An intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities
By combining modules such as IoT sensing, predictive pricing, supply and demand matching, coordinated management and control, and health assessment, the problem of insufficient utilization of idle resources in university asset allocation has been solved, and efficient sharing and healthy management of assets have been achieved.
Patent Information
- Application Number
- CN202611303700.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-26
- Publication Date
- 2026-11-17
AI Technical Summary
Existing university asset allocation schemes are unable to accurately perceive the actual idle status and time fragmentation of equipment, resulting in a disconnect between supply and demand, making it difficult to effectively revitalize idle resources.
The system employs an IoT sensing module to collect asset operation data and generate asset digital feature sets; a predictive pricing module to calculate floating price coefficients; a supply and demand matching module to perform multi-objective optimization matching; a linkage control module to monitor actual load and operation trajectory; a health assessment module to assess asset loss; and a dynamic settlement module to perform credit scoring and settlement.
It enables precise revitalization of idle assets in universities, improves equipment turnover rate, reduces maintenance interference, ensures healthy asset operation, and drives reliable liquidation through multimodal data.
Smart Images

Figure CN123117641A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of smart campus information technology, and in particular to an intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities. Background Technology
[0002] With the expansion of the scale of higher education institutions and the increase in scientific research investment, various departments have purchased a large number of high-value teaching and experimental equipment and scientific research assets. After completing the specific course scheduling tasks of their respective majors, these assets are usually idle for a long time, which objectively provides the material basis and potential demand for opening and sharing with other departments or scientific research teams within the university.
[0003] Existing asset allocation schemes mostly adopt a crude reservation method of manual registration and fixed rates, which makes it difficult to accurately perceive the actual idle status and time fragmentation of equipment. Furthermore, they lack the means to dynamically adjust prices according to the probability of idleness, resulting in a disconnect between supply and demand information and low matching efficiency, making it difficult to truly revitalize idle resources. Summary of the Invention
[0004] To overcome the above shortcomings, this invention provides an intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities. It aims to improve the existing asset scheduling schemes, which mostly adopt a crude reservation method of manual registration and fixed rate, making it difficult to accurately perceive the actual idle status of equipment and time fragmentation.
[0005] This invention provides the following technical solution: an intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities, comprising:
[0006] The IoT sensing module collects asset operation data and generates a set of digital features for assets, including asset number, idle time window, equipment specifications and historical usage data.
[0007] The predictive pricing module calculates the idle probability based on the historical usage data, calculates the floating price coefficient in combination with the initial or received feedback parameters, and constructs a shared resource pool containing the idle time window, idle probability and floating price coefficient.
[0008] The supply and demand matching module extracts the demand characteristics of the usage request, removes the locked time windows and occupied time windows in the shared resource pool, matches the demand characteristics with the remaining idle time windows, outputs a scheduling instruction containing the asset number, reservation time period and transaction fee rate, and marks the reservation time period as the occupied time window;
[0009] The linkage control module issues an unlocking signal based on the scheduling instruction, monitors the actual load and operation trajectory, and generates a performance log containing the actual load, the operation trajectory, resource consumption data, and the transaction fee rate.
[0010] The health assessment module assesses the asset wear and tear based on the actual load. When the asset wear and tear reaches a preset threshold, it writes a locking instruction back to the shared resource pool to generate the locking time window and outputs the wear and tear coefficient.
[0011] The dynamic settlement module generates a settlement result based on the performance log and the loss coefficient, updates the credit score based on the operation trajectory, and encapsulates the settlement result, the credit score, the loss coefficient, and the resource consumption data into feedback parameters and sends them to the predictive pricing module.
[0012] Preferably, the IoT sensing module specifically includes the following steps:
[0013] The equipment specifications of the asset and the historical current waveform data collected by the sensing devices are used as the historical usage data;
[0014] Input the historical current waveform data into the load identification model to extract the standby period and operating cycle;
[0015] The difference set between the standby time period and the university class schedule data is calculated, and the unoccupied time period is extracted as the idle time window;
[0016] The idle time window, the equipment specifications, the historical usage data, and the operating cycle are integrated, and the corresponding asset number is configured to generate the asset digital feature set.
[0017] Preferably, the predictive pricing module specifically includes the following steps:
[0018] Based on the historical usage data, the idle frequency is extracted, and a Markov transition matrix is constructed to calculate the idle probability;
[0019] Obtain the time decay difference between the current time and the start time of the idle time window;
[0020] Substitute the difference between the idle probability and the time decay into the decay function to calculate the floating price coefficient;
[0021] The idle probability, the idle time window, and the floating price coefficient are bound to the asset digital feature set to construct a shared resource pool.
[0022] Preferably, the calculation of the floating price coefficient specifically includes the following steps:
[0023] The resource consumption data in the feedback parameters is converted into energy consumption cost per unit time, and summed with the depreciation amount per unit time corresponding to the loss coefficient in the feedback parameters to obtain the cost benchmark value.
[0024] Substitute the difference between the idle probability and the time decay into the decay function to calculate the floating price coefficient;
[0025] When the rate corresponding to the floating price coefficient is lower than the cost benchmark value, the rate will be clamped to the cost benchmark value;
[0026] When the current time reaches the start time and the idle time window is not yet marked as occupied, the idle time window is marked as a zero-rate period and pushed to the user's terminal.
