal the smaller of the total near-expiry supply and total demand. This core constraint is designed to force the near-expiry batches of inventory to be prioritized and consumed to the maximum extent possible during the allocation process. By limiting the total allocation amount to the smaller of the total near-expiry supply and total demand, this application ensures that when near-expiry inventory is sufficient, it is used first to meet all demand; and when near-expiry inventory is insufficient to meet all demand, all near-expiry inventory is allocated. This mechanism effectively prevents near-expiry reagents from becoming expired and wasting due to untimely allocation, thereby achieving optimized utilization of inventory resources.

[0082] In some of the embodiments described above in this application, a dispatch network including supply and demand nodes is proposed, and the dispatch plan is calculated with the priority consumption of near-expiry batches of inventory as a constraint. However, in actual operation, if only the priority consumption of near-expiry batches is considered without effective optimization of logistics costs, the dispatch path may be too long, increasing transportation costs and time, thereby reducing overall dispatch efficiency and affecting the experimental progress.

[0083] In this regard, refer to Figure 5 As shown, this application further proposes a method to solve for the shortest total logistics dispatch path under certain constraints. Specifically, the method first obtains the path distance value between each supply node and each demand node in the dispatch network. These path distance values ​​can be the actual road distance and shortest travel time calculated by a Geographic Information System (GIS), or the logistics cost estimated based on historical transportation data, used to quantify the logistics costs of different dispatch paths.

[0084] Based on this, an objective function is constructed, which is the sum of the products of the dispatch quantity and the corresponding path distance value on all paths in the dispatch network where dispatch occurs. This objective function aims to minimize the overall logistics workload or cost.

[0085] Subsequently, under the premise of satisfying the first constraint, the allocation of reagents from each supply node to each demand node is adjusted to find an optimal allocation scheme that minimizes the objective function. This solution process is typically implemented using optimization algorithms such as linear programming, integer programming, network flow algorithms (e.g., minimum cost maximum flow algorithm), or heuristic algorithms. The algorithm iteratively adjusts the amount of reagents allocated from each supply node to each demand node until an optimal solution is found that satisfies both the constraint of prioritizing the consumption of near-expiry batches of inventory and minimizes the total logistics cost.

[0086] Finally, based on the solved allocation scheme, an allocation instruction is generated. This allocation instruction is an executable logistics operation checklist, clearly including the warehouse identifier for the originating warehouse, the laboratory identifier for the receiving laboratory, the reagent batch number, and the allocation quantity, which can directly guide logistics operations.

[0087] Through the above technical solution, while ensuring the priority consumption of near-expiry batches of inventory, the system can further obtain the path distance values ​​between each node and use them to construct an objective function to solve for the allocation scheme with the shortest total logistics path. This effectively avoids increased logistics costs and time delays caused by blind allocation, significantly improving the economy and efficiency of reagent allocation. This solution not only ensures the timely consumption of near-expiry reagents and reduces losses, but also reduces overall operating costs by optimizing logistics routes, ensuring the smooth progress of experimental plans, thereby achieving refined and intelligent reagent inventory management.

[0088] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A chemical reagent inventory scheduling method based on experimental demand forecasting, characterized in that, Includes the following steps: Obtain historical reagent consumption data and future experimental plan data for each laboratory. Historical reagent consumption data includes fields such as project identifier, reagent name, actual reagent consumption quantity, project execution period, and consumption record timestamp. Future experimental plan data includes fields such as planned project identifier, planned laboratory, reagent names involved in the plan, and planned execution period duration. This process generates future cycle consumption forecasts for different reagents for each laboratory based on historical consumption data and future experimental plan data. This includes: obtaining the project identifier field, reagent name field, actual reagent consumption quantity field, and project execution cycle field from the historical consumption data; grouping the historical consumption data according to the project identifier field; for each project identifier, summing the actual reagent consumption quantity corresponding to each reagent name within the project execution cycle corresponding to the project identifier, forming a project-level historical consumption record; matching the planned execution project identifier in the future experimental plan data with the project identifier field in the project-level historical consumption record; if a match is successful, extracting the sum of the actual reagent consumption quantities from the successfully matched project-level historical consumption record as the future cycle consumption forecast for the corresponding reagent. If the planned execution project identifier contained in the future experimental plan data does not match the same project identifier field in the project-level historical consumption record, then obtain the project type label corresponding to the planned execution project identifier; match the project type label with the project type label corresponding to each project identifier in the project-level historical consumption record, and extract the average value of the actual consumption quantity of reagents in all project-level historical consumption records with the same project type label, as the future cycle consumption prediction quantity of the corresponding reagent. Obtain the batch information and current inventory of reagents in the central warehouse and regional warehouse nodes. The batch information includes the production date and shelf life, and calculate the remaining validity period of each reagent batch. Based on the preset shelf-life warning threshold, the remaining effective time information of each reagent batch is judged. Reagent batches with a remaining effective time less than the shelf-life warning threshold are marked as near-expiry batches, and the warehouse node location and near-expiry inventory quantity of the near-expiry batches are extracted. For each reagent, the predicted future consumption of each laboratory is compared with the current inventory of the reagent in the corresponding regional warehouse node of each laboratory to determine the reagent shortage that the local inventory cannot meet the predicted consumption. In response to the existence of reagent shortage, the reagent shortage is taken as the demand, and the near-expiry inventory of batches located in other regional warehouse nodes is taken as the priority supply source, and an allocation network including supply nodes and demand nodes is constructed. In the allocation network, allocation schemes are calculated based on a preset allocation optimization model. The allocation optimization model takes the priority consumption of near-expiry batches of inventory as a constraint, and solves the allocation instruction with the shortest total logistics allocation path under the constraint condition. The allocation instruction includes the warehouse to be transferred out, the laboratory to be transferred in, the reagent batch, and the allocation quantity.

