Medical material bidding inquiry management system

Abstract

The application relates to the technical field of data processing, and discloses a medical material bidding management system, which comprises the following modules: a demand analysis module for extracting a benchmark bidding upper limit; a dynamic benchmark calculation module for determining a dynamic correction coefficient according to the change rate of the ratio of material consumption to stock and outputting a dynamic bidding upper limit boundary; a bidding broadcast module for issuing a bidding instruction stream; a comprehensive evaluation module for receiving a response data set; and a data intervention feedback module for calculating a weight attenuation factor according to the difference between an average bid price offset rate and the dynamic correction coefficient, adjusting the index weight distribution, and outputting an optimization control instruction. The application adopts dynamic boundary adjustment and index weight feedback, eliminates communication deadlock caused by a fixed threshold under a variable load working condition, improves evaluation calculation instability caused by abnormal convergence of multi-party data, and improves the output continuity of the scheduling decision system.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, and in particular relates to a medical supplies procurement and pricing management system. Background Technology

[0002] Currently, with the expansion of distributed interconnection topology, the ratio between the rate of material consumption and the current inventory exhibits nonlinear instantaneous disturbances. The acceleration of this data flow causes abrupt changes in boundary tension in the parameter processing path, and the evolution slope of its data characteristics has a close positive correlation with the marginal cost of the actual response on the supply side. This leads to the data processing system having an adaptive feedback adjustment mechanism based on the characteristics of changes in the state of the original materials. However, the physical carrying capacity of the distributed nodes in the existing system is limited, making it difficult to support concurrent throughput under extreme operating conditions. Software control methods also have shortcomings. For example, Chinese invention patent application CN117974007A discloses an intelligent price inquiry method, device, medium, and product based on goods procurement. The scheme implicitly relies on an ideal environment where the market supply elasticity is stable and the data input time sequence is completely continuous. It constructs a static benchmark price by selecting the lowest price and conducts multi-dimensional price comparison. Under abrupt operating conditions, this control method lacks dynamic coupling between the parameter boundary and the evaluation matrix and the real-time state, making it unable to adaptively track the surge in material acquisition costs under external disturbances.

[0003] When faced with sudden changes in external input flow or local supply disruptions, existing processing architectures relying on static stability cannot adaptively track soaring material acquisition costs due to their fixed price limit parameters. This causes the system's command flow to exceed the price limits of multiple supply nodes, leading to the underlying communication protocol directly intercepting the command flow and preventing the system from obtaining effective feedback datasets. This results in a deadlock in data interaction. Furthermore, the original static evaluation matrix maintains a fixed operator weight distribution under extreme pressure, failing to detect sudden deterioration in the timing performance indicators of supply nodes. Conventional improvement paths, such as directly increasing fixed limit parameters or adding computing servers to divert data interaction flows, are prone to causing abnormal concentration of response data returned by distributed nodes near the adjusted boundaries. This leads to multiple solutions or non-convergence in the multi-indicator decision matrix when processing highly similar multidimensional data. This not only fails to accurately isolate nodes with high performance variance and latency risks but also reduces the overall stability of the data processing system due to excessive concentration of computing power and increased communication overhead.

[0004] Therefore, the technical problem to be solved by this invention is how to construct a two-way adaptive control mechanism for price limits and weight matrices that can respond in real time to external material change data streams and dynamically reconstruct decision topology, thereby providing efficient and reliable industrial data integration services, eliminating system communication deadlock when external nonlinear disturbances occur, and ensuring the calculation convergence accuracy and continuous output performance of multi-index joint decision matrix. Summary of the Invention

[0005] This invention aims to solve the problems of communication deadlock and instability in decision matrix calculation caused by fixed thresholds under changing operating conditions using conventional techniques.

[0006] In this technical solution, a medical supplies procurement and pricing management system includes:

[0007] The demand analysis module is used to extract the benchmark demand parameter set for medical supplies to be procured. The benchmark demand parameter set includes the benchmark inquiry upper limit.

[0008] The dynamic benchmark calculation module is used to collect the real-time consumption and current physical inventory of medical supplies to be purchased within a set area, calculate the real-time consumption-to-inventory ratio, and calculate the dynamic correction coefficient based on the rate of change of the real-time consumption-to-inventory ratio within a set time window. When the dynamic correction coefficient is greater than the set stability threshold, the dynamic correction coefficient is used to calculate the multiplier compensation for the benchmark inquiry upper limit and output the dynamic inquiry upper limit boundary.

[0009] The inquiry broadcast module is used to generate an inquiry instruction stream that includes a dynamic upper limit boundary for inquiry, and to distribute the inquiry instruction stream to multiple registered supplier nodes;

[0010] The comprehensive evaluation module is used to receive response datasets containing real-time quotation parameters from multiple supplier nodes. Based on the weight decay factor, the comprehensive evaluation module reduces the weight of real-time quotation parameters in the internal decision evaluation matrix, and simultaneously raises the delivery delay variance parameter of the corresponding supplier node in the historical equivalent fluctuation period to the primary judgment weight. It also outputs optimization control instructions based on the internal decision evaluation matrix.