[0027] Preferably, the supply and demand matching module specifically includes the following steps:
[0028] Extract the expected duration and rate cap from the aforementioned demand characteristics;
[0029] Remove the locked time window and the occupied time window from the shared resource pool to obtain the remaining idle time window;
[0030] The desired duration and the remaining idle time window are input into a multi-objective optimization model for solution.
[0031] Solutions with required rates higher than the upper limit of the rate are eliminated, and Pareto non-dominated solution set is selected;
[0032] Extract the asset number and reservation time period corresponding to the target solution from the Pareto non-dominated solution set, generate the transaction fee rate by combining it with the floating price coefficient, and output the scheduling instruction.
[0033] Preferably, the linkage control module specifically includes the following steps:
[0034] The vibration sequence of the asset during the scheduled time period is collected as the operation trajectory, and the power sequence is collected as the actual load.
[0035] Integrating the power sequence yields the power consumption value, which serves as resource consumption data.
[0036] The feature vector of the vibration sequence is time-aligned with the envelope of the power sequence.
[0037] The verified operation trajectory, actual load, resource consumption data, and transaction fee rate are hashed and encapsulated into a performance log.
[0038] Preferably, the health assessment module specifically includes the following steps:
[0039] The actual load is sliced into load amplitudes over multiple cycles;
[0040] Based on the preset load-lifetime loss function, calculate the step loss increment corresponding to each load amplitude;
[0041] The asset loss amount is obtained by integrating and summing the incremental losses of each step.
[0042] When the asset loss reaches the preset threshold, a locking instruction is generated and written back to the shared resource pool to generate the locking time window.
[0043] The ratio of the incremental step loss generated in the current period to the baseline loss value is extracted as the loss coefficient.
[0044] Preferably, generating the locking time window specifically includes the following steps:
[0045] The stepped loss increment is decomposed and integrated according to frequency band to obtain the component loss amount of each component, and the component loss amounts are weighted and synthesized into the asset loss amount;
[0046] Calculate the loss rate of each of the aforementioned sub-items of loss, extract the components whose loss rate reaches the preset loss line as components to be maintained, and map the maintenance duration;
[0047] The period of time when the idle probability is higher than the preset probability line is extracted from the shared resource pool as the maintenance window. The maintenance duration and the maintenance window are encapsulated into the locking instruction and written back.
[0048] Preferably, the dynamic settlement module specifically includes the following steps:
[0049] The settlement result is calculated based on the transaction fee rate, the resource consumption data, and the loss coefficient.
[0050] The operation trajectory is matched with a standard trajectory library using an algorithm to calculate the normalized distance;
[0051] The normalized distance is converted into a dimensionless normalized index according to a preset mapping relationship;
[0052] The updated credit score is obtained by weighting the aforementioned standardized indicators with the user's historical credit score.
[0053] The settlement result, the credit score, the loss coefficient, and the resource consumption data are encapsulated into the feedback parameters and sent to the predictive pricing module.
[0054] The feedback parameter is used to instruct the predictive pricing module to add an upward adjustment coefficient when calculating the floating price coefficient for the next scheduling for users whose credit scores are lower than a preset credit threshold.
[0055] The present invention has the following beneficial effects:
[0056] 1. In this invention, idle time windows are accurately extracted by collecting underlying asset operation data. Floating prices are dynamically generated by combining historical idle status with the time decay difference. A multi-objective optimization algorithm is used to effectively reduce time fragmentation rate while ensuring user budget constraints, thus effectively revitalizing idle assets in universities and improving the overall equipment turnover rate.
[0057] 2. In this invention, the physical step loss of the equipment is accurately calculated based on the actual instantaneous power sequence collected. When the component loss reaches the threshold, the period with a high probability of being idle is automatically retrieved as the maintenance window, thereby minimizing the interference of maintenance on normal teaching scheduling and shared use while ensuring the healthy operation of the assets.
[0058] 3. In this invention, the normalization distance between the operation trajectory and the standard trajectory is extracted to quantify the user's behavior standardization. The behavior evaluation is converted into a credit score and directly linked to the billing rate of subsequent scheduling. With the help of hash anti-counterfeiting verification of multimodal physical data, reliable settlement driven by actual consumption and operation behavior is realized. Attached Figure Description
[0059] Figure 1 This is an architecture diagram of an intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities, as proposed in this invention.
[0060] Figure 2 This is a schematic diagram of the dynamic pricing and resource pool construction of the predictive pricing module of the present invention;
[0061] Figure 3 This is a schematic diagram of the multi-objective optimization scheduling of the supply and demand matching module of the present invention. Detailed Implementation
[0062] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0063] This invention provides an intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities, such as... Figures 1-3 As shown, it includes:
[0064] The IoT sensing module collects asset operation data and generates a set of digital features for assets, including asset number, idle time window, equipment specifications and historical usage data.
[0065] Furthermore, the IoT sensing module specifically includes the following steps:
[0066] Obtain the equipment specifications of the asset, as well as the historical current waveform data collected by the sensing equipment, as historical usage data;
[0067] Input historical current waveform data into the load identification model to extract standby periods and operating cycles;
[0068] The difference between the standby time period and the university's class schedule data is calculated, and the unoccupied time period is extracted as the idle time window;
[0069] By integrating idle time windows, equipment specifications, historical usage data, and operating cycles, and configuring corresponding asset numbers, an asset digital feature set is generated.
