Green product life cycle analysis system based on big data

Abstract

The application relates to the technical field of green product analysis, and discloses a green product whole life cycle analysis system based on big data. The system comprises data acquisition, preprocessing, environmental impact assessment, resource consumption optimization, correlation modeling and result generation and visual interaction modules. The system collects data of each stage in the whole life cycle and related element data such as environment, resources and carbon emissions, carries out preprocessing such as abnormal value elimination and entropy weight assignment, dynamically assesses environmental impact through a fuzzy membership function, optimizes multi-objective resource consumption through a genetic algorithm, determines key stages through a grey correlation degree and generates analysis results in combination with a multi-objective decision model, and finally presents the results in a three-dimensional graph and a dynamic heat map. The system realizes efficient analysis and comprehensive evaluation of whole life cycle data of green products, and provides scientific support for product environmental performance optimization and decision making.

Description

Technical Field

[0001] This invention relates to the field of green product analysis technology, specifically a green product lifecycle analysis system based on big data. Background Technology

[0002] With the increasing severity of global environmental problems, the concept of sustainable development has become a global consensus, and the research and development and promotion of green products have become important ways to address resource shortages and environmental pollution. Green product life cycle analysis, as a core method for assessing product environmental performance, aims to comprehensively evaluate a product's environmental impact and resource consumption throughout its entire life cycle, from raw material acquisition, production and processing, distribution and sales, use and maintenance to recycling, providing a scientific basis for the green design, production, and management of products. However, traditional green product life cycle analysis faces many challenges and is insufficient to meet the current complex needs of environmental governance and industrial upgrading.

[0003] From a data perspective, traditional analytical methods suffer from limited data acquisition channels, relying primarily on manual collection and data from a few monitoring points. This results in narrow data coverage, poor timeliness, and an inability to comprehensively reflect the true situation at each stage of a product's entire lifecycle. Furthermore, the data types are often limited, focusing only on certain environmental impact factors or resource consumption indicators. They lack comprehensive consideration of multi-dimensional and interconnected factors, including environmental impact factor data (such as pollutant emission type, emission concentration, and environmental restoration costs), resource consumption factor data (water resource consumption intensity, energy consumption intensity, and raw material utilization rate), carbon emission factor data (direct carbon emissions, indirect carbon emissions, and carbon offset), policy and regulatory factors, and user behavior factors. In addition, raw data often contains numerous outliers, noise, and missing values. Traditional data preprocessing methods struggle to efficiently remove outliers, filter noise, and imput missing data, and lack a scientific weighting mechanism, significantly compromising the accuracy of subsequent analyses.

[0004] In terms of analytical methods, traditional environmental impact assessments mostly employ static assessment models, which cannot dynamically track changes in environmental impact at each stage of the product lifecycle. For example, pollutant emissions and resource consumption exhibit significant dynamic characteristics at different stages, making it difficult for traditional models to capture these changes in real time and conduct comprehensive assessments. Regarding resource consumption optimization, traditional methods typically focus on optimizing only a single objective, such as simply reducing total resource consumption or carbon emissions, neglecting the need for synergistic optimization among multiple objectives. This makes it difficult to maximize resource utilization efficiency and optimize environmental benefits. In terms of correlation modeling, traditional methods lack sufficient correlation analysis of data from different lifecycle stages, failing to accurately identify stages that play a key role in the product's environmental performance and resource consumption, resulting in a lack of targeted optimization measures.

[0005] From a technical perspective, traditional analytics systems lack deep integration with big data technologies, making it difficult to handle massive amounts of data across the entire product lifecycle. With the development of technologies such as the Internet of Things and sensors, the data generated throughout the lifecycle of green products is experiencing explosive growth, and traditional data storage, management, and analysis technologies are insufficient in terms of processing speed and computing power. Furthermore, traditional systems have weak visualization and interactive functions, failing to present complex analytical results in an intuitive and dynamic manner, hindering decision-makers from quickly understanding and applying the analytical findings.

[0006] At the policy level, countries are imposing increasingly stringent regulatory requirements on green products. For example, the EU's Ecodesign Directive (ErP) and China's green product certification system both require companies to provide detailed environmental impact reports covering the entire product lifecycle. Traditional analytical methods, unable to meet policy requirements for data comprehensiveness, accuracy, and timeliness, may expose companies to compliance risks. Simultaneously, intensified market competition forces companies to continuously improve the green competitiveness of their products, but traditional analytical methods struggle to provide precise decision support, thus hindering companies' innovation capabilities in green product research and development and management. Summary of the Invention

[0007] The purpose of this invention is to provide a green product lifecycle analysis system based on big data to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a green product lifecycle analysis system based on big data, the system comprising:

[0009] The data acquisition module is used to acquire data on the entire lifecycle of green products and related elements.

[0010] The data preprocessing module performs multi-dimensional preprocessing on the green product lifecycle data and related element data acquired by the data acquisition module;

[0011] The environmental impact assessment module is used to dynamically assess the environmental impact of green products throughout their entire life cycle.

[0012] The resource consumption optimization module is used to perform multi-objective optimization of resource consumption throughout the entire life cycle of green products;

[0013] The correlation modeling and result generation module is used to perform correlation modeling on data from each stage of the green product's entire life cycle, and to generate full life cycle analysis results by combining a multi-objective decision-making model.

[0014] Preferably, the full life cycle data includes: raw material acquisition stage data, production and processing stage data, distribution and sales stage data, use and maintenance stage data, and recycling and disposal stage data. The related element data includes: environmental impact factor data, resource consumption factor data, carbon emission factor data, policy and regulatory factor data, and user behavior factor data. The environmental impact factor data includes: pollutant emission type data, emission concentration data, and environmental restoration cost data.

[0015] Preferably, the multi-dimensional preprocessing of the green product lifecycle data and related element data acquired by the data acquisition module includes:

[0016] Outlier removal and noise filtering are performed on the acquired raw data;

[0017] Missing data is imputed, and the data is weighted based on the entropy weight method.

[0018] The processed data is divided into time series based on the life cycle stages: raw material acquisition time series, production and processing time series, distribution and sales time series, use and maintenance time series, and recycling and processing time series.

[0019] Preferably, the dynamic assessment of the environmental impact of green products throughout their entire life cycle includes:

[0020] The environmental impact factor data is decomposed into resource consumption subsequence, pollution emission subsequence, and environmental restoration cost subsequence;

[0021] The fuzzy membership function is constructed to fuzzify each subsequence. The calculation formula is as follows:

[0022] ;

[0023] in, Indicates the first Environmental impact factors, Indicates the first One evaluation indicator, Indicates the first The first type of environmental impact factor Fuzzy membership degree of each evaluation indicator As a fuzzy factor, Indicates the first The first type of environmental impact factor The actual measured values ​​of each evaluation indicator As the baseline value;

[0024] A comprehensive environmental impact score is generated by fusing fuzzy evaluation results using a weighted average operator.