[0011] The data intervention feedback module, which connects the dynamic benchmark calculation module and the comprehensive evaluation module, is used to calculate the average price deviation rate within the response dataset and the price deviation difference between the average price deviation rate and the dynamic correction coefficient; when the price deviation difference is less than the set warning tolerance, a weight decay factor is generated.

[0012] Preferably, the data intervention feedback module extracts all real-time pricing parameters from the response dataset received by the comprehensive evaluation module, calculates the arithmetic mean of all real-time pricing parameters, and calculates the ratio of the arithmetic mean to the dynamic inquiry upper limit boundary output by the dynamic benchmark calculation module, outputting the average pricing deviation rate; the data intervention feedback module subtracts the average pricing deviation rate, which is a dimensionless coefficient, from the dynamic correction coefficient, which is also a dimensionless coefficient, and outputs the pricing deviation difference.

[0013] Preferably, the dynamic benchmark calculation module includes a time-series decay compensation module; the dynamic benchmark calculation module monitors the data update time interval between real-time consumption and current physical inventory; when the data update time interval is greater than a set heartbeat cycle, the time-series decay compensation module extracts the last retained real inventory ratio and matches it with a set time-series decay algorithm operator, multiplies the last retained real inventory ratio with a set time-series decay coefficient using the time-series decay algorithm operator to output a composite consumption ratio, and uses the composite consumption ratio to replace the real-time consumption inventory ratio to calculate the dynamic correction coefficient.

[0014] Preferably, the data intervention feedback module retrieves historical business records from the database, searches for historical dynamic correction coefficient values ​​whose absolute difference with the currently calculated dynamic correction coefficient is less than the set interval tolerance, and determines the historical time period in which the historical dynamic correction coefficient value is located as the historical equivalent fluctuation time period.

[0015] Preferably, the comprehensive evaluation module assigns a first weight to the real-time quotation parameter and a second weight to the delivery delay variance parameter, with the first weight being greater than the second weight. When the data intervention feedback module generates a weight decay factor, the comprehensive evaluation module uses the weight decay factor to decay the first weight and simultaneously increases the second weight to the primary judgment level, thus reconstructing the operator distribution in the internal decision evaluation matrix.

[0016] Preferably, the system also includes an edge degradation control module, which is connected to the comprehensive evaluation module. The edge degradation control module monitors the system's computing resource utilization rate and the communication network bandwidth load status. When at least one of the following conditions is met—the system's computing resource utilization rate exceeding a set computing threshold and the communication network bandwidth load status exceeding a set communication threshold—the edge degradation control module issues a weight convergence command to the comprehensive evaluation module. In response to the weight convergence command, the comprehensive evaluation module shuts down the multi-dimensional high-order tensor computation flow in its internal decision evaluation matrix, and reduces the multi-index evaluation to a single-index veto ranking based on the delivery delay variance parameter.

[0017] Preferably, both the dynamic benchmark calculation module and the demand analysis module are connected to the regional public health supplies consumption monitoring network through an internal high-concurrency streaming data interface. They collect the consumption rate of the medical supplies to be purchased at each demand node and the current physical inventory at a streaming data reading cycle of less than or equal to 10ms, providing a data chain extraction source for real-time consumption and current physical inventory.

[0018] Preferably, the inquiry broadcast module includes a directed acyclic graph state transition module; the directed acyclic graph state transition module reconstructs the inquiry process topology according to the dynamic inquiry upper limit boundary, and sends the generated inquiry instruction stream containing the dynamic inquiry upper limit boundary to multiple supplier nodes according to the predetermined concurrency control path.

[0019] Preferably, the comprehensive evaluation module uses the weight decay factor to decay the first weight as follows: multiply the first weight by the weight decay factor to obtain the decayed bid weight value; the weight decay factor is a value between 0 and 1, and the magnitude of the weight decay factor decreases linearly as the bid offset difference decreases.

[0020] Preferably, the system also includes a human-machine collaborative interaction interface module, which is connected to the output end of the comprehensive evaluation module. The human-machine collaborative interaction interface module receives the optimization control instructions output by the comprehensive evaluation module, converts the optimal matching node sequence corresponding to the optimization control instructions and the trigger status of the data intervention feedback module into graphical signals and outputs them for display.

[0021] Compared with existing technologies, the medical supplies procurement and pricing management system of this invention has the following advantages:

[0022] 1. In the management of medical supplies procurement and price inquiry, by collecting material change data streams within a preset physical area and calculating the ratio of real-time consumption to current inventory, when the slope of the ratio's sliding time window exceeds a preset stable threshold, an adaptive state adjustment factor is generated. This adjustment factor is used as a compensation parameter to apply to the limit parameters in the benchmark demand parameter set and outputs a dynamic adjustment boundary. This enables the data processing system to automatically broaden the parameter processing path through internal data closure under nonlinear abrupt changes in the external input stream, eliminating the lack of distributed instruction stream feedback and system decision deadlock caused by static hard threshold interception.

[0023] 2. By establishing a collaborative feedback control loop between the dynamic benchmark calculation unit and the game evaluation unit, the average data offset rate of the response dataset returned by the multi-party distributed nodes is calculated in real time. When the deviation between the average data offset rate and the state adjustment factor is less than the preset threshold and the multi-party data coordination is inconsistent, the operator weight distribution inside the multi-index joint decision matrix is ​​adaptively reconstructed. The computing power weight of the current processing feature in the comprehensive scoring is reduced based on the data offset rate. At the same time, the timing performance stability index of the corresponding node in the same historical pressure period is raised to the priority judgment level. This eliminates the interference of the concentrated enrichment of multiple nodes near the data boundary on the accuracy of the calculation and ensures that the instruction flow is directed to the entity node with high deterministic timing response capability.

