Renewable resource recovery data management system based on Internet of Things

Abstract

The invention relates to the technical field of industrial platform data analysis, in particular to a renewable resource recovery data management system based on the Internet of Things, which comprises the steps of synchronously acquiring weight, spectrum and microwave characteristic data of a tested resource through a data acquisition module, executing moisture weight decoupling operation by utilizing a resource matching module, and performing data analysis; then, an inventory evolution module tracks performance loss of resources in real time by using an evolution model of an integrated long and short-term memory network, generates a resource attenuation weight, and constructs a two-dimensional decision matrix containing scheduling priority and preprocessing strength grade instructions in combination with real-time inventory saturation; and finally, the clearance scheduling module executes dynamic pruning and bidirectional optimization through a resource scheduling model, accurately allocates the loading share and the access sequence of each node, and generates a dynamic instruction set containing a delivery sequence. According to the method, feeding homogenization is realized through industrial data multi-dimensional collaborative analysis.

Description

Technical Field

[0001] This invention relates to the field of industrial platform data analysis technology, specifically to an Internet of Things-based data management system for the recycling of renewable resources. Background Technology

[0002] With the widespread application of IoT technology in the field of renewable resources, the amount of material attribute data generated in the recycling process has exploded. In the current recycling management process, the management system usually collects data such as the weight of resources, basic categories and storage time at recycling nodes in real time, and uses this as a basis to perform preliminary planning of inventory statistics and transportation logistics.

[0003] However, due to their wide range of sources and strong physical heterogeneity, the internal material composition and moisture content of recycled resources often undergo complex dynamic evolution with changes in storage environment and time and space. At the same time, the automated production lines of terminal processing plants have strict stability requirements for the input quality of raw materials. Any lag or deviation in data from a single dimension will lead to fluctuations in composition during the feeding process.

[0004] Therefore, how to achieve accurate alignment and coordinated scheduling of heterogeneous data on both the supply and demand sides of renewable resources in a complex and volatile environment, so as to ensure the homogeneity of material input during the resource transfer process, is a key technical problem in the field of renewable resource recycling data management.

[0005] To address this, an Internet of Things-based data management system for recycling renewable resources is proposed. Summary of the Invention

[0006] The purpose of this invention is to provide an Internet of Things (IoT)-based data management system for recycled resources, which achieves homogenization of feed through multi-dimensional collaborative analysis of industrial data. This includes: synchronously acquiring the weight, spectral, and microwave characteristic data of the tested resources through a data acquisition module; performing moisture weight decoupling calculations using a resource matching module to output a standard dry-basis asset weight with cross-environment consistency; subsequently, a stock evolution module using an evolution model integrating a long short-term memory network to track resource performance degradation in real time, generating resource decay weights, and constructing a two-dimensional decision matrix including scheduling priority and preprocessing intensity level instructions based on real-time stock saturation; finally, a waste disposal scheduling module performing dynamic pruning and bidirectional optimization through a resource scheduling model to accurately allocate the loading share and access order of each node, generating a dynamic instruction set including delivery order.

[0007] To achieve the above objectives, the present invention provides the following technical solution: The Internet of Things-based data management system for recycled resource recovery includes: The data acquisition module obtains the real-time total weight, spectral reflectance sequence, microwave phase shift, microwave amplitude attenuation, ambient temperature, and humidity of the tested paper. The resource matching module extracts features from microwave phase offset and microwave amplitude attenuation, and extracts feature peaks from spectral reflectance sequences to obtain the deep moisture content and fiber grade characteristics of the paper package. It also performs moisture weight decoupling calculations in conjunction with environmental humidity compensation to output standard dry basis asset weight data. The inventory evolution module inputs the standard dry basis asset weight, paper package depth moisture content, fiber grade characteristics, and environmental temperature and humidity into the inventory evolution model to obtain resource decay weights. Based on the resource decay weights and fiber grade characteristics, quality grade data is generated, and a decision matrix is ​​constructed by combining the real-time inventory saturation of the recycling nodes. The waste collection and scheduling module retrieves the material feeding quality threshold requirements of the production line of the terminal processing plant, inputs the decision matrix, standard dry basis asset weight, and quality grade data of each recycling node into the resource scheduling model, and outputs a set of delivery sequence instructions that includes logistics trigger time, allocation path coordinates, and meets the material feeding homogenization requirements of the terminal processing plant.

[0008] Preferably, the acquisition of the real-time total weight, spectral reflectance sequence, microwave phase shift, microwave amplitude attenuation, ambient temperature, and humidity of the tested paper includes: scanning and collecting data on the surface of the tested paper using a near-infrared array sensor to generate the spectral reflectance sequence; transmitting a penetrating signal to the tested paper using microwave transceiver components symmetrically arranged within the recycling chamber; obtaining the microwave phase shift and microwave amplitude attenuation by analyzing the phase deviation and energy loss of the penetrating signal after passing through the paper medium; obtaining the real-time total weight by sensing the force on the tested paper using a weighing unit; and obtaining the real-time ambient temperature and humidity values ​​using an environmental sensing unit integrated inside the chamber.

[0009] Preferably, the process of outputting standard dry basis asset weight data includes: extracting the slope of the phase change of the microwave signal propagating in the paper packaging medium as a function of frequency to obtain the group delay parameter; calculating the unit decibel ratio of the microwave phase offset to the microwave amplitude attenuation; combining the unit decibel ratio with the group delay parameter into a normalized vector combination; and obtaining the paper packaging deep moisture content data by mapping the modulus of the normalized vector. The spectral reflectance sequence is transformed using a standard normal variable to eliminate background baseline noise caused by optical path differences; local extreme points of the transformed sequence are located using a sliding window, and the wavelength coordinates and peak intensities corresponding to the local extreme points are extracted, and the wavelength coordinates and peak intensities are used to construct the fiber grade feature vector; The intermediate weight value is obtained by subtracting the product of the real-time total weight and the moisture content data of the paper package depth from the real-time total weight; the corresponding balance moisture compensation operator is retrieved from the preset fiber balance moisture mapping table according to the environmental humidity value, and the intermediate weight value is multiplied by the balance moisture compensation operator to output the standard dry basis asset weight data.

[0010] Preferably, the inventory evolution model includes: an environmental stress layer, which receives real-time environmental temperature and humidity as input, and generates environmental partial pressure difference characteristics by calculating the difference between the saturated vapor pressure and the actual vapor pressure at the current temperature and humidity; The moisture migration layer inputs the historical moisture content time series constructed from the deep moisture content of the paper package and the environmental partial pressure difference feature into the long short-term memory network, transforms the discrete moisture content fluctuations into a continuous internal energy state latent vector, and performs mapping calculation with the environmental partial pressure difference feature to output the non-equilibrium potential difference feature. The structural calibration layer inputs the fiber grade characteristics and the non-equilibrium potential difference characteristics into the multilayer perceptron to learn the effective diffusion resistance distribution of different fiber grades at different aging stages. The potential difference characteristics and resistance coefficients are weighted and coupled to generate effective evolution rate characteristics. The boundary constraint layer receives the standard dry-based asset weight and the effective evolution rate feature as input, takes the average of the effective evolution rate at the current sampling time and the previous sampling time, multiplies it with the sampling step size, and outputs the instantaneous performance loss value. The spatiotemporal accumulation layer receives the instantaneous performance loss value as the current input, obtains the historical loss accumulation value stored at the previous sampling time, performs a first-order recursive superposition operation to generate the total asset performance loss value, maps the total asset performance loss value to [0, 1] through nonlinear mapping logic, and outputs the resource attenuation weight.

[0011] Preferably, the process of generating the quality grade data includes: using the resource attenuation weight to perform a proportional deduction calculation on the fiber grade characteristics to determine the residual performance quantification value of the paper under the current environmental influence; aligning the residual performance quantification value with multiple preset quality evaluation benchmarks to determine the quality management interval in which the residual performance quantification value is located; extracting the grade label corresponding to the quality management interval to confirm the quality grade data of the recycled resources.

[0012] Preferably, the process of constructing the decision matrix includes: extracting the real-time inventory saturation of the recycling node as the first decision dimension and the resource decay weight as the second decision dimension to construct a two-dimensional attribute mapping space; establishing orthogonal orthogonal decision boundary lines in the two-dimensional attribute mapping space based on preset storage capacity critical constraints and resource quality thresholds, and dividing the mapping space into multiple rectangular attribute management areas through the orthogonal decision boundary lines; wherein each rectangular attribute management area corresponds to a combination of storage urgency and resource loss status determined by a saturation interval defined by the orthogonal decision boundary lines and a preset resource decay weight interval; configuring a corresponding physical intervention parameter sequence for each rectangular attribute management area; the physical intervention parameter sequence includes a scheduling priority reflecting the transfer timeliness and a preprocessing intensity level reflecting the quality maintenance requirements; performing regional topology matching between the real-time coordinates of the current recycled resource in the two-dimensional attribute mapping space and the rectangular attribute management area to determine the target area to which the recycled resource belongs, and extracting the regional associated physical intervention parameter sequence to generate a decision matrix including resource survival control status and preprocessing intensity level instructions.

[0013] Preferably, the processing procedure of the waste collection and scheduling module includes: real-time retrieval of the material input quality threshold requirements of the terminal processing plant's production line as the target guidance benchmark for global scheduling; real-time synchronization of standard dry basis asset weight and quality grade data of each recycling node, combined with the real-time status in the decision matrix, and multi-dimensional data alignment based on timestamps to construct a dynamic scheduling task pool; monitoring the resource survival control status in the decision matrix, and automatically extracting the decision matrix and material input quality threshold as physical boundary constraints when the status changes and / or the storage capacity reaches the threshold, activating the resource scheduling model; and outputting a real-time dynamic instruction set including logistics trigger time, allocation path coordinates, and delivery order based on the model calculation results.