[0025] Preferably, the multi-objective optimization of resource consumption throughout the entire life cycle of green products includes:

[0026] The objective function for resource optimization is set as minimizing total resource consumption, maximizing resource utilization efficiency, and minimizing carbon emission intensity.

[0027] Construct chromosome coding rules to map resource allocation schemes to chromosome gene sequences;

[0028] The chromosome population is iteratively optimized through crossover, mutation, and selection operations until the preset convergence conditions are met.

[0029] Preferably, the correlation modeling of data at each stage of the green product's entire lifecycle includes:

[0030] Extract the feature sequences of data from each stage, including the feature sequences of raw material acquisition, production and processing, distribution and sales, use and maintenance, and recycling and processing.

[0031] The grey relational degree between each feature sequence and the target sequence is calculated using the following formula:

[0032] ;

[0033] in, Indicates the first Grey correlation degree between feature sequences and target sequences at each life cycle stage Indicates the first The absolute difference between the feature sequence and the target sequence at corresponding time points in each life cycle stage. Indicates all The minimum value in Indicates all The maximum value in, The resolution coefficient;

[0034] Key lifecycle stages are determined by ranking them according to their relevance.

[0035] Preferably, the multi-objective decision model includes:

[0036] Construct a decision matrix, using environmental impact score, resource optimization scheme, and grey relational degree as decision variables;

[0037] The closeness of each scheme to the ideal solution is calculated using the ideal point method;

[0038] The solution with the highest degree of similarity was selected as the optimal lifecycle analysis result.

[0039] Preferably, the data in the entire life cycle includes: raw material acquisition stage data, mining energy consumption data, and transportation distance data; production and processing stage data includes: process energy consumption data, waste generation data, and equipment efficiency data; and distribution and sales stage data includes: logistics carbon emission data, warehousing cost data, and sales channel type data.

[0040] Preferably, the resource consumption factor data includes: water resource consumption intensity data, energy consumption intensity data, and raw material utilization rate data; the carbon emission factor data includes: direct carbon emission data, indirect carbon emission data, and carbon sink offset data.

[0041] Preferably, the system further includes:

[0042] The visualization and interaction module is used to visualize the analysis results through 3D maps and dynamic heat maps;

[0043] The visualization and interactive module includes a three-dimensional map comprising: a life cycle stage relationship network diagram, a resource consumption intensity distribution map, and a carbon emission spatiotemporal evolution map; and a dynamic heat map used to display the changing trends of environmental impact scores in different regions.

[0044] Compared with the prior art, the beneficial effects of the present invention are:

[0045] The system's data acquisition module comprehensively collects data on the entire lifecycle of green products, including related elements. This covers all stages, from raw material acquisition and production / processing to distribution, sales, use, maintenance, and recycling, while also incorporating multi-dimensional data on environmental impact, resource consumption, carbon emissions, policies and regulations, and user behavior. This design overcomes the limitations of traditional methods with insufficient data coverage, ensuring analysis is based on complete data sources and providing a solid foundation for subsequent evaluation and optimization. For example, data from the raw material acquisition stage includes details such as raw material types, mining energy consumption, and transportation distances; data from the production / processing stage includes indicators such as process energy consumption, waste generation, and equipment efficiency. This allows the system to trace the environmental performance and resource consumption of products from source to end.

[0046] The data preprocessing module effectively improves data quality through multi-dimensional processing, including outlier removal, noise filtering, missing data imputation, and entropy-based weight allocation. Outlier and noise removal prevents spurious data from interfering with the analysis results, missing data imputation ensures data integrity, and the entropy-based weight allocation objectively assigns weights based on the data's inherent variability, making it more scientific than traditional subjective weighting methods. The processed data is divided into time series according to its lifecycle stages, providing a standardized and orderly data structure for subsequent dynamic analysis and correlation modeling. This clearly presents the temporal evolution characteristics of the data at each stage, facilitating the analysis of trends and relationships between data at different stages.

[0047] The environmental impact assessment module decomposes environmental impact factor data into subsequences of resource consumption, pollution emissions, and environmental restoration costs. It then uses fuzzy membership functions to fuzzify each subsequence and generates a comprehensive score using a weighted average operator, enabling dynamic assessment of the environmental impact of green products throughout their entire lifecycle. Fuzzy mathematics effectively addresses uncertainties and fuzziness in environmental assessments, such as determining whether pollutant emission concentrations exceed standards or whether environmental restoration costs are reasonable. By calculating fuzzy membership degrees, qualitative issues are quantified, making the assessment results more realistic. The dynamic assessment mechanism tracks changes in the environmental impact of products at different lifecycle stages in real time, promptly identifying environmental risks and providing real-time feedback for enterprises to adjust production processes and optimize environmental protection measures.

[0048] The resource consumption optimization module employs multiple objectives: minimizing total resource consumption, maximizing resource utilization efficiency, and minimizing carbon emission intensity. It iteratively optimizes resource allocation schemes using chromosome encoding, crossover, mutation, and selection operations in a genetic algorithm. This multi-objective optimization model overcomes the limitations of traditional single-objective optimization, seeking the optimal balance between resource consumption, utilization efficiency, and carbon emissions to achieve a win-win situation for both economic and environmental benefits. For example, by optimizing resource allocation, energy consumption intensity can be reduced while raw material utilization is improved, and carbon emissions can be reduced while production efficiency is maintained, making resource utilization more scientific and efficient for enterprises.

[0049] The correlation modeling and results generation module accurately identifies key lifecycle stages that play a crucial role in the environmental impact and resource consumption of products by extracting data feature sequences from each stage and calculating grey relational degrees. Grey relational analysis is suitable for analyzing complex systems with small samples and multiple factors, revealing the degree of correlation between factors without requiring extensive historical data. By identifying key stages, companies can concentrate optimization resources on the links that have the greatest impact on the green performance of products. For example, if the resource consumption in the raw material acquisition stage has the highest correlation with the target sequence, priority should be given to optimizing energy consumption in the raw material mining and transportation processes, improving the targeting and effectiveness of optimization measures. The multi-objective decision-making model constructs a decision matrix and uses the ideal point method to select the optimal solution, comprehensively considering multiple dimensions such as environmental impact score, resource optimization scheme, and grey relational degrees to ensure the scientific validity and reliability of the final analysis results, providing companies with comprehensive and objective decision-making basis.

[0050] The visualization and interactive module presents the analysis results intuitively through 3D maps (such as life cycle stage relationship network diagrams, resource consumption intensity distribution maps, and carbon emission spatiotemporal evolution maps) and dynamic heat maps. The 3D maps display the interrelationships between different life cycle stages, the spatial distribution of resource consumption, and the temporal evolution trend of carbon emissions in a three-dimensional way, while the dynamic heat maps show real-time changes in environmental impact scores in different regions, making complex analysis results easy to understand and interpret. This visualization method not only helps enterprise managers quickly grasp the green performance status of products but also provides policymakers with intuitive decision-making references, promoting the scientific formulation and effective implementation of green product-related policies.

[0051] Furthermore, based on a big data technology architecture, the system possesses powerful data storage, processing, and computing capabilities, enabling it to efficiently handle massive amounts of data throughout the entire product lifecycle and meet the needs of real-time analysis and dynamic evaluation. Simultaneously, the system can flexibly adjust evaluation indicators and optimization targets according to changes in policies, regulations, and market demands, demonstrating strong adaptability and scalability. This provides robust technical support for the continuous improvement of green products and industrial upgrading, contributing to the coordinated progress of economic development and environmental protection. Attached Figure Description

[0052] Figure 1 This is a schematic diagram illustrating the working principle of the big data-based green product lifecycle analysis system described in this invention.

[0053] Figure 2 A flowchart illustrating the data preprocessing process;

[0054] Figure 3 A flowchart for dynamic environmental impact assessment;

[0055] Figure 4 This is a flowchart illustrating the process of grey relational analysis. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Please see Figures 1-4 This invention provides a green product lifecycle analysis system based on big data, the system comprising:

[0058] Data Acquisition Module: Used to acquire data on the entire lifecycle of green products and related factors. The lifecycle data covers the stages of raw material acquisition, production and processing, distribution and sales, use and maintenance, and recycling. Related factor data includes data on environmental impact factors, resource consumption factors, carbon emission factors, policy and regulatory factors, and user behavior factors.

[0059] Data preprocessing module: Performs multi-dimensional preprocessing on the acquired full lifecycle data and related element data. Specific operations include removing outliers and filtering noise from the raw data, imputing missing data, assigning weights to the data based on the entropy weight method, and dividing the processed data into time series according to the lifecycle stages: raw material acquisition, production and processing, distribution and sales, use and maintenance, and recycling.

[0060] Environmental Impact Assessment Module: Dynamically assesses the environmental impact of green products throughout their entire life cycle.

[0061] Resource consumption optimization module: Performs multi-objective optimization of resource consumption throughout the entire life cycle of green products.

[0062] The correlation modeling and result generation module performs correlation modeling on data from each stage of the green product's entire lifecycle and generates full lifecycle analysis results by combining them with a multi-objective decision-making model.

[0063] The present invention will be further described below with reference to Examples 1 to 5:

[0064] Example 1:

[0065] The system's data acquisition module needs to comprehensively collect data on the entire lifecycle of green products and related elements. The lifecycle data covers the entire chain from raw material acquisition to recycling: raw material acquisition data includes raw material types (e.g., different materials such as metals, plastics, and glass), energy consumption during mining (e.g., electricity and fuel consumed by mining equipment), and transportation distance (the mileage of raw materials from the mining site to the production plant); production and processing data involves energy consumption during production processes (e.g., electricity consumption of production line equipment, energy loss during heating or cooling processes), waste generation (the amount of waste such as scraps, wastewater, and exhaust gas generated during production), and equipment efficiency (running time, output per unit time, and failure rate of production equipment); distribution and sales data includes carbon emissions from logistics (transportation...). Data on vehicle fuel consumption and carbon emissions, warehousing costs (warehouse rental fees, energy costs of warehousing equipment, inventory management costs, etc.), and sales channel types (different sales models such as online e-commerce platforms, offline physical stores, and dealer distribution); data on the usage and maintenance stage includes the frequency of user product use (e.g., daily and weekly usage time), maintenance operation records (maintenance cycles, replacement of parts), and energy consumption (electricity or fuel consumption during product operation); data on the recycling and processing stage covers the recycling rate of waste products (the proportion of recycled amount to scrapped amount), processing methods (reuse, remanufacturing, resource recovery, or harmless disposal), and energy consumption during the processing (energy consumption in recycling, sorting, dismantling, and reprocessing).

[0066] Related data need to collect multiple types of influencing factors: Environmental impact factor data includes pollutant emission types (such as specific pollutant types like sulfur dioxide, nitrogen oxides, and heavy metal ions), emission concentrations (the content of pollutants per unit volume or unit mass), and environmental restoration costs (the financial investment required for pollution control and ecological restoration); resource consumption factor data includes water consumption intensity (the amount of water consumed in the production or use of a unit of product), energy consumption intensity (the amount of standard coal consumed per unit of output or unit of product), and raw material utilization rate (the ratio of the actual amount of raw materials used in production to the total amount of raw materials purchased); carbon emission factor data includes direct carbon emissions (from product production, transportation, and use, etc.). The data includes direct carbon dioxide emissions from various stages of the process, indirect carbon emissions (such as carbon emissions from upstream stages like raw material production and power supply), and carbon offsets (carbon emissions offset through carbon sink projects such as afforestation). Policy and regulatory factors include national and local environmental protection standards (such as pollutant emission standards and energy consumption quota standards), resource utilization regulations (such as raw material recycling policies and waste management regulations), and green product certification systems. User behavior factors include user habits (such as frequent power on / off cycles and excessive packaging requirements), maintenance preferences (self-maintenance or professional maintenance), and disposal choices (actively sending to recycling points or discarding at will).

[0067] The data acquisition module collects data through multiple channels: for equipment operation data during the production and processing stage, it collects and transmits the data to the system in real time by installing sensors (such as current sensors, temperature sensors, flow meters, etc.) on the production equipment; for data during the raw material acquisition and distribution and sales stages, it obtains purchase records, transportation orders, warehousing data, etc., by connecting to the enterprise's supply chain management system (such as ERP system); for data related to environmental impact and resource consumption, it obtains data from industry databases (such as the pollutant emission data platform published by the environmental protection department and the energy statistical yearbook) and government open data platforms (such as the data center of the Ministry of Ecology and Environment and the data release system of the National Bureau of Statistics); and for user behavior data, it collects user operation logs through the smart terminals built into the product (such as the networking module of smart home appliances), or obtains data through questionnaires, user feedback platforms, etc.

[0068] The data preprocessing module performs multi-dimensional processing on the collected raw data. First, outlier removal and noise filtering are performed: statistical analysis methods are used to identify outlier data points, such as box plot detection, where data exceeding 1.5 times the interquartile range are considered outliers and removed; a moving average filtering algorithm is used to smooth the time series data, reducing the interference of random noise. For missing data, an imputation strategy is selected based on the data type and degree of missingness: if raw material purchase quantity data is missing for a certain period and the data has temporal continuity, linear interpolation is performed using the average of adjacent time periods; if user maintenance habit data has categorical missing values, imputation is performed using the mode by statistically analyzing the common maintenance methods of users of this type of product.

[0069] After data cleaning, the data is weighted using the entropy weighting method. The core logic of the entropy weighting method is to measure the information value of each indicator by calculating its information entropy. The lower the information entropy of an indicator, the greater its data variation, and the higher its weight in influencing the analysis results. The specific steps are as follows: First, standardize the data for each indicator to eliminate dimensional differences; then calculate the information entropy of each indicator using the following formula: (in For the first In the nth sample The proportion of each indicator Finally, the weights are calculated based on the information entropy. This allows us to determine the importance ranking of each data indicator in subsequent analysis.

[0070] The final step in data preprocessing is to divide the data into time series according to lifecycle stages. Data from the raw material acquisition stage (such as the types of raw materials purchased in different months, corresponding mining energy consumption, and transportation distance) are organized chronologically into a raw material acquisition time series. Data from the production and processing stage, such as process energy consumption, waste generation, and equipment efficiency, are divided into production and processing time series by production batch or date. Data from the distribution and sales stage, such as logistics carbon emissions, warehousing costs, and sales channels, are organized into distribution and sales time series by sales cycle (e.g., monthly, quarterly). Data from the use and maintenance stage, such as user usage frequency, energy consumption, and maintenance records, are organized into use and maintenance time series by user usage timeline. Data from the recycling and processing stage, such as recycling rate, processing method, and processing energy consumption, are organized into recycling and processing time series by recycling batch or year. Through these processes, the originally scattered raw data is transformed into a structured, time-series dataset, providing standardized input data for the environmental impact assessment module, resource consumption optimization module, and correlation modeling module. This ensures that each module can conduct analysis based on a unified and clean data foundation, improving the accuracy and reliability of the overall system analysis results.

[0071] Example 2:

[0072] After the data preprocessing module completes multi-dimensional preprocessing of the green product's entire lifecycle data and related element data, the environmental impact assessment module needs to conduct a dynamic assessment of the environmental impact of the green product's entire lifecycle. This assessment process is based on the structured time series data output by the data preprocessing module, and through the decomposition, fuzzification, and comprehensive scoring of environmental impact factor data, it achieves a quantitative analysis of the product's environmental impact throughout its entire lifecycle.

[0073] The environmental impact assessment module decomposes environmental impact factor data into three subsequences: resource consumption, pollution emissions, and environmental restoration costs. The resource consumption subsequence covers various data related to resource utilization, such as energy consumption during raw material acquisition, process energy consumption during production and processing, and energy consumption during use and maintenance. The pollution emissions subsequence includes data related to pollutants generated at each stage of the life cycle, such as exhaust gas emission concentrations during production and processing, pollutant emissions from logistics activities during distribution and sales, and wastewater discharge types during recycling and treatment. The environmental restoration cost subsequence involves cost data required to reduce or eliminate environmental impacts, such as the treatment costs for pollutant emissions during production and the restoration costs for ecological damage caused by product disposal. This decomposition process divides complex environmental impact factor data into more targeted dimensions, facilitating subsequent refined analysis.

[0074] The module needs to construct fuzzy membership functions to fuzzify each subsequence. The purpose of fuzzy membership functions is to transform the actual measured values ​​of specific assessment indicators into fuzzy membership degrees that reflect their degree of environmental impact, thus resolving the uncertainties and fuzziness in environmental impact assessments. Taking an assessment indicator under a certain type of environmental impact factor as an example, assuming this indicator is the emission concentration of a pollutant during the production and processing stage, its actual measured value is... The baseline value is (Benchmark values ​​can refer to national or industry-set pollutant emission standards, average emission levels of similar products, or internal environmental protection targets set by the enterprise), fuzzy factor The shape and sensitivity of the membership function are determined based on the characteristics of this indicator and historical data experience. This is achieved through a formula.

[0075] ;

[0076] The fuzzy membership degree of the index was calculated. Its value range is within arrive Between. When the actual measured value equal to the baseline value When the fuzzy membership degree is This indicates that the indicator's environmental impact is at an ideal level; with... Deviation from benchmark value As the degree of fuzzy membership increases, the degree of environmental impact gradually decreases, indicating an increase in the degree of environmental impact. For example, if the baseline value for the emission concentration of a certain pollutant is... When the actual measured value is lower than the benchmark value, it indicates that the emissions are better than the standard, and the fuzzy membership degree is close to or equal to the benchmark value. When the actual measured value exceeds the benchmark value, the greater the excess, the smaller the fuzzy membership degree, reflecting the greater the negative impact of the indicator on the environment.

[0077] For indicators in the resource consumption subsequence, such as process energy consumption in the production and processing stage, the baseline value is... It can be set to the industry average energy consumption level or the energy consumption limit set by the company, based on the actual measured value. This represents the actual energy consumption data for this stage. Through fuzzy membership function calculation, indicators with lower energy consumption will obtain higher fuzzy membership degrees, indicating higher resource utilization efficiency and a greater positive impact on the environment; conversely, indicators with higher energy consumption will have lower fuzzy membership degrees, reflecting greater resource waste and environmental pressure. In the environmental restoration cost subsequence, the benchmark value... The actual measured value can be determined based on historical cost data of similar environmental remediation projects or industry average cost levels. This represents the restoration cost required to address the environmental impact at each stage of the product's lifecycle. Lower costs correlate with higher fuzzy membership, indicating less difficulty and cost in environmental restoration and a relatively lower overall environmental impact. Conversely, higher costs correlate with lower fuzzy membership, suggesting more severe environmental damage and more complex and costly restoration efforts.

[0078] After fuzzification of all assessment indicators in each subsequence, the environmental impact assessment module fuses the fuzzy evaluation results using a weighted average operator to generate a comprehensive environmental impact score. The weighted average operator is calculated in conjunction with the weights of each indicator determined by the entropy weight method in the data preprocessing module. The entropy weight method assigns an objective weight value to each assessment indicator by analyzing the degree of data variation, avoiding the subjectivity of manually setting weights. For example, if the emission concentration of a certain pollutant in the pollution emission subsequence has a high weight calculated by the entropy weight method during the data preprocessing stage, it indicates that the indicator is important in the environmental impact assessment, and its fuzzy membership degree accounts for a large proportion in the comprehensive score calculation. The specific fusion process is as follows: multiply the fuzzy membership degree of each assessment indicator by its corresponding weight to obtain the weighted value of the indicator; then sum the weighted values ​​of all indicators to finally obtain the comprehensive environmental impact score. This score is a comprehensive quantitative indicator that can fully reflect the overall environmental impact of green products throughout their entire life cycle. A higher score indicates a better environmental performance and a smaller environmental impact; a lower score indicates a heavier environmental burden in the production, distribution, use, and recycling stages, requiring targeted improvements and optimizations.

[0079] In practice, the environmental impact assessment module conducts environmental impact assessments for each stage of the green product's life cycle, generating environmental impact scores for each stage, and also provides a comprehensive assessment of the environmental impact throughout the entire life cycle. For example, for the raw material acquisition stage, it assesses the environmental impact of indicators such as mining energy consumption and transportation distance, generating an environmental impact score for that stage; for the production and processing stage, it assesses indicators such as process energy consumption, waste generation, and pollutant emission concentrations, obtaining a score for that stage. By analyzing the scores for each stage, key stages with significant environmental impact can be identified, providing a basis for subsequent resource consumption optimization and life cycle stage correlation modeling. Furthermore, the module can also perform dynamic comparative analysis of environmental impacts over different time periods based on time series data, such as comparing the environmental impact trends of products in different production batches and different years, to promptly identify fluctuations in environmental performance or the effectiveness of improvements, providing data support for enterprises to adjust production processes, optimize supply chain management, and formulate environmental protection strategies.