[0024] 3. By deploying adaptive degradation control logic based on data update timing intervals, the update time difference of the input stream is continuously monitored. When the update time difference is greater than the set period and the data input from multiple parties is discontinuous, the last valid stock ratio data is automatically extracted and matched with the preset timing decay algorithm operator. The synthetic compensation parameter is calculated using the existing processing units in the system. This compensation parameter replaces the missing input stream data in the system state judgment, maintains the continuous output of the decision calculation stream, and avoids the system from entering a suspended state due to instantaneous concurrency or signal loss. Attached Figure Description

[0025] Figure 1 This is a flowchart of the data interaction and feedback control process of the medical supplies procurement and pricing management system of this invention;

[0026] Figure 2 This is a diagram showing the network topology and node deployment architecture of the medical supplies procurement and pricing management system of this invention. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0028] A medical supplies procurement and pricing management system includes:

[0029] The demand analysis module is used to extract the benchmark demand parameter set for medical supplies to be procured. The benchmark demand parameter set includes the benchmark inquiry upper limit.

[0030] The dynamic benchmark calculation module is used to collect the real-time consumption and current physical inventory of medical supplies to be purchased within a set area, calculate the real-time consumption-to-inventory ratio, and calculate the dynamic correction coefficient based on the rate of change of the real-time consumption-to-inventory ratio within a set time window. When the dynamic correction coefficient is greater than the set stability threshold, the dynamic correction coefficient is used to calculate the multiplier compensation for the benchmark inquiry upper limit and output the dynamic inquiry upper limit boundary.

[0031] The inquiry broadcast module is used to generate an inquiry instruction stream that includes a dynamic upper limit boundary for inquiry, and to distribute the inquiry instruction stream to multiple registered supplier nodes;

[0032] The comprehensive evaluation module is used to receive response datasets containing real-time quotation parameters from multiple supplier nodes. Based on the weight decay factor, the comprehensive evaluation module reduces the weight of real-time quotation parameters in the internal decision evaluation matrix, and simultaneously raises the delivery delay variance parameter of the corresponding supplier node in the historical equivalent fluctuation period to the primary judgment weight. It also outputs optimization control instructions based on the internal decision evaluation matrix.

[0033] The data intervention feedback module, which connects the dynamic benchmark calculation module and the comprehensive evaluation module, is used to calculate the average price deviation rate within the response dataset and the price deviation difference between the average price deviation rate and the dynamic correction coefficient; when the price deviation difference is less than the set warning tolerance, a weight decay factor is generated.

[0034] Preferably, the data intervention feedback module extracts all real-time pricing parameters from the response dataset received by the comprehensive evaluation module, calculates the arithmetic mean of all real-time pricing parameters, and calculates the ratio of the arithmetic mean to the dynamic inquiry upper limit boundary output by the dynamic benchmark calculation module, outputting the average pricing deviation rate; the data intervention feedback module subtracts the average pricing deviation rate, which is a dimensionless coefficient, from the dynamic correction coefficient, which is also a dimensionless coefficient, and outputs the pricing deviation difference.

[0035] Preferably, the dynamic benchmark calculation module includes a time-series decay compensation module; the dynamic benchmark calculation module monitors the data update time interval between real-time consumption and current physical inventory; when the data update time interval is greater than a set heartbeat cycle, the time-series decay compensation module extracts the last retained real inventory ratio and matches it with a set time-series decay algorithm operator, multiplies the last retained real inventory ratio with a set time-series decay coefficient using the time-series decay algorithm operator to output a composite consumption ratio, and uses the composite consumption ratio to replace the real-time consumption inventory ratio to calculate the dynamic correction coefficient.

[0036] Preferably, the data intervention feedback module retrieves historical business records from the database, searches for historical dynamic correction coefficient values ​​whose absolute difference with the currently calculated dynamic correction coefficient is less than the set interval tolerance, and determines the historical time period in which the historical dynamic correction coefficient value is located as the historical equivalent fluctuation time period.

[0037] Preferably, the comprehensive evaluation module assigns a first weight to the real-time quotation parameter and a second weight to the delivery delay variance parameter, with the first weight being greater than the second weight. When the data intervention feedback module generates a weight decay factor, the comprehensive evaluation module uses the weight decay factor to decay the first weight and simultaneously increases the second weight to the primary judgment level, thus reconstructing the operator distribution in the internal decision evaluation matrix.

[0038] Preferably, the system also includes an edge degradation control module, which is connected to the comprehensive evaluation module. The edge degradation control module monitors the system's computing resource utilization rate and the communication network bandwidth load status. When at least one of the following conditions is met—the system's computing resource utilization rate exceeding a set computing threshold and the communication network bandwidth load status exceeding a set communication threshold—the edge degradation control module issues a weight convergence command to the comprehensive evaluation module. In response to the weight convergence command, the comprehensive evaluation module shuts down the multi-dimensional high-order tensor computation flow in its internal decision evaluation matrix, and reduces the multi-index evaluation to a single-index veto ranking based on the delivery delay variance parameter.

