AI-assisted green lighting scheme evaluation and intelligent decision method and system

Abstract

The application provides an AI-assisted green lighting scheme evaluation and intelligent decision method and system, relates to the field of intelligent lighting, and comprises the following steps: identifying lighting demands of different spatial positions, dividing a region into multiple sub-regions, and determining evaluation weights; constructing an index coupling relationship graph to determine an optimization sequence; identifying public controllable factors to generate a preliminary adjustment scheme; predicting a change amount and optimizing the scheme through propagation calculation; and finally generating a control instruction for each sub-region. The application realizes accurate allocation of lighting resources, improves energy utilization efficiency, and meets specific lighting demands of different regions.

Description

TECHNICAL FIELD

[0001] The present application relates to intelligent lighting technology, and in particular to an AI-assisted green lighting scheme evaluation and intelligent decision-making method and system. BACKGROUND

[0002] With the increasing global demand for energy saving and emission reduction and sustainable development, green lighting has become an important part of building energy saving. Green lighting not only focuses on reducing energy consumption, but also pays attention to multi-dimensional evaluation indicators such as lighting quality, visual comfort and human health impact. Traditional lighting systems mainly rely on pre-set lighting schemes and simple on-off control, and cannot meet the differentiated lighting needs of different spaces, different time periods and different users. In recent years, with the development of Internet of Things technology and artificial intelligence, intelligent lighting systems have begun to be widely used in buildings, collecting environmental data through various sensors and adjusting lighting through intelligent algorithms to achieve an optimal balance between lighting effect and energy consumption.

[0003] However, the existing method usually evaluates the evaluation area as a whole, ignoring the lighting demand differences of different spatial positions in the same area, resulting in uneven distribution of lighting resources, excessive lighting in some areas and insufficient lighting in other areas, which not only wastes energy but also affects user experience. Traditional lighting evaluation indicators are often considered independently, without fully considering the complex coupling relationship between indicators. For example, there are both constraints and synergistic effects between illuminance and glare, color temperature and energy consumption. The optimization of a single indicator may lead to the deterioration of other indicators, making it difficult to achieve overall optimization of multiple indicators. Existing intelligent lighting control systems mostly use trial-and-error adjustment methods, lack of forward-looking prediction of the impact of adjustment, and cannot evaluate the overall effect of the adjustment scheme before executing the control, resulting in frequent adjustment, system shock, and increased energy consumption and user experience. SUMMARY

[0004] The embodiment of the present application provides an AI-assisted green lighting scheme evaluation and intelligent decision-making method and system, which can solve the problems in the prior art.

[0005] In a first aspect, the embodiment of the present application provides an AI-assisted green lighting scheme evaluation and intelligent decision-making method, comprising:

[0006] Identifying the lighting demand differences of different spatial positions in the evaluation area, dividing the evaluation area into multiple sub-areas, determining the evaluation weight for each sub-area, and generating a multi-dimensional evaluation indicator based on the evaluation weight;

[0007] Constructing a coupling relationship graph containing the constraint relationship and synergistic enhancement relationship between indicators based on the multi-dimensional evaluation indicator, determining an indicator optimization sequence according to the constraint relationship, and performing influence source reverse tracing on the to-be-optimized indicators based on the indicator optimization sequence to generate a set of controllable factors;

[0008] identify a common controllable factor from the set of controllable factors that can simultaneously affect multiple to-be-optimized indicators, trigger a synergistic enhancement relationship in the coupling relationship graph by adjusting the common controllable factor, and generate a preliminary adjustment scheme;

[0009] substitute the preliminary adjustment scheme into the coupling relationship graph for propagation calculation to obtain a predicted change amount, when the predicted change amount shows that there is indicator degradation, cut off the degradation propagation path in the coupling relationship graph, re-identify the common controllable factor, and generate a final adjustment scheme;

[0010] generate a control instruction for multiple sub-regions based on the final adjustment scheme, and send the control instruction to the lighting device of the corresponding sub-region.

[0011] identify the lighting demand differences of different spatial positions in the to-be-evaluated region, divide the to-be-evaluated region into multiple sub-regions, determine an evaluation weight for each sub-region, and generate a multi-dimensional evaluation indicator based on the evaluation weight, including:

[0012] obtain lighting demand parameters of multiple spatial position points in the to-be-evaluated region;

[0013] construct a demand difference matrix between the spatial position points based on the lighting demand parameters, and traverse the demand difference matrix to identify spatial position points with a deviation amount exceeding a preset deviation threshold as boundary position points;

[0014] divide the to-be-evaluated region into multiple sub-regions with the boundary position points as boundaries, and aggregate the lighting demand parameters of the spatial position points in each sub-region to obtain sub-region characteristic parameters;

[0015] statistically analyze the deviation degree distribution of the lighting demand parameters of the spatial position points in the sub-region relative to the sub-region characteristic parameters, calculate a demand fluctuation index based on the deviation degree distribution, and take the ratio of the demand fluctuation index to a preset reference fluctuation index as a fluctuation coefficient;

[0016] weight and combine the fluctuation coefficient and the sub-region characteristic parameters to obtain an evaluation weight, and calculate an energy efficiency evaluation value, a light quality evaluation value, and an environmental impact evaluation value for each sub-region based on the evaluation weight to form a multi-dimensional evaluation indicator.