[0070] Specifically, the IoT sensing module first reads the asset number from the university's asset management ledger and establishes a one-to-one correspondence between the asset and the equipment using RFID tags or QR codes. Simultaneously, it acquires equipment specification information, including rated power, rated voltage, equipment type, and installation location. The current sensing device operates according to a sampling period. Historical current waveform data is collected, and the sampling period is... Desirable Sliding window length Desirable By sampling points, the historical current sequence is obtained:
[0071] ;
[0072] in, Represents historical current waveform data. These represent the current sample values collected in chronological order. This indicates the total number of sampling points. Asset number, equipment specification information, and historical current sequence together serve as the foundational data for subsequent feature fusion.
[0073] After inputting historical current waveform data into the load identification model, characteristic quantities are first calculated for the current sequence within each sliding window. The root mean square current, standard deviation of ripple, and crazing factor are expressed as follows:
[0074] ;
[0075] ;
[0076] ;
[0077] in, Indicates the first The root mean square current within a sliding window This indicates the number of sampling points within the sliding window. Represents historical current sequence In the The corresponding sliding window within the first Each sample value, Indicates the first Standard deviation of current fluctuation within a sliding window Indicates the first The average current within each sliding window Indicates the first Peak factor within a sliding window, Indicates the first Peak current within a sliding window.
[0078] To ensure the uniformity of dimensions among the input terms in the classification function, the root mean square current and the standard deviation of fluctuation are standardized to obtain:
[0079] ;
[0080] ;
[0081] in, and Let represent the mean and standard deviation of the root mean square current training sample set, respectively. and Let represent the mean and standard deviation of the training sample set, respectively. , All values are non-zero. The standardized features and peak factors are input into the classification discriminant function to obtain the... The probability of running each sliding window:
[0082] ;
[0083] in, Indicates the first The probability that a sliding window is in a running state. , , and These are the model parameters obtained from training labeled standby samples and running samples. If the following conditions are met:
[0084] ;
[0085] Then determine the first Each sliding window is in running state; if the following conditions are met:
[0086] ;
[0087] Then determine the first Each sliding window is in standby mode. The threshold value is... Desirable The device will be continuously classified as being in standby mode for a duration of not less than [amount missing]. The interval is extracted as the standby time period, the Desirable The operating cycle is obtained by merging consecutive intervals that are continuously identified as being in operation. To avoid misjudgments caused by short-term fluctuations, time intervals shorter than [a certain value] are used. The adjacent standby segments are merged, the Desirable .
[0088] The extracted standby time periods are mapped to university class schedule data at a unified time granularity, and then the difference is calculated. The class schedule data is represented as follows:
[0089] ;
[0090] Standby time is represented as:
[0091] ;
[0092] The idle time window is represented as:
[0093] ;
[0094] in, This represents the set of class scheduling periods corresponding to the university's class schedule data. Indicates the first The start and end times of each scheduled class period. Indicates the number of class slots. Indicates a set of standby periods. Indicates the first The start and end times of each standby period. Indicates the number of standby periods. This represents a set consisting of multiple unoccupied time windows. Through set difference operations, intervals within the standby period that do not overlap with any scheduled time period are retained as unoccupied periods and used as shareable idle time windows.
[0095] By integrating idle time windows, equipment specification information, historical current waveform data, and operating cycles, a set of digital asset features is formed.
[0096] ;
[0097] in, Represents the set of digital features of assets. Indicates the asset number. This indicates equipment specifications. Represents historical current waveform data. Indicates the operating cycle. This represents the set of idle time windows. Ultimately, it will... It is bound to the corresponding asset digital feature set and written to the asset shared database for subsequent use by the prediction pricing module and the supply and demand matching module.
[0098] Through the above processing, it is possible to identify equipment standby periods and idle time windows from the actual operation data of assets, and associate them with asset number, equipment specifications and historical usage data to form the basic features required for subsequent pricing and scheduling.
[0099] The predictive pricing module calculates the idle probability based on historical usage data, and calculates the floating price coefficient by combining the initial or received feedback parameters, and constructs a shared resource pool containing idle time windows, idle probability and floating price coefficient;
[0100] Furthermore, the predictive pricing module specifically includes the following steps:
[0101] Based on historical usage data, extract the idle frequency and construct a Markov transition matrix to calculate the idle probability;
[0102] Get the time decay difference between the current time and the start time of the idle time window;
[0103] Substitute the difference between the idle probability and the time decay into the decay function to calculate the floating price coefficient;
[0104] By binding the idle probability, idle time window, and floating price coefficient to the asset digital feature set, a shared resource pool is constructed.
[0105] Furthermore, calculating the floating price coefficient specifically includes the following steps:
[0106] The resource consumption data in the feedback parameters is converted into energy consumption cost per unit time, and summed with the depreciation amount per unit time corresponding to the loss coefficient in the feedback parameters to obtain the cost benchmark value.
[0107] Substitute the difference between the idle probability and the time decay into the decay function to calculate the floating price coefficient;
[0108] When the rate corresponding to the floating price coefficient is lower than the cost benchmark value, the rate will be clamped to the cost benchmark value;
[0109] When the current time reaches the start time and the idle time window is not yet marked as occupied, the idle time window will be marked as a zero-rate period and pushed to the user's client.
[0110] Specifically, the predictive pricing module takes the asset digital feature set formed in the previous step as input, reads the asset number, historical current waveform data, operating cycle, and idle time window, and uses the corresponding historical time slices in the historical current waveform data as the asset's usage status sequence. Each time slice is marked as either idle or occupied, with idle state denoted as 1 and occupied state as 0, resulting in the following status sequence:
[0111] ;
[0112] in, This represents the total number of time slices divided according to a uniform time granularity within a historical statistical period. Indicates the first One time slice is idle. Indicates the first Each time slice is in an occupied state. The idle frequency is extracted based on this state sequence and denoted as:
[0113] ;
[0114] in, This indicates the frequency of the asset's idle time during the historical statistical period.