[0039] Preferably, both the dynamic benchmark calculation module and the demand analysis module are connected to the regional public health supplies consumption monitoring network through an internal high-concurrency streaming data interface. They collect the consumption rate of the medical supplies to be purchased at each demand node and the current physical inventory at a streaming data reading cycle of less than or equal to 10ms, providing a data chain extraction source for real-time consumption and current physical inventory.

[0040] Preferably, the inquiry broadcast module includes a directed acyclic graph state transition module. The directed acyclic graph state transition module reconstructs the inquiry process topology based on the dynamic inquiry upper limit boundary and sends the generated inquiry instruction stream containing the dynamic inquiry upper limit boundary to multiple supplier nodes according to a predetermined concurrency control path.

[0041] Preferably, the comprehensive evaluation module uses the weight decay factor to decay the first weight as follows: multiply the first weight by the weight decay factor to obtain the decayed bid weight value; the weight decay factor is a value between 0 and 1, and the magnitude of the weight decay factor decreases linearly as the bid offset difference decreases.

[0042] Preferably, the system also includes a human-machine collaborative interaction interface module, which is connected to the output end of the comprehensive evaluation module. The human-machine collaborative interaction interface module receives the optimization control instructions output by the comprehensive evaluation module, converts the optimal matching node sequence corresponding to the optimization control instructions and the trigger status of the data intervention feedback module into graphical signals and outputs them for display.

[0043] Example 1: In the regional medical supplies procurement and pricing management system, the system monitors the physical status of various high-frequency consumable medical supplies in real time. When the ratio of the real-time consumption to the physical inventory of a specific surgical consumable within a preset time window shows a non-linear increasing trend, the demand analysis module extracts the benchmark procurement parameter set of the medical supplies to be procured, which includes historical benchmark price limit parameters. The dynamic benchmark calculation module collects the consumption rate of supplies and the current physical inventory at a reading cycle of 10ms through the internal high-concurrency streaming data interface, and calculates the real-time consumption-to-inventory ratio.

[0044] To address the mismatch between the time constant of physical inventory counting and the millisecond-level sampling period under high-frequency data acquisition conditions, the current method of extracting physical inventory levels is transformed into tracking the digital warehouse transaction flow data stream: a high-concurrency streaming data interface reads electronic transaction flow data from the warehouse management database and extracts the current inventory balance at the sampling moment as the real-time inventory proxy quantity; to eliminate sudden fluctuation noise caused by discrete physical inbound and outbound operations, the system uses a moving average filter window containing 5 sampling periods to update the real-time inventory proxy quantity, establishing a continuous inventory feature chain at the digital level that meets the accuracy of dynamic slope calculation. Based on this, to eliminate high-frequency sampling... To address the cross-scale contradictions between discrete changes in physical warehousing, a time window is set that spans a historical sampling period of at least 3600 seconds. The evolution slope is not calculated by directly differentiating adjacent 10ms discrete sampling points, but rather by performing least squares linear fitting on multiple sets of continuous inventory characteristic chain data accumulated within the set time window. The slope of the fitted line is then extracted as the time-series evolution slope, thereby transforming discrete step disturbances in and out of warehousing into a smooth and continuous rate of change reflecting the long-term consumption trend of materials on the overall time scale. This ensures that the demand tension coefficient calculated at a millisecond-level interface reading frequency has engineering feasibility and physical self-consistency.

[0045] The dynamic benchmark calculation module monitors the data update interval between real-time consumption and current physical inventory. When the data update interval exceeds the set heartbeat cycle, the time-series decay compensation module in the dynamic benchmark calculation module extracts the last retained true inventory ratio and matches it with a preset time-series decay algorithm operator. The time-series decay algorithm operator multiplies the last retained true inventory ratio by the set time-series decay coefficient, outputting a composite consumption ratio. This composite consumption ratio is used to replace the real-time consumption inventory ratio in subsequent calculations. This module calculates the evolution slope of the real-time consumption inventory ratio within the sliding time window. When the evolution slope exceeds the set stable benchmark threshold, it is mapped to a demand tension coefficient, which the system uses as a supplement. The compensation parameter is applied to the historical benchmark price limit parameters in the benchmark procurement parameter set to calculate the dynamic inquiry upper limit boundary. The inquiry broadcast module uses the directed acyclic graph state transition module to generate an inquiry instruction stream containing the dynamic inquiry upper limit boundary and distributes the inquiry instruction stream to multiple supplier nodes. The directed acyclic graph state transition module reconstructs the inquiry process topology based on the dynamic inquiry upper limit boundary. The specific actions include: comparing the historical price benchmark of each supplier node with the dynamic inquiry upper limit boundary; if the historical price benchmark exceeds the boundary, setting the weight coefficient of the edge pointing to the corresponding supplier node address in the process topology adjacency matrix to a zero value, cutting off the corresponding data distribution path, and sending the inquiry instruction stream containing the boundary along the remaining active edge weight path.