[0080] The dynamic assessment process of the environmental impact assessment module has the following characteristics: First, it is data-driven, based entirely on real data collected by the data acquisition module and the results processed by the data preprocessing module, ensuring the objectivity and credibility of the assessment results. Second, it is multi-dimensional and comprehensive, constructing an assessment system from three dimensions: resource consumption, pollution emissions, and environmental restoration costs, comprehensively covering the main aspects of environmental impact. Third, it is dynamically adaptable, continuously assessing environmental impact based on the progression of the product life cycle and real-time data updates, promptly reflecting changes in the product's environmental performance at different stages. Through the operation of this module, the system can provide scientific environmental impact assessment basis for the design, production, distribution, use, and recycling of green products, helping enterprises achieve green development goals and promoting the formation of sustainable production and consumption patterns.

[0081] Example 3:

[0082] The resource consumption optimization module performs multi-objective optimization of resource consumption throughout the entire lifecycle of green products. Its core lies in iteratively optimizing resource allocation schemes by setting multi-objective functions, constructing chromosome coding rules, and executing genetic algorithms. This process is based on structured data output from the data preprocessing module, combined with the resource consumption characteristics of green products at each stage of raw material acquisition, production and processing, distribution and sales, use and maintenance, and recycling, to form a scientific optimization strategy.

[0083] The module needs to define objective functions for resource optimization, specifically including minimizing total resource consumption, maximizing resource utilization efficiency, and minimizing carbon emission intensity. Minimizing total resource consumption aims to reduce the total input of various resources (such as water, energy, and raw materials) throughout the product's lifecycle by optimizing resource allocation in procurement, production, and transportation to reduce unnecessary resource waste. Maximizing resource utilization efficiency focuses on improving resource conversion efficiency, such as increasing raw material utilization and reducing the energy consumption ratio of production equipment, so that a unit of resource input can produce more products or services. Minimizing carbon emission intensity involves optimizing the energy structure and improving production processes to reduce carbon emissions per unit of product or unit of output value, aligning with low-carbon development goals. These three objective functions are interconnected and mutually restrictive, forming the core framework of multi-objective optimization, aiming to achieve a balance between the economic and environmental benefits of resource utilization.

[0084] The module constructs chromosome encoding rules, mapping resource allocation schemes to chromosome gene sequences. In genetic algorithms, chromosomes represent the encoded solutions to optimization problems; here, real-number encoding is used, with each gene bit corresponding to a resource allocation variable. For example, in the raw material acquisition stage, gene bits can correspond to the purchase quantity of different raw materials and the energy consumption quota of mining equipment; in the production and processing stage, gene bits can represent the energy consumption threshold of the production process and the proportion of waste recycling; in the distribution and sales stage, gene bits can involve the selection of logistics transportation routes (encoded by mileage or carbon emission coefficients) and the energy management parameters of warehousing facilities; in the use and maintenance stage, gene bits can represent the parameter values ​​for optimizing product energy consumption; and in the recycling and processing stage, gene bits can correspond to energy consumption control indicators in the dismantling process of waste products and the utilization rate target of recycled materials. This encoding method transforms complex resource allocation schemes into calculable and operable chromosome gene sequences, providing a foundation for subsequent genetic algorithm operations.

[0085] After completing chromosome encoding, the module iteratively optimizes the chromosome population through crossover, mutation, and selection operations. Crossover is the primary method for generating new individuals in genetic algorithms; it generates offspring chromosomes by exchanging partial genes from two parent chromosomes. Specifically, two parent chromosomes are first randomly selected from the population. Then, the crossover point is randomly determined, and gene segments on either side of the crossover point are exchanged to form two new offspring chromosomes. For example, suppose the parent chromosomes... The gene sequence is , the parent chromosome The gene sequence is If the intersection point is determined to be the first If there are 1 gene locus, then the offspring chromosomes generated after crossover will be... The gene sequence is offspring chromosomes The gene sequence is Crossover can fully utilize the superior gene combinations of the parent chromosomes to produce offspring with stronger adaptability.

[0086] Mutation is a technique in genetic algorithms that introduces new mutated individuals. It involves randomly changing the value of a gene locus on a chromosome to disrupt the local optimum and maintain population diversity. The specific steps of mutation are: randomly select a chromosome, randomly determine the gene locus to mutate, and then randomly generate a new value within the range of that gene locus to replace the original gene value. For example, if the original value of a gene locus representing the energy consumption threshold of a production process on a chromosome is 500 kWh, and its range is 300-800 kWh, after mutation, the value of that gene locus might become 650 kWh. The probability of mutation is usually low, generally between 0.1% and 1%, to avoid excessive mutation that could slow down the algorithm's convergence or destroy a favorable gene structure.

[0087] Selection is the survival-of-the-fittest mechanism in genetic algorithms. It evaluates each chromosome in the population using a fitness function, selecting individuals with higher fitness to advance to the next generation and eliminating those with lower fitness. The fitness function is constructed based on a defined objective function, and its specific form is as follows:

[0088] ;

[0089] in, This represents the fitness value of a chromosome. These are the weight coefficients of the three objective functions: minimizing total resource consumption, maximizing resource utilization efficiency, and minimizing carbon emission intensity, respectively, and they satisfy... The weighting coefficient is determined based on the company's actual needs and policy guidance. For example, when a company focuses on low-carbon development, the weighting coefficient can be appropriately increased. The weights; The normalized value of total resource consumption (by mapping the actual total resource consumption to...) The range is obtained, and the smaller the value, the less the total resource consumption. This is a normalized value for resource utilization efficiency (the larger the value, the higher the resource utilization efficiency). This is a normalized value for carbon emission intensity (the smaller the value, the lower the carbon emission intensity). By calculating the fitness function, the merits of the resource allocation scheme represented by each chromosome can be quantified, ensuring that the population evolves towards a better solution.

[0090] During the iterative optimization process, the module first randomly generates an initial chromosome population, typically set at 50-200 individuals to balance computational efficiency and search space. Then, the fitness of each chromosome in the initial population is calculated, and a selection operation is performed based on the fitness value, retaining individuals with higher fitness as parents. Next, crossover and mutation operations are performed on the parent individuals to generate the offspring population. The fitness calculation, selection, crossover, and mutation operations are then repeated on the offspring population, iterating in this cycle until a preset convergence condition is met. The convergence condition typically includes the number of iterations reaching a set value (e.g., 100 or 200 generations) or the fitness value no longer changing significantly over several consecutive generations (e.g., the change in fitness value is less than 0.1% over 10 consecutive generations).

[0091] In the process of optimizing resource consumption, the module needs to consider the resource correlation characteristics between different lifecycle stages. For example, the procurement volume in the raw material acquisition stage directly affects the raw material inventory and production plan in the production and processing stage, which in turn affects the energy consumption and waste generation in the production stage; the choice of logistics routes in the distribution and sales stage not only affects carbon emissions in the transportation process, but may also affect resource consumption in the warehousing stage through inventory strategies. Therefore, in chromosome encoding and genetic algorithm operations, it is necessary to ensure that the resource allocation variables in each stage have logical consistency and constraint compatibility to avoid contradictory resource allocation schemes (such as situations where the raw material procurement volume is insufficient to support the production plan).

[0092] Furthermore, the module also needs to address conflicts in multi-objective optimization. For example, minimizing total resource consumption and maximizing resource utilization efficiency may have a synergistic effect, while minimizing carbon emission intensity may require increased investment in low-carbon technologies, thereby increasing total resource consumption to some extent. By adjusting the weight coefficients of the objective function and the operating parameters of the genetic algorithm, the module can weigh different objectives and generate a set of Pareto optimal solutions that meet specific needs, allowing decision-makers to select the optimal solution based on the actual situation.

[0093] The operation of the resource consumption optimization module relies on accurate data and weighting information provided by the data preprocessing module. For example, the weights of resource consumption factors determined by the entropy weight method (such as the weights of indicators like water resource consumption intensity and energy consumption intensity) affect the importance of each objective function in the fitness function, thereby guiding the genetic algorithm to optimize towards indicators with higher weights. Simultaneously, the optimization schemes generated by the module (such as resource allocation parameters for each stage, energy consumption thresholds, and carbon emission control targets) will serve as input data for the correlation modeling and result generation module, used for generating and deciding on the full life-cycle analysis results.

[0094] Example 4:

[0095] The correlation modeling and results generation module performs correlation modeling on data from each stage of the green product's entire lifecycle. The aim is to identify lifecycle stages that play a crucial role in the product's environmental impact and resource consumption by uncovering the inherent connections between data at each stage. This process, based on time-series data output from the data preprocessing module, achieves in-depth analysis of the entire lifecycle data through three core steps: feature sequence extraction, grey relational analysis calculation, and key stage identification.

[0096] The module needs to extract feature sequences from data at each stage of the green product's lifecycle. Feature sequence extraction involves filtering key variables from the raw data that effectively reflect the core attributes and changing patterns of each stage, thereby reducing data dimensionality and highlighting key information. Specifically, feature sequence extraction at the raw material acquisition stage focuses on data such as raw material types, mining energy consumption, and transportation distance. Dimensionality reduction techniques such as principal component analysis (PCA) are used to transform multidimensional data into a few comprehensive indicators, such as extracting the principal component of "resource acquisition intensity" (comprehensive mining energy consumption and transportation distance) and the indicator of "raw material diversity." At the production and processing stage, "production efficiency characteristics" (such as unit product energy consumption and equipment failure rate) and "environmental load characteristics" (waste generation density) are extracted from data such as process energy consumption, waste generation, and equipment efficiency. Feature sequences at the distribution and sales stage include logistics carbon emission intensity, warehousing cost ratio, and sales channel response speed. At the use and maintenance stage, features such as user usage frequency, maintenance cost ratio, and energy consumption volatility are extracted. At the recycling and processing stage, key indicators such as recycling rate, processing energy consumption intensity, and recycled material utilization rate are extracted. The extraction of feature sequences needs to be combined with industry characteristics and product features. The core feature variables of each stage should be determined through data mining algorithms or expert experience to ensure that the extracted feature sequences can accurately represent the main attributes of the corresponding stage.

[0097] After extracting the feature sequences at each stage, the module needs to calculate the grey relational degree between each feature sequence and the target sequence. Grey relational analysis is a method used to measure the degree of correlation between sequences, suitable for system analysis with limited data and incomplete or unclear information. The target sequence can be set according to the analysis requirements, for example, using the comprehensive environmental impact score of the entire life cycle as the target sequence, or selecting a single indicator such as total resource consumption or carbon emission intensity as the target sequence. The formula for calculating the grey relational degree is:

[0098] ;

[0099] in, Indicates the first The grey relational degree between the feature sequences and the target sequence at each life cycle stage, with values ​​ranging from [value range missing]. arrive Between these values, a larger value indicates a higher degree of correlation; For the first The absolute difference between the feature sequence and the target sequence at the corresponding time point in each life cycle stage reflects the degree of local difference between the two. For all The minimum value in the sequence represents the minimum difference between the feature sequence and the target sequence at each stage. For all The maximum value in the range represents the largest difference; The resolution coefficient typically ranges from 100 to 1000. arrive Used to adjust the resolution of difference information, generally taken as... To balance resolution and data stability.

[0100] Taking the correlation calculation between the characteristic sequence of raw material acquisition stage and the target sequence of the comprehensive environmental impact assessment throughout the entire life cycle as an example: Assume that at a certain point in time... The characteristic value of "resource acquisition intensity" in the raw material acquisition stage is The environmental impact score of the target sequence is Then the absolute difference at that time point By iterating through all time points, the absolute difference sequence between the feature sequence and the target sequence for that stage is calculated. to determine and . (The sentence is incomplete and lacks context. It appears to be a fragment from a larger text.) Substituting into the formula, the grey relational degree of this stage can be obtained. Similarly, the grey relational degree between the feature sequences and the target sequence at each stage of production and processing, distribution and sales, use and maintenance, and recycling can be calculated. .

[0101] In the calculation of grey relational analysis, it is important to pay attention to the dimensionless processing of the data. Since the units of the indicators in the feature sequences and target sequences at each stage may differ (e.g., energy consumption is in kWh, cost is in yuan, and the score is a dimensionless number), directly calculating the absolute difference will lead to results influenced by the units of measurement. Therefore, all sequences need to be standardized before calculation. Common methods include initialization (dividing each data point in the sequence by the first data point), mean normalization (dividing each data point in the sequence by the mean), or normalization (mapping the data to a normal value). (Interval), to ensure that indicators of different dimensions are comparable.

[0102] After calculating the grey relational degree for each stage, the module sorts the stages according to their relational degree values ​​to determine the critical lifecycle stages. The correlation degree ranking results intuitively reflect the degree of influence of each stage on the target sequence: the stage with the highest correlation degree indicates that its characteristic sequence and the target sequence have the most consistent trends, and its impact on the target is the most significant; this is the critical lifecycle stage. For example, if the grey relational degree of the production and processing stage... The highest value indicates that the characteristics of this stage (such as process energy consumption and waste generation) are most closely related to the environmental impact or resource consumption throughout the entire life cycle, and it is a key link affecting the green performance of products. After identifying the key stage, enterprises can focus on optimizing this stage, such as improving production processes to reduce energy consumption and increasing waste recycling rates, thereby more efficiently improving the overall environmental performance and resource utilization efficiency of products.

[0103] In practical applications, the correlation modeling process needs to consider multi-objective correlation analysis. For example, when environmental impact score, total resource consumption, and carbon emission intensity are simultaneously considered as multiple objective sequences, the module needs to calculate the grey relational degree between the characteristic sequence of each stage and each objective sequence, forming a correlation degree matrix. By comprehensively analyzing the correlation degree ranking under each objective, the key stages under different dimensions are determined, providing a more comprehensive basis for multi-objective optimization. For example, for a certain product, the key stage under the environmental impact dimension is the production and processing stage, while under the resource consumption dimension, the key stage is the raw material acquisition stage. The company can formulate phased optimization strategies based on the priority of different objectives.

[0104] Furthermore, grey relational analysis is dynamic, capable of calculating the correlation degree of each stage in real time based on the updates of time series data, reflecting changes in key stages of the product lifecycle. For example, when product upgrades lead to a significant reduction in energy consumption during the use and maintenance stage, the correlation degree between this stage and the environmental impact target sequence may decrease, while the correlation degree increases during the production and processing stage due to the introduction of new processes. The module can capture these changes in a timely manner, providing support for enterprises to dynamically adjust their management strategies.

[0105] The results of correlation modeling will serve as one of the inputs to the multi-objective decision-making model. In the correlation modeling and result generation module, the grey relational degree values ​​at each stage, along with data such as environmental impact scores and resource optimization schemes, constitute a decision matrix. Using multi-objective decision-making methods such as the ideal point method, the final life-cycle analysis results are generated. Therefore, the accuracy of correlation modeling directly affects the scientific nature of the decision-making process, requiring assurance of the rationality of feature sequence extraction, the standardization of data preprocessing, and the precision of grey relational degree calculation.

[0106] Example 5:

[0107] The multi-objective decision-making model in the correlation modeling and results generation module needs to integrate the results of environmental impact assessment, resource consumption optimization, and correlation modeling to form the final life-cycle analysis scheme. This process is based on the quantitative data output by each module, and through constructing a decision matrix, calculating the closeness of the schemes, and selecting the optimal solution, it achieves the transformation from multi-dimensional data to scientific decision-making. The implementation method is detailed below with specific examples.

[0108] Suppose that the life cycle analysis of a certain green home appliance involves three different design schemes (Scheme A, Scheme B, and Scheme C), each exhibiting different characteristics in terms of environmental impact, resource consumption, and stage correlation. A multi-objective decision-making model first needs to construct a decision matrix, using environmental impact scores, resource optimization indicators (such as total resource consumption, resource utilization efficiency, and carbon emission intensity), and grey relational analysis as decision variables. For example:

[0109] Environmental impact score: Calculated through the environmental impact assessment module, Option A scores 85 points, Option B scores 78 points, and Option C scores 90 points (the higher the score, the smaller the environmental impact).

[0110] Resource optimization metrics:

[0111] Total resource consumption (unit: tons of standard coal): Option A: 1200, Option B: 1150, Option C: 1300;

[0112] Resource utilization efficiency (percentage): Option A is 75%, Option B is 82%, and Option C is 70%;

[0113] Carbon emission intensity (unit: kg / 10,000 yuan of output value): Option A is 450, Option B is 420, and Option C is 480.

[0114] Grey relational degree: Taking the environmental impact of the whole life cycle as the target sequence, the average relational degree of the characteristic sequence of each stage in each scheme is calculated. Scheme A is 0.72, Scheme B is 0.68, and Scheme C is 0.75 (the higher the relational degree, the stronger the stage relational relationship).

[0115] The above data was organized into a decision matrix, with the horizontal column representing the decision variables (6 items in total: environmental impact score, total resource consumption, resource utilization efficiency, carbon emission intensity, and grey relational degree), and the vertical column representing the various options (3 rows). Since different variables have different dimensions and trends (e.g., environmental impact score is a positive indicator, the higher the value, the better; total resource consumption is a negative indicator, the lower the value, the better), the data needs to be standardized to eliminate these dimensional differences. For example, negative indicators such as total resource consumption and carbon emission intensity can be normalized by taking their reciprocals or maximizing them to transform them into positive indicators, ensuring all variables are aligned (the higher the value, the better).

[0116] After standardization, the ideal point method is used to calculate the closeness between each scheme and the ideal solution. The ideal solution is the optimal combination of variables in the decision matrix (e.g., the standardized value corresponding to the maximum environmental impact score of 90, the minimum total resource consumption of 1150, the maximum resource utilization efficiency of 82%, the minimum carbon emission intensity of 420, and the maximum grey relational degree of 0.75). The negative ideal solution is the worst combination of variables. The Euclidean distance between each scheme and the ideal solution and the negative ideal solution is calculated using the closeness formula:

[0117] ;

[0118] The proximity score ranges from [0,1], and the larger the value, the closer the solution is to the ideal solution.

[0119] Taking Scheme A as an example: calculate its distance from the ideal solution (such as the square root of the sum of the squared difference between the environmental impact score and 90, the square root of the sum of the squared difference between the standardized value of total resource consumption and the corresponding optimal value), and its distance from the negative ideal solution (the square root of the sum of the squared differences between each item and the worst value). Substitute these values ​​into the formula to obtain the closeness value. Similarly, calculate the closeness of Schemes B and C. Assuming the calculation results are a closeness value of 0.65 for Scheme A, 0.78 for Scheme B, and 0.59 for Scheme C, then Scheme B has the highest closeness value and is selected as the optimal scheme.

[0120] In practice, the weighting of decision variables needs to be combined with the entropy weighting method results from the data preprocessing module. For example, if the environmental impact score has a weight of 0.3, total resource consumption has a weight of 0.2, resource utilization efficiency has a weight of 0.2, carbon emission intensity has a weight of 0.2, and grey relational degree has a weight of 0.1 in the entropy weighting method, then when calculating the distance, the differences between the variables need to be weighted to highlight the impact of indicators with higher weights on the decision. For instance, although the environmental impact score of Scheme B is lower than that of Scheme C, its total resource consumption and carbon emission intensity are better, and the resource-related indicators have higher weights, making its overall relevance higher.

[0121] The following points should be noted when applying multi-objective decision-making models:

[0122] Comprehensiveness of decision variables: It should cover the core outputs of each module of the system, such as the comprehensive score of the environmental impact assessment module, the multi-objective solution of the resource consumption optimization module (total resource consumption, utilization efficiency, carbon emission intensity), and the grey relational degree of the correlation modeling module, to ensure that the decision is based on multi-dimensional analysis of the entire life cycle.

[0123] Accuracy of data standardization: Differences in the dimensions of different variables can lead to decision-making biases, so it is necessary to select an appropriate standardization method based on the nature of the variables. For example, rating variables can be directly normalized, while variables with physical dimensions (such as energy consumption and emissions) need to be standardized using extreme value methods or Z-score methods.

[0124] Objectivity of weight setting: Data-driven weight calculation methods such as entropy weighting are used to avoid the subjectivity of manually setting weights. For example, if the data on the raw material acquisition stage of a product varies greatly (such as significant differences in mining energy consumption among different suppliers), the entropy weight of the correlation index at that stage may be higher and play a more important role in decision-making.