[0014] Preferably, the resource scheduling model includes: an attribute mapping layer: which performs feature mapping on the standard dry basis asset weight and quality grade data of each recycling node, and converts them into a supply state vector in a high-dimensional feature space; Demand guidance layer: Converts the input terminal processing plant production line material quality threshold demand into a target center vector, and establishes the convergence guidance direction of the supply state vector in the feature space; Constraint Transformation Layer: Parses the input decision matrix, transforms the resource survival control state into boundary constraint operators that act on the algorithm logic, and achieves dynamic pruning by removing node branches that do not meet the control state in real time from the dynamic scheduling task pool, thereby narrowing the search range of candidate solutions. Bidirectional optimization layer: Under the constraints of the boundary constraint operator, the homogenization deviation is measured by the spatial distance between the supply state vector and the target center vector, and the candidate solution that minimizes both path cost and quality deviation is locked through iterative calculation. Instruction generation layer: Extract the quality deviation of each recycling node in the candidate scheme, allocate the loading share using the deviation offset logic including the stability adjustment benchmark; achieve linear alignment between total quality and feeding requirements by matching the mixing ratio of positive and negative deviation resources, and after simulation verification and error fine-tuning, generate a delivery sequence instruction set including logistics trigger time, allocation path coordinates and meeting the feeding homogenization requirements of the terminal processing plant.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. By synergistically extracting microwave penetration characteristics and infrared spectral peaks, combined with environmental humidity mapping compensation logic, a deep separation of moisture and dry weight in the physical properties of the tested paper was achieved. This effectively corrected the error of paper moisture absorption and weight gain caused by fluctuations in environmental humidity, and transformed the underlying physical sensing signals into standard dry-basis data that can reflect the actual asset value. This significantly reduced the fluctuation range of the original data due to environmental interference, provided the management system with a highly consistent data starting point, effectively alleviated the differences in data sources caused by material heterogeneity, and laid a reliable quantitative foundation for the accurate alignment of supply and demand.

[0016] 2. By constructing an inventory evolution model integrating a long short-term memory network and a structural calibration layer, we achieved real-time dynamic tracking of the performance loss of recycled resources caused by temperature and humidity stress during the warehousing process. This transformed discrete sensor data into continuous resource attenuation weights, quantified the quality evolution law of resources over time and space, significantly reduced the information deviation between static inbound labels and actual outbound quality, and enabled the management system to perceive the physical performance status of supply-side resources in real time. This provides timely dynamic parameter support for real-time optimization of scheduling decisions under complex and fluctuating environments.

[0017] 3. By mapping real-time inventory saturation and resource decay weights to a two-dimensional attribute space and establishing orthogonal decision boundary lines, the system achieves rapid transformation from multi-dimensional heterogeneous data to standardized decision instructions. This transforms complex logistics management requirements into a decision matrix that includes scheduling priority and preprocessing intensity, enabling refined regional topology matching of the status of recycling nodes. It optimizes the system's management path when handling large-scale heterogeneous resources, helps improve the response speed and accuracy of management decisions, and provides standardized logical support for achieving cross-node, multi-objective collaborative scheduling.

[0018] 4. By using dynamic pruning logic in the resource scheduling model and an adaptive weight allocation algorithm based on quality deviation offsetting logic, precise coupling between fluctuating resources at the recycling end and the high-precision production needs of the processing plant is achieved. By utilizing the dynamic balance compensation mechanism of quality attributes, the mixing ratio of resources of different qualities is dynamically adjusted and transformed into specific logistics triggering time and delivery sequence instructions. This scheduling method effectively offsets the interference of the physical heterogeneity of recycled resources on the stability of terminal feeding, making the instantaneous component ratio of delivered resources closer to the preset production threshold, thus providing a systematic management means for the goal of homogenization of feeding in the process of large-scale resource circulation. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the Internet of Things-based data management system for recycling of renewable resources according to the present invention. Figure 2 This is a schematic diagram of the inventory evolution model of the present invention; Figure 3 This is a schematic diagram of the resource scheduling model of the present invention. Detailed Implementation

[0020] The technical solutions of 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.

[0021] Please see Figures 1 to 3 This invention provides an Internet of Things-based data management system for the recycling of renewable resources, and the technical solution is as follows.

[0022] Example 1: The IoT-based data management system for recycled resource recovery has the following structure: Figure 1 As shown, it includes: The data acquisition module obtains the real-time total weight, spectral reflectance sequence, microwave phase shift, microwave amplitude attenuation, ambient temperature, and humidity of the tested paper. The resource matching module extracts features from microwave phase offset and microwave amplitude attenuation, and extracts feature peaks from spectral reflectance sequences to obtain the deep moisture content and fiber grade characteristics of the paper package. It also performs moisture weight decoupling calculations in conjunction with environmental humidity compensation to output standard dry basis asset weight data. The inventory evolution module inputs the standard dry basis asset weight, paper package depth moisture content, fiber grade characteristics, and environmental temperature and humidity into the inventory evolution model to obtain resource decay weights. Based on the resource decay weights and fiber grade characteristics, quality grade data is generated, and a decision matrix is ​​constructed by combining the real-time inventory saturation of the recycling nodes. The waste collection and scheduling module retrieves the material feeding quality threshold requirements of the production line of the terminal processing plant, inputs the decision matrix, standard dry basis asset weight, and quality grade data of each recycling node into the resource scheduling model, and outputs a set of delivery sequence instructions that includes logistics trigger time, allocation path coordinates, and meets the material feeding homogenization requirements of the terminal processing plant.

[0023] Furthermore, the acquisition of the real-time total weight, spectral reflectance sequence, microwave phase shift, microwave amplitude attenuation, ambient temperature, and humidity of the tested paper includes: scanning and collecting data on the surface of the tested paper using a near-infrared array sensor to generate the spectral reflectance sequence; transmitting a penetrating signal to the tested paper using microwave transceiver components symmetrically arranged within the recycling chamber; obtaining the microwave phase shift and microwave amplitude attenuation by analyzing the phase deviation and energy loss of the penetrating signal after passing through the paper medium; obtaining the real-time total weight by sensing the force on the tested paper using a weighing unit; and obtaining the real-time ambient temperature and humidity values ​​using an environmental sensing unit integrated inside the chamber.

[0024] In the spectral data acquisition stage, a near-infrared array sensor is used to scan the surface of the paper being tested. This near-infrared array sensor is installed above the inlet of the recycling bin or on top of the conveyor device, and its sensing wavelength range covers 900nm to 1700nm. When the paper being tested passes through the sensing area, the sensor performs a linear scan of the paper surface at a preset sampling frequency. By capturing the diffuse reflection intensity of light in a specific wavelength band by the paper fibers, a continuous spectral reflectance sequence is generated. This sequence contains characteristic information about the type of paper fibers, coatings, and surface humidity. In the deep physical feature acquisition stage, microwave transceivers symmetrically arranged on the inner walls of both sides of the recycling bin are used. The component performs penetrating detection. The microwave transceiver includes a microwave transmitting antenna and a microwave receiving antenna. The transmitting antenna transmits a penetrating microwave signal of a preset frequency to the paper being tested. When the microwave signal penetrates a paper medium of a certain thickness, its phase and amplitude will change due to the influence of polar molecules (such as moisture) inside the paper. By comparing and analyzing the signal waveforms of the transmitting and receiving ends, the phase deviation and energy loss after penetrating the paper medium can be analyzed, thereby obtaining the microwave phase shift and microwave amplitude attenuation respectively. This penetrating sensing method can effectively obtain the average moisture content characteristics inside the paper package and avoid data distortion caused by only detecting surface moisture.

[0025] During the physical weight acquisition phase, a weighing unit installed at the bottom of the recycling bin senses the force applied to the paper being tested. This weighing unit typically consists of multiple high-precision pressure sensors. By sensing the pressure exerted on the paper and its carrier at the bottom of the bin, the analog pressure signal is converted into a digital weight signal. The force value is extracted in real time, and zero-point drift calibration is performed to obtain the real-time total weight of the paper being tested. During the environmental baseline acquisition phase, an environmental sensing unit integrated in a non-contact area inside the bin acquires real-time environmental status data. The environmental sensing unit includes an integrated digital thermometer and hygrometer to monitor the real-time ambient temperature and humidity values ​​of the microenvironment inside the recycling bin.

[0026] The real-time total weight, spectral reflectance sequence, microwave phase shift, microwave amplitude attenuation, and ambient temperature and humidity values ​​collected above are uploaded to the central processing unit in real time via the industrial bus or wireless transmission network of the data acquisition module, so that the resource matching module can perform subsequent feature extraction and weight decoupling calculations.

[0027] After acquiring the spectral reflectance sequence, the microwave phase offset, and the microwave amplitude attenuation, the data acquisition module further includes performing multi-scale wavelet packet decomposition logic: using preset wavelet basis functions to perform time-frequency domain decomposition on the original signal, separating the high-frequency noise component reflecting environmental interference and the low-frequency characteristic component reflecting the internal structure of the material; performing energy suppression and signal reconstruction on the noise frequency range through a threshold shrinkage algorithm, eliminating baseline drift and shot noise interference caused by the dynamic electromagnetic environment in the recovery bin, and obtaining the reconstructed spectral reflectance sequence and microwave characteristic waveform, which are then input to the resource matching module.

