A building carbon emission reduction optimization method
Patent Information
- Application Number
- CN202610935822.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-26
- Publication Date
- 2026-09-15
Smart Images
Figure CN122760101A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of building cycle carbon emission technology, and in particular to an optimization method for reducing building carbon emissions. Background Technology
[0002] As a key area for carbon emissions in society, the construction industry has a high proportion of carbon emissions throughout its entire life cycle, a long management chain, and complex influencing factors. Traditional carbon emission accounting and emission reduction scheme design, which rely on manual experience, generally suffer from pain points such as fragmented data, insufficient accounting accuracy, single optimization dimensions, and weak multi-objective coordination capabilities, making it difficult to support the demand for refined and intelligent low-carbon building construction.
[0003] Therefore, providing an optimized method for reducing carbon emissions from buildings to overcome the difficulties of existing technologies is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0004] In view of this, the present invention provides a building carbon emission reduction optimization method, which outputs an optimal low-carbon solution that can be implemented under multiple constraints such as building structural performance, safety and quality, cost budget, and construction period, providing quantitative support and systematic tools for low-carbon decision-making in the building design, construction and operation stages.
[0005] To achieve the above objectives, the present invention adopts the following technical solution: An optimization method for reducing building carbon emissions includes the following steps: Collect data related to building carbon emissions to form a dataset, which includes basic data and process data of building projects; Knowledge graph ontology design and graph construction are based on the relationships between entities and attributes defined in the dataset. Carbon emission identification and factor matching are performed within the framework of knowledge graphs. Machine learning models are used to intelligently extrapolate from the knowledge graphs to assess the impact of different building material usage, equipment energy consumption, and construction processes on carbon emissions. An optimization algorithm is used on the simulation results to find the optimal building design scheme with minimizing carbon emissions as the main constraint.
[0006] Optionally, data sources include engineering documents, monitoring data, material databases, statistical reports, industry standards, technical specifications, and literature; data formats include text, images, and tables; data acquisition adopts an interface-based integration approach, compatible with the standard-defined PI outputs of the Dominant Integrator Model (IM), Enterprise Resource Planning (ERP), Project Management System (PM), Supply Chain Management (SCM), and on-site IoT data platforms, forming a unified and structured time-series dataset.
[0007] Optionally, constructing a knowledge graph includes: Extract entities and attributes and determine the relationships between entities to complete the knowledge graph ontology modeling. Entities include activity entities, material and energy entities, and influence and rule entities. Relationships between entities include causal influence relationships, spatiotemporal combination relationships, and rule mapping relationships. Entities are represented by nodes, and relationships between entities are represented by edges. Explicit attributes and semantics are added to each node and edge to connect nodes at each stage. By storing nodes and edges in a suitable graph database and using visualization tools to display the knowledge graph structure, a knowledge graph of building carbon emissions is formed.
[0008] Optional intelligent inference includes: Entity identification and standardization are performed on the material, equipment, energy and process data of building projects, and a multi-level optimization strategy is used to accurately retrieve and match the carbon emission factors corresponding to each entity from the knowledge graph; Based on the relationship between processes and energy, the system automatically retrieves the unit shift energy consumption factor corresponding to the equipment model, and then dynamically calculates the energy consumption and carbon emissions of each process by combining time and the number of equipment, forming a full life cycle carbon emission baseline that includes total carbon emissions, sub-items of carbon emissions, and key carbon sources. The graph inference engine synchronously infers all possible emission reduction measures upstream of key carbon sources and checks the prerequisites and applicable conditions of the measures, thereby generating the complete feasible solution space for the corresponding emission reduction measures.
[0009] Optionally, finding the optimal architectural design solution includes: The association rules in the building carbon emission knowledge graph are used to dynamically assemble a multi-objective optimization function and establish a multi-objective optimization model. In this model, one or more of the following variables are used as decision variables: material substitution variables, path selection variables, energy configuration variables, and process equipment combination variables. The objective functions are to minimize carbon emissions, costs, and construction period. A multi-objective particle swarm optimization algorithm is used to solve the multi-objective optimization model to obtain a set of Pareto optimal solutions; By using a multi-attribute decision-making method combined with user-preset preference weights, an optimal compromise solution is recommended from the set of solutions, and a specific list of measures is provided to form the optimal architectural design scheme.