[0017] construct a coupling relationship graph containing a constraint relationship and a synergistic enhancement relationship between indicators based on the multi-dimensional evaluation indicator, determine an indicator optimization sequence according to the constraint relationship, and perform influence source reverse tracing on the to-be-optimized indicators based on the indicator optimization sequence to generate a set of controllable factors, including:

[0018] collect change history data of the multi-dimensional evaluation indicator, calculate a response time delay and a change synchronization between indicators based on the history data, and generate a constraint relationship matrix and a synergistic enhancement relationship matrix, respectively;

[0019] The multi-dimensional evaluation index is constructed as an index node, the response time value is extracted from the constraint relationship matrix to form a constraint relationship edge between the index nodes, the change synchronization value is extracted from the synergistic enhancement relationship matrix to form a synergistic enhancement relationship edge between the index nodes, the constraint relationship edge and the synergistic enhancement relationship edge are superimposed between the index nodes to form a coupling relationship graph;

[0020] The constraint relationship edge is extracted from the coupling relationship graph to construct a constraint relationship link, the number of constraint relationship edges connected by the index nodes in the constraint relationship link is calculated to obtain a constraint relationship strength, and the constraint relationship strength is sorted to construct an index optimization sequence;

[0021] The index node with the maximum constraint relationship strength is selected from the index optimization sequence as a to-be-optimized index, and the to-be-optimized index is taken as a starting point to trace back along the constraint relationship link to identify a source node, and an influence factor tree is constructed;

[0022] The number of constraint relationship edges and synergistic enhancement relationship edges of the end node of the influence factor tree in the coupling relationship graph is counted to obtain an influence propagation coefficient, the execution condition of the end node is counted to obtain an execution cost coefficient, the influence propagation coefficient and the execution cost coefficient are multiplied to obtain a regulation and control cost coefficient, and the end node with a regulation and control cost coefficient less than a preset cost threshold is screened to form a regulatable factor set.

[0023] The constraint relationship edge is extracted from the coupling relationship graph to construct a constraint relationship link, the number of constraint relationship edges connected by the index nodes in the constraint relationship link is calculated to obtain a constraint relationship strength, and the constraint relationship strength is sorted to construct an index optimization sequence including:

[0024] The index node pairs in the coupling relationship graph are traversed, the constraint relationship edge weight values between the index node pairs are obtained, and an adjacency matrix of the index nodes is constructed;

[0025] A weight value dense distribution area is identified in the adjacency matrix to form an index node group, and the variance and mean of the weight values in the index node group are calculated to obtain a node group difference value;

[0026] The constraint relationship edges are grouped and aggregated based on the node group difference value to form a constraint relationship edge group, the transmission path length value and the branch number value of the constraint relationship edges in the constraint relationship edge group are calculated, and a link complexity value of the constraint relationship link is constructed;

[0027] The distribution probability value of the link complexity value at different levels of the constraint relationship link is calculated, the distribution probability value is used to construct a cumulative density function, the inflection point position value is extracted from the cumulative density function to divide a segmented interval of the constraint relationship strength, and the constraint relationship link is sorted according to the constraint relationship strength value in the segmented interval from high to low to generate an index optimization sequence.

[0028] identify a common controllable factor that can simultaneously affect multiple to-be-optimized indicators from the set of controllable factors, trigger a synergistic enhancement relationship in the coupling relationship graph by adjusting the common controllable factor, and generate a preliminary adjustment scheme including:

[0029] Calculate the number of to-be-optimized indicators connected in the coupling relationship graph for each controllable factor in the set of controllable factors to obtain an indicator coverage value, calculate the number of synergistic enhancement relationship edges connected for each controllable factor to obtain a synergy value, and multiply the indicator coverage value by the synergy value to obtain a comprehensive evaluation value;

[0030] Filter the controllable factors whose comprehensive evaluation values exceed a preset evaluation threshold as common controllable factors, and calculate the ratio of the comprehensive evaluation value of each common controllable factor to the sum of the comprehensive evaluation values of all common controllable factors to obtain an adjustment weight value;

[0031] Multiply the adjustment weight value of each common controllable factor by the propagation coefficient of the corresponding synergistic enhancement relationship edge to obtain an adjustment gain value, and filter the common controllable factor with the largest adjustment gain value as the dominant adjustment factor;

[0032] Generate a reference adjustment amount based on the adjustment weight value of the dominant adjustment factor, distribute the reference adjustment amount to the remaining common controllable factors according to the proportion of the adjustment weight value to obtain a linkage adjustment amount, adjust the parameters of the common controllable factors according to the linkage adjustment amount, and generate a preliminary adjustment scheme.

[0033] Substitute the preliminary adjustment scheme into the coupling relationship graph for propagation calculation to obtain a predicted change amount, when the predicted change amount shows that there is an indicator degradation, cut off the degradation propagation path in the coupling relationship graph, and then re-identify the common controllable factor to generate a final adjustment scheme including:

[0034] Substitute the adjustment parameters of the common controllable factors in the preliminary adjustment scheme into the corresponding positions in the coupling relationship graph to construct an initial state table, perform multiple rounds of propagation calculation based on the propagation rules of the synergistic enhancement relationship edges, record the numerical changes and propagation path information of the to-be-optimized indicators after each round of calculation, and generate a predicted change amount table;

[0035] Filter the to-be-optimized indicators with negative numerical changes from the predicted change amount table as degradation indicators, and extract the propagation paths corresponding to the degradation indicators to construct a propagation path table;

[0036] Calculate the combination value of the product of the propagation coefficients of each path in the propagation path table and the path length, and filter the propagation path with the largest combination value as the to-be-cut path;

[0037] Remove the synergistic enhancement relationship edge on the to-be-cut path in the coupling relationship graph to form an optimized graph, count the number of connection edges of each public controllable factor in the optimized graph to obtain a node connection value, calculate the sum of the propagation coefficients of the connection edges to obtain a propagation influence value, and multiply the node connection value by the propagation influence value to obtain a regulation evaluation value;

[0038] Select the public controllable factor with the maximum regulation evaluation value as a master control factor, calculate a regulation reference amount based on the regulation evaluation value of the master control factor, distribute the regulation reference amount to the public controllable factors in proportion to the regulation evaluation value through the synergistic enhancement relationship edge in the optimized graph, and generate a final adjustment scheme.