[0028] Specifically, in practical applications, due to the alternating electromagnetic field generated by the operation of high-power motors at the recycling node site, and the multipath interference of the metal recycling bin on the electromagnetic waves, the raw signal acquired by the data acquisition module is often superimposed with complex non-stationary noise. In order to ensure the accuracy of subsequent feature extraction, this embodiment introduces a multi-scale wavelet packet decomposition preprocessing process after acquiring the spectral reflection sequence, microwave phase shift, and microwave amplitude attenuation. First, the system uses a preset wavelet basis function (Daubechies wavelet basis db4 in this embodiment) to decompose the spectral signal acquired by the near-infrared array sensor and the raw microwave signal generated by the microwave transceiver component into multiple orthogonal frequency band subspaces. During the decomposition process, the algorithm maps the signal to approximate coefficients that cover the low-frequency trend of the material fiber structure, as well as sudden high-frequency detail coefficients that reflect the motor start-stop or environmental vibration. Second, for the high-frequency noise components obtained by decomposition, the system uses a threshold shrinkage algorithm to perform energy suppression. By setting scientific hard or soft threshold functions, the system can accurately identify and reduce the particle noise hidden in the detail coefficients and the baseline drift caused by the internal environment of the recycling bin. This step effectively reduces the energy amplitude of background electromagnetic interference while preserving the physical characteristics of the material. Subsequently, the system performs inverse wavelet transform on the processed coefficients of each frequency band to restore the reconstructed spectral reflection sequence and the reconstructed microwave characteristic waveform with a high signal-to-noise ratio. The reconstructed waveform removes non-characteristic glitches and offsets, making the waveform characteristics more significant and directly improving the data quality input to the resource matching module.

[0029] By coordinating signal reconstruction, closed-loop calibration, and logistics timeliness, the purity of underlying data and the long-term accuracy of the system are enhanced, and the unloading smoothness is improved. This helps to improve the collaborative efficiency and flow stability of resource allocation throughout the entire process while achieving homogenization of material input.

[0030] By using a multi-source sensing matrix to simultaneously perceive the physical properties and environmental parameters of the tested paper, the single-dimensional physical weighing is transformed into a multi-dimensional data description that includes surface fiber characteristics, deep moisture content, and environmental influencing factors. This acquisition mode effectively alleviates the problem of incomplete data representation caused by the physical heterogeneity of recycled resources. Through the coordinated use of penetrating microwave signals and real-time environmental temperature and humidity, the interference of external environmental fluctuations on the determination of material properties is significantly reduced. This not only provides highly timely and cross-environmentally consistent raw input for subsequent asset value quantification and attribute decoupling, but also supports the accurate alignment of heterogeneous data on both the supply and demand sides at the perception level, thereby helping to achieve the goal of homogenized material feeding management in a dynamic circulation environment.

[0031] Furthermore, the process of outputting standard dry basis asset weight data includes: extracting the slope of the phase change of the microwave signal propagating in the paper packaging medium as a function of frequency to obtain the group delay parameter; calculating the unit decibel ratio of the microwave phase offset to the microwave amplitude attenuation; combining the unit decibel ratio with the group delay parameter into a normalized vector; and obtaining the paper packaging deep moisture content data by mapping the magnitude of the normalized vector. The spectral reflectance sequence is transformed using a standard normal variable to eliminate background baseline noise caused by optical path differences; local extreme points of the transformed sequence are located using a sliding window, and the wavelength coordinates and peak intensities corresponding to the local extreme points are extracted, and the wavelength coordinates and peak intensities are used to construct the fiber grade feature vector; The intermediate weight value is obtained by subtracting the product of the real-time total weight and the moisture content data of the paper package depth from the real-time total weight; the corresponding balance moisture compensation operator is retrieved from the preset fiber balance moisture mapping table according to the environmental humidity value, and the intermediate weight value is multiplied by the balance moisture compensation operator to output the standard dry basis asset weight data.

[0032] Specifically, the resource matching module receives the multi-dimensional raw signals transmitted by the data acquisition module and performs feature extraction and weight decoupling through the following logical steps to achieve a quantitative conversion of the physical asset value of the tested paper: In the process of outputting standard dry basis asset weight data, the system first extracts the slope of the phase change of the microwave signal propagating within the paper-packaged medium as a function of frequency, thereby calculating the group delay parameter to characterize the equivalent thickness and density characteristics of the medium. Simultaneously, the system analyzes the energy loss of microwaves during transmission through the medium to calculate the unit decibel ratio of the microwave phase shift to the microwave amplitude attenuation. This ratio reflects the modulation intensity of the microwave signal by the moisture polarity per unit volume. Subsequently, the system performs a normalized vector combination of the unit decibel ratio and the group delay parameter, utilizing the normalized vector... The modulus value of the quantity is mapped to a pre-stored modulus-moisture content mapping curve, thereby outputting the deep moisture content data of the paper package. This mapping method, which uses group delay for density compensation, ensures that the moisture content data can reflect the true deep moisture content state inside the paper package. The modulus-moisture content mapping curve is obtained in the following way: multiple groups of paper package samples with different densities and fiber grades are pre-selected, and their true moisture content is determined by oven drying as the true value. At the same time, the normalized vector modulus value generated by the corresponding microwave signal is recorded. The modulus value and the true moisture content value are fitted by the least squares method to generate a continuous modulus-moisture content mapping curve.

[0033] The normalized vector combination is specifically as follows: First, since the unit decibel ratio and the group delay parameter belong to different categories of physical quantities, the system uses max-min normalization logic to map the original values ​​of the two parameters to a dimensionless interval between 0 and 1. Second, the normalized unit decibel ratio is used as a vertical feature component, and the normalized group delay parameter is used as a horizontal feature component. Thus, in the feature coordinate system, the two discrete physical features are combined and defined as a two-dimensional normalized vector pointing to a specific coordinate point. Subsequently, the system calculates the geometric span of the two-dimensional normalized vector from the origin in the coordinate space (i.e., calculates the arithmetic square root of the sum of the squares of the two components) to obtain a comprehensive feature modulus reflecting the internal physical properties of the resource. Finally, the system retrieves a preset modulus-moisture content mapping curve through this modulus, thereby outputting the paper package deep moisture content data with density compensation characteristics.

[0034] In the specific process of constructing fiber grade feature vectors, the system performs the following standardized data dimensionality reduction and feature mapping operations: The system first performs a standard normal variable transformation on the acquired spectral reflectance sequence. This process eliminates optical path variations and baseline noise caused by differences in surface physical structure among different tested samples by subtracting the sequence mean and dividing by the standard deviation, making the spectral data of different batches comparable.

[0035] The system uses a sliding window of a preset size to perform local extremum retrieval in the transformed sequence. The size of the sliding window is set to 11 consecutive data points (corresponding to a wavelength width of about 8 nanometers), and the moving step size is 1 data point. The system compares the values ​​of the center point of the window with those of the neighboring points to identify the wavelength position that reflects the fiber absorption characteristics.

[0036] To ensure the fixed dimension of the feature vector and the computational stability of the subsequent model, the system precisely selects the first 8 significant extrema from all retrieved local extrema, in descending order of peak intensity. If fewer than 8 extrema are detected in a sample, zeros are padded at the end of the vector.

[0037] The final fiber-grade feature vector is set to a fixed 16-dimensional structure. This vector is formed by sequentially concatenating the first 8 dimensions of wavelength coordinate data (ranging from 900 nm to 1700 nm) with the last 8 dimensions of peak intensity data (normalized values).

[0038] The feature points captured by this vector directly correspond to the key components of paper fibers: for example, the peak near 970 nm reflects the moisture content of the paper; the feature points near 1200 nm reflect the polysaccharide structure of cellulose and hemicellulose; and the feature near 1450 nm is related to the hydrogen bond stretching vibration inside the fiber. The combination of these wavelengths and intensities together constitutes the core basis for determining the quality grade of the fiber.

[0039] During the output of standard dry basis asset weight data, the system performs a two-stage operation of moisture stripping and environmental compensation. First, the system subtracts the product of the real-time total weight and the paper package depth moisture content data from the real-time total weight to deduct the free water content from the physical total, obtaining the intermediate weight value. To offset the dynamic interference of external environmental humidity on fiber weight, the system searches for and matches the corresponding equilibrium moisture compensation operator in a preset fiber equilibrium moisture mapping table based on the environmental humidity value. This mapping table presets the moisture absorption weight gain ratio of fibers under different humidity baselines. Finally, the system multiplies the intermediate weight value by the equilibrium moisture compensation operator and outputs the result. The standard dry basis asset weight data, as an asset indicator with cross-environmental consistency, effectively solves the asset accounting deviation in the process of recycling resource transfer. The fiber balance moisture mapping table adopts a multi-dimensional structure, consisting of a set of sub-tables corresponding to different environmental temperature values. Before performing compensation calculations, the system first locates the target temperature sub-table based on the real-time environmental temperature value obtained by the environmental sensing unit, and then retrieves the corresponding balance moisture compensation operator in the sub-table based on the real-time environmental humidity value. In this embodiment, the fiber balance moisture mapping table preset by the system is stored in the database associated with the resource matching module. Table 1 shows some typical data of this mapping table: Table 1. Example of fiber equilibrium moisture mapping

[0040] The resource matching module also includes establishing a self-calibrating closed-loop feedback link: real-time acquisition of the true value of the material input quality sampling inspection output from the terminal processing plant, and calculation of the multi-dimensional residual between the standard dry basis asset weight output value and the true value of the sampling inspection; based on the characteristic distribution of the residual, dynamically adjusting the correction coefficient of the equilibrium moisture compensation operator and the microwave-moisture content mapping curve using the gradient descent algorithm to adaptively compensate for physical sensing errors.

[0041] In actual operation, due to sensor aging, sudden environmental changes, or long-term drift of material types, fixed mapping parameters may cause the asset quantification error to gradually increase. In order to ensure the long-term reliability of the system, the resource matching module has established a self-calibration mechanism connected to the feedback link of the terminal processing plant.