[0125] To illustrate with another example: A car manufacturer analyzes the entire lifecycle options for electric vehicles, adding "policy and regulatory compliance score" (e.g., compliance with new energy vehicle subsidy policies and carbon emission regulations) and "user behavior impact index" (e.g., the impact of user charging frequency on battery life) as decision variables. When constructing the decision matrix, these two new variables are incorporated, and standardized values ​​and proximity are recalculated. If option D has the highest policy and regulatory score (due to compliance with the latest subsidy policies) and the best user behavior impact index (high charging infrastructure compatibility), even if its total resource consumption is slightly higher than option E, it may still be the optimal solution due to the high weighting of policy and user dimensions, demonstrating the adaptability of the decision model to external factors.

[0126] The optimal solution generated by the decision-making model needs to be fed back to each module for verification. For example, if the critical life cycle stage of the optimal solution is the production and processing stage (based on the correlation modeling results), it is necessary to check whether the energy consumption threshold setting of the resource consumption optimization module for this stage is reasonable, and whether the environmental impact assessment module's assessment of pollution emission indicators for this stage is comprehensive. If it is found that there is a data gap in the calculation of logistics carbon emissions for a certain solution in the distribution and sales stage, it is necessary to return to the data acquisition module to collect supplementary data such as transportation routes and vehicle types for this stage, and re-preprocess and analyze them to ensure the completeness of the decision-making basis.

[0127] The visualization and interactive module plays a crucial role in this process. For example, the variables in the decision matrix are displayed as a dynamic heatmap, and the performance of different options is compared using color depth (e.g., green indicates superiority, red indicates inferiority), helping decision-makers quickly identify the strengths and weaknesses of each option. The lifecycle stage relationship network diagram in the 3D map can show the correlation strength of each stage in the optimal solution (e.g., a thicker line connecting the production and processing and recycling stages indicates a significant synergistic optimization effect), providing intuitive guidance for the implementation of the solution.

[0128] Another characteristic of multi-objective decision-making models is their iterative nature. As the product lifecycle progresses (e.g., after entering the sales stage and acquiring actual user behavior data), the variable values ​​in the decision matrix can be updated, the proximity can be recalculated, and the optimal solution can be dynamically adjusted. For example, after a home appliance product is launched, it is found that the actual energy consumption of users is higher than the design expectation, leading to an increase in the grey relational degree in the use and maintenance stage. This necessitates a reassessment of the proximity of each solution and may trigger optimization of energy-saving design in the use stage.

[0129] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0130] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A big data-based green product life cycle analysis system, characterized in that, include: The data acquisition module is used to acquire data on the entire lifecycle of green products and related elements. The data preprocessing module performs multi-dimensional preprocessing on the green product lifecycle data and related element data acquired by the data acquisition module; The environmental impact assessment module is used to dynamically assess the environmental impact of green products throughout their entire life cycle. The resource consumption optimization module is used to perform multi-objective optimization of resource consumption throughout the entire life cycle of green products; The correlation modeling and result generation module is used to perform correlation modeling on data from each stage of the green product's entire life cycle, and to generate full life cycle analysis results by combining a multi-objective decision-making model. The full lifecycle data includes: raw material acquisition stage data, production and processing stage data, distribution and sales stage data, use and maintenance stage data, and recycling and disposal stage data. The related element data includes: environmental impact factor data, resource consumption factor data, carbon emission factor data, policy and regulatory factor data, and user behavior factor data. The environmental impact factor data includes: pollutant emission type data, emission concentration data, and environmental restoration cost data. The dynamic assessment of the environmental impact of green products throughout their entire life cycle includes: The environmental impact factor data were decomposed into pollutant emission type subsequence, emission concentration subsequence, and environmental restoration cost subsequence. The fuzzy membership function is constructed to fuzzify each subsequence. The calculation formula is as follows: ； wherein, represents the environmental impact factor, represents the evaluation index, represents the environmental impact factor, fuzzy membership degree of the evaluation index, is a fuzzy factor, represents the evaluation index, is a reference value; A comprehensive environmental impact score is generated by fusing fuzzy evaluation results using a weighted average operator. The aforementioned correlation modeling of data at each stage of the green product lifecycle includes: Extract the feature sequences of data from each stage, including the feature sequences of raw material acquisition, production and processing, distribution and sales, use and maintenance, and recycling and processing. The grey relational degree between each feature sequence and the target sequence is calculated using the following formula: ； in, Indicates the first Grey correlation degree between feature sequences and target sequences at each life cycle stage Indicates the first The absolute difference between the feature sequence and the target sequence at corresponding time points in each life cycle stage. Indicates all The minimum value in Indicates all The maximum value in, The resolution coefficient; Determine key lifecycle stages based on correlation ranking; The multi-objective decision model includes: Construct a decision matrix, using environmental impact score, resource optimization scheme, and grey relational degree as decision variables; The closeness of each scheme to the ideal solution is calculated using the ideal point method; The solution with the highest degree of similarity was selected as the optimal lifecycle analysis result.

2. The big data-based green product lifecycle analysis system according to claim 1, characterized in that, The process of preprocessing the green product lifecycle data and related element data acquired by the data acquisition module in multiple dimensions includes: Outlier removal and noise filtering are performed on the acquired raw data; Missing data is imputed, and the data is weighted based on the entropy weight method. The processed data is divided into time series based on the life cycle stages: raw material acquisition time series, production and processing time series, distribution and sales time series, use and maintenance time series, and recycling and processing time series.

3. The big data-based green product lifecycle analysis system according to claim 2, characterized in that, The multi-objective optimization of resource consumption throughout the entire life cycle of green products includes: The objective function for resource optimization is set as minimizing total resource consumption, maximizing resource utilization efficiency, and minimizing carbon emission intensity. Construct chromosome coding rules to map resource allocation schemes to chromosome gene sequences; The chromosome population is iteratively optimized through crossover, mutation, and selection operations until the preset convergence conditions are met.

4. The big data-based green product lifecycle analysis system according to claim 3, characterized in that, The full lifecycle data includes the following data: raw material acquisition stage data includes raw material type data, mining energy consumption data, and transportation distance data; production and processing stage data includes process energy consumption data, waste generation data, and equipment efficiency data; and distribution and sales stage data includes logistics carbon emission data, warehousing cost data, and sales channel type data.

5. The big data-based green product lifecycle analysis system according to claim 4, characterized in that, The resource consumption factor data includes: water resource consumption intensity data, energy consumption intensity data, and raw material utilization rate data; the carbon emission factor data includes: direct carbon emission data, indirect carbon emission data, and carbon sink offset data.

6. The big data-based green product lifecycle analysis system according to claim 5, characterized in that, The system also includes: The visualization and interaction module is used to visualize the analysis results through 3D maps and dynamic heat maps; The visualization and interactive module includes a three-dimensional map comprising: a life cycle stage relationship network diagram, a resource consumption intensity distribution map, and a carbon emission spatiotemporal evolution map; and a dynamic heat map used to display the changing trends of environmental impact scores in different regions.

Images