[0039] Based on the final adjustment scheme, generate regulation instructions for the multiple sub-regions, and send the regulation instructions to the lighting devices of the corresponding sub-regions, comprising:

[0040] Obtain the adjustment parameters of the public controllable factors in the final adjustment scheme and the historical adjustment data of the sub-region lighting devices, calculate the change trend of the lighting parameters in the historical adjustment data, and construct a frequency domain response matrix between the lighting parameters;

[0041] Based on the frequency domain response matrix, analyze the frequency components of the lighting parameter changes, extract the resonant frequency and the resonance frequency in the lighting parameters, and generate frequency regulation instructions for the multiple sub-regions;

[0042] Collect the output spectrum data of the multiple sub-region lighting devices when executing the frequency regulation instructions, calculate the energy transfer coefficient between adjacent wavelength segments, perform coupling operation on the energy transfer coefficient and the frequency regulation instructions to obtain a spectrum compensation amount, adjust the execution period of the frequency regulation instructions based on the spectrum compensation amount, and generate energy balance instructions for the multiple sub-regions;

[0043] Monitor the driving circuit temperature and the light emitting element temperature of the multiple sub-region lighting devices when executing the energy balance instructions, and construct a gradient field distribution of temperature change;

[0044] Calculate the heat accumulation area and the heat diffusion rate in the gradient field, modify the energy balance instructions, generate regulation instructions for the multiple sub-regions, and send the regulation instructions to the lighting devices of the corresponding sub-regions.

[0045] The second aspect of the embodiment of the application provides a green lighting scheme evaluation and intelligent decision system based on AI assistance, comprising:

[0046] The first unit is configured to identify the lighting demand differences of different spatial positions in the to-be-evaluated region, divide the to-be-evaluated region into multiple sub-regions, determine an evaluation weight for each sub-region, and generate a multi-dimensional evaluation index based on the evaluation weight;

[0047] A second unit is configured to construct a coupling relationship graph containing constraint relationships and synergistic enhancement relationships among the indicators based on the multi-dimensional evaluation indicators, determine an indicator optimization sequence according to the constraint relationships, perform influence source reverse tracking on the to-be-optimized indicators based on the indicator optimization sequence, and generate a set of controllable factors;

[0048] A third unit is configured to identify a common controllable factor capable of simultaneously affecting multiple to-be-optimized indicators from the set of controllable factors, trigger the synergistic enhancement relationships in the coupling relationship graph by adjusting the common controllable factor, and generate a preliminary adjustment scheme.

[0049] A fourth unit is configured to substitute the preliminary adjustment scheme into the coupling relationship graph to perform propagation calculation and obtain a predicted change amount, re-identify the common controllable factor after cutting off a deterioration propagation path in the coupling relationship graph when the predicted change amount shows that there is indicator deterioration, and generate a final adjustment scheme.

[0050] A fifth unit is configured to generate a control instruction for multiple sub-regions based on the final adjustment scheme, and send the control instruction to a lighting device in the corresponding sub-region.

[0051] In a third aspect, an electronic device is provided, including:

[0052] a processor;

[0053] a memory for storing processor-executable instructions;

[0054] The processor is configured to invoke the instructions stored in the memory to execute the method described above.

[0055] In a fourth aspect, a computer-readable storage medium is provided, which stores computer program instructions. When the computer program instructions are executed by a processor, the method described above is implemented.

[0056] In the embodiment, by dividing the to-be-evaluated region into multiple sub-regions and generating multi-dimensional evaluation indexes based on evaluation weights, the lighting demand differences of different spatial positions can be accurately identified, the targeted optimization of the lighting scheme is realized, and the lighting quality and user experience are effectively improved. By constructing a coupling relationship graph containing the constraint relationship and synergistic enhancement relationship between indexes, the internal correlation mechanism of the complex lighting system is deeply modeled, and the synergistic enhancement relationship is triggered by identifying the public controllable factors, thereby maximizing the reduction of energy consumption while ensuring the lighting quality, and significantly improving the energy utilization efficiency of the lighting system. On the basis of the coupling relationship graph, the propagation algorithm is performed and the deterioration propagation path is cut off, thereby effectively preventing the index deterioration problem that may occur in the optimization process, ensuring the balanced development of the performance indexes of the lighting system, and realizing the intelligent automatic adjustment of the entire lighting system by generating differentiated adjustment instructions for multiple sub-regions, thereby greatly improving the stability and adaptability of the system operation. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 A flowchart of an AI-assisted green lighting scheme evaluation and intelligent decision-making method according to an embodiment of the present application is shown in

[0058] Figure 2 A flowchart of an index optimization sequence generation according to an embodiment of the present application is shown in DETAILED DESCRIPTION

[0059] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in conjunction with the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0060] The technical solutions of the present application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described in some embodiments.

[0061] Figure 1 A flowchart of an AI-assisted green lighting scheme evaluation and intelligent decision-making method according to an embodiment of the present application is shown in Figure 1 The method comprises:

[0062] identifying the lighting demand differences of different spatial positions in the to-be-evaluated region, dividing the to-be-evaluated region into multiple sub-regions, determining the evaluation weights for each sub-region, and generating multi-dimensional evaluation indexes based on the evaluation weights;

[0063] construct a coupling relationship graph containing the constraint relationship and the synergistic enhancement relationship among the indicators based on the multi-dimensional evaluation indicators, and determine an indicator optimization sequence according to the constraint relationship, perform influence source reverse tracing on the to-be-optimized indicators based on the indicator optimization sequence, and generate a set of controllable factors;

[0064] identify a public controllable factor capable of simultaneously affecting multiple to-be-optimized indicators from the set of controllable factors, trigger the synergistic enhancement relationship in the coupling relationship graph by adjusting the public controllable factor, and generate a preliminary adjustment scheme;

[0065] substitute the preliminary adjustment scheme into the coupling relationship graph for propagation calculation to obtain a predicted change amount, when the predicted change amount shows that there is an indicator degradation, re-identify the public controllable factor after cutting off the degradation propagation path in the coupling relationship graph, and generate a final adjustment scheme;

[0066] generate a control instruction for multiple sub-regions based on the final adjustment scheme, and send the control instruction to the lighting device of the corresponding sub-region.