[0046] Each supplier node returns a response dataset containing real-time pricing parameters. The game evaluation unit receives the response dataset. The data intervention feedback module, which connects the dynamic benchmark calculation module and the comprehensive evaluation module, calculates the average price deviation rate within the response dataset. The average price deviation rate is defined as the ratio of the arithmetic mean of all real-time pricing parameters to the dynamic inquiry upper limit boundary. The data intervention feedback module subtracts the average price deviation rate from the demand tension coefficient, outputting the price deviation difference. When the price deviation difference is less than the set warning tolerance, the data intervention feedback module generates a weight decay factor and feeds it back to the comprehensive evaluation module. The comprehensive evaluation module, based on the weight decay factor, adjusts the real-time... The weight of the pricing parameter in the internal decision evaluation matrix decreases according to a set decay step. Simultaneously, historical business records are retrieved from the database, and historical dynamic correction coefficient values ​​whose absolute difference from the current dynamic correction coefficient is less than a set tolerance range are identified. The historical time period containing this value is defined as the historical equivalent fluctuation period. Furthermore, the comprehensive evaluation module prioritizes the delivery delay variance parameter of the corresponding supplier node within the historical equivalent fluctuation period as the primary judgment weight. Based on the reconstructed internal decision evaluation matrix, optimization control instructions are output. Specifically, under normal operating conditions, the internal decision evaluation matrix consists of real-time pricing parameters, delivery delay variance, historical fulfillment rate, and network response latency. The system consists of a four-dimensional tensor data stream composed of four independent feature terms; a multi-dimensional higher-order tensor computation stream calculates the comprehensive competitiveness feature vector of each supplier node by performing multilinear principal component analysis and higher-order singular value decomposition on the above four-dimensional feature tensor; when the edge degradation control module issues a weight convergence command to the comprehensive evaluation module due to overload of computing resources or communication bandwidth, the comprehensive evaluation module responds to the weight convergence command by directly shutting down the iterative solution matrix operator in the above four-dimensional higher-order tensor computation stream, forcibly reducing the order and dimension of the evaluation matrix from a four-dimensional higher-order tensor to a single-dimensional vector containing only the delivery delay variance feature, thereby achieving efficient ranking output based on a single-index veto system using the delivery delay variance parameter. To avoid the limitation of a single variance parameter failing to characterize the absolute delay magnitude, the internal decision evaluation matrix synchronously retrieves the average delivery delay mean within the historical equivalent fluctuation period. A comprehensive risk assessment index is constructed by multiplying the mean and variance parameter. When the optimization control command is output, the system calls a hash de-identification algorithm to hide the supplier's real name and physical warehouse address code. Only the Internet Protocol address and device port number of the optimal matching supplier node that has passed verification are written into the main procurement queue register, driving the communication gateway to open the corresponding channel. The optimization control command outputs the optimal matching node sequence through the human-machine collaborative interaction interface module, thereby guiding the system to accurately identify supply chain paths with real delivery capabilities under extreme operating conditions.

[0047] Example 2: This experiment aims to verify the decision convergence performance of the medical supplies procurement and pricing management system in the context of high-frequency sudden consumption coupled with drastic fluctuations in supply-side prices. The experimental platform is built on a distributed stream computing architecture and is equipped with a real-time data acquisition interface with high-concurrency data processing capabilities. Its sampling frequency is set to 100Hz to ensure that the system response latency is less than 50ms when the input stream data generates nonlinear disturbances. The experiment selects three types of medical supplies with different supply elasticities as test objects: routine consumable supplies, shortage risk supplies, and highly volatile emergency supplies. After the experiment starts, the system extracts the benchmark procurement parameter set by the demand analysis module. This parameter set includes historical benchmark price limit parameters. The dynamic benchmark calculation module collects the consumption rate of supplies and the current physical inventory at a period of 10ms through the internal high-concurrency stream data interface to calculate the real-time consumption-to-inventory ratio.

[0048] The dynamic benchmark calculation module monitors the data update time interval between real-time consumption and current physical inventory. When this data update time interval exceeds the set heartbeat cycle, the time-series decay compensation module in the dynamic benchmark calculation module extracts the last retained real inventory ratio and matches it with a preset time-series decay algorithm operator. The time-series decay algorithm operator multiplies the last retained real inventory ratio with the set time-series decay coefficient to output a composite consumption ratio. This composite consumption ratio is then used to replace the real-time consumption inventory ratio in subsequent calculations. This module calculates the evolution slope of the real-time consumption inventory ratio within the sliding time window. When the evolution slope exceeds the set stable benchmark threshold of 0.15 / h, it is mapped to a demand tension coefficient. The system uses this demand tension coefficient as a compensation parameter to apply to the historical benchmark price limit parameter in the benchmark procurement parameter set to calculate the dynamic inquiry upper limit boundary. The inquiry broadcast module uses the directed acyclic graph state flow module to generate an inquiry instruction stream containing the dynamic inquiry upper limit boundary and distributes the inquiry instruction stream to multiple supplier nodes.