[0115] Furthermore, a Markov transition matrix is constructed based on the state transitions of adjacent time slices. Let the idle state be denoted as I and the occupied state as B, then the state transition matrix is expressed as:
[0116] ;
[0117] in, Indicates from state Transition to state The probability of , and:
[0118] ;
[0119] in, Represents the sequence of historical states from state Transition to state Number of times, Represents the smoothing term, which can be taken as... Based on the transition matrix, the probability that the asset will be idle in the next time window is calculated and denoted as:
[0120] ;
[0121] in, Represents the state vector The component corresponding to idle state I in the prediction is the probability of being in an idle state at the start of the idle time window. The fusion coefficient between historical frequency and transfer prediction can be taken as... ,and When the current state is idle, When the current state is occupied, ,but:
[0122] ;
[0123] in, This indicates the number of time slices between the current time and the start time of the idle time window.
[0124] Get current time With the start time of the idle time window The time decay difference is defined as:
[0125] ;
[0126] in, A timestamp in seconds representing the current time. A timestamp in seconds representing the start time of the idle time window. The unit is hours. To make the pricing change as time approaches, the difference between the idle probability and the time decay is substituted into the decay function to obtain the floating price coefficient:
[0127] ;
[0128] in, Indicates the floating price coefficient. Indicates the minimum fluctuation coefficient. Indicates the maximum fluctuation coefficient. This represents the time decay factor, expressed in hours. Desirable , Desirable , Desirable This floating price coefficient is used to reflect the impact of the probability of asset idling and the proximity of time on the shared price.
[0129] When the module receives feedback parameters from the dynamic settlement module, if no feedback parameters are received on the first run, the initial feedback parameters are used for calculation. In the initial feedback parameters, resource consumption data can be estimated based on the equipment's rated power and the scheduled duration, and the loss coefficient can be taken as... Let the resource consumption data be... The loss coefficient is The resource consumption data is converted into energy consumption cost per unit time, and then summed with the depreciation cost per unit time corresponding to the loss coefficient to obtain the cost baseline value:
[0130] ;
[0131] in, Indicates the cost baseline value. This represents the unit electricity price, which can be taken as... Yuan per kilowatt-hour, This represents the base value of the asset used to calculate depreciation expense within the current billing period. This indicates the reservation duration, in hours. The floating price coefficient is converted to the corresponding rate, denoted as:
[0132] ;
[0133] in, This indicates a pre-configured reference rate based on equipment specifications, which can be taken as... Yuan per hour, This indicates the converted floating rate. If Then Clamp to This is to ensure that the sharing fee rate is not lower than the cost benchmark value.
[0134] After completing the above calculations, the idle probability will be... Idle Time Window With floating price coefficient Binding to the asset's digital feature set to form a shared resource pool record:
[0135] ;
[0136] in, Represents shared resource pool records, Indicates the asset number. This represents a set of idle time windows consisting of multiple unoccupied time windows. For usage requests that are allowed for public sharing within the school, if the current time has reached the start time of an idle time window and the time window has not yet been marked as occupied, the remaining time period of the idle time window will be marked as a zero-rate period and pushed to the user's client to activate idle assets.
[0137] Through the above processing, historical idle patterns, real-time time decay, and cost feedback can be linked to form a shared resource pool with dynamically updated prices, providing a flexible and cost-effective pricing basis for subsequent supply and demand matching.
[0138] The supply and demand matching module extracts the demand characteristics of usage requests, removes locked time windows and occupied time windows from the shared resource pool, matches the demand characteristics with the remaining idle time windows, outputs a scheduling instruction containing asset number, reservation time period and transaction fee rate, and marks the reservation time period as occupied time window;
[0139] Furthermore, the supply and demand matching module specifically includes the following steps:
[0140] Extract the expected duration and maximum rate from the demand characteristics;
[0141] Remove locked and occupied time windows from the shared resource pool to obtain the remaining free time windows;
[0142] Input the expected duration and the remaining idle time window into the multi-objective optimization model for solution;
[0143] Solutions with required rates higher than the rate cap are eliminated, and Pareto non-dominated solution set is selected;
[0144] Extract the asset number and reservation time slot corresponding to the target solution from the Pareto non-dominated solution set, generate the transaction fee rate by combining it with the floating price coefficient, and output the scheduling instruction.
[0145] Specifically, the supply and demand matching module receives usage requests from users and extracts the demand characteristics of the requests, including the expected duration. With rate cap , of which the expected duration The unit is hourly, and the rate cap is [not specified]. The unit is yuan per hour. Simultaneously, it reads the shared resource pool built by the pricing prediction module in the previous step. Obtain the asset number. Collection of idle time windows and the corresponding floating rates Obtain the set of locked time windows for the health assessment module's write-back. and the set of occupied time windows generated by historical scheduling Perform a set elimination operation on the set of idle time windows to obtain the set of remaining idle time windows:
[0146] ;
[0147] in, This represents the set of remaining free time windows that are actually available after the elimination operation. Iterate through... Within each time window, select those with a duration greater than or equal to the expected duration. The time window is used as a candidate window to form a candidate solution set. For each candidate solution, by default, the starting time of the candidate window is taken as the starting point, and a continuous interval equal to the expected duration is extracted as the reservation time period corresponding to that candidate solution.