2. The chemical reagent inventory scheduling method based on experimental demand forecasting according to claim 1, characterized in that: Also includes: Retrieve the historical project execution cycle duration corresponding to the successfully matched project-level historical consumption records; Obtain the planned execution period duration corresponding to the planned execution item identifier in the future experiment plan data; If the planned execution period is inconsistent with the historical project execution period, the ratio of the planned execution period to the historical project execution period is calculated. The sum of the actual reagent consumption in the successfully matched project-level historical consumption records is multiplied by the ratio to obtain the corrected sum of actual reagent consumption. The corrected sum of actual reagent consumption is then used as the future cycle consumption forecast for the corresponding reagent.

3. The chemical reagent inventory scheduling method based on experimental demand forecasting according to claim 1, characterized in that: Set shelf-life warning thresholds based on reagent type, including: For room temperature stable reagents, the shelf life warning threshold is set to a remaining effective time of less than or equal to 90 days. For refrigeration-sensitive reagents, the shelf-life warning threshold is set to a remaining effective time of less than or equal to 30 days. For reagents that require cryopreservation, the shelf life warning threshold is set to a remaining effective time of less than or equal to 60 days.

4. The chemical reagent inventory scheduling method based on experimental demand forecasting according to claim 1, characterized in that: If a reagent batch has been marked as near expiration, and the sum of the predicted future consumption of this reagent by all laboratories is zero, then the near expiration mark on the reagent batch will be removed, and the reagent batch will be marked as a batch to be disposed of, and will not participate in the construction of the allocation network.

5. The chemical reagent inventory scheduling method based on experimental demand forecasting according to claim 1, characterized in that: Constructing a dispatch network that includes supply nodes and demand nodes, including: Iterate through all laboratories marked as having reagent shortages, identify each laboratory with a reagent shortage as a demand node, and assign a corresponding reagent shortage value to each demand node; Traverse all warehouse nodes, filter out other warehouse nodes that store near-expiry batches and whose warehouse nodes do not belong to the regional warehouse nodes corresponding to the laboratory that caused the reagent shortage, determine the other warehouse nodes as supply nodes, and assign a corresponding value for the near-expiry inventory to each supply node. A directed connection path is established between each supply node and each demand node, and the set of all directed connection paths constitutes the allocation network.

6. The chemical reagent inventory scheduling method based on experimental demand forecasting according to claim 5, characterized in that: The allocation optimization model uses the quantity of near-expiry batches of inventory as a constraint, including: The total amount of near-expiry inventory held by all supply nodes in the statistical allocation network is taken as the total near-expiry supply. The total reagent shortage corresponding to all demand nodes in the statistical allocation network is taken as the total demand. The first constraint is to require that the sum of the allocation quantities from all supply nodes to all demand nodes in the allocation network equals the smaller of the total supply and demand for near-expiry batches, so as to force the inventory quantity of near-expiry batches to be consumed first and to the maximum extent during the allocation process.

7. The chemical reagent inventory scheduling method based on experimental demand forecasting according to claim 6, characterized in that: Solving for the dispatch instruction that minimizes the total logistics dispatch path while satisfying constraints includes: Obtain the path distance between each supply node and each demand node in the allocation network; Construct an objective function, which is the sum of the products of the number of dispatches and the corresponding path distance values ​​on all paths in the dispatch network where dispatches occur; Under the premise of satisfying the first constraint, the allocation scheme that minimizes the objective function is solved by adjusting the allocation quantity from each supply node to each demand node. Based on the allocation plan, an allocation instruction is generated, which includes the warehouse identifier, the receiving laboratory identifier, the reagent batch number, and the allocation quantity.

Images