[0049] Each supplier node returns a response dataset containing real-time pricing parameters. The game evaluation unit receives the response dataset, and the data intervention feedback module, which connects the dynamic benchmark calculation module and the comprehensive evaluation module, calculates the average pricing deviation rate within the response dataset. The average pricing deviation rate is defined as the ratio of the arithmetic mean of all real-time pricing parameters to the dynamic inquiry upper limit boundary. The data intervention feedback module introduces a damping adjustment factor to offset the calculation deviation; this factor is determined by the following formula: ,in, As the damping adjustment factor, Let be the system's response time constant. This represents the change in the ratio of material inventory to consumption. For time increments, the data intervention feedback module subtracts the average price deviation rate from the demand tension coefficient, outputting the price deviation difference. When the price deviation difference is less than the set warning tolerance of 0.05, the data intervention feedback module generates a weight decay factor and feeds it back to the comprehensive evaluation module. Based on the weight decay factor, the comprehensive evaluation module reduces the weight of the real-time price parameter in the internal decision evaluation matrix according to a set decay step. Simultaneously, the comprehensive evaluation module retrieves historical business records from the database, searches for historical dynamic correction coefficient values ​​whose absolute difference from the current dynamic correction coefficient is less than the set interval tolerance, and determines the historical time period containing this value as the historical equivalent fluctuation period. Furthermore, the comprehensive evaluation module elevates the delivery delay variance parameter of the corresponding supplier node within the historical equivalent fluctuation period to the primary judgment weight, and outputs optimization control instructions based on the reconstructed internal decision evaluation matrix. Comparative experimental data shows that under the condition of nonlinear mutation of input stream data, the control group using the static threshold mode, when the evolution slope exceeds 0.20 / h, suffers from a decrease in response data integrity to 65.0% due to the fixed price limit causing supplier quotations to generally exceed the benchmark upper limit, and the absence of distributed instruction stream feedback. The sample group of this invention, by dynamically adjusting the upper limit boundary of the inquiry, automatically increases the upper limit boundary of the dynamic inquiry by 15.0% when the evolution slope reaches 0.25 / h, and the integrity of supplier response data remains above 98.5%, without communication deadlock. When the adjustment range of the key indicator weight parameter exceeds 0.30, the optimization control command output by the decision evaluation matrix shows a significant convergence trend, and the system's optimization decision time is reduced by 45.2ms compared to before the adjustment. This phenomenon indicates that under extreme conditions, the dynamic correction mechanism effectively mitigates the interference of price fluctuations on the stability of the decision matrix.

[0050] Example 3: In a multi-node procurement decision-making scenario within a medical supplies procurement and pricing management system, the system needs to balance the competitiveness of supplier quotations with the reliability of contract fulfillment in real time to address the challenge of drastic price fluctuations caused by short-term supply and demand imbalances. The demand analysis module extracts the benchmark procurement parameter set for the medical supplies to be procured, including historical benchmark price limits. The dynamic benchmark calculation module collects real-time consumption data of medical supplies through a network interface. With physical inventory data Calculate the real-time consumption-to-stock ratio ,in The system is internally configured with a dynamic benchmark calculation module to perform real-time data monitoring. This module reads the timestamp difference of the data. ,when When the time decay compensation module exceeds 100ms, it calls the last stock ratio of the historical cache. The synthesis consumption ratio is calculated according to the following formula. : in This is the time-series decay coefficient, with a value of 0.05.

[0051] The dynamic benchmark calculation module calculates the slope of the real-time consumption-to-stock ratio over a sliding time window. Calculate the demand tension coefficient Specifically, the demand tension coefficient The quantitative calculation formula is as follows: ,in, The preset demand elasticity gain coefficient is set to 0.85. The stable baseline threshold is 0.15 / h; and when the calculated... When it is less than 0, the system will automatically Setting it to 0 establishes a deterministic quantitative mapping path between the evolution slope and the demand tension coefficient. When the rate exceeds the stability benchmark threshold set at 0.15 / h, the system initiates dynamic price inquiry at the upper limit boundary. The adjustment, The calculation logic is as follows The system will As parameter constraints for optimization control instructions, these are distributed to supplier nodes, and each supplier node returns a response dataset containing real-time quotation parameters. The comprehensive evaluation module receives the response dataset through the data intervention feedback module and calculates the average quote offset rate. ,in ,in, The number of supplier nodes participating in the bidding. For the first Real-time quote parameters from each supplier.

[0052] The data intervention feedback module is based on the price deviation difference. Perform logical judgment, where γ is the scaling factor of the inverse of the price elasticity of the current material. At that time, the system establishes the weight decay factor. , The calculation process is executed by the logic decision unit: initialization Secondly, retrieve historical business records from the database to obtain the historical delivery delay variance for the time period corresponding to the current dynamic correction coefficient. ,like , The attenuation is set to 0.5 to make the dimensionless average quote offset rate... Demand tension coefficient reflecting external consumption trends To ensure comparability within the same dimension under extremely concentrated conditions, the data intervention feedback module introduces a benchmark transformation mapping rule before performing subtraction: when the quotations from multiple supplier nodes are highly concentrated and converge to the dynamic inquiry upper limit boundary. And led to When the value approaches 0, the system automatically adds the demand tension coefficient used in the difference calculation. Scale by multiplying by the inverse of the current material's price elasticity, then normalize and decay to the same level. Within the same surface fluctuation range, this ensures that even under high-stress conditions with extremely tight external demand, the price deviation difference remains constant. It can still accurately detect collaborative fraud risks with abnormally consistent data from multiple parties through dimensionality reduction, triggering the weight decay factor. Once operational, the comprehensive evaluation module will provide real-time quote parameters. Initial weights in the internal decision evaluation matrix Adjusted to The system executes an optimization control strategy, when the delivery delay variance of a certain supplier node... The failure tolerance exceeds the system's set failure tolerance of 0.20 and the corresponding If the optimization threshold is still not reached after weight adjustment, the system sets its status to unavailable through a logic switch. Finally, the optimization control command outputs the optimal matching node sequence, guiding the material procurement and supply decision to converge on the supply chain path with better performance capability. This decision-making process maps external data fluctuations to the dynamic adjustment of internal weight parameters, ensuring supply continuity while simultaneously filtering out price deviations and performance risks.