[0067] In an optional implementation, the lighting demand differences of different spatial positions in the to-be-evaluated region are identified, the to-be-evaluated region is divided into multiple sub-regions, the evaluation weight of each sub-region is determined, and the multi-dimensional evaluation indicators are generated based on the evaluation weight, including:

[0068] obtain lighting demand parameters of multiple spatial position points in the to-be-evaluated region;

[0069] construct a demand difference matrix between the spatial position points based on the lighting demand parameters, and identify a spatial position point with a deviation amount exceeding a preset deviation threshold as a boundary position point by traversing the demand difference matrix;

[0070] divide the to-be-evaluated region into multiple sub-regions with the boundary position point as a boundary, and perform aggregation operation on the lighting demand parameters of the spatial position points in each sub-region to obtain a sub-region characteristic parameter;

[0071] statistically analyze the deviation degree distribution of the lighting demand parameters of the spatial position points in the sub-region relative to the sub-region characteristic parameter, calculate a demand fluctuation index based on the deviation degree distribution, and take the ratio of the demand fluctuation index to a preset reference fluctuation index as a fluctuation coefficient;

[0072] perform weighted combination operation on the fluctuation coefficient and the sub-region characteristic parameter to obtain the evaluation weight, calculate the energy efficiency evaluation value, the light quality evaluation value, and the environmental impact evaluation value of each sub-region based on the evaluation weight, and construct the multi-dimensional evaluation indicators.

[0073] The embodiment firstly acquires the lighting demand parameters of multiple spatial position points in the to-be-evaluated region. In a specific example, the to-be-evaluated region is an office environment, and 20 spatial position points are selected, and the lighting demand parameters of each position point are recorded, including the required illuminance value (unit: lux), the color temperature requirement (unit: Kelvin), the color rendering index requirement (unit: Ra), and the glare limit value (unit: UGR). For example, the position point P1 in the working area has the demand parameters of 500 lux of illuminance, 4000 Kelvin of color temperature, 85 Ra of color rendering index, and 19 UGR of glare limit value; and the position point P15 in the rest area has the demand parameters of 300 lux of illuminance, 3500 Kelvin of color temperature, 80 Ra of color rendering index, and 22 UGR of glare limit value. These data are acquired by photometric measuring instruments or preset parameter tables based on the types of functional areas.

[0074] Based on the lighting demand parameters, a demand difference matrix between the spatial position points is constructed. For each pair of spatial position points P i and P j , the differences in the lighting demand parameters are calculated and composed into a difference vector. For example, the illuminance difference between the position points P1 and P15 is 200 lux, the color temperature difference is 500 Kelvin, the color rendering index difference is 5 Ra, and the glare limit value difference is 3 UGR. In order to comprehensively consider the differences in the parameters, the difference vector is normalized by dividing each parameter value by the corresponding standard deviation. In this example, the standard deviations of the illuminance, color temperature, color rendering index, and glare limit value are 100 lux, 300 Kelvin, 3 Ra, and 2 UGR, respectively. The square root of the sum of the squares of the elements of the normalized difference vector is taken as the comprehensive deviation quantity.

[0075] The demand difference matrix is traversed, and the spatial position points with the deviation quantity exceeding a preset deviation threshold are identified as the boundary position points. In this example, the preset deviation threshold is set to 1.5. Through traversal, it is found that the comprehensive deviation quantities between the position points P8 and P9 and between the position points P12 and P13 are 1.78 and 1.65, respectively, both of which exceed the preset threshold, and therefore P8, P9, P12, and P13 are marked as the boundary position points.

[0076] The to-be-evaluated region is divided into multiple sub-regions with these boundary position points as boundaries. In this example, the office is divided into three sub-regions: the sub-region A contains the position points P1 to P8, the sub-region B contains the position points P9 to P12, and the sub-region C contains the position points P13 to P20. The lighting demand parameters of the spatial position points in each sub-region are aggregated to obtain the characteristic parameters of the sub-region. The aggregation operation adopts a weighted average method, and the weights are allocated according to the use frequency or importance of the position points. For example, the weight of the working position point in the sub-region A is 0.7, and the weight of the passage position point is 0.3. The calculated characteristic parameters of the sub-region A are 480 lux of illuminance, 4200 Kelvin of color temperature, 86 Ra of color rendering index, and 18 UGR of glare limit value.

[0077] The deviation distribution of the lighting demand parameters of the spatial position points in the statistical sub-area from the characteristic parameters of the sub-area is calculated. The deviation of each position point is calculated by subtracting the characteristic parameter of the sub-area from the demand parameter of the position point, and then dividing by the characteristic parameter of the sub-area. Based on the deviation distribution, the demand fluctuation index is calculated, which is the standard deviation of the deviations of all position points. In sub-area A, the standard deviation of the deviations of the illuminance parameter is 0.08, the standard deviation of the deviations of the color temperature is 0.05, the standard deviation of the deviations of the color rendering index is 0.03, and the standard deviation of the deviations of the glare limit value is 0.06. The average of these standard deviations is taken as the demand fluctuation index, and the demand fluctuation index of sub-area A is 0.055. The ratio of the demand fluctuation index to the preset reference fluctuation index (0.05 in this example) is taken as the fluctuation coefficient, and the fluctuation coefficient of sub-area A is 1.1.

[0078] The evaluation weight is calculated by weighted combination of the fluctuation coefficient and the characteristic parameters of the sub-area. Specifically, the dimensions of the characteristic parameters of the sub-area (illuminance, color temperature, color rendering index, and glare limit value) are multiplied by different coefficients and then added together, and then multiplied by the fluctuation coefficient. In this example, the illuminance coefficient is 0.4, the color temperature coefficient is 0.3, the color rendering index coefficient is 0.2, and the glare limit value coefficient is 0.1. The evaluation weight of sub-area A is calculated to be 290.4.

[0079] Based on the evaluation weight, the energy efficiency evaluation value, the light quality evaluation value, and the environmental impact evaluation value are calculated for each sub-area. The energy efficiency evaluation value is calculated as the ratio of the lighting power density to the evaluation weight, and the lighting power density in sub-area A is 10 watts per square meter, so the energy efficiency evaluation value is 10 / 290.4 = 0.034. The light quality evaluation value is calculated by measuring the compliance of the actual lighting parameters with the characteristic parameters of the sub-area, for example, the compliance rate of the actual illuminance of sub-area A with the characteristic parameter of 480 lux is 95.8%, and similarly the compliance rates of other parameters are calculated and weighted to obtain the light quality evaluation value of 92.5%. The environmental impact evaluation value is calculated based on the environmental friendliness of the lamp material, the life cycle assessment, and the energy consumption, and sub-area A uses LED lamps, with an environmental friendliness score of 85, a life cycle score of 90, and an energy consumption score of 88, and the weighted average of the environmental impact evaluation value is 87.5.