[0042] First, before the recycled resources arrive at the terminal processing plant and enter the production line, the system obtains the true value of the furnace quality sampling inspection of the batch of resources through the industrial interface. This true value is usually obtained by the factory laboratory through oven drying or high-precision sample analysis. Subsequently, the system performs deviation tracing calculation. The calculation module extracts the standard dry basis asset weight output by the resource matching module when the batch of resources leaves the warehouse and the sampling inspection true value fed back by the factory. By calculating the algebraic difference between the two, a multi-dimensional residual sequence is generated. This residual reflects the cumulative prediction deviation caused by environmental humidity compensation error or microwave mapping deviation. Next, the system uses the gradient descent algorithm to perform parameter optimization. If the residual distribution shows a significant systematic shift, the algorithm uses the residual as the objective function and adjusts the correction coefficient in the equilibrium moisture compensation operator through iterative calculation. For example, when the actual dry basis weight is detected to be continuously lower than the predicted value, the system will reduce the weight of the compensation operator by a preset step size until the residual converges to within the preset error tolerance.

[0043] During the self-calibration closed-loop feedback process, the preset step size is not a fixed value, but a dynamic analytical result based on the residual strength and operator sensitivity. First, the system retrieves multiple sets of historical data recorded by the resource matching module within a preset period and calculates the average absolute value of the residual between the standard dry basis asset weight and the terminal feedback true value. By analyzing the distribution characteristics of this residual under different environmental humidity values, the system determines the systematic deviation level of the current sensor system. Second, the system performs a sensitivity test on the compensation operator. The calculation module calculates the displacement of the output standard dry basis asset weight when the balance moisture compensation operator undergoes a unit increment change by simulating disturbances, thereby establishing the influence factor of the operator on the final quality judgment. The system defines this influence factor as the step size benchmark to ensure that the adjustment action does not cause a drastic jump in asset value. Subsequently, the preset step size is obtained by multiplying the learning rate and the normalized residual. The system presets a convergence increment in the range [0.001, 0.01] as the base learning rate and multiplies it by the normalized residual of the current batch. When the residual is large, the calculated preset step size is increased accordingly to accelerate error convergence; when the residual is close to zero, the step size is automatically reduced to ensure that the retrieval accuracy of the equilibrium moisture compensation operator in the fiber equilibrium moisture mapping table can smoothly approach the true value. Finally, to prevent numerical oscillations during the adjustment process, after obtaining the preset step size, the system compares it with the preset discrete step size reference in the mapping table (such as temperature step size 0). The system compares the calculated step size with the compensation change rate corresponding to 0.1℃. If the calculated step size exceeds the preset multiple of the benchmark (e.g., 1.5 times), the system will execute a truncation rate limit and force the minimum index spacing of the mapping table sub-table to be used as the current execution step size. Through this closed-loop self-calibration process, the system dynamically corrects the systematic deviation caused by changes in the physical sensing environment, improves the long-term accuracy of asset accounting, and ensures that the balance moisture compensation operator and microwave-moisture content mapping curve can continuously adapt to the complex flow environment, providing more deterministic preliminary data support for homogenized feeding.

[0044] By mapping the microwave group delay characteristics to the decibel ratio and extracting the spectral extreme points after standard normal variable transformation, a deep analysis of the deep moisture content and fiber grade characteristics of recycled resources was achieved. The coupling relationship of physical parameters was used to offset the interference of material density and surface noise. Furthermore, a balance moisture compensation operator was introduced and multiplied with the intermediate weight value to reduce the random impact of environmental humidity fluctuations on asset weight determination. This improved the physical accuracy and management consistency of standard dry basis asset weight data, providing reliable quantitative support for the accurate alignment of heterogeneous data on both the supply and demand sides.

[0045] Furthermore, the inventory evolution model includes an environmental stress layer, which receives real-time environmental temperature and humidity as input, and generates environmental partial pressure difference characteristics by calculating the difference between the saturated vapor pressure and the actual vapor pressure at the current temperature and humidity. The moisture migration layer inputs the historical moisture content time series constructed from the deep moisture content of the paper package and the environmental partial pressure difference feature into the long short-term memory network, transforms the discrete moisture content fluctuations into a continuous internal energy state latent vector, and performs mapping calculation with the environmental partial pressure difference feature to output the non-equilibrium potential difference feature. The structural calibration layer inputs fiber grade characteristics and the non-equilibrium potential difference characteristics into a multilayer perceptron to learn the effective diffusion resistance distribution of different fiber grades at different aging stages. The potential difference characteristics and resistance coefficients are weighted and coupled to generate effective evolution rate characteristics.

[0046] The boundary constraint layer receives the standard dry-based asset weight and the effective evolution rate feature as input, takes the average of the effective evolution rate at the current sampling time and the previous sampling time, multiplies it with the sampling step size, and outputs the instantaneous performance loss value. The spatiotemporal accumulation layer receives the instantaneous performance loss value as the current input and obtains the historical loss accumulation value stored at the previous sampling time. It performs a first-order recursive superposition operation to generate the total asset performance loss value. Through nonlinear mapping logic, it maps the total asset performance loss value to [0, 1] and outputs the resource attenuation weight. The specific process is as follows: Figure 2 As shown.

[0047] Specifically, the environmental stress layer serves as the input interface for the model. The system acquires real-time environmental temperature and humidity values ​​collected by the environmental sensing unit. Within this layer, the system pre-stores a saturated vapor pressure reference table (in this embodiment, pressure reference values ​​are recorded within the range of 0℃ to 50℃, with a step size of 0.1℃). The system first retrieves the corresponding saturated vapor pressure based on the current temperature. Then, it calculates the actual vapor pressure in the atmosphere by multiplying the real-time environmental humidity value by the saturated vapor pressure. By calculating the difference between the saturated vapor pressure and the actual vapor pressure, an environmental partial pressure difference feature is generated. This partial pressure difference is defined within the model as the escape potential energy that induces moisture migration from inside the fiber to the outside or the absorption of external moisture, and is encapsulated as an environmental partial pressure difference feature to provide driving parameters for subsequent layers.

[0048] The moisture migration layer utilizes a long short-term memory network to process moisture content data with temporal evolution characteristics. The system inputs the historical moisture content time series constructed from the deep moisture content of the paper package over the past 12 hours, along with the environmental partial pressure difference feature, into an LSTM network with a hidden layer dimension of 128. The LSTM, through its internal memory gating mechanism, maps the discrete fluctuation data to a 128-dimensional high-dimensional feature space, generating an internal energy state latent vector containing historical trend information. This vector encodes the dynamic evolution law of moisture distribution inside the fiber, reflecting the resistance characteristics of the fiber pore structure to moisture diffusion. This 128-dimensional latent vector is then spatially mapped and coupled with the current environmental partial pressure difference feature through a fully connected layer, compressing the high-dimensional physical state into a scalarized non-equilibrium potential difference feature.

[0049] The structural calibration layer utilizes a multilayer perceptron with a three-layer fully connected architecture to dynamically calibrate the physical resistance of resources. It incorporates two hidden layers with 64 and 32 neurons respectively to handle complex material structural constraints. The system uses fiber grade features (including spectral feature points) and non-equilibrium potential difference features as joint inputs. It leverages pre-trained diffusion model parameters stored within the multilayer perceptron for fiber density, lignin content, and porosity to map abstract grade features into continuous physical parameters, thereby automatically fitting the corresponding physical parameters for the material at the current aging stage. The effective diffusion resistance distribution, i.e., the resistance coefficient, is determined by factors such as the density of the paper's fibers. For example, the model outputs a higher resistance coefficient (around 0.85) for densely fibrous paper, and a lower coefficient to reflect weaker barrier properties. The system uses the non-equilibrium potential difference feature as the evolutionary driving force and performs a weighted multiplication operation with the reciprocal of the resistance coefficient. This weighted coupling process follows the non-equilibrium migration law in physical logic, where the evolution rate is proportional to the strength of the driving force and inversely proportional to the diffusion resistance. Ultimately, it outputs an effective evolution rate feature that can accurately quantify the rate of quality degradation.

[0050] The boundary constraint layer is responsible for performing numerical integration. The system obtains the effective evolution rate of the current sampling time and the previous sampling time, calculates the average of the two, and multiplies it by the standard dry-base asset weight and the sampling step size (e.g., once every 10 minutes) to obtain the instantaneous performance loss value in the current period. The spatiotemporal accumulation layer performs a first-order recursive superposition operation, that is, superimposing the current instantaneous loss onto the historical loss accumulation value stored in the system register to generate the total asset performance loss value. Finally, the system uses nonlinear normalization logic (in this embodiment, the Sigmoid function) to map the total loss value to the interval [0, 1] and outputs the resource decay weight. The closer the weight value is to 1, the closer the resource is to its original quality. The closer it is to 0, the more severe the loss.

[0051] In order to enable the inventory evolution model to accurately capture the physical evolution of the tested resources under different environmental stresses, this embodiment constructs a training dataset through offline experiments. First, recycled paper samples covering different fiber grades (corresponding to the fiber grade feature vectors) are pre-selected and placed in a simulation chamber with high-precision temperature and humidity control capabilities. During the accelerated aging evolution cycle of up to 360 hours, the system synchronously records the environmental partial pressure difference characteristics and the time series data of the deep moisture content of the paper package obtained by microwave sensing at a sampling step of 10 minutes. These multi-dimensional time series data constitute the feature space of the model input layer.

[0052] To obtain target labels for supervised learning, at each sampling step, core physical indicators such as residual tensile strength and breaking length of paper fibers are simultaneously measured using high-precision laboratory instruments. The system normalizes these measured physical performance indicators to measured loss values ​​between 0 and 1 through a preset standardized mapping logic. These measured loss values ​​serve as the Ground Truth during the training phase and directly correspond to the resource decay weights at the model output. By using measured physical failure data as labels, the weights output by the model are ensured to have clear physical evolutionary meaning.