[0010] Optionally, a segmented hybrid coding strategy can be used to set different codes for the optimization variables, forming hybrid coded particles; Initialize the particle population, including: randomly generate a predetermined number of particles, and for each newly generated particle, perform a legality check and repair operation based on the knowledge graph rules, until all particles in the generated population meet all the hard constraints set by the knowledge graph, forming a legal initial particle population; For each particle in the population, demapping and evaluation are performed, including: mapping the particle's encoding to specific entity content, calculating the optimization target value of the emission reduction scheme and the comprehensive constraint violation degree of the particle based on the entity corresponding content in the knowledge graph, forming a particle swarm consisting of three objective function values and one constraint violation degree; The Pareto dominance rule is used to update the individual historical best position and global leader position of the particles; Based on the current particle position, the individual's historical best position, and the position of the global leader selected from the external archive, the discretized velocity and position are updated. The update stops when the preset maximum number of iterations is reached, forming a set of Pareto optimal solutions.
[0011] Optional Pareto dominance rules include: Both the defining solution and the dominant solution satisfy all hard constraints, and the defining solution is not inferior to the dominant solution on the three optimization objectives, and is strictly superior to the dominant solution on at least one optimization objective; The solution is defined to satisfy all hard constraints, while the dominant solution does not. Neither the defining solution nor the dominating solution satisfies all hard constraints, but the constraint violation degree of the defining solution is less than that of the dominating solution.
[0012] Optionally, obtaining the best compromise includes: Provide an interactive preference collection interface to guide decision-makers in inputting their judgments on the relative importance of the three optimization objectives; Based on the decision-maker's input, the system uses fuzzy set theory or analytic hierarchy process to transform qualitative judgments into quantitative weight vectors: Obtain the Pareto optimal solution set output by the multi-objective optimization algorithm and construct the original evaluation matrix. The evaluation matrix is standardized using the linear scaling method, and the normalized score of each scheme on each objective is calculated. The comprehensive relative closeness of each scheme is calculated by combining the approximation ideal solution ranking method, and the best compromise scheme is obtained.
[0013] As can be seen from the above technical solution, compared with the prior art, the present invention provides a method for optimizing building carbon emission reduction, which has the following beneficial effects: 1) This invention integrates scattered carbon emission knowledge and data through a domain knowledge graph, realizes the structured and reasonable carbon emission logic, and the multi-level factor matching strategy ensures the accuracy of the calculation, solving the problems of data fragmentation and rough calculation in traditional methods; 2) This invention embeds the rule constraints of the knowledge graph into the entire optimization algorithm process, ensuring that all optimization schemes meet the hard constraints of the project; the hybrid encoding strategy adapts to multiple types of decision variables, and Pareto optimization takes into account the balance of multiple objectives and the diversity of solutions; 3) The interactive preference setting of this invention, combined with the multi-attribute decision-making method, not only ensures the objectivity of the optimization, but also fully incorporates the actual needs of the decision-maker's project, and the output solution has stronger feasibility and executability. Attached Figure Description
[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0015] Figure 1 This is a flowchart of the building carbon emission reduction optimization method disclosed in this invention; Figure 2 This is a library of carbon emission factors disclosed in this invention; Figure 3 This is an overall simulation diagram of the carbon emission reduction plan disclosed in this invention; Figure 4 This is a diagram illustrating the effect of the carbon emission reduction plan disclosed in this invention. Detailed Implementation
[0016] 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.
[0017] Reference Figure 1 As shown, this invention discloses a method for optimizing building carbon emission reduction, comprising the following steps: Collect data related to building carbon emissions to form a dataset, which includes basic data and process data of building projects; Knowledge graph ontology design and graph construction are based on the relationships between entities and attributes defined in the dataset. Carbon emission identification and factor matching are performed within the framework of knowledge graphs. Machine learning models are used to intelligently extrapolate from the knowledge graphs to assess the impact of different building material usage, equipment energy consumption, and construction processes on carbon emissions. An optimization algorithm is used on the simulation results to find the optimal building design scheme with minimizing carbon emissions as the main constraint.
[0018] Furthermore, the data sources include engineering documents, monitoring data, materials databases, statistical reports, industry standards, technical specifications, and literature; the data formats include text, images, and tables; data acquisition adopts an interface-based integration approach, compatible with the standard definition PI outputs of the Dominant Integrator Model (IM), Enterprise Resource Planning (ERP), Project Management System (PM), Supply Chain Management (SCM), and on-site IoT data platforms, forming a unified and structured time-series dataset.
[0019] Specifically, engineering document data includes design drawings, bill of quantities, construction organization design, technical disclosure documents, etc., covering various forms such as text, images, and tables, and carrying core information such as building structural parameters, material usage, and construction technology; The monitoring data consists of real-time time-series data from the IoT sensor network at the construction site, including equipment energy consumption, material consumption, environmental parameters, etc., which can reflect the dynamic carbon emission characteristics of the construction process. The materials database includes publicly available industry databases of building material carbon emission factors, equipment energy consumption parameters, and supply chain material information, providing basic parameter benchmarks for carbon emission accounting. Statistical report data: including monthly project progress reports, energy consumption statistics reports, cost accounting reports, etc., reflecting summary information on the phased operation of the project; Industry standards and technical specifications: including national standards such as the "Building Carbon Emission Calculation Standard" and local low-carbon technical regulations, providing a basis for the compliance of carbon emission accounting rules and emission reduction measures; The literature covers cutting-edge emission reduction technologies and performance parameters of new low-carbon materials, expanding the feasible boundaries of emission reduction measures.
[0020] Furthermore, for different forms of raw data, differentiated preprocessing techniques are first employed to perform data preprocessing, including: For text-based data, a fine-tuned domain-pre-trained language model is used to achieve named entity recognition and relation extraction, automatically extracting key information such as building material names, equipment models, process names, and technical parameters; For drawings and image data, structured information such as component dimensions, quantities, and equipment identification are extracted by combining optical character recognition (OCR) and computer vision target detection technologies. For tabular data, unstructured tables are converted into standardized two-dimensional data tables through header semantic recognition and cell mapping technology. For IoT time-series data, the 3σ principle is used for outlier detection, and missing values are handled by linear interpolation and K-nearest neighbor interpolation algorithms. Time-series alignment of multi-source data is completed based on a unified timestamp to ensure the consistency of the data in the time dimension.
[0021] The preprocessed data is organized into a three-tiered structure to form a standardized dataset, including: Basic data: Stores the static attributes of building projects, including building area, structural form, geographical location, design service life, and functional positioning; Process data: Stores dynamic time-series data, including building material usage, equipment shift data, energy consumption data, and construction progress data at each stage; Rule data: Stores rule-related information such as carbon emission factors, industry standard clauses, emission reduction technology parameters, and constraint thresholds.
[0022] Furthermore, constructing a knowledge graph includes: Extract entities and attributes and determine the relationships between entities to complete the knowledge graph ontology modeling. Entities include activity entities, material and energy entities, and influence and rule entities. Relationships between entities include causal influence relationships, spatiotemporal combination relationships, and rule mapping relationships. Entities are represented by nodes, and relationships between entities are represented by edges. Explicit attributes and semantics are added to each node and edge to connect nodes at each stage. By storing nodes and edges in a suitable graph database and using visualization tools to display the knowledge graph structure, a knowledge graph of building carbon emissions is formed.
[0023] Furthermore, the activities cover various process nodes throughout the entire building lifecycle, including scheme selection activities in the design phase, sub-project work activities in the construction phase, and equipment operation and maintenance activities in the operation phase. Entity attributes include activity name, duration, phase, spatial location, and responsible unit; Material energy entities include various building materials, construction and operation equipment, and various energy types. Entity attributes include specifications, unit carbon emission factor, market unit price, energy efficiency parameters, and supply radius.
[0024] The impacts and rules entities comprise four main subcategories: carbon emission accounting rules, low-carbon emission reduction measures, industry standard clauses, and building performance constraints. Entity attributes include rule number, scope of application, constraint threshold, emission reduction efficiency, implementation cost, and technological maturity.
[0025] Causal relationships describe the logic of carbon emission transmission and interaction between entities; spatiotemporal combinations describe the temporal logic and spatial subordination of active entities; rule mapping relationships describe the corresponding constraint relationship between standard specifications and entity parameters.
[0026] Furthermore, intelligent inference includes: Entity identification and standardization are performed on material, equipment, energy, and process data in construction projects. A multi-level optimization strategy is used to accurately retrieve and match carbon emission factors corresponding to each entity from a knowledge graph. The corresponding carbon emission factor database is as follows: Figure 2 As shown; Based on the relationship between processes and energy, the system automatically retrieves the unit shift energy consumption factor corresponding to the equipment model, and then dynamically calculates the energy consumption and carbon emissions of each process by combining time and the number of equipment, forming a full life cycle carbon emission baseline that includes total carbon emissions, sub-items of carbon emissions, and key carbon sources. The graph inference engine synchronously infers all possible emission reduction measures upstream of key carbon sources and checks the prerequisites and applicable conditions of the measures, thereby generating a complete solution space for the corresponding emission reduction measures.
[0027] Furthermore, an entity alignment algorithm maps non-standard entity names in the project to standard entity nodes in the knowledge graph, unifying data standards and eliminating accounting errors caused by naming differences. Specifically, the multi-level optimization strategy involves: First-level exact match: Prioritizes searching for carbon emission factors that are completely consistent with the entity name, specifications, and technical parameters, and directly uses the corresponding values to ensure the highest accuracy; Second-level similarity matching: When no exact match is found, the entity's attributes are converted into feature vectors. The similarity with similar entities in the graph is calculated using cosine similarity. The three candidate factors with the highest similarity are selected, and recommended values are given based on the project scenario. Level 3 reasoning completion: When candidate factors are still missing, the carbon emission factor value range of the entity is calculated by reasoning through the rule mapping relationship of the knowledge graph and the attribute pattern of similar entities, and the recommended value is output and the confidence level is marked to ensure the completeness of carbon emission accounting.
[0028] Furthermore, based on the knowledge graph's "process-equipment-energy-carbon emission" relationship path, the energy consumption and carbon emissions of each process are dynamically calculated, expressed as follows: , in, This refers to the total carbon emissions from a single process. Let i be the number of devices of type i. This represents the cumulative shift duration for the corresponding equipment. This is the unit shift carbon emission factor for this type of equipment. Based on this, we trace the implicit carbon emissions upstream in the production and transportation stages of building materials, and extend the carbon emissions of equipment operation downstream in the operation stage. Finally, we form a carbon emission baseline covering the entire life cycle of building materials, on-site construction, operation and maintenance. The results output the total carbon emissions, the carbon emission ratio of each stage, and the contribution of each category of carbon sources. By ranking the contribution, we can accurately locate the key carbon source nodes with the highest proportion, and clarify the core direction for subsequent emission reduction optimization.
[0029] Furthermore, the search for the optimal architectural design solution includes: The association rules in the building carbon emission knowledge graph are used to dynamically assemble a multi-objective optimization function and establish a multi-objective optimization model. In this model, one or more of the following variables are used as decision variables: material substitution variables, path selection variables, energy configuration variables, and process equipment combination variables. The objective functions are to minimize carbon emissions, costs, and construction period. A multi-objective particle swarm optimization algorithm is used to solve the multi-objective optimization model to obtain a set of Pareto optimal solutions; By using a multi-attribute decision-making method combined with user-preset preference weights, an optimal compromise solution is recommended from the set of solutions, and a specific list of measures is provided to form the optimal architectural design scheme.
[0030] Furthermore, a segmented hybrid coding strategy is adopted to set different codes for the optimization variables, forming hybrid-coded particles; Initialize the particle population, including: randomly generating a predetermined number of particles; for each newly generated particle, perform a legality verification and repair operation based on knowledge graph rules: call the knowledge graph query interface to verify whether there are semantic conflicts in the scheme combinations implicit in the particle encoding; if a conflict is detected, forcibly reset the conflict encoding bit to a feasible value that satisfies all association constraints, until all particles in the generated population satisfy all hard constraints set by the knowledge graph, forming a legal initial particle population.
[0031] For each particle in the population, demapping and evaluation are performed, including: mapping the particle's encoding to specific entity content, including specific material alternatives, equipment selection lists, transportation route plans, and energy allocation strategies; calculating the optimization target value of the emission reduction plan and the comprehensive constraint violation degree of the particle based on the entity corresponding content in the knowledge graph; forming a particle swarm consisting of three objective function values and one constraint violation degree. The objective function values include: total carbon emissions: the sum of carbon emissions from material production, transportation, construction equipment, and on-site energy consumption; total additional cost: the sum of the cost increments of each alternative plan compared to the baseline plan; schedule impact value: the cumulative number of days that material replacement, equipment replacement, and process adjustment affect the schedule of the critical path; and the comprehensive constraint violation degree is a scalar that quantifies the violated and uncompromising rigid constraints in the plan.
[0032] The Pareto dominance rule is used to update the individual historical best position and global leader position of the particles; Based on the current particle position, the individual's historical best position, and the position of the global leader selected from the external archive, the discretized velocity and position are updated. The update stops when the preset maximum number of iterations is reached, forming a set of Pareto optimal solutions.
[0033] Furthermore, the encoding settings include: Material substitution coding segment: Integer coding is used, and the value of each bit represents the index of a specific substitution scheme selected from the "low-carbon alternative materials" entity set in the knowledge graph; Equipment selection coding segment: Integer coding is used, and the value of each bit represents the index of a specific equipment model selected for a specified process from the "Low-carbon / Energy-saving Construction Equipment" entity set in the knowledge graph; Transportation route coding segment: Integer coding is used, and the value of each bit represents the index of a transportation scheme selected for a specific building material from the "supplier-transportation mode" combination entity set in the knowledge graph; Energy configuration coding segment: Real number coding is used, and the value of each bit represents the proportion of renewable energy substitution in the corresponding construction period or area.
[0034] Hard constraints are set as building structural performance constraints, mandatory requirements of industry standards, safety and quality standards, and upper limits of resource supply, all of which must be met during the optimization process.
[0035] Furthermore, the Pareto dominance rule includes: Both the defining solution and the dominant solution satisfy all hard constraints, and the defining solution is not inferior to the dominant solution on the three optimization objectives, and is strictly superior to the dominant solution on at least one optimization objective; The solution is defined to satisfy all hard constraints, while the dominant solution does not. Neither the defining solution nor the dominating solution satisfies all hard constraints, but the constraint violation degree of the defining solution is less than that of the dominating solution.
[0036] The update rules include: Individual historical best (pbest) update: If the new position of a particle dominates its historical pbest, or if the two do not dominate each other, one is selected according to a random criterion based on crowding, then pbest is updated. Global external archive maintenance: An external archive with a fixed capacity is established to store all non-dominated solutions found by the algorithm so far (i.e., the Pareto front approximation set). When a new solution is added, if it is dominated by a solution in the archive, it is not added; if it dominates several old solutions in the archive, the dominated old solutions are deleted; if the archive reaches its capacity limit, an adaptive grid method is used to delete a solution located in the densest grid cell to maintain the distribution of the solution set.
[0037] Furthermore, obtaining the best compromise solution includes: Provide an interactive preference collection interface to guide decision-makers in inputting their judgments on the relative importance of the three optimization objectives; Based on the decision-maker's input, the system uses fuzzy set theory or analytic hierarchy process to transform qualitative judgments into quantitative weight vectors: Obtain the Pareto optimal solution set output by the multi-objective optimization algorithm and construct the original evaluation matrix. The evaluation matrix is standardized using the linear scaling method, and the normalized score of each scheme on each objective is calculated. The comprehensive relative closeness of each scheme is calculated by combining the approximation ideal solution ranking method, and the best compromise scheme is obtained.
[0038] Furthermore, the particle codes corresponding to the selected optimal compromise are fully decoded, mapping the abstract decision variables they contain to specific, actionable combinations of emission reduction measures. The decoding process calls the entity query interface of the knowledge graph to restore the index values in the codes to entity information with complete attributes. Different entities are set to different ranges to obtain the corresponding planned carbon emissions, such as... Figure 3 and Figure 4 As shown.
[0039] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for building carbon emission reduction optimization, characterized in that, Includes the following steps: Collect data related to building carbon emissions to form a dataset, which includes basic data and process data of building projects; Knowledge graph ontology design and graph construction are based on the relationship between entities and attributes defined in the dataset. Carbon emission identification and factor matching are performed within the framework of knowledge graphs. Machine learning models are used to intelligently extrapolate from the knowledge graphs to assess the impact of different building material usage, equipment energy consumption, and construction processes on carbon emissions. An optimization algorithm is used on the simulation results to find the optimal building design scheme with minimizing carbon emissions as the main constraint.
2. The building carbon emission reduction optimization method according to claim 1, characterized in that, Data sources include engineering documents, monitoring data, material databases, statistical reports, industry standards, technical specifications, and literature; data formats include text, images, and tables; data acquisition adopts an interface-based integration approach, compatible with the standard-defined PI outputs of the Dominant Integrator Model (IM), Enterprise Resource Planning (ERP), Project Management System (PM), Supply Chain Management (SCM), and on-site IoT data platforms, forming a unified structured time-series dataset.
3. The method for optimizing building carbon emission reduction according to claim 1, characterized in that, Building a knowledge graph includes: Extract entities and attributes and determine the relationships between entities to complete the knowledge graph ontology modeling. Entities include activity entities, material and energy entities, and influence and rule entities. Relationships between entities include causal influence relationships, spatiotemporal combination relationships, and rule mapping relationships. Entities are represented by nodes, and relationships between entities are represented by edges. Explicit attributes and semantics are added to each node and edge to connect nodes at each stage. By storing nodes and edges in a suitable graph database and using visualization tools to display the knowledge graph structure, a knowledge graph of building carbon emissions is formed.
4. The building carbon emission reduction optimization method according to claim 1, characterized in that, Intelligent simulation includes: Entity identification and standardization are performed on the material, equipment, energy and process data of building projects, and a multi-level optimization strategy is used to accurately retrieve and match the carbon emission factors corresponding to each entity from the knowledge graph; Based on the relationship between processes and energy, the system automatically retrieves the unit shift energy consumption factor corresponding to the equipment model, and then dynamically calculates the energy consumption and carbon emissions of each process by combining time and the number of equipment, forming a full life cycle carbon emission baseline that includes total carbon emissions, sub-items of carbon emissions, and key carbon sources. The graph inference engine synchronously infers all possible emission reduction measures upstream of key carbon sources and checks the prerequisites and applicable conditions of the measures, thereby generating a complete solution space for the corresponding emission reduction measures.
5. The building carbon emission reduction optimization method according to claim 4, characterized in that, Finding the optimal architectural design solution includes: The association rules in the building carbon emission knowledge graph are used to dynamically assemble a multi-objective optimization function and establish a multi-objective optimization model. In this model, one or more of the following variables are used as decision variables: material substitution variables, path selection variables, energy configuration variables, and process equipment combination variables. The objective functions are to minimize carbon emissions, costs, and construction period. A multi-objective particle swarm optimization algorithm is used to solve the multi-objective optimization model to obtain a set of Pareto optimal solutions; By using a multi-attribute decision-making method combined with user-preset preference weights, an optimal compromise solution is recommended from the set of solutions, and a specific list of measures is provided to form the optimal architectural design scheme.
6. The method for optimizing building carbon emission reduction according to claim 5, characterized in that, A segmented hybrid coding strategy is adopted to set different codes for the optimization variables, forming hybrid-coded particles; Initialize the particle population, including: randomly generate a predetermined number of particles, and for each newly generated particle, perform a legality check and repair operation based on the knowledge graph rules, until all particles in the generated population meet all the hard constraints set by the knowledge graph, forming a legal initial particle population; For each particle in the population, demapping and evaluation are performed, including: mapping the particle's encoding to specific entity content, calculating the optimization target value of the emission reduction scheme and the comprehensive constraint violation degree of the particle based on the entity corresponding content in the knowledge graph, forming a particle swarm consisting of three objective function values and one constraint violation degree; The Pareto dominance rule is used to update the individual historical best position and global leader position of the particles; Based on the current particle position, the individual's historical best position, and the position of the global leader selected from the external archive, the discretized velocity and position are updated. The update stops when the preset maximum number of iterations is reached, forming a set of Pareto optimal solutions.
7. The building carbon emission reduction optimization method according to claim 6, characterized in that, Pareto dominance rules include: Both the defining solution and the dominant solution satisfy all hard constraints, and the defining solution is not inferior to the dominant solution on the three optimization objectives, and is strictly superior to the dominant solution on at least one optimization objective; The solution is defined to satisfy all hard constraints, while the dominant solution does not. Neither the defining solution nor the dominating solution satisfies all hard constraints, but the constraint violation degree of the defining solution is less than that of the dominating solution.
8. The method for optimizing building carbon emission reduction according to claim 5, characterized in that, The best compromise solutions include: Provide an interactive preference collection interface to guide decision-makers in inputting their judgments on the relative importance of the three optimization objectives; Based on the decision-maker's input, the system uses fuzzy set theory or analytic hierarchy process to transform qualitative judgments into quantitative weight vectors: Obtain the Pareto optimal solution set output by the multi-objective optimization algorithm and construct the original evaluation matrix. The evaluation matrix is standardized using the linear scaling method, and the normalized score of each scheme on each objective is calculated. The comprehensive relative closeness of each scheme is calculated by combining the approximation ideal solution ranking method, and the best compromise scheme is obtained.