[0053] Example 4: In a scenario involving the procurement and price inquiry of high-frequency emergency medical supplies, when the system faces a sudden supply disruption signal, the demand analysis module reads the physical inventory change trends of each warehouse node. To effectively avoid material shortages caused by single-point supply chain disruptions, the system constructs a risk warning indicator based on real-time supply status, defined as the Material Supply Stability Index. This indicator is calculated by aggregating the real-time response rate and historical delivery compliance rate of global suppliers. To eliminate biases caused by different units of measurement in various feature dimensions, the system performs unit-free processing on the raw data. The system calculates the material supply stability index. The specific path is as follows: Collect the total number of responses from all candidate suppliers within the preset time period. Total number of valid responses And calculate the response rate. Secondly, call the delivery delay variance parameter. The exponent is obtained using weighted linear combination logic. According to the formula definition: in The contribution weight is set to 0.60. This is the variance penalty weight, with a value of 0.40.

[0054] During the process of the system executing the procurement optimization control command, when When the risk trigger threshold is set below 0.45, the system automatically activates the defense strategy for the procurement optimization link. Specifically, the defense strategy involves the system forcibly expanding the instruction distribution scope of the inquiry broadcast module to the reserved secondary supplier candidate pool and automatically removing suppliers from the decision evaluation matrix. All supplier nodes with an index below 0.30, when When the system recovers to a value above 0.60, it triggers a reset procedure for the instruction queue, reverting the optimization strategy to the normal optimal configuration. When processing material-related information, the system consistently implements a privacy protection engineering strategy. Before collecting real-time consumption data and current physical inventory data at each storage node, the system anonymizes the physical location identifier and medical supply number using a hash mapping function. This processing logic is completed locally on the terminal server, ensuring that the anonymized feature data cannot be reverse-analyzed to reveal the specific physical storage location or sensitive material structure information of a particular medical institution before being uploaded to the centralized price inquiry management system. This mechanism ensures the security of procurement data from the underlying architecture without affecting the convergence accuracy of the procurement decision matrix.

[0055] Example 5: During the deployment of the medical supplies procurement and pricing management system, to ensure the system's ability to respond to sudden supply disruptions and price fluctuations, standardized offline parameter calibration and benchmark construction procedures must be performed before formal operation. The technical team retrieves historical turnover cycle data from various levels of material nodes to construct a physical layer logistics response benchmark library. The system performs offline stress tests, inputting continuous simulated commands of surging material demand into the optimization decision engine to monitor the system's response to different response delays. The convergence of multi-objective matching under certain conditions is determined by the system's response delay. The fluctuation range is defined, and a defensive boundary is set for the optimization algorithm. When the system detects the network response latency of the logistics node, a defensive boundary is established. When the time exceeds 150ms, the system activates the degradation adjustment logic of the optimization strategy, switching the original optimization model based on real-time price competition to a static ranking model based on delivery reliability priority. This adjustment logic is determined by the threshold decision function obtained from the offline calibration below: ,in, For safety stock adjustment, This is the weighting coefficient for material turnover. This represents the real-time inventory level of the current physical node. This represents the average daily rate of material consumption. This is the logistics delay sensitivity constant, with a value of 0.25; The maximum tolerable delay allowed for optimal decision-making.

[0056] When the calculation results When the risk level is below the preset stockout risk threshold, the system automatically resets the weight allocation strategy in the comprehensive evaluation module. The data intervention feedback module feeds back the delivery risk index of this node to the inquiry broadcast module, causing the inquiry instruction to be distributed to the backup supplier node with higher logistics response bandwidth. This procedure couples and maps the response characteristics of the physical layer logistics link with the upper-layer procurement decision logic, ensuring that the generation and execution process of optimization control instructions remains closed-loop and stable under complex network disturbances, and avoiding the risk of data deadlock and procurement failure caused by relying solely on price indicators for decision-making under extreme logistics conditions.

[0057] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.