[0080] These three evaluation values constitute a multi-dimensional evaluation index, providing detailed reference for the optimization and improvement of the lighting system. For example, sub-area B has a low energy efficiency evaluation value, and high-efficiency lamps can be considered for replacement; sub-area C has a low light quality evaluation value, and the lighting configuration needs to be adjusted to improve visual comfort.

[0081] In an alternative embodiment, a coupling relationship graph containing the constraint relationship and synergistic enhancement relationship between indicators is constructed based on multi-dimensional evaluation indicators, and the indicator optimization sequence is determined according to the constraint relationship. The influence source is traced in reverse based on the indicator optimization sequence to generate a set of controllable factors, including:

[0082] The change history data of the multi-dimensional evaluation indicators is collected, and the response time delay and change synchronization between the indicators are calculated based on the historical data to generate a constraint relationship matrix and a synergistic enhancement relationship matrix, respectively.

[0083] The multi-dimensional evaluation indicators are constructed as indicator nodes, the response time delay values between the indicator nodes are extracted from the constraint relationship matrix to form constraint relationship edges, and the change synchronization values between the indicator nodes are extracted from the synergistic enhancement relationship matrix to form synergistic enhancement relationship edges. The constraint relationship edges and the synergistic enhancement relationship edges are superimposed between the indicator nodes to form a coupling relationship graph.

[0084] The constraint relationship edges are extracted from the coupling relationship graph to construct a constraint relationship link, and the constraint relationship strength is calculated by counting the number of constraint relationship edges connected by the indicator nodes in the constraint relationship link. The constraint relationship strength is sorted to construct an indicator optimization sequence.

[0085] The indicator node with the largest constraint relationship strength is selected as the to-be-optimized indicator from the indicator optimization sequence. The influence source node is identified by tracing in reverse along the constraint relationship link starting from the to-be-optimized indicator, and an influence factor tree is constructed.

[0086] The influence propagation coefficient is calculated by counting the number of constraint relationship edges and synergistic enhancement relationship edges of the end nodes of the influence factor tree in the coupling relationship graph. The execution cost coefficient is calculated by counting the execution conditions of the end nodes. The control cost coefficient is calculated by multiplying the influence propagation coefficient and the execution cost coefficient. The end nodes with a control cost coefficient less than a preset cost threshold are selected to form a set of controllable factors.

[0087] Firstly, the historical data of multi-dimensional evaluation indicators is collected. The change data of energy efficiency evaluation value, light quality evaluation value and environmental impact evaluation value in the lighting system within nearly six months is collected, and the change of the three indicators in different time periods is recorded every day. Based on the historical data, the response time delay and change synchronization between indicators are calculated. Specifically, for any two indicators A and B, when indicator A changes, the time required for indicator B to start responding to the change is calculated to obtain the response time delay value. For example, when the energy efficiency evaluation value increases from 0.034 to 0.028, the light quality evaluation value decreases from 92.5% to 90.2% in 25 minutes, and the response time delay of the energy efficiency evaluation value to the light quality evaluation value is 25 minutes. The response time delay between all indicator pairs is normalized to form a 3x3 constraint relationship matrix. At the same time, the synchronization degree of the change of any two indicators is calculated, that is, the proportion of the same change of indicator B during the change of indicator A, to obtain the change synchronization value. For example, among the 80 records of the increase of the energy efficiency evaluation value, the environmental impact evaluation value also increases 68 times, and the change synchronization of the two is 0.85. The change synchronization of all indicator pairs is calculated to form a 3x3 synergistic enhancement relationship matrix.

[0088] The multi-dimensional evaluation indicators are constructed as indicator nodes, including energy efficiency evaluation value node, light quality evaluation value node and environmental impact evaluation value node. The response time delay value is extracted from the constraint relationship matrix, and when the response time delay value is less than the preset delay threshold, a constraint relationship edge is established between the corresponding two indicator nodes. In this embodiment, the preset delay threshold is 30 minutes, so a constraint relationship edge is established between the energy efficiency evaluation value node and the light quality evaluation value node, and the weight value of the edge is 1-normalized response time delay value. The change synchronization value is extracted from the synergistic enhancement relationship matrix, and when the change synchronization value is greater than the preset synchronization threshold, a synergistic enhancement relationship edge is established between the corresponding two indicator nodes. In this embodiment, the preset synchronization threshold is 0.7, so a synergistic enhancement relationship edge is established between the energy efficiency evaluation value node and the environmental impact evaluation value node, and the weight of the edge is the change synchronization value. Finally, the constraint relationship edge and the synergistic enhancement relationship edge are superimposed between the indicator nodes to form a complete coupling relationship graph.

[0089] The constraint relationship edge is extracted from the coupling relationship graph to construct a constraint relationship link. The number of constraint relationship edges connected to each node is marked by traversing the indicator nodes to obtain the constraint relationship strength. In this embodiment, the energy efficiency evaluation value node is connected to 2 constraint relationship edges, the light quality evaluation value node is connected to 3 constraint relationship edges, and the environmental impact evaluation value node is connected to 1 constraint relationship edge. The constraint relationship strength is sorted from large to small to construct an indicator optimization sequence, that is, light quality evaluation value, energy efficiency evaluation value, environmental impact evaluation value.

[0090] The index node with the strongest constraint relationship in the index optimization sequence is selected as the to-be-optimized index, that is, the light quality evaluation value node. Starting from the to-be-optimized index, the source nodes are identified by reverse tracing along the constraint relationship link. In this embodiment, it is found through reverse tracing that the factors affecting the light quality evaluation value include lighting power, light source color temperature, glare control, and lighting layout, which constitute the first layer nodes of the influence factor tree. Continuing to reverse trace, it is found that the lighting power is affected by the lamp type and the dimming system, the light source color temperature is affected by the lamp type and the color temperature adjustment system, the glare control is affected by the reflector design and the installation angle, and the lighting layout is affected by the number of lamps and the installation position, which constitute the end nodes of the influence factor tree.

[0091] The influence propagation coefficient is obtained by counting the number of constraint relationship edges and synergistic enhancement relationship edges of the end nodes of the influence factor tree in the coupling relationship graph. In this embodiment, the lamp type is connected to 3 constraint relationship edges and 2 synergistic enhancement relationship edges, and the influence propagation coefficient is 5; the dimming system is connected to 2 constraint relationship edges and 1 synergistic enhancement relationship edge, and the influence propagation coefficient is 3. The execution cost coefficient is obtained by counting the conditions for the end nodes to perform adjustment, considering the time, manpower, and material resources required for adjustment. For example, replacing the lamp type requires disassembling the old lamp, purchasing a new lamp, and installing and debugging, and the execution cost coefficient is 0.9; adjusting the dimming system only requires software parameter modification, and the execution cost coefficient is 0.2. The regulation and control cost coefficient is obtained by multiplying the influence propagation coefficient and the execution cost coefficient, and the regulation and control cost coefficient of the lamp type is 5*0.9=4.5, and the regulation and control cost coefficient of the dimming system is 3*0.2=0.6. The regulatable factor set is constituted by screening the end nodes with regulation and control cost coefficients less than the preset cost threshold, and the preset cost threshold is 3.0, so the nodes with regulation and control cost coefficients less than 3.0, such as the dimming system, the color temperature adjustment system, and the reflector design, are included in the regulatable factor set.

[0092] The present embodiment systematically analyzes the complex correlations between various evaluation indexes in the lighting system by constructing the coupling relationship graph, and realizes the scientific determination of the index optimization sequence and the accurate tracing of the influence factors through the quantitative processing of the constraint relationship and the synergistic enhancement relationship. The regulatable factors are screened by combining the influence propagation coefficient and the execution cost coefficient, which considers the influence range of adjustment on the whole system and also takes into account the feasibility of actual execution, and the generated adjustment scheme has high efficiency and operability.

[0093] As shown in FIG. 1, Figure 2 The index optimization sequence generation flowchart of the present embodiment is shown.

[0094] In an alternative embodiment, the constraint relationship edges are extracted from the coupling relationship graph to construct the constraint relationship link, the number of constraint relationship edges connected by the index nodes in the constraint relationship link is counted to obtain the constraint relationship strength, and the constraint relationship strength is sorted to construct the index optimization sequence, including:

[0095] The constraint relationship edge weight values between the index node pairs in the coupling relationship graph are obtained by traversing the index node pairs in the coupling relationship graph, and an adjacency matrix of the index nodes is constructed.

[0096] An index node group is formed by identifying a weight value dense distribution area in the adjacency matrix, and a difference degree value between the index node groups is obtained by calculating the variance and mean of the weight values in the index node group.

[0097] The constraint relationship edges are grouped and aggregated based on the difference degree value between the index node groups to form a constraint relationship edge group, a transmission path length value and a branch number value of the constraint relationship edges in the constraint relationship edge group are calculated, and a link complexity value of the constraint relationship link is constructed.

[0098] The distribution probability value of the link complexity value at different levels of the constraint relationship link is calculated, the distribution probability value is used to construct a cumulative density function, an inflection point position value is extracted from the cumulative density function to divide a segmented interval of the constraint relationship strength, and the constraint relationship links are sorted according to the constraint relationship strength values in the segmented interval from high to low to generate an index optimization sequence.

[0099] First, the constraint relationship edge weight values between the index node pairs in the coupling relationship graph are obtained by traversing the index node pairs in the coupling relationship graph, and an adjacency matrix of the index nodes is constructed. In the green lighting scheme evaluation system, the adjacency matrix includes multiple evaluation dimensions such as energy efficiency indicators, light quality indicators, environmental impact indicators, user satisfaction indicators, and maintenance cost indicators. In the adjacency matrix, the value of each cell represents the constraint degree of the row index pair to the column index. The weight value of the constraint relationship edge is calculated according to the correlation and time sequence characteristics of the index changes in the historical data. In actual calculation, when the energy efficiency index improves, the probability of the light quality index decreasing is 0.75, and the delay time is 20 minutes, so the constraint relationship weight value of the energy efficiency index to the light quality index is 0.75. In this way, a complete 5x5 adjacency matrix is constructed, and the non-zero elements in the matrix represent the index pairs with constraint relationships.

[0100] The weight value dense distribution area in the adjacency matrix is identified to form an index node group. By scanning the adjacency matrix, when the weight value difference of adjacent cells is less than a preset threshold 0.1, the index nodes corresponding to these cells are classified into the same node group. The weight value between the energy efficiency index and the maintenance cost index is 0.68, the weight value between the energy efficiency index and the environmental impact index is 0.72, and the weight value between the maintenance cost index and the environmental impact index is 0.65. The weight value difference of the three indexes is less than 0.1, so the three indexes are classified into one node group. Similarly, the light quality index and the user satisfaction index form another node group. The variance and mean of the weight value in each node group are calculated. The weight value variance of node group 1 is 0.0012, and the mean is 0.68. The weight value variance of node group 2 is 0.0008, and the mean is 0.82. The difference value between the two node groups is the difference between the means divided by the sum of the variances, and the difference value is 12.7.

[0101] Based on the difference value between the node groups, the constraint relationship edges are grouped and aggregated to form a constraint relationship edge group. When the difference value is greater than a preset difference threshold 10, it is considered that the constraint relationship between the two node groups is significantly different, and should be divided into different constraint relationship edge groups. The difference value between node group 1 and node group 2 is 12.7, which is greater than the preset threshold, so the constraint relationship between them is divided into an independent constraint relationship edge group. For each constraint relationship edge group, the transmission path length value and the branch number value of the constraint relationship edge in it are calculated. The transmission path length refers to the number of constraint relationship edges passed from the starting index node to the terminal index node; the branch number refers to the number of other index nodes connected to an index node. In node group 1, the transmission path length from the energy efficiency index to the environmental impact index is 2, passing through the maintenance cost index; the branch number of the energy efficiency index is 2, connecting the maintenance cost index and the environmental impact index respectively. The weighted sum of the transmission path length value and the branch number value is taken as the link complexity value of the constraint relationship link, and the weights are 0.6 and 0.4 respectively. In this example, the link complexity value of the constraint relationship link is 0.6*2+0.4*2=2.0.

[0102] The link complexity value restricts the distribution probability value of the link at different levels of the constraint relationship. All constraint relationship links are sorted according to the link complexity value from low to high, and are divided into multiple levels, each level containing links with similar complexity values. The link complexity value ranges from 1.0 to 4.5, which is evenly divided into 7 levels, and each level has a width of 0.5. The number of links in each level is counted and divided by the total number of links to obtain the distribution probability value of the level. In the example, the proportion of links with complexity values between 1.5 and 2.0 is 0.35, i.e. the distribution probability value is 0.35. Based on the distribution probability value, a cumulative density function is constructed, which represents the probability that the link complexity value is less than or equal to a certain value. The calculation method of the cumulative density function is to sequentially accumulate the distribution probability values of each level. At a complexity value of 2.0, the cumulative probability is 0.48; at a complexity value of 2.5, the cumulative probability is 0.76. The inflection point position value is extracted from the cumulative density function, i.e. the position where the slope of the curve changes significantly, as the boundary of the segmentation interval for dividing the constraint relationship strength. In the lighting system example, 2.5 is a significant inflection point, the cumulative probability rises rapidly from 0.48 to 0.76, and the slope changes by more than 50%, so 2.5 is used as a segmentation interval boundary. Finally, the constraint relationship links are sorted according to the constraint relationship strength value within the segmentation interval from high to low to generate the index optimization sequence. In this example, the constraint relationship strength of the light quality index is 3.8, the energy efficiency index is 3.2, the user satisfaction index is 2.7, the environmental impact index is 2.1, and the maintenance cost index is 1.6, so the index optimization sequence is: light quality index, energy efficiency index, user satisfaction index, environmental impact index, and maintenance cost index.

[0103] In this embodiment, by constructing the constraint relationship link and the constraint relationship strength sorting mechanism, the complex relationship between the multi-dimensional evaluation indexes in the green lighting system is accurately quantified and systematically analyzed. This method overcomes the limitations of traditional index optimization, which relies only on experience or simple weighting. Through a series of data mining methods such as adjacency matrix construction, node group identification, link complexity calculation, and cumulative density function analysis, the implicit constraint rules and influence paths between indexes are revealed. This index optimization sequence generation method based on the constraint relationship graph not only improves the accuracy and systematicness of lighting scheme evaluation, but also provides a data-driven optimization path for intelligent decision-making, effectively solving the problem of determining the priority of indexes in multi-objective optimization.

[0104] In an optional implementation, a public controllable factor that can simultaneously affect multiple to-be-optimized indexes is identified from the set of controllable factors, a synergistic enhancement relationship in the coupling relationship graph is triggered by adjusting the public controllable factor, and a preliminary adjustment scheme is generated, including:

[0105] An index coverage value is calculated by counting the number of to-be-optimized indexes connected to each regulatable factor in the coupling relationship graph, a synergy value is calculated by counting the number of synergistic enhancement relationship edges connected to each regulatable factor, and a comprehensive evaluation value is obtained by multiplying the index coverage value and the synergy value;

[0106] A regulatable factor whose comprehensive evaluation value exceeds a preset evaluation threshold is selected as a public regulatable factor, and a regulation weight value is obtained by calculating the ratio of the comprehensive evaluation value of each public regulatable factor to the sum of the comprehensive evaluation values of all public regulatable factors;

[0107] A regulation gain value is obtained by multiplying the regulation weight value of each public regulatable factor and the propagation coefficient of the corresponding synergistic enhancement relationship edge, and a leading regulation factor is selected by screening the public regulatable factor with the largest regulation gain value;

[0108] A benchmark regulation amount is generated based on the regulation weight value of the leading regulation factor, a linkage regulation amount is obtained by distributing the benchmark regulation amount to the remaining public regulatable factors according to the proportion of the regulation weight values, and the parameters of the public regulatable factors are adjusted according to the linkage regulation amount to generate a preliminary adjustment scheme.

[0109] In one embodiment, a complete set of regulatable factors and a set of to-be-optimized indexes are first obtained. The set of regulatable factors includes all adjustable parameters, including lighting power density, lamp color temperature, color rendering index, illuminance uniformity, lamp layout, light source type, and the like. By querying the connection relationship between each factor and the to-be-optimized indexes in the coupling relationship graph, the number of connections is counted to obtain the index coverage value. Taking the lighting power density factor as an example, the factor is directly connected to three to-be-optimized indexes, i.e., the energy efficiency evaluation index, the environmental impact evaluation index, and the maintenance cost evaluation index, and the index coverage value is 3. The lamp color temperature factor is connected to two to-be-optimized indexes, i.e., the light quality evaluation index and the user satisfaction evaluation index, and the index coverage value is 2. Meanwhile, the number of synergistic enhancement relationship edges connected to each regulatable factor is counted to obtain the synergy value. The lighting power density factor forms a synergistic enhancement relationship between the energy efficiency evaluation index and the environmental impact evaluation index, and also forms a synergistic enhancement relationship between the energy efficiency evaluation index and the maintenance cost evaluation index, so the synergy value is 2.

[0110] The lamp color temperature factor forms a synergistic enhancement relationship between the light quality evaluation index and the user satisfaction evaluation index, and the synergy value is 1. The comprehensive evaluation value is obtained by multiplying the index coverage value and the synergy value, the comprehensive evaluation value of the lighting power density factor is 3x2=6, and the comprehensive evaluation value of the lamp color temperature factor is 2x1=2.

[0111] The controllable factors whose comprehensive evaluation values exceed the preset evaluation threshold are selected as public controllable factors. The preset evaluation threshold is set to 3, so the lighting power density factor is selected as a public controllable factor, while the luminaire color temperature factor is not selected. It is found that the comprehensive evaluation value of the illuminance uniformity factor is 4, the comprehensive evaluation value of the luminaire layout factor is 5, and the comprehensive evaluation value of the light source type factor is 7, all of which exceed the preset evaluation threshold, and are selected as public controllable factors. The adjustment weight value is calculated by taking the ratio of the comprehensive evaluation value of each public controllable factor to the sum of the comprehensive evaluation values of all public controllable factors. The sum of the comprehensive evaluation values of all public controllable factors is 6+4+5+7=22, the adjustment weight value of the lighting power density factor is 6÷22=0.27, the adjustment weight value of the illuminance uniformity factor is 4÷22=0.18, the adjustment weight value of the luminaire layout factor is 5÷22=0.23, and the adjustment weight value of the light source type factor is 7÷22=0.32.

[0112] The adjustment gain value of each public controllable factor is obtained by multiplying the adjustment weight value of the public controllable factor by the propagation coefficient of the corresponding synergistic enhancement relationship edge. The propagation coefficient of the synergistic enhancement relationship edge represents the strength and influence range of the relationship, which is calculated by analyzing historical data. The propagation coefficient of the synergistic enhancement relationship edge corresponding to the lighting power density factor is 1.2, the propagation coefficient of the synergistic enhancement relationship edge corresponding to the illuminance uniformity factor is 0.9, the propagation coefficient of the synergistic enhancement relationship edge corresponding to the luminaire layout factor is 1.1, and the propagation coefficient of the synergistic enhancement relationship edge corresponding to the light source type factor is 1.3. The adjustment gain values of the factors are calculated, the adjustment gain value of the lighting power density factor is 0.27×1.2=0.324, the adjustment gain value of the illuminance uniformity factor is 0.18×0.9=0.162, the adjustment gain value of the luminaire layout factor is 0.23×1.1=0.253, and the adjustment gain value of the light source type factor is 0.32×1.3=0.416. The public controllable factor with the largest adjustment gain value is selected as the dominant adjustment factor. In this embodiment, the adjustment gain value of the light source type factor is the largest, which is 0.416, so it is selected as the dominant adjustment factor.

[0113] The reference adjustment amount is generated based on the adjustment weight value of the dominant adjustment factor. In the lighting system, the adjustment range of the light source type factor is a discrete level value between 0 and 100, the current value is 65, and it needs to be increased to a higher level according to the optimization target. The calculation of the reference adjustment amount takes into account the adjustment weight value, the gap between the current value and the maximum value, and the preset adjustment step. The preset adjustment step is 10, and the reference adjustment amount of the light source type factor is (100-65)×0.32×10÷100=1.12, which is rounded up to 2 level units. The reference adjustment amount is distributed to the remaining public controllable factors in proportion to the adjustment weight value to obtain the linkage adjustment amount.

[0114] The proportion of the adjustment weight values of each factor is lighting power density: illumination uniformity: luminaire layout = 0.27: 0.18: 0.23, and the linkage adjustment amount is calculated considering the adjustment range and current value of each factor. The adjustment range of the lighting power density factor is 5 to 20 watts per square meter, the current value is 12 watts per square meter, and according to the proportion, it should be reduced by (12-5) x 0.27 ÷ 0.32 x 2 ÷ (100-65) = 0.84 watts per square meter, and rounded down to 1 watt per square meter.

[0115] The adjustment range of the illumination uniformity factor is 0.4 to 0.9, and the current value is 0.65. According to the proportion, it should be increased by (0.9-0.65) x 0.18 ÷ 0.32 x 2 ÷ (100-65) = 0.06, and rounded up to 0.1.

[0116] The adjustment range of the luminaire layout factor is 1 to 10 layout levels, and the current value is 6. According to the proportion, it should be increased by (10-6) x 0.23 ÷ 0.32 x 2 ÷ (100-65) = 0.16, and rounded up to 1 level unit. According to the linkage adjustment amount, the parameters of the public controllable factors are adjusted to generate a preliminary adjustment scheme: the light source type is improved by 2 level units to 67, the lighting power density is reduced by 1 watt per square meter to 11 watts per square meter, the illumination uniformity is improved by 0.1 to 0.75, and the luminaire layout is improved by 1 level unit to 7 level layout.

[0117] In this embodiment, by introducing the public controllable factor identification and collaborative adjustment mechanism, intelligent decision-making under the multi-dimensional optimization goal of the green lighting system is realized. The influence breadth (index coverage value) and influence depth (collaboration value) of the controllable factors are combined to establish a comprehensive evaluation system, breaking through the limitations of single factor independent adjustment in traditional methods. Through the linkage adjustment mechanism of the leading adjustment factor driving other factors, the synergistic enhancement relationship in the coupling relationship graph is fully utilized to realize the multiplier effect of multi-factor collaborative adjustment. This adjustment amount calculation method based on weight distribution not only ensures the balance and system stability of the adjustment, but also maximizes the utilization efficiency of limited adjustment resources.

[0118] In an alternative embodiment, the preliminary adjustment scheme is substituted into the coupling relationship graph for propagation calculation to obtain a predicted change amount. When the predicted change amount shows that there is an index degradation, the public controllable factor is re-identified after cutting off the degradation propagation path in the coupling relationship graph, and a final adjustment scheme is generated, including:

[0119] The adjustment parameters of the public controllable factors in the preliminary adjustment scheme are substituted into the corresponding positions in the coupling relationship graph to construct an initial state table. Based on the propagation rules of the synergistic enhancement relationship edges, multiple rounds of propagation calculation are performed, and the numerical changes and propagation path information of the to-be-optimized indexes after each round of calculation are recorded to generate a predicted change amount table.

[0120] Screening the to-be-optimized indicators with negative numerical changes from the predicted change scale as deterioration indicators, extracting the propagation paths corresponding to the deterioration indicators to construct a propagation path table;

[0121] