Multi-cycle recycling resource multiplier effect optimization regulation system

Abstract

The application discloses a multi-period recycling resource multiplier effect optimization regulation system, and belongs to the technical field of resource recycling, which comprises the following steps: obtaining multi-period resource data by hierarchical sampling according to preset resource types and period standards, and obtaining pretreated multi-period resource data after processing; constructing an initial stock-flow dynamic model based on the law of conservation of matter, calculating the deviation of the pretreated multi-period resource data from the model predicted stock and correcting the deviation to obtain a corrected model; calculating core indexes by using the corrected model, evaluating resource utilization states and generating regulation scenarios; optimizing scenario parameters by using a genetic algorithm, screening out an optimal regulation strategy combination with the highest comprehensive benefit, and generating regulation instructions; receiving regulation effect data according to an execution period, and using the regulation effect data for model iteration correction after verification to form a closed-loop optimization; and the method improves the resource recycling efficiency, enhances the multiplier effect of the total resource supply, and realizes dynamic optimization regulation of multi-period resource utilization.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of resource recycling, and in particular relates to a resource multiplier effect optimization and regulation system for multi-period recycling. BACKGROUND

[0002] In the field of resource recycling, accurately assessing resource utilization status and optimizing and regulating resource multiplier effect are crucial for improving the ability of sustainable resource utilization. In existing resource recycling evaluation and regulation methods, the dynamic coupling relationship between the supply of recycled resources and the total demand is not quantified accurately in the accounting and evaluation stage, and the dynamic changes in the quality of renewable resources in the multi-period recycling process are not fully considered, resulting in poor scenario simulation and regulation execution effect based on this, and the resource multiplier effect is difficult to be effectively optimized.

[0003] Specifically, when using the stock-flow dynamic model, the existing method is not perfect in modifying the model, and the quality characteristics of renewable resources in the multi-period recycling process are not combined, so that the quantification result of the dynamic coupling relationship between the supply of recycled resources and the total demand deviates from the actual situation, and the accuracy of evaluation indexes such as sustainable utilization index, replacement amount of accumulated primary resources, and recycling resource multiplier effect is affected, resulting in a lack of reliable basis for subsequent scenario simulation, and the generated regulation strategy lacks pertinence and effectiveness. SUMMARY

[0004] In view of the deficiencies of the prior art, the application provides a resource multiplier effect optimization and regulation system for multi-period recycling, which obtains multi-period resource data by hierarchical sampling according to preset resource types and period standards, and obtains preprocessed multi-period resource data after processing. Based on the law of conservation of mass, an initial stock-flow dynamic model is constructed, the deviation between the preprocessed multi-period resource data and the model predicted stock is calculated by the least square method and is corrected to obtain a corrected model. The core index is calculated using the corrected model, the resource utilization status is evaluated, and a regulation scenario is generated. Genetic algorithm is used to optimize the scenario parameters, and the optimal regulation strategy combination with the highest comprehensive benefit is selected to generate regulation instructions. The regulation effect data is received according to the execution period, and is used for model iterative correction after verification to form a closed loop optimization. This method improves the efficiency of resource recycling, enhances the multiplier effect, and realizes dynamic optimization and regulation of multi-period resource utilization.

[0005] To achieve the above purpose, the application provides the following technical scheme:

[0006] The multi-period cyclic resource multiplier effect optimization regulation system comprises a data acquisition module, an accounting and evaluation module, a scenario simulation module and a regulation execution module; the data acquisition module comprises a period data acquisition unit and a feedback data receiving unit; the accounting and evaluation module comprises a model correction unit and a state evaluation unit; the scenario simulation module comprises a scenario generation unit and a strategy calculation unit; the regulation execution module comprises an instruction generation unit and an effect acquisition unit;

[0007] The period data acquisition unit acquires multi-period resource data, which is transmitted to the model correction unit. The model correction unit generates a corrected stock-flow dynamic model based on the multi-period resource data and outputs the model to the state evaluation unit to generate a resource utilization state evaluation result. Based on the resource utilization state evaluation result, the scenario generation unit generates a regulation scenario, and the strategy calculation unit generates an optimal regulation strategy combination. The instruction generation unit outputs a regulation instruction according to the optimal regulation strategy combination, and the effect acquisition unit collects regulation effect data and transmits the data back to the model correction unit through the feedback data receiving unit, forming a closed-loop optimization.

[0008] Specifically, the period data acquisition unit comprises a time sequence acquisition subunit and a multi-dimensional preprocessing subunit.

[0009] According to a preset resource type list and period division standard, the time sequence acquisition subunit acquires multi-period resource data according to a hierarchical sampling rule, and the multi-dimensional preprocessing subunit fills in missing values and corrects abnormal values of the acquired multi-period resource data, and then outputs the data to the model correction unit. The multi-period resource data includes resource stock data, flow data, loss data and associated influence data in each period. The flow data includes consumption and scrap data.

[0010] Specifically, the feedback data receiving unit comprises a real-time receiving subunit and a data verification subunit.

[0011] The real-time receiving subunit receives regulation effect data uploaded by the effect acquisition unit according to a regulation instruction execution period, and the data verification subunit performs integrity verification and consistency verification on the regulation effect data. The feedback data set that passes the verification is output to the model correction unit after being associated with the corresponding regulation instruction ID and period identifier. The regulation effect data includes resource utilization efficiency change value, cycle shortening rate and multiplier effect coefficient.

[0012] Specifically, the model correction unit comprises an initial model construction subunit and a dynamic iterative correction subunit.

[0013] The initial model construction subunit constructs an initial stock-flow dynamic model based on the law of conservation of mass and outputs the model to the dynamic iterative correction subunit.

[0014] The dynamic iterative correction subunit iteratively corrects the initial stock-flow dynamic model based on the preprocessed multi-period resource data and the feedback data set, including:

[0015] The least square method is used to calculate the deviation value of the resource data of each period and the predicted stock of the initial stock-flow dynamic model, and when the deviation value is greater than the preset deviation threshold, the parameter correction is triggered.

[0016] After completing 3 period data iterations, the AIC information criterion is used to verify the fitting degree of the initial stock-flow dynamic model, and when the AIC value drop value is greater than or equal to the preset maximum fluctuation value, the current initial stock-flow dynamic model parameter is saved as a reference, and the corrected stock-flow dynamic model is generated and output to the state assessment unit.

[0017] Specifically, when the initial model construction subunit constructs the initial stock-flow dynamic model based on the law of conservation of mass, the initial stock-flow dynamic model is:

[0018] The input resource quantity is multiplied by the conversion efficiency coefficient to obtain the converted effective resource input quantity; the converted effective resource input quantity is subtracted from the product of the output resource quantity and the loss coefficient to obtain the actual loss amount in the resource output process; the actual loss amount in the resource output process is further subtracted from the product of the current stock and the natural decay rate to obtain the stock change amount; and finally, the stock change amount is added to the product of the associated system influence quantity and the coupling coefficient to obtain the change rate of the resource stock.

[0019] Specifically, when the model correction unit generates the corrected stock-flow dynamic model, the Weibull distribution model and the Logistic distribution model are used to correct the initial stock-flow dynamic model, including:

[0020] A1: Based on the product resource life distribution data in the multi-period resource data, the maximum likelihood estimation method is used to adjust the shape parameter and the scale parameter of the Weibull distribution model;

[0021] A2: Based on the cyclic resource supply quantity data and the total demand quantity data in the multi-period resource data, the cyclic resource supply quantity prediction deviation and the total demand quantity prediction deviation of the previous period are used as correction factors to adjust the growth rate parameter and the saturation capacity parameter of the Logistic distribution model.

[0022] Specifically, the specific steps of A1 include:

[0023] A1.1: Obtain and preprocess resource life distribution data, extract resource life distribution data from multi-period resource data; the resource life distribution data is the duration record of resources from being put into use to failure in different periods;

[0024] A1.2: Classify the resource life distribution data, distinguish complete life data and censored data, and perform outlier rejection and missing value filling on the classified resource life distribution data, and output the preprocessed resource life distribution data; the censored data is the life record of resources that have not failed at the end of the observation period;

[0025] A1.3: Define the core function of Weibull distribution; the Weibull distribution is used to describe the probability distribution characteristics of resource life, including shape parameter and scale parameter; the shape parameter reflects the shape of resource life distribution; the scale parameter reflects the typical level of resource life;

[0026] A1.4: Define the probability density function and survival function of Weibull distribution;

[0027] A1.5: Construct the likelihood function, based on the preprocessed resource life distribution data and the defined probability density function and survival function of Weibull distribution, construct the likelihood function; the likelihood function is the product of the probability density function values corresponding to all complete life distribution data and the product of the survival function values corresponding to all censored data, which is used to reflect the fitting degree of the current shape parameter and scale parameter to the resource life distribution data, and output the initial likelihood function;

[0028] A1.6: Convert the initial likelihood function to log-likelihood function, take the natural logarithm of the initial likelihood function output by A1.5, convert the original product relationship to summation relationship, and obtain the log-likelihood function;

[0029] A1.7: Construct the likelihood equation set, take the partial derivative of the output log-likelihood function with respect to the shape parameter and scale parameter respectively, and set both partial derivatives to zero to obtain the likelihood equation set composed of two equations;

[0030] A1.8: Solve the likelihood equation set to obtain the estimated values of the shape parameter and the scale parameter, solve the output likelihood equation set by using numerical iteration method, including: first set the initial values of the shape parameter and the scale parameter, substitute the initial values into the likelihood equation set to calculate the residual, solve the parameter correction amount by matrix operation to update the parameters, and repeat the iteration until the parameter value change amount of two iterations is less than the preset convergence threshold, and output the estimated value of the shape parameter and the estimated value of the scale parameter;

[0031] A1.9: verifying the validity of the parameter estimates, based on the output shape parameter estimate and scale parameter estimate, verifying the validity of the parameter estimates, including: by checking whether the log-likelihood function is maximum at the parameter estimates, and using a goodness-of-fit test to compare the cumulative distribution of the Weibull distribution based on the parameter estimates with the actual life data;

[0032] If the difference is within the preset acceptance range, the parameter estimates are determined as valid shape parameters and valid scale parameters;

[0033] If not, return to A1.2 to reprocess the data or adjust the initial value of iteration and repeat A1.3-A1.9, output the valid shape parameters and valid scale parameters.

[0034] Specifically, the specific steps of the A2 include:

[0035] A2.1: obtaining the preprocessed multi-period resource data, and based on the preprocessed multi-period resource data, extracting the actual supply amount, the supply amount prediction value, the actual demand amount and the demand total amount prediction value of the previous period, calculating the cycle resource supply amount prediction deviation of the previous period and the demand total amount prediction deviation of the previous period;

[0036] The cycle resource supply amount prediction deviation of the previous period is the actual supply amount of the previous period minus the corresponding supply amount prediction value;

[0037] The demand total amount prediction deviation of the previous period is the actual demand amount of the previous period minus the corresponding demand total amount prediction value;

[0038] A2.2: based on the output cycle resource supply amount prediction deviation of the previous period and the demand total amount prediction deviation of the previous period, converting them into standardized correction factors to obtain the supply amount correction factor and the demand total amount correction factor; the supply amount correction factor is the cycle resource supply amount prediction deviation of the previous period divided by the absolute value of the actual supply amount of the previous period; the demand total amount correction factor is the demand total amount prediction deviation of the previous period divided by the absolute value of the actual demand amount of the previous period;

[0039] A2.3: defining the Logistic distribution model parameters, and explicitly defining that the Logistic distribution model contains the growth rate parameter and the saturation capacity parameter; the growth rate parameter is used to determine the speed of the cycle resource amount approaching the saturation capacity; the saturation capacity parameter is used to represent the maximum limit value that the cycle resource amount can reach;

[0040] A2.4: establishing parameter correction association rules, based on the supply and demand relationship of the cycle resource, establishing the correction association rules of the previous period prediction deviation on the growth rate parameter and the saturation capacity parameter; the correction association rules include:

[0041] If the supply amount prediction deviation of the previous cycle is positive or the total demand prediction deviation is positive, the growth rate parameter is increased, and vice versa;

[0042] If the supply amount prediction deviation of the previous cycle is positive and lasts for multiple cycles or the total demand prediction deviation is positive, the saturation capacity parameter is increased, and vice versa;

[0043] A2.5: Based on the supply amount correction factor and the total demand correction factor, and the correction association rule, a parameter correction formula is constructed; the parameter correction formula includes a growth rate parameter correction formula and a saturation capacity parameter correction formula;

[0044] A2.6: The adjusted growth rate parameter and the adjusted saturation capacity parameter are substituted into the Logistic distribution model to replace the initial growth rate parameter and the initial saturation capacity parameter, and an updated Logistic distribution model is output;

[0045] A2.7: Based on the updated Logistic distribution model, the supply amount and the total demand of the current cycle are predicted, and the actual supply amount and the actual total demand of the current cycle are collected, the supply amount prediction deviation and the total demand prediction deviation of the current cycle are calculated, and they are compared with the supply amount prediction deviation and the total demand prediction deviation of the previous cycle;

[0046] If the absolute value of the deviation of the current cycle is less than that of the previous cycle and within the preset acceptable range, the updated Logistic distribution model is determined as an effective model;

[0047] If it is not within the preset acceptable range, return to A2.2 to re-determine the correction factor, repeat A2.3 to A2.7, until an effective Logistic distribution model is obtained.

[0048] Specifically, the state evaluation unit includes an index calculation subunit and a grade evaluation subunit;

[0049] The index calculation subunit receives the corrected stock-flow dynamic model, and calculates resource utilization core indexes based on the resource stock-flow simulation results of each cycle output by the corrected stock-flow dynamic model; the resource utilization core indexes include resource recycling rate, accumulated primary resource consumption, and resource multiplier effect contribution value;

[0050] The grade evaluation subunit receives resource utilization core indexes, determines the weight of each index by using the analytic hierarchy process, and generates a resource recycling state evaluation result by fuzzy comprehensive evaluation; the resource recycling state evaluation result includes three grades of upgraded utilization, original grade utilization and secondary utilization, and corresponding improvement indexes, and is output to the scenario generation unit.

[0051] Specifically, the strategy calculation unit includes a parameter optimization subunit and a strategy combination subunit.

[0052] The parameter optimization subunit receives the regulation and control scenarios, and optimizes the regulation and control parameters under each regulation and control scenario based on the corrected stock-flow dynamic model by using a genetic algorithm; the regulation and control parameters include a resource recycling threshold, a regeneration utilization input coefficient and a cross-cycle allocation ratio.

[0053] The strategy combination subunit receives the optimal regulation and control parameters of each scenario, combines the grade evaluation in the resource recycling state evaluation result, screens out a parameter combination that meets the constraint boundary and has the highest comprehensive benefit, and generates an optimal regulation and control strategy combination; the optimal regulation and control strategy combination contains at least three regulation and control measures and corresponding implementation priorities, and is output to the instruction generation unit.

[0054] Compared with the prior art, the beneficial effects of the present application are:

[0055] 1. The present application proposes a multi-period recycling resource multiplier effect optimization regulation and control system, and optimizes and improves the architecture, operation steps and flow; the system has the advantages of simple flow, low investment and operation cost, and low production cost.

[0056] 2. The present application proposes a multi-period recycling resource multiplier effect optimization regulation and control system, which can accurately reflect the actual state of resource recycling utilization by obtaining multi-period resource data through hierarchical sampling and preprocessing, combining the law of conservation of mass and dynamically correcting the stock-flow model; based on the model calculation core index and the evaluation of resource utilization state, the regulation and control scenarios can be generated, and the optimal strategy can be screened out by optimizing the parameters through a genetic algorithm, so that the scientificity and accuracy of resource regulation and control are realized; at the same time, by receiving and verifying the regulation and control effect data according to the period, the data is used for model iteration correction, forming a complete closed loop, which can continuously improve the resource utilization efficiency, reduce the input of primary resources, enhance the resource multiplier effect, ensure the dynamic optimization of multi-period resource recycling utilization process, and effectively improve the sustainability and comprehensive benefit of resource utilization. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 It is the architecture diagram of the multi-period recycling resource multiplier effect optimization regulation and control system of the present application.

[0058] Figure 2The principle flow chart of the resource multiplier effect optimization regulation system for multi-period recycling utilization of the application. DETAILED DESCRIPTION

[0059] Embodiment 1

[0060] Please refer to Figure 1 An embodiment provided by the application is a multi-period recycling utilization resource multiplier effect optimization regulation system, comprising:

[0061] a data acquisition module, an accounting and evaluation module, a scenario simulation module, and a regulation execution module; the data acquisition module comprises a period data acquisition unit and a feedback data receiving unit; the accounting and evaluation module comprises a model correction unit and a state evaluation unit; the scenario simulation module comprises a scenario generation unit and a strategy calculation unit; the regulation execution module comprises an instruction generation unit and an effect acquisition unit.

[0062] The period data acquisition unit acquires multi-period resource data, which is transmitted to the model correction unit. The model correction unit generates a corrected stock-flow dynamic model based on the multi-period resource data and outputs it to the state evaluation unit to generate a resource utilization state evaluation result. Based on the resource utilization state evaluation result, the scenario generation unit generates a regulation scenario, and the strategy calculation unit generates an optimal regulation strategy combination. The instruction generation unit outputs a regulation instruction according to the optimal regulation strategy combination, and the effect acquisition unit collects regulation effect data and returns it to the model correction unit through the feedback data receiving unit, forming a closed-loop optimization.

[0063] Further, the scenario generation unit generates regulation scenarios including life regulation scenarios, cyclic growth scenarios, peak optimization scenarios, and component reduction scenarios. The life regulation scenarios are generated by setting different resource service life parameters. The cyclic growth scenarios are generated by setting different recycling rates. The component reduction scenarios are generated by setting the reduction proportion of each component in the resource, and the scenario generation unit generates regulation scenarios according to the weak links in the resource utilization state evaluation result.

[0064] Further, the scenario generation unit comprises a constraint condition setting subunit and a scenario library construction subunit.

[0065] The constraint condition setting subunit receives the resource utilization state evaluation result, extracts the key improvement indicators therein as constraint boundaries, and the constraint boundaries include an upper limit of resource loss rate, a lower limit of recycling period, and a lower limit of multiplier effect coefficient.

[0066] The scenario library construction subunit generates regulation scenarios based on the constraint boundaries, and the regulation scenarios include benchmark scenarios, reinforced recycling scenarios, and efficiency priority scenarios.

[0067] The benchmark scenario maintains the current resource circulation parameters unchanged, simulates the resource state change in the next 3 cycles;

[0068] The reinforced circulation scenario aims to shorten the circulation period, and sets the circulation period shortening rate to at least increase by 15%;

[0069] The efficiency priority scenario aims to improve the multiplier effect, and sets the multiplier effect coefficient to at least increase by 20%;

[0070] The scenario generation unit outputs the generated regulation scenario to the strategy calculation unit.

[0071] It needs to be explained that the resource multiplier effect optimization regulation of multi-cycle circulation is a systematic resource management strategy, the core of which is to build a multi-cycle resource circulation path, combined with real-time monitoring, dynamic evaluation and intelligent regulation means, so that resources are efficiently utilized in multiple circulation periods, so as to fully play the multiplier effect of resources, realize the maximization of resource value and sustainable utilization. Specifically, first, classify resources, and divide them into three levels of upgraded utilization, primary utilization and secondary utilization according to different resource circulation utilization states, such as secondary utilization of recycled resources with reduced quality, such as food-grade PET plastic bottles recycled into trays; then monitor resource circulation information through Internet of Things and other technologies, evaluate circulation efficiency and other indicators using big data and artificial intelligence, and improve resource multiplier effect. The resource multiplier effect refers to the amplification effect of the total value and resource output rate of resources through multiple recycling of resources. Simply put, a resource can produce a certain value and benefit through one cycle of utilization, and the total value and benefit produced will be amplified like a multiplier after multiple cycles of utilization, realizing one resource, multiple value. For example, a plastic waste, in primary utilization, is simply processed into plastic products of the same quality, producing a certain value; in upgraded utilization, it is deeply processed into high-performance plastic products, further amplifying the value; in secondary utilization, it can be made into hard plastic containers such as trays, thus realizing the bottom protection of resource value. The total value of the resource in this entire process far exceeds that of single utilization, which embodies the resource multiplier effect.

[0072] For example, the long-cycle recycling of copper / gold / indium metals in electronic waste is regulated. First, according to the preset resource type list, such as copper / gold / indium-containing waste circuit boards, chips, display screen modules, etc., and the division standard of every half a year, the metal resource data of multiple cycles is collected through full-process tracking: including: 1) the copper / gold / indium-containing waste inventory in the disassembly stage, such as copper-containing circuit board fragments, gold-containing pin waste, and indium-containing glass substrate scraps generated by disassembly; 2) the copper / gold / indium inventory contained in the electronic products in use, such as copper foil in the mainboard of the smart phone in use, gold-plated layer in the computer connector, and indium tin oxide film in the flat panel display screen; 3) the copper / gold / indium carrier inventory contained in the recycled electronic waste, such as scrapped server mainboard, old television display screen module, and retired communication equipment cable; at the same time, the flow data of metal resources in each cycle, the loss amount in the processing process and the influence data are recorded, wherein the flow data includes the amount of copper / gold / indium-containing waste generated by disassembly, the actual content of copper / gold / indium in recycled electronic waste, and the circulation amount of different grades of regenerated copper / gold / indium after purification processing, such as cathode copper, brass, high-purity gold powder, and industrial-grade indium ingot; the loss amount in the processing process includes the anode slime loss of copper in electrolytic purification, the residual loss of gold in chemical dissolution, the volatilization loss of indium in high-temperature treatment, and the metal carrier crushing loss during sorting; the influence data includes the average update cycle of copper / gold / indium-containing electronic products, the willingness of consumers to deliver electronic waste, and the demand difference of copper / gold / indium of different purities. After processing these data, some missing actual service life records of indium-containing display screens are filled in, and the purity detection deviation of abnormal batches of copper waste is corrected to obtain pre-processed multi-cycle metal resource data, which completely covers the long-cycle chain of in-use metal-containing electronic products-scraped electronic waste-disassembled metal-containing waste-purification processing of regenerated metal-reapplication.

[0073] Then, based on the law of conservation of mass, an initial stock-flow dynamic model is constructed, in which the model clearly divides four core stocks: the stock of disassembled classified metal-containing waste, the total amount of copper / gold / indium in the in-use electronic products in social circulation, the stock of copper / gold / indium in the electronic waste that has been retired but not yet recycled, and the stock of available metal materials after recycling into the purification / processing link; and the corresponding flow relationships are defined, including: the inflow rate of mixed metal waste into different processing links, such as the inflow rate of copper / gold / indium-containing waste into the high-purity purification link to produce primary / upgrade raw materials, and the inflow rate into the simple processing link to produce secondary raw materials; the outflow rate of recycled copper / gold / indium into different application fields, such as the primary utilization outflow rate of high-purity recycled copper / gold / indium for precision electronic components, the upgrade utilization outflow rate of high-performance recycled copper / gold / indium for high-end products, and the secondary utilization outflow rate of ordinary purity recycled copper / gold / indium for basic components; the inflow rate of retired electronic products into the stock of metals to be recycled; and the inflow rate of metals to be recycled into the purification link, i.e., the metal recycling rate. Combined with the pretreated full-chain data, the actual metal stock is calculated using the least squares method, such as the deviation between the actual output of high-purity copper / gold / indium in any period and the model predicted stock, focusing on the stripping rate of copper / gold / indium in in-use electronic products, the processing conversion rate of copper / gold / indium of different purity, and other parameters to obtain a revised model that better fits the actual long-period cycle.

[0074] Using the revised model, the core indicators are calculated, including: 1) metal recycling rate, i.e., the proportion of the actual amount of pure copper / gold / indium recovered from electronic waste in any period to the total amount of corresponding metal in all retired electronic products in that period; 2) accumulated primary resource consumption, i.e., the total amount of primary metal that needs to be invested to meet the demand for copper / gold / indium in the electronic manufacturing industry in that period, minus the amount of recycled metal; and 3) multiplier effect contribution value, i.e., the ratio of the value created by recycled metal through primary / upgrade utilization to the recycling and processing cost.

[0075] According to the evaluation results, a regulation scenario is generated to improve the proportion of high-value utilization of copper / gold / indium, and the regulation parameters are optimized by genetic algorithm to screen out the strategy combination with the highest comprehensive benefit: first, for the sorting and purification link, introduce AI intelligent sorting system and high-precision spectral detection equipment to improve the sorting purity and efficiency of copper / gold / indium-containing waste, and provide high-quality raw materials for primary / upgrading utilization; second, for the copper processing link, develop high-purity recycled copper / gold / indium for primary utilization of precision electronic components to reduce the proportion of downgrading utilization; third, for the recovery link of gold and indium, establish a directional recovery network, such as setting up an old machine recycling point with electronic product manufacturers, and optimize the purification process of copper / gold / indium simultaneously to improve the overall output rate of high-purity metals and determine the priority of implementation. The regulation instructions are generated in the preset format and issued to the enterprise disassembly workshop, purification workshop and cooperating electronic manufacturers for execution.

[0076] At the same time, during the execution of the instructions, the regulation effect data is collected periodically, such as the metal recognition accuracy of the intelligent sorting system, the pass rate of high-purity recycled copper / gold / indium in precision components, and the improvement range of the directional recovery network, etc. After checking the data integrity and consistency, they are fed back to the model correction link, focusing on updating the processing and conversion efficiency of different grades of metals, the loss coefficient of each link and other parameters for the next iteration of the model. Through this closed loop of continuous optimization, the primary / upgrading utilization ratio of copper / gold / indium is improved synchronously, the overall recovery efficiency of high-purity is improved, the value loss in the long-term cycle of metals is reduced, and the multiplier effect of scarce metal resources is significantly enhanced.

[0077] The periodic data collection unit includes a time sequence collection subunit and a multi-dimensional preprocessing subunit;

[0078] According to the preset resource type list and periodic division standard, the time sequence collection subunit acquires multi-period resource data according to the hierarchical sampling rule, and the multi-dimensional preprocessing subunit preprocesses the acquired multi-period resource data; the multi-period resource data includes resource stock data, flow data, loss data and associated influence data in each period; the flow data includes consumption and scrap data.

[0079] Further, the time sequence collection subunit executes:

[0080] (1) According to the N resource identification codes in the resource type list, the collection period is divided into T equal length period windows according to the periodic division standard;

[0081] (2) For each type of resource, sample points are extracted according to the hierarchical sampling rule: first, divide the resource space into levels, and then perform secondary sampling in each level according to the resource flow rate interval;

[0082] (3) In each cycle window, four basic data are collected in real time by the Internet of Things terminal device: inventory data are obtained by using high-precision weighing sensors, flow data (consumption and scrap data) are collected by using turbine flow meters, loss data are determined by using infrared spectrum analyzers, and correlation influence data are obtained from the correlation system through the industrial bus protocol.

[0083] Further, the execution steps of the multi-dimensional preprocessing subunit include:

[0084] (1) The collected raw data are subjected to four-order processing: first-order outlier rejection, second-order time-space alignment by sliding time window interpolation, third-order unit unification, and fourth-order feature fusion to generate an eight-dimensional feature vector containing time-space labels;

[0085] (2) A cycle-resource two-dimensional matrix storage structure is established, the matrix row vector represents the cycle sequence, the column vector contains the four basic data after preprocessing, and a time stamp check code is attached.

[0086] Embodiment 2

[0087] Please refer to Figure 2 The overall implementation process of the multi-cycle recycling resource multiplier effect optimization and control system in this embodiment includes:

[0088] (1) The time sequence sampling subunit of the cycle data acquisition unit acquires multi-cycle resource data by using hierarchical sampling according to the preset resource type list and cycle division standard, and then the multi-dimensional preprocessing subunit performs missing value filling and outlier correction on the multi-cycle resource data to output the preprocessed multi-cycle resource data;

[0089] (2) The initial model construction subunit constructs an initial inventory-flow dynamic model based on the law of conservation of mass, and uses the least squares method to calculate the deviation between the preprocessed multi-cycle resource data and the predicted inventory of the initial inventory-flow dynamic model and corrects it to obtain a corrected inventory-flow dynamic model;

[0090] (3) The core indicators are calculated based on the corrected inventory-flow dynamic model to evaluate the resource utilization state, and the control scenarios are generated according to the resource recycling state evaluation results;

[0091] (4) The parameters of the control scenarios are optimized using a genetic algorithm, the strategy combination with the highest comprehensive benefit is selected, the optimal control strategy combination is generated, and the control instructions are generated in the preset format;

[0092] (5) The real-time receiving subunit of the feedback data receiving unit receives the regulation effect data uploaded by the effect collecting unit according to the regulation instruction execution cycle, the data verification subunit performs integrity and consistency verification thereon, generates a feedback data set that passes the verification, and outputs the feedback data set to the dynamic iterative correction subunit of the model correction unit, for the next iteration correction of the stock-flow dynamic model, forming a closed-loop optimization.

[0093] The feedback data receiving unit comprises a real-time receiving subunit and a data verification subunit.

[0094] The regulation effect data uploaded by the effect collecting unit is received by the real-time receiving subunit according to the regulation instruction execution cycle, and the regulation effect data is subjected to integrity verification and consistency verification by the data verification subunit, a feedback data set that passes the verification is generated, and the corresponding regulation instruction ID and cycle identifier are associated; the regulation effect data comprises a resource utilization efficiency change value, a cycle period shortening rate, and a multiplier effect coefficient.

[0095] Further, the working process of the real-time receiving subunit is as follows:

[0096] (1) A dual-channel receiving mechanism is established: the main channel receives the effect data packet in real time through the MQTT protocol, and the standby channel synchronously backs up the data through the blockchain node;

[0097] (2) At the end of each regulation instruction execution cycle, a data packet reorganization operation is triggered: discrete effect data is indexed according to the [instruction ID + cycle identifier + timestamp] triplet to form a structured effect record.

[0098] Further, the data verification subunit implements three-level verification:

[0099] (1) The first-level verification checks the data integrity: whether each effect record contains the required six fields is verified;

[0100] (2) The second-level verification implements logical consistency verification: invalid data is filtered through a pre-set constraint rule library;

[0101] (3) The third-level verification performs space-time correlation verification: the current effect data is correlated with the same-instruction data of the previous three cycles in a sliding window, and data deviating from the mean value by 2.5 times the standard deviation is removed.

[0102] The model correction unit comprises an initial model construction subunit and a dynamic iterative correction subunit.

[0103] The initial model construction subunit constructs an initial stock-flow dynamic model based on the law of conservation of mass and outputs the model to the dynamic iterative correction subunit.

[0104] Further, assuming the object of action is copper, the construction process of the initial stock-flow dynamic model includes:

[0105] (1) Clear the boundaries and goals of the model, including: according to the hierarchical division of resources, such as the basic layer, the advanced layer and the core layer, and different resource recycling states, such as upgraded utilization, primary utilization, secondary utilization, determine the resource range covered by the model, and clearly include the recycling links, such as recycling, processing, recycling and related systems; The modeling goal focuses on simulating the impact of resource stock changes on multiplier effects over multiple periods, and quantifying the effect of flow parameters on recycling efficiency;

[0106] For example, the upgrading of brass to cathode copper is upgraded utilization, the direct reuse of cathode copper with purity is primary utilization, and the processing of cathode copper to brass is secondary utilization.

[0107] (2) After determining the boundaries and goals, identify the core stock variables, flow variables and conversion relationships;

[0108] Among them, the stock is the amount of resources accumulated in the system, which can reflect the state of the system, and needs to be split according to the recycling links of resources. In the basic layer resource, there is a stock of cathode copper to be processed in the secondary utilization process, which is used to produce brass, and a stock of brass products produced in the secondary utilization process, such as brass pipe fittings; In the advanced layer resource, there is a stock of cathode copper in the electrolytic refining state in the primary utilization process, and a stock of high-purity cathode copper produced in the primary utilization process; In the core layer resource, there is a stock of waste brass to be purified for upgrading utilization, and a stock of cathode copper produced in the upgrading utilization, and also considering the cumulative stock of copper loss that cannot be recycled in each cycle, such as copper residue in the anode mud.

[0109] Among them, the flow is the rate of changing the stock, which is divided into inflow and outflow, and is related to the specific process link. The inflow refers to the input rate of cathode copper into the secondary utilization system for smelting brass, the input rate of crude copper into the primary utilization electrolysis link, and the input rate of brass scrap into the upgrading utilization purification link. The outflow refers to the loss rate of brass slag produced in the secondary utilization process, the elimination rate of substandard cathode copper in the primary utilization process, and the output rate of cathode copper in the upgrading utilization process. The conversion efficiency coefficient can be used to establish the relationship between flow and stock, for example, the output rate of cathode copper in the primary utilization depends on the stock of crude copper to be refined and the electrolytic conversion efficiency, and the output rate of cathode copper in the upgrading utilization depends on the stock of brass scrap to be purified and the purification conversion efficiency.

[0110] (3) Based on the law of conservation of matter, the correlation between variables is sorted out, and the feedback relationship is identified. Positive feedback will form an enhancement cycle, such as the increase in the inventory of cathode copper produced by upgrading utilization will increase product acceptance, thereby increasing the recycling input of brass scrap, and further increasing the inventory of upgrading utilization; negative feedback will form a balance cycle, such as when the inventory of cathode copper in the electrolytic refining state in primary utilization is too high, the electrolytic tank capacity will tend to be saturated, thereby reducing the refining rate and keeping the inventory stable. The influence relationship between these variables will form a loop, which will jointly affect the dynamic change of the system;

[0111] (4) Establish the dynamic change rule of inventory-flow;

[0112] Further, the change rate of resource inventory in any period is jointly determined by the upstream inflow rate, the downstream outflow rate, the loss rate, and the cross-period allocation rate. The upstream inflow rate refers to the copper resource amount entering the corresponding utilization state in the current period from the recycling link, such as the inflow of cathode copper into secondary utilization, the inflow of crude copper into primary utilization, and the inflow of brass scrap into upgrading utilization. The downstream outflow rate refers to the copper resource amount entering the next link from the corresponding utilization state in the current period, such as the outflow of brass products produced by secondary utilization, the outflow of high-purity cathode copper produced by primary utilization, and the outflow of cathode copper produced by upgrading utilization. The loss rate is the amount of copper that cannot continue to circulate in each utilization state in the current period, which is related to the inventory and loss coefficient of the corresponding state, such as the smelting loss coefficient of secondary utilization and the purification slag loss coefficient of upgrading utilization. The cross-period allocation rate is the amount of copper resource transferred between different utilization states, such as the transfer of substandard cathode copper in primary utilization to secondary utilization, i.e. processing into brass, or the transfer of high-purity brass in degradation utilization to upgrading utilization, i.e. purification into cathode copper. At the same time, initial values are assigned to parameters such as loss coefficient and conversion efficiency for each state based on historical data or industry standards of different utilization states.

[0113] (5) Check whether key inventory or flow is missed through boundary testing to ensure that the full process of multi-period circulation including upgrading utilization, primary utilization, and secondary utilization is covered; perform reasonableness check, such as through extreme value testing, when the input rate of brass scrap in upgrading utilization is zero, the inventory of brass scrap to be purified in this state should gradually decrease with the outflow of cathode copper and its own loss, thereby verifying the equation logic, and finally outputting the initial inventory-flow dynamic model to the dynamic iteration correction sub-unit, thereby realizing accurate quantification and control of resource multiplier effect.

[0114] The dynamic iteration correction sub-unit iteratively corrects the initial inventory-flow dynamic model based on the pre-processed multi-period resource data and feedback data set, including:

[0115] The least square method is used to calculate the deviation value of the resource data of each period and the predicted stock of the initial stock-flow dynamic model. When the deviation value is greater than the preset deviation threshold, the parameter correction is triggered. The least square method is prior art content in the field and is not the creative scheme of the present application, and thus is not described herein.

[0116] After each 3-cycle data iteration is completed, the initial stock-flow dynamic model fitting degree is verified by the AIC information criterion. When the AIC value drop value is greater than or equal to the preset fluctuation maximum value, the current initial stock-flow dynamic model parameter is saved as a reference, the corrected stock-flow dynamic model is generated, and is output to the state evaluation unit. Specifically, the method comprises the following steps:

[0117] Firstly, the data range of three periods is determined. After completing three consecutive resource cycle periods, the start and end times of the three periods are determined to ensure that the full process of resource flow from entering the cycle to completing the flow is covered in each period. At the same time, the actual resource data in the three periods is extracted from the system database, including the resource inventory, flow, and loss at each time node. These data need to correspond completely with the prediction dimensions of the initial inventory-flow dynamic model to ensure that each actual data can find the corresponding prediction index in the model. Then, the prediction data of the initial inventory-flow dynamic model in the three periods is obtained. The initial inventory-flow dynamic model is called, and the initial parameters such as conversion efficiency and loss coefficient of the three periods are input to generate the prediction results of resource inventory, flow, and loss at each time node. The prediction results are aligned with the actual data of the same period according to the time stamp to form a corresponding data set of actual value-prediction value, ensuring that the time granularity of the data is consistent and avoiding calculation bias caused by time misalignment. Then, the AIC value of the initial model in the current three periods is calculated. First, the likelihood function value of the initial inventory-flow dynamic model is calculated based on the actual value-prediction value data set. The likelihood function is used to measure the fitting degree of the model to the actual data. The closer the prediction value is to the actual value, the larger the likelihood function value. When calculating the likelihood function, it is assumed that the prediction error follows a normal distribution. The error value is the difference between the actual value and the prediction value. The overall likelihood is calculated through the distribution probability of the error. Then, the number of parameters to be estimated in the initial inventory-flow dynamic model is counted. These parameters include conversion efficiency coefficient, loss coefficient, cross-period allocation ratio, and all variables that need to be dynamically adjusted in model operation. Finally, according to the definition of AIC information criterion, the current AIC value is calculated to ensure the fitting effect of the model to the data and avoid overfitting caused by too many parameters. Then, the baseline AIC value of the last verification is determined. If it is the first 3-period verification, the last baseline AIC value is the AIC value when the model is initially constructed. If it has been verified for many times, the last baseline AIC value is the AIC value saved during the last 3-period verification. The difference between the current AIC value and the last baseline AIC value is calculated, which is the AIC drop value. When the current AIC value is less than the last baseline value, the drop value is positive. The larger the value, the more significant the improvement of the model fitting degree. Then, the AIC drop value is compared with the preset maximum fluctuation value, which is a threshold set based on historical verification data and business requirements to judge whether the model has improved significantly. The value should be set according to the normal fluctuation range of AIC value in the past 3-period iteration to ensure that the model needs to be corrected only when the improvement of the model fitting degree exceeds the normal fluctuation. If the current AIC drop value is greater than or equal to the maximum fluctuation value, it means that the fitting effect of the model in the three periods has improved significantly compared with the previous one, reaching the correction standard.Finally, save the current model parameters and generate the revised model, when the revision condition is met, extract all parameters used by the initial model in the current 3 period operation, such as conversion efficiency coefficient, loss coefficient, etc., mark these parameters as reference parameters, associate the corresponding period identifier and timestamp and store them in the system parameter library to ensure traceability, at the same time, update the parameter configuration of the initial inventory-flow dynamic model based on these reference parameters, generate the revised inventory-flow dynamic model, and use the revised inventory-flow dynamic model as the initial model of the next 3 period iteration for subsequent resource prediction and regulation, in addition, record the full flow data of this verification, including actual data range, AIC calculation process, comparison result of drop value, saved parameter list, etc., form a verification log to provide reference for subsequent iteration.

[0118] Further, the calculation process of the AIC value is: first, calculate the natural logarithm of the likelihood function of the model to the actual data, multiply the logarithm result by negative two, and then add the number of parameters to be estimated in the model multiplied by two, and finally the obtained value is the AIC value, wherein the natural logarithm of the likelihood function is used to reflect the fitting tightness of the model to the actual data, and the number of parameters is used to control the complexity of the model, and the combination of the two can not only reflect the fitting effect of the model, but also avoid the overfitting of the model to local data due to too many parameters.

[0119] When the initial model construction subunit constructs the initial inventory-flow dynamic model based on the law of conservation of mass, the initial inventory-flow dynamic model is:

[0120] The input resource quantity is multiplied by the conversion efficiency coefficient to obtain the effective resource input quantity after conversion; the effective resource input quantity after conversion is subtracted from the product of the output resource quantity and the loss coefficient to obtain the actual loss quantity in the resource output process; the actual loss quantity in the resource output process is further subtracted from the product of the current inventory and the natural decay rate to obtain the inventory change quantity; finally, the inventory change quantity is added to the product of the associated system influence quantity and the coupling coefficient to obtain the change rate of the resource inventory.

[0121] When the model revision unit generates the revised inventory-flow dynamic model, the Weibull distribution model and the Logistic distribution model are used to revise the initial inventory-flow dynamic model, including:

[0122] A1: based on the product resource life distribution data in the multi-period resource data, the maximum likelihood estimation method is used to adjust the shape parameter and the scale parameter of the Weibull distribution model;

[0123] Wherein, the shape parameter reflects the shape of the life distribution, such as whether to accelerate failure over time; the scale parameter reflects the typical level of life.

[0124] A2: Based on the cycle resource supply data and the total demand data in the multi-period resource data, the cycle resource supply prediction deviation and the total demand prediction deviation of the previous period are taken as correction factors to adjust the growth rate parameter and the saturation capacity parameter of the Logistic distribution model.

[0125] The specific steps of A1 include:

[0126] A1.1: Obtain and preprocess resource life distribution data, extract resource life distribution data from multi-period resource data; the resource life distribution data is the duration record of resources from being put into use to failure in different periods;

[0127] A1.2: Classify the resource life distribution data, distinguish complete life data and truncated data, and perform outlier rejection and missing value filling on the classified resource life distribution data, and output the preprocessed resource life distribution data, wherein the complete life data is the life record of resources that have failed within the observation period, and the truncated data is the life record of resources that have not failed at the end of the observation period;

[0128] A1.3: Define the core function of Weibull distribution; the Weibull distribution is used to describe the probability distribution characteristics of resource life, including shape parameter and scale parameter, wherein the shape parameter reflects the shape of resource life distribution, and the scale parameter reflects the typical level of resource life;

[0129] A1.4: Define the probability density function and survival function of Weibull distribution; the probability density function is used to describe the failure probability density of resources at a certain life time, and its characteristic is that when the life is close to the scale parameter, the change trend of the failure probability density is determined by the shape parameter; the survival function is used to describe the probability that the resource life exceeds a predetermined time, and its characteristic is that it monotonically decreases with the extension of life, and is the integral of the probability density function from the current life to infinity;

[0130] A1.5: Construct the likelihood function, based on the preprocessed resource life distribution data and the defined probability density function and survival function of Weibull distribution, construct the likelihood function; the likelihood function is the product of the probability density function values corresponding to all complete life distribution data and the product of the survival function values corresponding to all truncated data, which is used to reflect the fitting degree of the current shape parameter and scale parameter to the resource life distribution data, and outputs the initial likelihood function, wherein the likelihood function is the prior art content in the field, and is not the inventive scheme of the present application, which will not be described here;

[0131] A1.6: converting the initial likelihood function into a log-likelihood function, taking a natural logarithm of the output initial likelihood function, converting the original product relationship into a summation relationship, and obtaining the log-likelihood function; the log-likelihood function is a sum of log values of probability density functions corresponding to all complete life data, and a sum of log values of survival functions corresponding to all censored data, and outputting the log-likelihood function;

[0132] A1.7: constructing a likelihood equation set, respectively taking partial derivatives of the output log-likelihood function with respect to the shape parameter and the scale parameter, and setting both partial derivatives to 0 to obtain a likelihood equation set composed of two equations; wherein when taking the partial derivative with respect to the shape parameter, the log life value of the complete life data, the influence term of the shape parameter in the probability density function, and the influence of the shape parameter on the survival function in the censored data are combined; when taking the partial derivative with respect to the scale parameter, the influences of the scale parameter on the probability density function and the survival function in the complete life data and the censored data are combined, and outputting the likelihood equation set;

[0133] A1.8: solving the likelihood equation set to obtain the estimated values of the shape parameter and the scale parameter, and solving the output likelihood equation set by using a numerical iteration method, including: setting initial values for the shape parameter and the scale parameter, substituting the initial values into the likelihood equation set to calculate the residual, solving the parameter correction amount by matrix operation to update the parameters, and repeating the iteration until the parameter value change amount of two iterations is less than a preset convergence threshold, and outputting the estimated value of the shape parameter and the estimated value of the scale parameter;

[0134] A1.9: verifying the effectiveness of the parameter estimates, verifying the effectiveness of the parameter estimates based on the output shape parameter estimate and scale parameter estimate, including: checking whether the log-likelihood function is the maximum value at the parameter estimate, and comparing the cumulative distribution difference between the Weibull distribution based on the parameter estimate and the actual life data by using the goodness-of-fit test;

[0135] If the difference is within the preset acceptance range, the parameter estimate is determined to be an effective shape parameter and an effective scale parameter;

[0136] If not, return to A1.2 to reprocess the data or adjust the iteration initial value and repeat A1.3-A1.9, and output the effective shape parameter and the effective scale parameter.

[0137] The specific steps of the A2 include:

[0138] A2.1: obtaining the preprocessed multi-period resource data, and extracting the actual supply amount, the supply amount prediction value, the actual demand amount, and the demand total amount prediction value of the previous period based on the preprocessed multi-period resource data, and calculating the cycle resource supply amount prediction deviation of the previous period and the demand total amount prediction deviation of the previous period;

[0139] The cycle resource supply amount prediction deviation of the previous period is the actual supply amount of the previous period minus the corresponding supply amount prediction value;

[0140] The demand total amount prediction deviation of the previous period is the actual demand amount of the previous period minus the corresponding demand total amount prediction value;

[0141] The cycle resource supply amount prediction value and the demand total amount prediction value are prediction results of the supply amount and the demand amount of the previous period for the current period.

[0142] A2.2: Based on the cycle resource supply amount prediction deviation of the previous period and the demand total amount prediction deviation of the previous period, the deviations are converted into standardized correction factors to obtain a supply amount correction factor and a demand total amount correction factor; the supply amount correction factor is the cycle resource supply amount prediction deviation of the previous period divided by the absolute value of the actual supply amount of the previous period; the demand total amount correction factor is the demand total amount prediction deviation of the previous period divided by the absolute value of the actual demand amount of the previous period;

[0143] It should be noted that if the actual supply amount or the actual demand amount of the previous period is zero, the denominator in the supply amount correction factor uses the historical average actual supply amount, and the denominator in the demand total amount correction factor uses the historical average actual demand amount.

[0144] A2.3: Define the Logistic distribution model parameters, and specify that the Logistic distribution model includes a growth rate parameter and a saturation capacity parameter; the growth rate parameter is used to determine the speed at which the cycle resource amount approaches the saturation capacity; the saturation capacity parameter is used to represent the maximum limit value that the cycle resource amount can reach;

[0145] The initial value of the growth rate parameter is set based on historical cycle resource growth data; the initial value of the saturation capacity parameter is set based on the maximum recyclable resource amount determined based on technical constraints and market constraints.

[0146] A2.4: Establish a parameter correction association rule, and based on the supply and demand relationship of the cycle resource, establish a correction association rule of the growth rate parameter and the saturation capacity parameter based on the prediction deviation of the previous period; the correction association rule includes:

[0147] If the cycle resource supply amount prediction deviation of the previous period is positive or the demand total amount prediction deviation is positive, the growth rate parameter is increased, and vice versa;

[0148] If the cycle resource supply amount prediction deviation of the previous period is positive and lasts for multiple periods or the demand total amount prediction deviation is positive, the saturation capacity parameter is increased, and vice versa;

[0149] A2.5: based on the supply amount correction factor and the total demand amount correction factor, and the correction association rule; the parameter correction formula includes a growth rate parameter correction formula and a saturation capacity parameter correction formula, wherein the growth rate parameter correction formula is that the adjusted growth rate parameter is equal to the initial growth rate parameter plus the product of the supply amount correction factor and the supply weight, and plus the product of the total demand amount correction factor and the demand weight, and the saturation capacity parameter correction formula is that the adjusted saturation capacity parameter is equal to the initial saturation capacity parameter plus the product of the supply amount correction factor and the supply saturation weight, and plus the product of the total demand amount correction factor and the demand saturation weight; the adjusted growth rate parameter and the adjusted saturation capacity parameter are calculated through the parameter correction formula, and the adjusted growth rate parameter and the adjusted saturation capacity parameter are output;

[0150] A2.6: the adjusted growth rate parameter and the adjusted saturation capacity parameter are substituted into the Logistic distribution model to replace the initial growth rate parameter and the initial saturation capacity parameter, and an updated Logistic distribution model is output;

[0151] A2.7: based on the updated Logistic distribution model, the supply amount and the total demand amount of the current cycle of the circulating resource are predicted, at the same time, the actual supply amount and the actual total demand amount of the current cycle of the circulating resource are collected, the supply amount prediction deviation and the total demand amount prediction deviation of the current cycle are calculated, and they are compared with the supply amount prediction deviation and the total demand amount prediction deviation of the previous cycle of the circulating resource;

[0152] If the absolute value of the deviation of the current cycle is smaller than that of the previous cycle and is within the preset acceptable range, the updated Logistic distribution model is determined as an effective model;

[0153] If it is not within the preset acceptable range, return to A2.2 to re-determine the correction factor, repeat A2.3 to A2.7, and until an effective Logistic distribution model is obtained.

[0154] The state evaluation unit includes an index calculation subunit and a grade evaluation subunit;

[0155] B1: the index calculation subunit receives the corrected stock-flow dynamic model, and calculates resource utilization core indexes based on the resource stock-flow simulation results of each cycle output by the corrected stock-flow dynamic model; the resource utilization core indexes include resource recycling rate, accumulated primary resource consumption, and multiplier effect contribution value;

[0156] Further, the specific steps of B1 include:

[0157] Step (1): receiving the corrected stock-flow dynamic model, extracting core data from the simulation results of resource stock-flow in each cycle output by the corrected stock-flow dynamic model; the core data includes renewable resource amount in each cycle, total resource consumption, actual cycle period, theoretical optimal cycle period, benefit generated by resource recycling, initial investment of resource recycling, wherein the renewable resource amount is the resource amount entering the utilization link again through recycling and recycling in each cycle, the total resource consumption is the total resource amount used for production or use in each cycle, the actual cycle period is the actual time from the input use to the completion of one cycle and the input use again, the theoretical optimal cycle period is the shortest feasible cycle time based on the technical level and resource characteristics, the benefit generated by resource recycling is the positive benefit generated by resource recycling, and the initial investment of resource recycling is the investment for resource recycling, processing and recycling;

[0158] Step (2): classifying and arranging the core data by cycle, and outputting a structured index calculation data set;

[0159] Step (3): based on the structured index calculation data set, extracting the circulating resource amount and the total resource consumption in each cycle, and calculating the resource recycling rate in each cycle; the resource recycling rate is the proportion of the circulating resource amount in the total resource consumption in any cycle;

[0160] It should be noted that in the calculation, the circulating resource amount is the resource amount actually participating in recycling in the cycle, the total resource consumption covers all resources input for use in the cycle, and the resource recycling rate in each cycle is output.

[0161] Step (4): based on the output structured index calculation data set, extracting the total resource demand and the resource recycling amount in each cycle, and calculating the accumulated virgin resource consumption in each cycle; the accumulated virgin resource consumption is the difference between the total resource demand and the resource recycling amount in any cycle, that is, the total amount of virgin resources input to meet the total resource demand under the premise of resource multi-cycle recycling;

[0162] Step (5): based on the output structured index calculation data set, extracting the benefit generated by resource recycling and the initial investment of circulating resource utilization in each cycle, and calculating the multiplier effect contribution value in each cycle; the multiplier effect contribution value is the ratio of the benefit generated by resource recycling to the initial investment of circulating resource utilization in a certain cycle;

[0163] It should be noted that in the calculation, the benefit generated by resource recycling includes direct economic benefit and indirect benefit, the initial investment of resource recycling includes fixed asset investment and operating cost, and long-term investment is allocated by cycle, and the multiplier effect contribution value in each cycle is output.

[0164] Step (6): The output of each cycle of resource recycling rate, each cycle of accumulated primary resource consumption, each cycle of multiplier effect contribution value is summarized by period to form a set of resource utilization core indicators of each cycle;

[0165] Step (7): The resource utilization core indicator set is checked for logical consistency and numerical reasonableness of each indicator, and if there is an exception, return to step (1) to extract key data or check the calculation logic of steps (2) to (4) and correct until the verification is passed. The verified resource utilization core indicator set of each cycle is output as the input of the state assessment unit for resource utilization state level assessment.

[0166] B2: The level assessment subunit receives resource utilization core indicators, determines the weight of each indicator using the analytic hierarchy process, and generates resource utilization state assessment results through fuzzy comprehensive evaluation; The resource recycling state assessment results include three levels of upgraded utilization, primary utilization and secondary utilization, and corresponding improvement indicators, and are output to the scenario generation unit. The analytic hierarchy process and fuzzy comprehensive evaluation are prior art in the art and are not the inventive scheme of the present application, and will not be described here.

[0167] The strategy calculation unit includes a parameter optimization subunit and a strategy combination subunit;

[0168] The parameter optimization subunit receives the control scenario, and based on the corrected stock-flow dynamic model, uses a genetic algorithm to optimize the control parameters under each control scenario; The control parameters include resource recovery threshold, regeneration utilization input coefficient, and cross-cycle allocation ratio;

[0169] The strategy combination subunit receives the optimal control parameters of each scenario, combines the level assessment in the resource recycling state assessment result, selects the parameter combination that meets the constraint boundary and has the highest comprehensive benefit, and generates the optimal control strategy combination; The optimal control strategy combination contains at least three control measures and corresponding implementation priorities, and is output to the instruction generation unit.

[0170] Further, the instruction generation unit includes an instruction encoding subunit and a delivery scheduling subunit;

[0171] The instruction encoding subunit receives the optimal control strategy combination and generates control instructions according to the preset instruction format; The control instructions include strategy ID, implementation cycle, responsible subject, operation parameter, and evaluation index; The operation parameter corresponds to the control parameter output by the strategy calculation unit;

[0172] The dispatching subunit transmits the control instruction to the corresponding terminal by encryption according to the execution terminal list associated with the resource type, and generates a dispatching log containing an instruction ID, a dispatching time and a terminal identification, and synchronizes to the effect collecting unit.

[0173] The embodiments of the present application are described above with reference to the drawings; however, the present application is not limited to the specific embodiments described above, but the specific embodiments described above are merely illustrative rather than restrictive, and a person of ordinary skill in the art can make changes, modifications, replacements and variations to the above-described embodiments without departing from the purpose of the present application and the scope of protection, and these are all within the protection of the present application.

Claims

1. A multi-cycle, cyclically utilized resource multiplier effect optimization and regulation system, characterized by, Comprise: Data acquisition module, accounting and evaluation module, scenario simulation module and control execution module; the data acquisition module comprises a periodic data acquisition unit and a feedback data receiving unit; the accounting and evaluation module comprises a model correction unit and a state evaluation unit; the scenario simulation module comprises a scenario generation unit and a strategy calculation unit; the control execution module comprises an instruction generation unit and an effect acquisition unit; Through the periodic data acquisition unit, multi-period resource data is obtained and transmitted to the model correction unit. The model correction unit generates a corrected stock-flow dynamic model based on the multi-period resource data and outputs it to the state evaluation unit to generate a resource utilization state evaluation result. Based on the resource utilization state evaluation result, the scenario generation unit generates a control scenario, and the strategy calculation unit generates an optimal control strategy combination. The instruction generation unit outputs control instructions according to the optimal control strategy combination, and the effect acquisition unit collects control effect data and returns it to the model correction unit through the feedback data receiving unit, forming a closed-loop optimization; The model correction unit comprises an initial model construction subunit and a dynamic iterative correction subunit; The initial model construction subunit constructs an initial stock-flow dynamic model based on the law of conservation of mass and outputs it to the dynamic iterative correction subunit; The dynamic iterative correction subunit iteratively corrects the initial stock-flow dynamic model based on preprocessed multi-period resource data and a feedback data set, including: Using the least squares method to calculate the deviation value of each period resource data and the initial stock-flow dynamic model prediction stock, when the deviation value is greater than the preset deviation threshold, triggering parameter correction; After completing 3 period data iterations, verify the initial stock-flow dynamic model fitting degree through AIC information criterion, when AIC value drop value is greater than or equal to the preset maximum fluctuation value, save the current initial stock-flow dynamic model parameters as the benchmark, generate the corrected stock-flow dynamic model and output to the state evaluation unit; When the model correction unit generates the corrected stock-flow dynamic model, the Weibull distribution model and the Logistic distribution model are used to correct the initial stock-flow dynamic model, including: A1: Based on the product resource life distribution data in the multi-period resource data, adjust the shape parameter and scale parameter of the Weibull distribution model using the maximum likelihood estimation method; A2: Based on the recycling resource supply amount data and demand total amount data in the multi-period resource data, the recycling resource supply amount prediction deviation and demand total amount prediction deviation of the previous period are used as correction factors to adjust the growth rate parameter and saturation capacity parameter of the Logistic distribution model.

2. The multi-cycle, cyclically-utilized resource multiplier effect optimization regulation system of claim 1, wherein, The periodic data acquisition unit comprises a time sequence acquisition subunit and a multi-dimensional preprocessing subunit; According to the preset resource type list and period division standard, the multi-period resource data is obtained by the timing sampling subunit according to the hierarchical sampling rule, and after the multi-period resource data is filled with missing values and corrected for abnormal values by the multi-dimensional preprocessing subunit, the multi-period resource data is output to the model correction unit; the multi-period resource data includes resource inventory data, flow data, loss data and associated influence data in each period; the flow data includes consumption and scrap data.

3. The multi-cycle, cyclically-utilized resource multiplier effect optimization regulation system of claim 2, wherein, The feedback data receiving unit includes a real-time receiving subunit and a data checking subunit; The real-time receiving subunit receives the control effect data uploaded by the effect collection unit according to the control instruction execution period, the data checking subunit checks the control effect data for integrity and consistency, generates a feedback data set that passes the check, and associates the corresponding control instruction ID and period identifier, and then outputs to the model correction unit; The control effect data includes resource utilization efficiency change value, cycle shortening rate, and multiplier effect coefficient.

4. The multi-cycle, cyclically-utilized resource multiplier effect optimization regulation system of claim 3, wherein, When the initial inventory-flow dynamic model is constructed based on the law of conservation of mass, the initial inventory-flow dynamic model is: The input resource quantity is multiplied by the conversion efficiency coefficient to obtain the converted effective resource input quantity; the converted effective resource input quantity is subtracted from the product of the output resource quantity and the loss coefficient to obtain the actual loss amount in the resource output process; the actual loss amount in the resource output process is further subtracted from the product of the current inventory and the natural decay rate to obtain the inventory change amount; finally, the inventory change amount is added to the product of the associated system influence amount and the coupling coefficient to obtain the change rate of the resource inventory.

5. The multi-cycle, cyclically-utilized resource multiplier effect optimization regulation system of claim 4, wherein, The specific steps of A1 include: A1.1: Obtain and preprocess resource life distribution data, and extract resource life distribution data from multi-period resource data; the resource life distribution data is the duration record of resources from being put into use to failure in different periods; A1.2: classify the resource life distribution data, distinguish complete life data and truncated data, and perform abnormal value elimination and missing value filling on the classified resource life distribution data, and output the preprocessed resource life distribution data; the truncated data is the life record of resources that have not failed at the end of the observation period; A1.3: define the core function of Weibull distribution; the Weibull distribution is used to describe the probability distribution characteristics of resource life, including shape parameter and scale parameter; the shape parameter reflects the shape of resource life distribution; the scale parameter reflects the typical level of resource life; A1.4: define the probability density function and survival function of Weibull distribution; A1.5: Construct a likelihood function based on the pre-processed resource life distribution data and the defined probability density function and survival function of the Weibull distribution to construct a likelihood function; the likelihood function is the product of the probability density function values corresponding to all complete life distribution data and the product of the survival function values corresponding to all censored data, used to reflect the fitting degree of the current shape parameter and scale parameter to the resource life distribution data, and output an initial likelihood function; A1.6: Convert the initial likelihood function to a log-likelihood function, take the natural logarithm of the initial likelihood function output in A1.5, convert the original product relationship to a summation relationship, and obtain a log-likelihood function; A1.7: Construct a likelihood equation set, respectively, the partial derivatives of the output log-likelihood function with respect to the shape parameter and the scale parameter, and set both partial derivatives to zero to obtain a likelihood equation set composed of two equations; A1.8: Solve the likelihood equation set to obtain the estimated values of the shape parameter and the scale parameter, and solve the output likelihood equation set by using a numerical iteration method, including: first, set initial values for the shape parameter and the scale parameter, calculate the residual by substituting the initial values into the likelihood equation set, solve the parameter correction amount by matrix operation to update the parameters, and repeat the iteration until the parameter value change of two iterations is less than a preset convergence threshold, and output the shape parameter estimate and the scale parameter estimate; A1.9: Verify the effectiveness of the parameter estimate, based on the output shape parameter estimate and scale parameter estimate, verify the effectiveness of the parameter estimate, including: checking whether the log-likelihood function is the maximum value at the parameter estimate, and comparing the cumulative distribution difference between the Weibull distribution based on the parameter estimate and the actual life data by goodness-of-fit test; If the difference is within the preset acceptance range, the parameter estimate is determined to be an effective shape parameter and an effective scale parameter; If not, return to A1.2 to reprocess the data or adjust the iteration initial value and repeat A1.3-A1.9 to output the effective shape parameter and the effective scale parameter.

6. The multi-cycle, cyclically-utilized resource multiplier effect optimization regulation system of claim 5, wherein, The specific steps of A2 include: A2.1: Obtain the pre-processed multi-period resource data, and based on the pre-processed multi-period resource data, extract the actual supply amount, the supply amount prediction value, the actual demand amount and the demand total amount prediction value of the previous period, and calculate the cycle resource supply amount prediction deviation of the previous period and the demand total amount prediction deviation of the previous period; The cycle resource supply amount prediction deviation of the previous period is the actual supply amount of the previous period minus the corresponding supply amount prediction value; The demand total amount prediction deviation of the previous period is the actual demand amount of the previous period minus the corresponding demand total amount prediction value; A2.2: Based on the output cycle resource supply amount prediction deviation of the previous period and the demand total amount prediction deviation of the previous period, convert them into standardized correction factors to obtain supply amount correction factors and demand total amount correction factors; the supply amount correction factor is the cycle resource supply amount prediction deviation of the previous period divided by the absolute value of the actual supply amount of the previous period; the demand total amount correction factor is the demand total amount prediction deviation of the previous period divided by the absolute value of the actual demand amount of the previous period; A2.3: defining the Logistic distribution model parameters, and explicitly indicating that the Logistic distribution model comprises a growth rate parameter and a saturation capacity parameter; the growth rate parameter is used to determine the speed at which the amount of circulating resources approaches the saturation capacity; and the saturation capacity parameter is used to represent the maximum limit value that the amount of circulating resources can reach; A2.4: establishing a parameter correction association rule, and based on the supply-demand relationship of the circulating resources, establishing a correction association rule of the growth rate parameter and the saturation capacity parameter according to the prediction deviation of the previous period; the correction association rule comprises: if the prediction deviation of the supply amount of the circulating resources in the previous period is positive or the prediction deviation of the total demand amount is positive, then the growth rate parameter is increased, and vice versa; if the prediction deviation of the supply amount of the circulating resources in the previous period is positive and lasts for multiple periods or the prediction deviation of the total demand amount is positive, then the saturation capacity parameter is increased, and vice versa; A2.5: based on the supply amount correction factor and the total demand amount correction factor, and the correction association rule, constructing a parameter correction formula; the parameter correction formula comprises a growth rate parameter correction formula and a saturation capacity parameter correction formula; A2.6: substituting the adjusted growth rate parameter and the adjusted saturation capacity parameter into the Logistic distribution model to replace the initial growth rate parameter and the initial saturation capacity parameter, and outputting an updated Logistic distribution model; A2.7: based on the updated Logistic distribution model, predicting the supply amount of the circulating resources and the total demand amount in the current period, collecting the actual supply amount of the circulating resources and the actual total demand amount in the current period, calculating the supply amount prediction deviation and the total demand amount prediction deviation in the current period, and comparing them with the supply amount prediction deviation and the total demand amount prediction deviation in the previous period; if the absolute value of the deviation in the current period is smaller than that in the previous period and within a preset acceptable range, then the updated Logistic distribution model is determined to be an effective model; if it is not within the preset acceptable range, then return to A2.2 to re-determine the correction factor, repeat A2.3 to A2.7, and until an effective Logistic distribution model is obtained.

7. The multi-cycle, cyclically-utilized resource multiplier effect optimization regulation system of claim 6, wherein, The state evaluation unit comprises an index calculation subunit and a grade evaluation subunit; The index calculation subunit receives the corrected stock-flow dynamic model, calculates resource utilization core indexes based on the stock-flow simulation results of each period output by the corrected stock-flow dynamic model; the resource utilization core indexes comprise resource recycling rate, accumulated primary resource consumption, and resource multiplier effect contribution value; The grade evaluation subunit receives the resource utilization core indexes, determines the weight of each index by using the analytic hierarchy process, and generates a resource recycling utilization state evaluation result through fuzzy comprehensive evaluation; the resource recycling utilization state evaluation result comprises three grades of upgraded utilization, primary utilization, and secondary utilization, and corresponding improvement indexes, and is output to the scenario generation unit.

8. The multi-cycle, cyclically-utilized resource multiplier effect optimization regulation system of claim 7, wherein, The strategy calculation unit comprises a parameter optimization subunit and a strategy combination subunit; The parameter optimization subunit receives regulation scenarios, and optimizes the regulation parameters under each regulation scenario based on the corrected stock-flow dynamic model and by using a genetic algorithm; the regulation parameters include a resource recycling threshold, a recycling input coefficient, and a cross-period allocation proportion; The strategy combination subunit receives the optimal regulation parameters of each scenario, combines the grade assessment in the resource recycling state assessment result, filters out a parameter combination that meets the constraint boundary and has the highest comprehensive benefit, and generates an optimal regulation strategy combination; the optimal regulation strategy combination contains at least three regulation measures and corresponding implementation priorities, and is output to the instruction generation unit.

Images