[0053] In designing the loss function for model training, the system employs a joint loss function to perform backpropagation optimization. The main loss function uses the mean squared error function to minimize the deviation between the resource decay weights predicted by the model and the laboratory-measured loss labels, ensuring the model's prediction accuracy in macro trends. Meanwhile, to ensure that the internal energy state latent vectors possess realistic physical evolution logic, a physical consistency constraint term based on Fick's law is introduced into the loss function. This constraint term penalizes non-equilibrium potential difference feature jumps that do not conform to the water migration rate law, forcing the latent vectors generated by LSTM to conform to the energy conservation and migration laws in the water diffusion process.

[0054] Finally, through the above training framework, the internal energy state latent vector, under the memory gating mechanism of the LSTM layer, can spontaneously capture the potential energy state inside the fiber through the historical moisture content sequence. The multilayer perceptron in the structure calibration layer automatically fits the effective diffusion resistance distribution by learning the difference in strength decay rate of samples of different fiber grades at different aging stages. This supervised learning method with physical constraints enables the model to output resource decay weights with high signal-to-noise ratio and definite physical meaning when processing unseen heterogeneous data, thus providing scientific support for the subsequent construction of decision matrices.

[0055] The data input and synchronization mechanism of the inventory evolution model is as follows: the system sets a unified global sampling clock with a sampling period of 10 minutes. Each data acquisition module (microwave phase offset, microwave amplitude attenuation, spectral reflectance sequence, ambient temperature and humidity) performs synchronous acquisition according to this global clock to ensure that all physical sensing signals are aligned on the time axis.

[0056] In the environmental stress layer, the system acquires real-time ambient temperature and relative humidity once every sampling period (10 minutes) and performs environmental stress calculation. If the environmental sensing unit fails to provide temperature and humidity data in a certain sampling period due to hardware failure or other reasons, the system adopts a forward filling strategy, that is, it temporarily continues to use the temperature and humidity values ​​of the previous sampling period and marks the data point as an interpolation state. After the hardware is restored, the data is filled in and the model is recalculated.

[0057] In the moisture migration layer, the system constructs a new historical moisture content time series every 10 minutes. This series contains moisture content data points from the past 12 hours, totaling 72 points (72 × 10 minutes = 12 hours). When the system has not accumulated 12 hours of data in the early stages of operation, a warm-up strategy is adopted. In the first 720 minutes, the system only inputs the existing valid data (less than 72 points) into the LSTM for rolling processing, while marking the sequence length as partial. After accumulating 12 hours, the system switches to the standard 72-point input mode.

[0058] The output (resource decay weight) of the inventory evolution model is also set to update every 10 minutes. That is, the system calculates a new resource decay weight value every 10 minutes and updates the corresponding resource status parameters stored in memory. This synchronization mechanism ensures that the input data of each level of the model are strictly aligned on the time axis.

[0059] By dynamically decoupling environmental stress and fiber aging logic through an inventory evolution model, the system achieves a digital transformation from static snapshots to dynamic quality evolution. By using a long short-term memory network to extract loss trends and combining diffusion resistance to calibrate the evolution rate, it can more objectively reflect the decay characteristics of fiber performance. This not only improves the real-time accuracy of asset accounting, but also provides a timely basis for decay weights for subsequent proportioning and scheduling, which helps to ensure the stability of material quality throughout the entire resource flow process.

[0060] Furthermore, the process of generating quality grade data includes: using the resource attenuation weight to perform a proportional deduction calculation on the fiber grade characteristics to determine the residual performance quantification value of the paper under the current environmental influence; aligning the residual performance quantification value with multiple preset quality evaluation benchmarks to determine the quality management interval in which the residual performance quantification value is located; extracting the grade label corresponding to the quality management interval to confirm it as quality grade data of recycled resources.

[0061] Specifically, firstly, the system performs a quantitative assessment of residual performance. The resource matching module acquires fiber grade characteristics (in this embodiment, this is represented by an initial quality score based on spectral feature intensity mapping, with a score range of 0-100) and resource attenuation weights (range of 0-1). The system uses these resource attenuation weights to perform a proportional deduction calculation on the fiber grade characteristics, i.e., by multiplying the two, it determines the quantitative value of the paper's residual performance under the current environmental influence. Subsequently, the system executes the alignment and judgment logic of the quality range. The system aligns the calculated quantitative value of residual performance with multiple preset quality evaluation benchmarks. The specific quality evaluation benchmark mapping table is shown in Table 2. The quality evaluation benchmarks are based on pre-set score boundaries according to industry standards. The system determines the specific numerical range into which the quantitative value of residual performance falls through numerical comparison, i.e., it determines the quality management range to which it belongs. Finally, the system completes the extraction and confirmation of grade labels. Based on the determined quality management range, the system extracts the corresponding grade labels from the grade database and confirms them as the quality grade data of the batch of recycled resources. This grade data is then used as the core decision-making basis and input into the subsequent decision matrix.

[0062] Table 2 Quality Evaluation Benchmark Mapping Table

[0063] By dividing the quantified residual performance value after dynamic deduction into four core management intervals, the system minimizes the classification complexity of warehousing and logistics while ensuring quality accuracy. This four-level differentiation mapping logic captures the real-time damage to resource value caused by the environment through secondary and tertiary labels, and avoids logistics stagnation and computational redundancy caused by excessively fine classification through reasonable benchmark settings. This provides a standardized data index for generating highly executable homogenized proportioning instructions, effectively solving the problem of quality-price alignment in the entire process of recycled resource circulation.

[0064] Further, the process of constructing the decision matrix includes: extracting the real-time inventory saturation of the recycling node as the first decision dimension and the resource decay weight as the second decision dimension to construct a two-dimensional attribute mapping space; establishing orthogonal orthogonal decision boundary lines in the two-dimensional attribute mapping space based on preset storage capacity critical constraints and resource quality thresholds, and dividing the mapping space into multiple rectangular attribute management areas through the orthogonal decision boundary lines; wherein each rectangular attribute management area corresponds to a combination of storage urgency and resource loss status determined by a saturation interval defined by the orthogonal decision boundary lines and a preset resource decay weight interval; configuring a corresponding physical intervention parameter sequence for each rectangular attribute management area; the physical intervention parameter sequence includes a scheduling priority reflecting the transfer timeliness and a preprocessing intensity level reflecting the quality maintenance requirements; performing regional topology matching between the real-time coordinates of the current recycled resource in the two-dimensional attribute mapping space and the rectangular attribute management area to determine the target area to which the recycled resource belongs, and extracting the regional associated physical intervention parameter sequence to generate a decision matrix including resource survival control status and preprocessing intensity level instructions.

[0065] Specifically, firstly, the decision module obtains the real-time inventory saturation of the recycling node and defines it as the first decision dimension reflecting the physical pressure of the warehouse; at the same time, it retrieves the resource decay weight output by the preceding inventory evolution model and defines it as the second decision dimension reflecting the remaining value of the resource. The system constructs a two-dimensional attribute mapping space in the logical layer by orthogonally projecting the above two dimensions. In this space, each point represents the dual state coordinates of inventory and quality of a specific resource at a specific node.

[0066] Secondly, based on preset storage capacity critical constraints and resource quality thresholds, the system establishes mutually perpendicular orthogonal judgment boundary lines within the two-dimensional attribute mapping space. These boundary lines act as logical rulers in the space. By dividing the continuous coordinate space into discrete rectangular attribute management areas, the system achieves semantic hierarchical classification of complex physical states. Each rectangular attribute management area logically locks a combination of storage urgency (determined by the saturation range) and resource depletion status (determined by the resource decay range).

[0067] The saturation span refers to a preset discrete numerical range of real-time inventory saturation for recycling nodes, used to characterize the physical pressure state of storage space. Specifically, the system divides the horizontal axis of saturation into multiple spans with different management meanings based on preset storage capacity critical constraints: Safety span ([0, 0.50]): Represents ample storage space, low storage urgency, and normal scheduling priority. Warning span ([0.50, 0.80]): Represents inventory approaching the critical point, increased storage urgency, and the system activates the resource scheduling model to optimize potential solutions. Storage capacity saturation span ([0.80, 1.0]): Represents storage space about to be exhausted, maximum storage urgency, corresponding to the forced circulation state in the decision matrix.

[0068] In constructing the decision matrix, this embodiment logically divides the resource decay weight (its value range is [0, 1]) into multiple resource decay weight regions. These regions are discrete numerical intervals established along the vertical axis of a two-dimensional attribute mapping space, based on the physical performance evolution of the fiber structure of recycled resources (such as waste paper). Each region represents a specific performance decay state and value retention level of the material under the current storage environment. Specifically, they are defined as follows: High value retention region ([0.85, 1.0]): Represents a material with an intact fiber structure, minimal impact from environmental stress, and at its most stable physical performance stage. Performance fluctuation region ([0.60, 0.85]): Represents that fluctuations in the internal moisture content of the material have induced initial fiber aging or strength decay, and the performance is in a dynamic rheological state. Severe loss region (e.g., [0, 0.60]): Represents that the total asset performance loss value has reached a critical point, and the industrial utilization value of the material faces a significant risk of decline.

[0069] Subsequently, for each rectangular attribute management area, the system pre-configures the corresponding physical intervention parameter sequence at the underlying level. The physical intervention parameter sequence is the basis for the execution layer to perform specific actions, including the scheduling priority reflecting the transfer timeliness and the preprocessing intensity level reflecting the resource quality maintenance requirements. During real-time operation, the system extracts the real-time coordinates of the currently recycled resources in space and uses spatial geometric relationships to match them with the topology of the execution area of ​​each rectangular attribute management area. Once the target area to which the resource belongs is determined, the system immediately extracts the pre-associated physical intervention parameter sequence within that area.