[0148] The candidate solution set, consisting of the expected duration and the remaining idle time window, is input into a multi-objective optimization model for solving. Two minimization objective functions are established for the multi-objective optimization model: the first objective function minimizes the usage cost, and the second objective function minimizes the time fragmentation rate. For the candidate solutions... Its objective function is expressed as:
[0149] ;
[0150] ;
[0151] in, Indicates candidate solutions Total booking fee, Indicates candidate solutions The corresponding floating rate, Indicates candidate solutions The resulting time fragmentation rate Indicates candidate solutions The total duration of the remaining idle time window, in hours. By minimizing... The system prioritizes matching the idle time window that is closest to the expected duration, thereby reducing the fragmented waste caused by the segmentation of asset time.
[0152] During the model solution process, a hard rate constraint filter is applied to the candidate solution set to eliminate floating rates. Greater than the rate cap Invalid solutions. For the remaining valid solution set after elimination, a Pareto non-dominated solution set is selected based on Pareto dominance relations. If no solution exists in the valid solution set... Make and If at least one condition is strictly less than the given condition, then the solution is determined. A solution is a Pareto nondominated solution, and all nondominated solutions together constitute the Pareto nondominated solution set.
[0153] To extract a unique objective solution from the Pareto nondominated solution set, the objective function values in the nondominated solution set are subjected to extremum normalization to eliminate dimensional differences, and the weighted Euclidean distance from each nondominated solution to the ideal point is calculated. The distance metric formula is as follows:
[0154] ;
[0155] in, Indicates a non-inferior solution Target distance, and Let represent the minimum and maximum costs of the non-dominated solution set, respectively. and These represent the minimum and maximum values of the time fragmentation rate, respectively. This represents a very small normal quantity used to prevent the denominator from being zero, and can be taken as... , and These represent the normalized weighted coefficients for cost and time fragmentation rate, respectively, and their sum is one. In one feasible implementation of this solution, to prioritize cost-effectiveness for users, 0.7 is acceptable. 0.3 is acceptable.
[0156] Select target distance The smallest non-dominated solution is taken as the final objective solution. The corresponding asset number is then extracted from this objective solution. The exact reservation time slot is matched, and a transaction fee rate is generated by combining this with the floating fee rate corresponding to the target solution. All of this information is then output as a scheduling instruction. Finally, the reservation time slot corresponding to the target solution is incorporated into the set of occupied time windows. In the middle, mark the occupied time slot after the reservation is completed.
[0157] Through the above processing, it is possible to balance the transaction costs and time resource continuity of asset sharing while ensuring budget constraints, and achieve optimal matching and conflict-avoiding scheduling among multiple assets.
[0158] The linkage control module issues unlocking signals based on scheduling instructions, monitors actual load and operation trajectory, and generates a performance log containing actual load, operation trajectory, resource consumption data and transaction fee rate;
[0159] Furthermore, the coordinated control module specifically includes the following steps:
[0160] Vibration sequences of assets during the scheduled time period are collected as operation trajectories, and power sequences are collected as actual loads.
[0161] The power consumption value is obtained by integrating the power sequence, and is used as resource consumption data.
[0162] The eigenvectors of the vibration sequence are time-aligned with the envelope of the power sequence.
[0163] The verified operation trajectory, actual load, resource consumption data, and transaction fee rate are hashed and encapsulated into a performance log.
[0164] Specifically, the linkage control module receives scheduling instructions from the supply and demand matching module, extracting the asset number, reservation time period, and transaction fee rate. When the current system time reaches the start time of the reservation time period, the module sends an unlock signal to the underlying IoT controller of the corresponding asset number according to the scheduling instruction, granting access to the asset. During the reservation time period, the module synchronously collects the physical and electrical operating status of the equipment through sensors deployed on the asset itself, collecting vibration sequences as operating trajectories at set sampling time intervals, and collecting instantaneous power sequences as actual loads. Let the sampling time interval be... Take 0.5 seconds; the total number of sampling points within the scheduled time period is: The resulting power sequence is represented as:
[0165] ;
[0166] The operation trajectory is represented as follows:
[0167] ;
[0168] in, Represents a power sequence. Indicates the first The instantaneous power value at each sampling point, in watts. Represents a vibration sequence. Indicates the first The combined vibration amplitude of the triaxial acceleration at each sampling point.
[0169] To accurately calculate the energy consumption costs incurred by users during usage, the collected power sequence is subjected to time-discrete integration and summation to calculate the power consumption value, which is then used as resource consumption data. The conversion calculation formula is as follows:
[0170] ;
[0171] in, This represents resource consumption data, in kilowatt-hours, and is a constant. The joule, a unit used to convert power multiplied by a second, into a kilowatt-hour.
[0172] To prevent data falsification or sensor separation during the performance process, the module performs time alignment verification between the feature vector of the vibration sequence and the envelope of the power sequence. The sequence is divided into segments according to a fixed-length time window. The time window can be 5 seconds long. The sum of squares of the vibration amplitudes within each segment is extracted as the short-time energy, forming a feature vector. The average power within each segment is extracted as the envelope value, forming an envelope vector. The Pearson correlation coefficient between the two is then calculated.
[0173] ;
[0174] in, This represents the time alignment correlation coefficient. Indicates the first The short-term energy of vibration within a time segment Indicates the first Power envelope values within a time segment This represents the average short-time energy of vibrations across all time segments. This represents the mean of the power envelope values across all time segments. If the correlation coefficient... If the value is greater than or equal to the preset alignment threshold, it is determined that the physical trajectory and the electrical load are highly synchronized on the time axis, and the verification is passed. The alignment threshold can be 0.65. If it is lower than the alignment threshold, the record is marked as an anti-counterfeiting abnormal state.