Claims

1. A medical supplies procurement and pricing management system, characterized in that, include: The demand analysis module is used to extract the benchmark demand parameter set for medical supplies to be procured. The benchmark demand parameter set includes the benchmark inquiry upper limit. The dynamic benchmark calculation module is used to collect the real-time consumption and current physical inventory of medical supplies to be purchased within a set area, calculate the real-time consumption-to-inventory ratio, and calculate the dynamic correction coefficient based on the rate of change of the real-time consumption-to-inventory ratio within a set time window. When the dynamic correction coefficient is greater than the set stability threshold, the dynamic correction coefficient is used to calculate the multiplier compensation for the benchmark inquiry upper limit and output the dynamic inquiry upper limit boundary. The inquiry broadcast module is used to generate an inquiry instruction stream that includes a dynamic upper limit boundary for inquiry, and to distribute the inquiry instruction stream to multiple registered supplier nodes; The comprehensive evaluation module is used to receive response datasets containing real-time quotation parameters returned by multiple supplier nodes; The comprehensive evaluation module reduces the weight of real-time quotation parameters in the internal decision evaluation matrix based on the weight decay factor, and simultaneously raises the delivery delay variance parameter of the corresponding supplier node in the historical equivalent fluctuation period to the primary judgment weight, and outputs optimization control instructions based on the internal decision evaluation matrix. The data intervention feedback module, which connects the dynamic benchmark calculation module and the comprehensive evaluation module, is used to calculate the average price deviation rate within the response dataset and the price deviation difference between the average price deviation rate and the dynamic correction coefficient. A weight decay factor is generated when the price deviation difference is less than the set warning tolerance.

2. The medical supplies procurement and pricing management system according to claim 1, characterized in that, The data intervention feedback module extracts all real-time quotation parameters from the response dataset received by the comprehensive evaluation module, calculates the arithmetic mean of all real-time quotation parameters, and calculates the ratio of the arithmetic mean to the dynamic inquiry upper limit boundary output by the dynamic benchmark calculation module, and outputs the average quotation offset rate. The data intervention feedback module subtracts the average price deviation rate (as a dimensionless coefficient) from the dynamic correction coefficient (also a dimensionless coefficient) and outputs the price deviation difference.

3. The medical supplies procurement and pricing management system according to claim 1, characterized in that, The dynamic benchmark calculation module includes a time-series decay compensation module. The dynamic benchmark calculation module monitors the data update time interval between real-time consumption and current physical inventory. When the data update time interval is greater than the set heartbeat cycle, the time-series decay compensation module extracts the last retained real inventory ratio and matches it with the set time-series decay algorithm operator. The time-series decay algorithm operator multiplies the last retained real inventory ratio with the set time-series decay coefficient to output the synthetic consumption ratio. The synthetic consumption ratio is then used to replace the real-time consumption inventory ratio to calculate the dynamic correction coefficient.

4. The medical supplies procurement and pricing management system according to claim 1, characterized in that, The data intervention feedback module retrieves historical business records from the database, searches for historical dynamic correction coefficient values ​​whose absolute difference from the currently calculated dynamic correction coefficient is less than the set interval tolerance, and determines the historical time period in which the historical dynamic correction coefficient value is located as the historical equivalent fluctuation time period.

5. A medical supplies procurement and pricing management system according to claim 1, characterized in that, The comprehensive evaluation module assigns a first weight to the real-time quotation parameter and a second weight to the delivery delay variance parameter, with the first weight being greater than the second weight. When the data intervention feedback module generates a weight decay factor, the comprehensive evaluation module uses the weight decay factor to decay the first weight and simultaneously increases the second weight to the primary judgment level, thus reconstructing the operator distribution in the internal decision evaluation matrix.

6. The medical supplies procurement and pricing management system according to claim 1, characterized in that, The system also includes an edge degradation control module, which is connected to the comprehensive evaluation module. The edge degradation control module monitors the system's computing resource utilization and communication network bandwidth load. When at least one of the following conditions is met—the system's computing resource utilization exceeding a set computing threshold or the communication network bandwidth load exceeding a set communication threshold—the edge degradation control module issues a weight convergence command to the comprehensive evaluation module. In response to the weight convergence command, the comprehensive evaluation module shuts down the multi-dimensional high-order tensor computation flow in its internal decision evaluation matrix, reducing the multi-index evaluation to a single-index veto ranking based on the delivery delay variance parameter.

7. The medical supplies procurement and pricing management system according to claim 1, characterized in that, Both the dynamic benchmark calculation module and the demand analysis module are connected to the regional public health supplies consumption monitoring network through an internal high-concurrency streaming data interface. They collect the consumption rate of the medical supplies to be purchased at each demand node and the current physical inventory at a streaming data reading cycle of less than or equal to 10ms, providing a data chain extraction source for real-time consumption and current physical inventory.

8. The medical supplies procurement and pricing management system according to claim 1, characterized in that, The inquiry broadcast module includes a directed acyclic graph state transition module; the directed acyclic graph state transition module reconstructs the inquiry process topology based on the dynamic inquiry upper limit boundary, and sends the generated inquiry instruction stream containing the dynamic inquiry upper limit boundary to multiple supplier nodes according to the predetermined concurrency control path.

9. A medical supplies procurement and pricing management system according to claim 5, characterized in that, The comprehensive evaluation module uses a weight decay factor to decay the first weight, and the rule is as follows: multiply the first weight by the weight decay factor to obtain the decayed bid weight value. The weight decay factor is a value between 0 and 1, and its magnitude decreases linearly as the price offset difference decreases.

10. A medical supplies procurement and pricing management system according to claim 1, characterized in that, The system also includes a human-machine collaborative interaction interface module, which is connected to the output end of the comprehensive evaluation module. The human-machine collaborative interaction interface module receives the optimization control instructions output by the comprehensive evaluation module, converts the optimal matching node sequence corresponding to the optimization control instructions and the trigger status of the data intervention feedback module into graphical signals and outputs them for display.

Images