[0070] Finally, the decision module logically encapsulates the matched scheduling priorities and preprocessing intensity levels to generate a decision matrix that includes resource survival control status (used to define the urgency of the resource's flow) and preprocessing intensity level instructions (used to standardize the intensity of maintenance actions). This decision matrix is ​​ultimately output as the decision result and directly sent to the logistics scheduling terminal and the station processing equipment, through this complete process from coordinate mapping to topology matching.

[0071] By mapping inventory saturation and resource decay weights to a two-dimensional attribute space, a quantitative correlation between physical pressure and quality risk is achieved. Orthogonal boundary partitioning and topology matching mechanisms mitigate the lag and subjective interference in the decision-making process. This process allows scheduling parameters to better align with the current resource availability, improving the balance between inventory safety and quality maintenance while enhancing the objectivity and real-time responsiveness of circulation decisions.

[0072] Furthermore, the processing procedure of the waste collection and scheduling module includes: real-time retrieval of the material feeding quality threshold requirements of the production line of the terminal processing plant as the target guidance benchmark for global scheduling; real-time synchronization of standard dry basis asset weight and quality grade data of each recycling node, combined with the real-time status in the decision matrix, and multi-dimensional data alignment based on timestamps to construct a dynamic scheduling task pool; monitoring the resource survival control status in the decision matrix, and automatically extracting the decision matrix and material feeding quality threshold as physical boundary constraints when the status changes and / or the storage capacity reaches the threshold, activating the resource scheduling model; and outputting the delivery sequence instruction set containing logistics trigger time, allocation path coordinates, and meeting the material feeding homogenization requirements of the terminal processing plant based on the model calculation results.

[0073] Specifically, firstly, the waste collection and scheduling module retrieves the material quality threshold requirements of the production line at the terminal processing plant in real time. These requirements cover the minimum tolerance of raw material fiber strength, average moisture content, and impurity content in the production process. The system sets this as the target guidance benchmark for global scheduling, which is used to reverse-screen material packages that meet the conditions across the entire network. This ensures that the overall quality of resources flowing from each recycling node accurately aligns with the production process requirements when they arrive at the terminal. Secondly, the system synchronizes the standard dry basis asset weight and corresponding quality grade data of each distributed recycling node in real time through the network layer and associates it with the real-time status of the node in the decision matrix. To eliminate the signal asynchrony problem caused by geographical distribution, the system uses a unified global clock pulse to perform alignment operations on the above multi-dimensional data based on timestamps. Specifically, the system encapsulates the material weight, grade, location, and control urgency at the same time node into structured data units and aggregates them to build a dynamic scheduling task pool, providing a real-time, high-fidelity data foundation for subsequent global optimization. Subsequently, the system continuously monitors the resource survival and control status recorded in the decision matrix. Once a change in the status of a node is detected (e.g., switching from regular turnover to forced removal), or when the real-time inventory of that node reaches the storage capacity threshold, the system automatically extracts the decision matrix parameters of that node and the material input quality threshold of the terminal, and transforms them into physical boundary constraints of the resource scheduling model. This process realizes the transformation from logical judgment to physical constraints, turning the necessity of management into hard boundary conditions for model calculation. Finally, under the constraints of the above physical boundary, the system activates the resource scheduling model to perform path and ratio optimization calculations. Based on the model calculation results, the removal scheduling module finally outputs a real-time dynamic instruction set containing multiple dimensions. This instruction set specifically includes: logistics trigger time, which specifies the precise time for the fleet to depart or the loading equipment to start; allocation path coordinates, which is the optimal logistics driving trajectory composed of a series of geographical latitude and longitude coordinates; and delivery order, which clarifies the order in which resources from different sources arrive at the processing plant. This instruction set is directly issued to the logistics execution terminal, driving the physical loading, unloading, and transfer operations.

[0074] By coupling terminal material feeding requirements with front-end resource availability in real time and utilizing a timestamp-based multidimensional data alignment mechanism, the system effectively solves the problems of information fragmentation and poor timeliness in the recycled resource supply chain, ensuring the authenticity of the dynamic scheduling task pool. By monitoring the decision matrix status and automatically extracting physical boundary constraints, the system can achieve quality-preserving clearing of high-value resources while ensuring storage capacity safety, significantly improving the configuration efficiency and delivery quality stability of recycled resources in the circulation network.

[0075] Furthermore, the resource scheduling model includes: an attribute mapping layer: which performs feature mapping on the input standard dry basis asset weight and quality grade data of each recycling node, converting them into a supply state vector in a high-dimensional feature space; Demand guidance layer: Converts the input terminal processing plant production line material quality threshold demand into a target center vector, and establishes the convergence guidance direction of the supply state vector in the feature space; Constraint Transformation Layer: Parses the input decision matrix, transforms the resource survival control state into boundary constraint operators that act on the algorithm logic, and achieves dynamic pruning by removing node branches that do not meet the control state in real time from the dynamic scheduling task pool, thereby narrowing the search range of candidate solutions. Bidirectional optimization layer: Under the constraints of the boundary constraint operator, the homogenization deviation is measured by the spatial distance between the supply state vector and the target center vector, and the candidate solution that minimizes both path cost and quality deviation is locked through iterative calculation. Instruction Generation Layer: Extracts the quality deviation of each recycling node in the candidate schemes, allocates loading shares using deviation offset logic including stability adjustment benchmarks; achieves linear alignment between total quality and material feeding requirements by matching the mixing ratio of positive and negative deviation resources, and after simulation verification and error fine-tuning, generates a delivery sequence instruction set including logistics trigger time, allocation path coordinates, and meeting the material feeding homogenization requirements of the terminal processing plant. The specific process is as follows: Figure 3 As shown.

[0076] Specifically, firstly, the system performs normalization processing on the raw data. The attribute mapping layer receives standard dry-based asset weight and quality grade data from each recycling node. The algorithm first retrieves the system's preset attribute dimension benchmarks. In this embodiment, the quality grades (level 1 to level 4) are mapped to preset grade weight distributions (level 1 corresponds to 0.95, level 2 corresponds to 0.80, level 3 corresponds to 0.65, and level 4 corresponds to 0.50). The standard dry-based asset weight is then divided by the node's historical maximum carrying capacity, thereby uniformly transforming the raw physical data of different dimensions into dimensionless normalized feature components within the interval [0, 1].

[0077] Secondly, the demand-guided layer establishes the gravitational center for algorithm optimization. The system acquires the material quality threshold requirements (such as the required fiber strength benchmark value) output by the terminal processing plant's production line in real time and converts them into a target center vector in the feature space. The role of this layer is to establish the convergence guide direction of the supply state vector in the space. All subsequent path selection and resource combination actions are driven by the core force of reducing the geometric deviation between each supply point vector and the target center vector. By calculating the cosine similarity between the supply state vector of each node and the target center vector, the system can initially lock in the candidate node set that meets the terminal production requirements.

[0078] Subsequently, the constraint transformation layer performs dynamic spatial pruning based on management logic. The system deeply analyzes the input decision matrix and transforms the resource survival control status recorded in it (such as forced transfer, delayed processing, or loss exceeding the limit) into boundary constraint operators that act on the algorithm optimization. Before performing optimization calculations, the system uses these operators in the dynamic scheduling task pool to physically remove node branches that do not meet the control requirements, i.e., dynamic pruning. Through this mechanism, the system directly blocks paths that do not meet the conditions for clearing, significantly narrowing the search range of candidate solutions for global optimization and ensuring the real-time and effectiveness of decision generation in large-scale recycling networks.

[0079] The constraint transformation layer of the resource scheduling model, when parsing the decision matrix, also includes executing multi-node time-of-arrival alignment discrimination logic: predicting the estimated arrival time window of each node based on the geographical coordinates, real-time road conditions and vehicle driving characteristics of each recycling node; setting a queuing length threshold for the unloading area of ​​the terminal processing plant and constructing a time overlap constraint operator that reflects the unloading frequency; and performing secondary dynamic pruning on low-priority nodes that cause logistics congestion conflicts by calculating the intersection of the estimated arrival time windows of each node.

[0080] During global logistics scheduling, if multiple recycling nodes with the capacity for collection trigger logistics commands simultaneously, it can easily lead to vehicle accumulation in the unloading area of ​​the terminal processing plant, resulting in logistics congestion and reduced material delivery timeliness. To address this issue, the constraint transformation layer of the resource scheduling model introduces multi-node timeliness alignment discrimination logic to achieve smooth processing of the logistics flow timeline. The system calls the geographical coordinates of each recycling node in real time and attaches an external real-time traffic interface to obtain traffic flow data. By combining the average travel speed of the collection vehicles and the loading operation time estimated based on the dry basis asset weight of the materials, the system performs comprehensive simulation calculations. This calculation not only calculates the theoretical travel time but also incorporates operational disturbances, thereby generating an estimated arrival time window for each recycling node that includes the earliest and latest arrival times. The system establishes a criterion for unloading conflicts, pre-setting a queue length threshold for the unloading area of ​​the terminal processing plant to define the maximum carrying capacity of the area per unit time. Based on this threshold, a time overlap constraint operator reflecting the unloading frequency is constructed. This operator is used to monitor the intersection span of the estimated arrival time windows between different nodes. When the arrival times of multiple nodes overlap on the time axis, and the overlap duration exceeds the terminal's preset average unloading cycle per vehicle, the system determines that there is a risk of logistics congestion conflict. When a congestion risk is determined, the system retrieves the decision matrix of each node, compares the scheduling priorities recorded therein, and performs secondary dynamic pruning for the node branches that cause conflicts and have lower priorities.

[0081] By coordinating signal reconstruction, closed-loop calibration, and logistics timeliness, this solution enhances the purity of underlying data and the long-term accuracy of the system, improves unloading smoothness, and helps to improve the collaborative efficiency and flow stability of resource allocation throughout the entire process while achieving homogenization of material input.