[0175] After completing the above verification, the module extracts the operation trajectory. Actual load Resource consumption data and the transaction fee rate determined in the previous step. Concatenate the above fields byte-by-byte in sequence and input them into a secure hash algorithm to calculate the digital digest:
[0176] ;
[0177] in, This represents the generated hash digest value. This indicates a data concatenation operation. Subsequently, the plaintext data before concatenation, the verification status identifier, and the hash digest value are jointly encapsulated to generate a performance log containing actual load, operation trajectory, resource consumption data, and transaction fee rate.
[0178] Through the above processing, accurate energy consumption during asset use was achieved, and the multimodal cross-calculation of physical and electrical data ensured the authenticity and tamper-proof nature of the performance data, providing authentic and reliable evidence for subsequent health assessments and credit settlement.
[0179] The health assessment module assesses asset wear based on actual load. When asset wear reaches a preset threshold, it writes a locking command back to the shared resource pool to generate a locking time window and outputs the wear coefficient.
[0180] Furthermore, the health assessment module specifically includes the following steps:
[0181] The actual load is sliced into load amplitudes over multiple cycles;
[0182] Based on the preset load-lifetime loss function, calculate the step loss increment corresponding to each load amplitude;
[0183] The asset loss amount is obtained by integrating and summing the incremental losses at each stage.
[0184] When the asset wear and tear reaches a preset threshold, a locking instruction is generated and written back to the shared resource pool to generate a locking time window.
[0185] The ratio of the incremental step loss generated in the current period to the baseline loss value is extracted as the loss coefficient.
[0186] Furthermore, generating the locking time window specifically includes the following steps:
[0187] The incremental step loss is decomposed and integrated according to frequency band to obtain the individual loss of each component, and the individual loss is weighted and synthesized into the asset loss.
[0188] Calculate the loss rate of each sub-item of loss, extract the components whose loss rate reaches the preset loss line as the components to be maintained and map the maintenance duration;
[0189] Extract time periods from the shared resource pool where the idle probability is higher than a preset probability line as maintenance windows, encapsulate the maintenance duration and maintenance window into a lock command and write it back.
[0190] Specifically, the health assessment module acquires the actual load generated by the linkage control module, i.e., the instantaneous power sequence. The power sequence was sliced using a rainflow counting algorithm to extract its internal fluctuation characteristics, resulting in a set of load amplitude values for multiple periods. ,in This represents the total number of periods obtained from slicing. Indicates the first The load amplitude for each cycle is expressed in watts. Based on the preset load-lifetime loss function, the step loss increment corresponding to each load amplitude is calculated:
[0191] ;
[0192] in, Indicates the first The step loss increment per cycle, dimensionless. This indicates the rated load reference of the equipment, in watts. This represents the stress index, which can be taken as 2. This represents the basic fatigue constant, which can be taken as 0.00001.
[0193] To accurately pinpoint the specific source of loss, the module extracts the above... The duration of each cycle is determined, and the cycles are divided into different frequency bands based on their duration. Each cycle is then classified and mapped to a specific frequency band. Among the key components, let the one belonging to the first... The set of periodic indices for each component is The increments of each step loss within the set are summed to obtain the first step loss. The incremental loss of each component in the current period Add this increment to the historical beginning-of-period cumulative value to obtain the current breakdown loss of each component. Then, the loss rate of each individual loss item is calculated:
[0194] ;
[0195] in, Indicates the first The loss rate of each component, dimensionless. This represents the preset total increment of the rated lifespan of this component. The loss rates of each component are weighted and summed to synthesize the asset loss of the equipment.
[0196] ;
[0197] in, This indicates the total amount of asset loss. Indicates the first The normalized weighting coefficients for each component are determined. Simultaneously, the sum of the incremental step losses generated in the current period is extracted and converted into an overall loss coefficient.
[0198] ;
[0199] in, This represents the loss coefficient output to the predictive pricing module and the dynamic settlement module. It represents the baseline loss increment rate corresponding to the same duration of operation of an asset under rated stable load, and is dimensionless.
[0200] When the assessed asset depreciation amount Reaching the preset threshold At that time, a locking command is generated and preparation is made to execute a write-back operation, wherein the preset threshold... A value of 0.85 is acceptable. At this point, the system iterates through the loss rates of each component. The extraction loss rate reaches the preset loss line. The component is designated as the component to be maintained, and the preset loss line is... A value of 0.90 is acceptable. Based on the specific type of the component to be maintained, the required maintenance duration is retrieved from the preset component maintenance mapping table. The unit is hours. The module reads from the shared resource pool. Retrieve idle probability from idle time window set Higher than the preset probability line The continuous time period is used as the maintenance window, and the preset probability line A value of 0.80 can be taken. Finally, the mapped maintenance duration will be... The maintenance window matching the search is encapsulated as a locking command, written back to the shared resource pool, and a locking time window is generated in the corresponding time period. .
[0201] Through the above processing, a precise quantitative mapping from the actual operating electrical load to the physical life depreciation of the equipment is achieved. By utilizing the predicted high-probability idle periods, maintenance plans can be seamlessly inserted, ensuring the healthy operation of assets while avoiding conflicts with normal class scheduling or shared use.
[0202] The dynamic settlement module generates settlement results based on performance logs and loss coefficients, updates credit scores based on operation trajectories, and encapsulates settlement results, credit scores, loss coefficients, and resource consumption data into feedback parameters and sends them to the predictive pricing module.
[0203] Furthermore, the dynamic settlement module specifically includes the following steps:
[0204] The settlement results are calculated based on the transaction fee rate, resource consumption data, and loss coefficient.
[0205] The operation trajectory is matched with a standard trajectory library using an algorithm to calculate the normalized distance;
[0206] The normalized distance is converted into a dimensionless normalized index according to a preset mapping relationship;
[0207] The updated credit score is obtained by weighting the standardized indicators with the user's historical credit score.
[0208] The settlement results, credit score, loss coefficient and resource consumption data are encapsulated as feedback parameters and sent to the predictive pricing module.
[0209] The feedback parameter is used to instruct the predictive pricing module to add an upward adjustment coefficient when calculating the floating price coefficient for the next scheduling for users whose credit scores are lower than a preset credit threshold.
[0210] Specifically, after the asset reservation period ends, the dynamic settlement module first parses the performance log generated by the linkage control module, extracting the transaction fee rate, resource consumption data, and operation trajectory. Then, based on the user identifier associated with this scheduling, it queries the database for the user's historical credit score. Simultaneously, it receives the current period loss coefficient output by the health assessment module. Based on the transaction fee rate, resource consumption data, and loss coefficient, the final settlement result is calculated using the following formula:
[0211] ;
[0212] in, This indicates the final settlement fee for a single reservation, in yuan. This indicates the transaction fee rate, in yuan per hour. This indicates the actual reservation duration, in hours. This represents the resource compensation unit price for actual electricity consumption, which can be taken as 0.85 yuan per kilowatt-hour. This represents actual resource consumption data, expressed in kilowatt-hours. This indicates that the base amount for loss compensation, calculated based on the original value of the asset, can be taken as 200 yuan. This represents a dimensionless loss coefficient. This settlement logic combines the basic transaction fee for time consumption with the actual physical power consumption and depreciation compensation, achieving accurate settlement based on actual physical costs.
[0213] To quantify the standardization of user actions, the module performs algorithmic matching between the operation trajectories in the performance log and a pre-defined standard trajectory library. Let the extracted operation trajectory be a vibration sequence. ,in This represents the total number of sampling points for the actual operation trajectory. Represents the first in the actual sequence Each vibration amplitude; retrieve the benchmark vibration sequence of similar assets under standard operation from the standard trajectory library. ,in This represents the total number of sampling points in the reference sequence. Represents the first in the benchmark sequence The vibration amplitude is calculated. The normalized distance between the two sequences is calculated using the dynamic time warping algorithm. The recursive formula for the elements of the cumulative distance matrix is as follows:
[0214] ;
[0215] in, Represents a sequence The former Each sampling point and sequence The former The minimum cumulative distance between each sampling point. The calculated global normalized distance is denoted as... The normalized distance is then converted into a dimensionless normalized index according to a preset exponential mapping relationship:
[0216] ;
[0217] in, This represents an operational standard indicator, with a value range from zero to one. This represents the penalty sensitivity coefficient, whose dimension is set to the reciprocal of the vibration amplitude to ensure that the exponential term is dimensionless; a value of 0.05 is acceptable. A smaller regularization distance indicates more stable and standardized user operation, and the corresponding standardization index approaches one.
[0218] Furthermore, the standardized indicators obtained this time are weighted and calculated with the user's historical credit score to obtain the updated credit score:
[0219] ;
[0220] in, This indicates the updated credit score. This indicates the user's historical credit score. This represents the historical credit weight, which can be set to 0.7. This step converts the performance of this operation into a percentage score and smoothly integrates it into the user's overall credit rating.
[0221] After completing the calculation of all settlement indicators, the module will release the settlement results. Updated credit score Loss coefficient With resource consumption data The parameters are jointly encapsulated as feedback parameters and sent to the predictive pricing module. These feedback parameters instruct the predictive pricing module to implement countermeasures against users with low credit scores. Below the preset credit threshold At that time, the upward adjustment coefficient is calculated:
[0222] ;
[0223] in, This represents the upward adjustment coefficient, which is dimensionless. This indicates a preset credit threshold, which can be set to 75 points. The penalty slope is represented by a unit set to the reciprocal of the fraction to ensure the overall formula is dimensionless; its numerical value can be 0.02. When processing the floating price coefficient for the user's next scheduling, the predictive pricing module multiplies the previously calculated floating price coefficient by the upward adjustment coefficient, thereby achieving targeted rate superposition.
[0224] Through the above processing, the system forms a financial and credit closed loop that dynamically rewards and punishes users based on actual physical consumption and behavioral patterns. This not only scientifically compensates for abnormal asset wear and energy costs, but also effectively constrains users' illegal operations through a dynamic fee rate countermeasure mechanism.
[0225] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. An intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities, characterized in that, include: The IoT sensing module collects asset operation data and generates a set of digital features for assets, including asset number, idle time window, equipment specifications and historical usage data. The predictive pricing module calculates the idle probability based on the historical usage data, calculates the floating price coefficient in combination with the initial or received feedback parameters, and constructs a shared resource pool containing the idle time window, idle probability and floating price coefficient. The supply and demand matching module extracts the demand characteristics of the usage request, removes the locked time windows and occupied time windows in the shared resource pool, matches the demand characteristics with the remaining idle time windows, outputs a scheduling instruction containing the asset number, reservation time period and transaction fee rate, and marks the reservation time period as the occupied time window; The linkage control module issues an unlocking signal based on the scheduling instruction, monitors the actual load and operation trajectory, and generates a performance log containing the actual load, the operation trajectory, resource consumption data, and the transaction fee rate. The health assessment module assesses the asset wear and tear based on the actual load. When the asset wear and tear reaches a preset threshold, it writes a locking instruction back to the shared resource pool to generate the locking time window and outputs the wear and tear coefficient. The dynamic settlement module generates a settlement result based on the performance log and the loss coefficient, updates the credit score based on the operation trajectory, and encapsulates the settlement result, the credit score, the loss coefficient, and the resource consumption data into feedback parameters and sends them to the predictive pricing module.
2. The intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities according to claim 1, characterized in that, The IoT sensing module specifically includes the following steps: The equipment specifications of the asset and the historical current waveform data collected by the sensing devices are used as the historical usage data; Input the historical current waveform data into the load identification model to extract the standby period and operating cycle; The difference set between the standby time period and the university class schedule data is calculated, and the unoccupied time period is extracted as the idle time window; The idle time window, the equipment specifications, the historical usage data, and the operating cycle are integrated, and the corresponding asset number is configured to generate the asset digital feature set.
3. The intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities according to claim 1, characterized in that, The predictive pricing module specifically includes the following steps: Based on the historical usage data, the idle frequency is extracted, and a Markov transition matrix is constructed to calculate the idle probability; Obtain the time decay difference between the current time and the start time of the idle time window; Substitute the difference between the idle probability and the time decay into the decay function to calculate the floating price coefficient; The idle probability, the idle time window, and the floating price coefficient are bound to the asset digital feature set to construct a shared resource pool.
4. The intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities according to claim 3, characterized in that, The calculation of the floating price coefficient specifically includes the following steps: The resource consumption data in the feedback parameters is converted into energy consumption cost per unit time, and summed with the depreciation amount per unit time corresponding to the loss coefficient in the feedback parameters to obtain the cost benchmark value. Substitute the difference between the idle probability and the time decay into the decay function to calculate the floating price coefficient; When the rate corresponding to the floating price coefficient is lower than the cost benchmark value, the rate will be clamped to the cost benchmark value; When the current time reaches the start time and the idle time window is not yet marked as occupied, the idle time window is marked as a zero-rate period and pushed to the user's terminal.
5. The intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities according to claim 1, characterized in that, The supply and demand matching module specifically includes the following steps: Extract the expected duration and rate cap from the aforementioned demand characteristics; Remove the locked time window and the occupied time window from the shared resource pool to obtain the remaining idle time window; The desired duration and the remaining idle time window are input into a multi-objective optimization model for solution. Solutions with required rates higher than the upper limit of the rate are eliminated, and Pareto non-dominated solution set is selected; Extract the asset number and reservation time period corresponding to the target solution from the Pareto non-dominated solution set, generate the transaction fee rate by combining it with the floating price coefficient, and output the scheduling instruction.
6. The intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities according to claim 1, characterized in that, The linkage control module specifically includes the following steps: The vibration sequence of the asset during the scheduled time period is collected as the operation trajectory, and the power sequence is collected as the actual load. Integrating the power sequence yields the power consumption value, which serves as resource consumption data. The feature vector of the vibration sequence is time-aligned with the envelope of the power sequence. The verified operation trajectory, actual load, resource consumption data, and transaction fee rate are hashed and encapsulated into a performance log.
7. The intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities according to claim 1, characterized in that, The health assessment module specifically includes the following steps: The actual load is sliced into load amplitudes over multiple cycles; Based on the preset load-lifetime loss function, calculate the step loss increment corresponding to each load amplitude; The asset loss amount is obtained by integrating and summing the incremental losses of each step. When the asset loss reaches the preset threshold, a locking instruction is generated and written back to the shared resource pool to generate the locking time window. The ratio of the incremental step loss generated in the current period to the baseline loss value is extracted as the loss coefficient.
8. The intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities according to claim 7, characterized in that, The generation of the locking time window specifically includes the following steps: The stepped loss increment is decomposed and integrated according to frequency band to obtain the component loss amount of each component, and the component loss amounts are weighted and synthesized into the asset loss amount; Calculate the loss rate of each of the aforementioned sub-items of loss, extract the components whose loss rate reaches the preset loss line as components to be maintained, and map the maintenance duration; The period of time when the idle probability is higher than the preset probability line is extracted from the shared resource pool as the maintenance window. The maintenance duration and the maintenance window are encapsulated into the locking instruction and written back.
9. The intelligent scheduling system for the digital revitalization and shared operation of idle assets in universities according to claim 1, characterized in that, The dynamic settlement module specifically includes the following steps: The settlement result is calculated based on the transaction fee rate, the resource consumption data, and the loss coefficient. The operation trajectory is matched with a standard trajectory library using an algorithm to calculate the normalized distance; The normalized distance is converted into a dimensionless normalized index according to a preset mapping relationship; The updated credit score is obtained by weighting the aforementioned standardized indicators with the user's historical credit score. The settlement result, the credit score, the loss coefficient, and the resource consumption data are encapsulated into the feedback parameters and sent to the predictive pricing module. The feedback parameter is used to instruct the predictive pricing module to add an upward adjustment coefficient when calculating the floating price coefficient for the next scheduling for users whose credit scores are lower than a preset credit threshold.