[0082] The bidirectional optimization layer first extracts the supply status vectors of all recycling nodes involved in the scheme, and then weights them according to the loading share of each node to synthesize the comprehensive feature center vector of the batch of materials. The initial value of the loading share is determined by the proportion of the inventory of each node to the total demand, and is dynamically fine-tuned by the algorithm in subsequent iterations. Subsequently, the system calculates the spatial distance between this vector and the target center vector pre-existing in the system. The target center vector is obtained by the system in real time through connection with the production execution system. According to the process standards of the paper type to be produced (such as target tensile index, allowable moisture content range, etc.), the ideal physical characteristics are mapped to ideal points in the feature space. The spatial distance is specifically represented as the sum of the absolute values ​​of the differences between various physical characteristic components. The smaller the distance, the higher the degree of material homogenization. Secondly, to ensure the economy of the solution, the system defines logistics cost as a logical set of physical displacement cost in the spatial dimension and time-related depreciation cost in the time dimension. Physical displacement cost is obtained by multiplying the real-time navigation mileage of the vehicle by the unit energy consumption coefficient. The unit energy consumption coefficient is pre-calculated based on the nominal energy consumption level of the vehicle and the real-time load ratio and stored in the database. Time-related depreciation cost is calculated by multiplying the effective evolution rate output by the aforementioned inventory evolution model with the expected transportation time to quantify the performance value loss caused by fiber aging during the transit period. The system uses a preset cost balance coefficient to weight and sum the two to generate a path cost component that reflects the overall logistics expenditure. The cost balance coefficient is pre-calibrated through marginal benefit analysis of historical logistics expenditure and material degradation loss to dynamically adjust the system's preference for transportation cost and quality preservation.

[0083] Then, the system uses a non-dominated sorting genetic algorithm with strategies to perform iterative evolution under the constraints of boundary constraint operators. The strategies specifically include an elite retention strategy and a crowding allocation strategy: the elite retention strategy ensures that the best non-dominated individuals in each generation are directly retained to the next generation, preventing the loss of high-quality scheduling schemes during evolution; the crowding allocation strategy calculates the distribution density of individuals in the solution space and prioritizes uniformly distributed individuals to ensure that the search process covers more diverse path combinations and avoids the algorithm getting trapped in local optima. During the evolution process, the algorithm continuously optimizes the performance indicators of candidate schemes through multiple rounds of crossover and mutation operations. When performing optimization calculations, the system introduces a boundary constraint operator transformed from the decision matrix in real time. This operator acts as a logical barrier in the evolution process, forcing the algorithm to search within the feasible domain of the dynamic scheduling task pool. Any illegal solution containing a node branch that has been removed through dynamic pruning will be identified and eliminated by the algorithm. Ultimately, the algorithm locks a set of candidate solutions in Pareto optimal state in the feature space. This ensures that, under the physical premise of minimizing logistics path overhead, the comprehensive physical characteristics of the mixed materials can be most accurately aligned with the ideal target center vector of the terminal processing plant, thereby generating the final delivery sequence instruction set.

[0084] The system retrieves the quality grade score (quantitative value from 0.50 to 1.0) of each recycling node in the candidate schemes and compares it with the material input quality threshold requirement fed back in real time by the terminal processing plant production line. The system automatically divides all nodes into two categories by calculating the algebraic difference between the node quality score and the requirement threshold: a high-quality surplus group with a positive difference and a low-quality deficit group with a negative difference. To prevent numerical overflow due to extremely small deviations during the calculation process, the system introduces a regularization constant for a stability adjustment benchmark (preset to 0.01 in this embodiment). The system calculates the negative correlation influence factor of the sum of the absolute value of the deviation of each node and the benchmark value as a preliminary quality influence coefficient. The closer the quality of a node is to the requirement, the higher its coefficient. The system then divides the coefficient of a single node by the value within the scheme. The sum of coefficients for all nodes yields the initial loading share percentage. The system executes a physical hedging logic to offset surpluses with deficits, dynamically adjusting the initial share. The adjustment process follows the linear hybrid law: the system calculates the average deviation span between the surplus group and the deficit group, using the material quality threshold requirement as the balance point. If the average quality of the surplus group is much higher than the threshold, the system will automatically lower the loading share percentage of the surplus group and proportionally increase the load of the deficit group. The system retrieves the standard dry basis asset weight of each recovery node in the task pool in real time. If the calculated loading exceeds the real-time inventory of that node, the system will automatically lock the loading share of that node to its inventory limit and proportionally transfer the remaining difference percentage to other nodes in the same group, ensuring that the adjustment result does not deviate from the actual physical inventory boundary.

[0085] Before the instructions are officially generated, the system enters a closed-loop simulation verification mode. The system simulates the instantaneous average value of the components after mixing each node according to the current share, and calculates the predicted residual with the target threshold. If the absolute value of the predicted residual is greater than the preset accuracy threshold (0.02 in this embodiment), the system will dynamically reduce the total percentage of the group with the larger deviation in a fixed step size (e.g., 1%), and simultaneously increase the percentage of the other side. The simulation operation is executed in a loop until the predicted residual is reduced to an acceptable range. The percentage locked at this time is the final loading share. Finally, the system encapsulates and generates a dynamic instruction set based on the final determined loading share, combined with the geographical location and real-time road conditions. The instruction set clarifies the access order of each node. This order is used to drive resources to form a specific physical stacking layer inside the vehicle. When unloading at the terminal, this spatial layering is transformed into a controlled time flow sequence to ensure that the material component ratio entering the production line is accurately aligned with the homogenization feeding requirements of the processing plant.

[0086] The dynamic pruning mechanism effectively reduces computational redundancy in large-scale networks, ensuring the real-time generation of scheduling decisions. By coordinating the control of loading share and access order, the system transforms static storage characteristics into a dynamic and stable feeding sequence at the physical level, significantly reducing the impact of natural fluctuations in raw materials on the stability of the production line and improving the level of refined control in the circulation of recycled resources.

[0087] By integrating dynamic resource quality assessment, inventory evolution early warning, and multi-dimensional scheduling optimization, this invention constructs a quantitative management system for the circulation process of recycled resources. Utilizing the logical coupling of resource attenuation weights and storage capacity pressure helps reduce decision-making biases caused by information lag, improves the quality certainty of discrete resources in storage and transportation, and the bidirectional optimization algorithm, while coordinating path costs, also considers the homogenized material input requirements of terminal processing plants, mitigating the impact of natural fluctuations in raw material quality on the stability of downstream production, and enhancing the collaborative efficiency and scientific nature of resource allocation throughout the entire process.

[0088] Example 2: In this embodiment, a waste paper recycling and circulation scenario is taken as an example, which includes two supply points (recycling station A and recycling station B), one restricted supply point (recycling station C), and one terminal processing plant (paper mill): First, the system executes the data sensing, physical feature extraction, and asset value quantification processes at the recycling nodes. At recycling station A, the system uses a near-infrared array sensor to perform surface scanning on the 16 tons of old corrugated cardboard boxes stored therein, generating characteristic peak data reflecting the fiber structure. In conjunction with the microwave transceiver component, the system obtains the microwave phase offset and amplitude attenuation through a penetrating signal. The resource matching module extracts the slope of the microwave phase change with frequency to obtain the group delay parameter. Combined with the signal attenuation ratio, the system calculates the deep moisture content of the cardboard package. The system also uses the measured ambient humidity to retrieve the fiber balance moisture mapping table to obtain the compensation operator. Through the above calculations, the system deducts moisture interference from the total weight of 16 tons, outputs the standard dry basis asset weight of recycling station A, and initially maps the quality level to a score of 0.95 (corresponding to the first-level resource weight).

[0089] Secondly, the system executes the inventory evolution simulation and real-time quality level determination process. The environmental perception unit detects that the local humidity of recycling station A is too high. Based on this, the environmental stress layer in the inventory evolution model calculates the saturated vapor pressure difference and generates the environmental partial pressure difference feature. The moisture migration layer uses a long short-term memory network to process the historical moisture content sequence. Combined with the fiber diffusion resistance output by the structural calibration layer, the spatiotemporal accumulation layer finally calculates the resource decay weight as 0.90. Subsequently, the system uses this decay weight to perform a proportional deduction on the initial quality score, calculates the residual performance quantification value of 0.855, and aligns it with the preset quality management range. The quality level data of recycling station A is dynamically corrected from the initial first level to the second level label, realizing the digital transformation from static snapshot to dynamic quality evolution.

[0090] Subsequently, the system executes the logical pruning process of decision matrix construction and dynamic scheduling task pool. The decision module extracts the real-time inventory saturation of recycling station A (0.80) as the first decision dimension and the resource decay weight (0.90) as the second decision dimension. In the two-dimensional attribute mapping space, topology matching is performed by orthogonally determining the boundary line. Because this point falls into the high-urgency circulation area, the system automatically associates the corresponding physical intervention parameters to generate a decision matrix containing high scheduling priority and forced clearance status. Before activating the resource scheduling model, the clearance scheduling module detects this status change and automatically removes recycling station C, which is in a closed state and does not meet the control conditions, in the dynamic scheduling task pool through the boundary restriction operator, thus realizing the dynamic pruning of candidate solutions.

[0091] Finally, the system executes a two-way optimization matching and dynamic instruction generation process based on terminal demand. The paper mill processing line issues a material feeding quality threshold requirement of 0.75. Based on this, the demand guidance layer establishes the target center vector. The two-way optimization layer uses the NSGA-II algorithm to find the Pareto optimal balance between path cost and quality deviation (Euclidean distance), locking in the combination scheme of recycling station A and recycling station B. The instruction generation layer compares the quality score of each station with the demand threshold. Station A (score 0.855) belongs to the high-quality surplus group with an absolute deviation of 0.105; Station B (score 0.60) belongs to the low-quality deficit group with an absolute deviation of 0.15. The system introduces a stability adjustment benchmark (regularization constant 0.01) to calculate the quality impact coefficient: the coefficient for station A is 8.7. 0, Station B coefficient is 6.25; following the linear mixing law to perform physical hedging to offset surplus against deficit, the preliminary calculation shows that the loading share of Station A is about 58% and that of Station B is about 42%. The system verifies the physical boundary in real time and confirms that the 16-ton inventory of Station A is sufficient to support this share demand, and there is no need to perform share shift. Before the formal issuance of instructions, the system enters the simulation verification mode. After calculation, the instantaneous average value of the mixed components is 0.748, which is precisely aligned with the target accuracy threshold (0.02) requirement. Based on this, the system plans the order in which vehicles visit Station A first and then Station B. Using the spatial stacking structure, it is transformed into a controlled time flow sequence at the unloading end. Finally, the dynamic instruction set generated by the encapsulation is issued to the execution end to drive the physical layer to complete the precise homogenization scheduling.

[0092] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A data management system for recycling renewable resources based on the Internet of Things, characterized in that, include: The data acquisition module obtains the real-time total weight, spectral reflectance sequence, microwave phase shift, microwave amplitude attenuation, ambient temperature, and humidity of the tested paper. The resource matching module extracts features from microwave phase offset and microwave amplitude attenuation, and extracts feature peaks from spectral reflectance sequences to obtain the deep moisture content and fiber grade characteristics of the paper package. It also performs moisture weight decoupling calculations in conjunction with environmental humidity compensation to output standard dry basis asset weight data. The inventory evolution module inputs the standard dry basis asset weight, paper package depth moisture content, fiber grade characteristics, and environmental temperature and humidity into the inventory evolution model to obtain resource decay weights. Based on the resource decay weights and fiber grade characteristics, quality grade data is generated, and a decision matrix is ​​constructed by combining the real-time inventory saturation of the recycling nodes. The waste collection and scheduling module retrieves the material feeding quality threshold requirements of the production line of the terminal processing plant, inputs the decision matrix, standard dry basis asset weight, and quality grade data of each recycling node into the resource scheduling model, and outputs a set of delivery sequence instructions that includes logistics trigger time, allocation path coordinates, and meets the material feeding homogenization requirements of the terminal processing plant.

2. The Internet of Things-based data management system for recycling renewable resources according to claim 1, characterized in that, The process of obtaining the real-time total weight, spectral reflectance sequence, microwave phase shift, microwave amplitude attenuation, ambient temperature, and humidity of the tested paper includes: scanning and collecting data on the surface of the tested paper using a near-infrared array sensor to generate the spectral reflectance sequence; transmitting a penetrating signal to the tested paper using microwave transceiver components symmetrically arranged within the recycling chamber; obtaining the microwave phase shift and microwave amplitude attenuation by analyzing the phase deviation and energy loss of the penetrating signal after passing through the paper medium; obtaining the real-time total weight by sensing the force on the tested paper using a weighing unit; and obtaining the real-time ambient temperature and humidity values ​​using an environmental sensing unit integrated inside the chamber.

3. The Internet of Things-based data management system for recycling renewable resources according to claim 2, characterized in that, The process of outputting standard dry basis asset weight data includes: extracting the slope of the phase change of the microwave signal propagating in the paper packaging medium as a function of frequency to obtain the group delay parameter; calculating the unit decibel ratio of the microwave phase offset to the microwave amplitude attenuation; combining the unit decibel ratio with the group delay parameter into a normalized vector; and obtaining the paper packaging deep moisture content data by mapping the modulus of the normalized vector. The spectral reflectance sequence is transformed using a standard normal variable to eliminate background baseline noise caused by optical path differences; local extreme points of the transformed sequence are located using a sliding window, and the wavelength coordinates and peak intensities corresponding to the local extreme points are extracted, and the wavelength coordinates and peak intensities are used to construct the fiber grade feature vector; The intermediate weight value is obtained by subtracting the product of the real-time total weight and the moisture content data of the paper package depth from the real-time total weight; the corresponding balance moisture compensation operator is retrieved from the preset fiber balance moisture mapping table according to the environmental humidity value, and the intermediate weight value is multiplied by the balance moisture compensation operator to output the standard dry basis asset weight data.

4. The Internet of Things-based data management system for recycling renewable resources according to claim 1, characterized in that, The inventory evolution model includes an environmental stress layer, which receives real-time environmental temperature and humidity as input, and generates environmental partial pressure difference characteristics by calculating the difference between the saturated vapor pressure and the actual vapor pressure at the current temperature and humidity. The moisture migration layer inputs the historical moisture content time series constructed from the deep moisture content of the paper package and the environmental partial pressure difference feature into the long short-term memory network, transforms the discrete moisture content fluctuations into a continuous internal energy state latent vector, and performs mapping calculation with the environmental partial pressure difference feature to output the non-equilibrium potential difference feature. The structural calibration layer inputs the fiber grade characteristics and the non-equilibrium potential difference characteristics into the multilayer perceptron to learn the effective diffusion resistance distribution of different fiber grades at different aging stages. The potential difference characteristics and resistance coefficients are weighted and coupled to generate effective evolution rate characteristics. The boundary constraint layer receives the standard dry-based asset weight and the effective evolution rate feature as input, obtains the average effective evolution rate of the current sampling time and the previous sampling time, performs cumulative multiplication with the sampling step size, and outputs the instantaneous performance loss value. The spatiotemporal accumulation layer receives the instantaneous performance loss value as the current input, obtains the historical loss accumulation value stored at the previous sampling time, performs a first-order recursive superposition operation to generate the total asset performance loss value, maps the total asset performance loss value to [0, 1] through nonlinear mapping logic, and outputs the resource attenuation weight.

5. The Internet of Things-based data management system for recycling renewable resources according to claim 1, characterized in that, The process of generating the quality grade data includes: using the resource attenuation weight to perform a proportional deduction calculation on the fiber grade characteristics to determine the residual performance quantification value of the paper under the current environmental influence; aligning the residual performance quantification value with multiple preset quality evaluation benchmarks to determine the quality management interval in which the residual performance quantification value is located; extracting the grade label corresponding to the quality management interval to confirm the quality grade data of the recycled resources.

6. The Internet of Things-based data management system for recycling renewable resources according to claim 1, characterized in that, The process of constructing the decision matrix includes: extracting the real-time inventory saturation of the recycling node as the first decision dimension and the resource decay weight as the second decision dimension to construct a two-dimensional attribute mapping space; establishing orthogonal orthogonal decision boundary lines in the two-dimensional attribute mapping space based on preset storage capacity critical constraints and resource quality thresholds, and dividing the mapping space into multiple rectangular attribute management areas through the orthogonal decision boundary lines; wherein each rectangular attribute management area corresponds to a combination of storage urgency and resource loss status determined by a saturation interval defined by the orthogonal decision boundary lines and a preset resource decay weight interval; configuring a corresponding physical intervention parameter sequence for each rectangular attribute management area; the physical intervention parameter sequence includes a scheduling priority reflecting the transfer timeliness and a preprocessing intensity level reflecting the quality maintenance requirements; performing regional topology matching between the real-time coordinates of the current recycled resource in the two-dimensional attribute mapping space and the rectangular attribute management area to determine the target area to which the recycled resource belongs, and extracting the regional associated physical intervention parameter sequence to generate a decision matrix including resource survival control status and preprocessing intensity level instructions.

7. The Internet of Things-based data management system for recycling renewable resources according to claim 1, characterized in that, The processing steps of the waste collection and scheduling module include: real-time retrieval of the material feeding quality threshold requirements of the terminal processing plant's production line as the target guidance benchmark for global scheduling; real-time synchronization of standard dry basis asset weight and quality grade data of each recycling node, combined with the real-time status in the decision matrix, and multi-dimensional data alignment based on timestamps to construct a dynamic scheduling task pool; monitoring the resource survival control status in the decision matrix, and automatically extracting the decision matrix and material feeding quality threshold as physical boundary constraints when the status changes and / or the storage capacity reaches the threshold, activating the resource scheduling model; and outputting a delivery sequence instruction set that includes logistics trigger time, allocation path coordinates, and meets the material feeding homogenization requirements of the terminal processing plant based on the model calculation results.

8. The Internet of Things-based data management system for recycling renewable resources according to claim 7, characterized in that, The resource scheduling model includes: an attribute mapping layer: which performs feature mapping on the standard dry basis asset weight and quality grade data of each recycling node, and converts them into a supply state vector in a high-dimensional feature space; Demand guidance layer: Converts the input terminal processing plant production line material quality threshold demand into a target center vector, and establishes the convergence guidance direction of the supply state vector in the feature space; Constraint Transformation Layer: Parses the input decision matrix, transforms the resource survival control state into boundary constraint operators that act on the algorithm logic, and achieves dynamic pruning by removing node branches that do not meet the control state in real time from the dynamic scheduling task pool, thereby narrowing the search range of candidate solutions. Bidirectional optimization layer: Under the constraints of the boundary constraint operator, the homogenization deviation is measured by the spatial distance between the supply state vector and the target center vector, and the candidate solution that minimizes both path cost and quality deviation is locked through iterative calculation. Instruction generation layer: Extract the quality deviation of each recycling node in the candidate scheme, allocate the loading share using the deviation offset logic including the stability adjustment benchmark; achieve linear alignment between total quality and feeding requirements by matching the mixing ratio of positive and negative deviation resources, and after simulation verification and error fine-tuning, generate a delivery sequence instruction set including logistics trigger time, allocation path coordinates and meeting the feeding homogenization requirements of the terminal processing plant.

Images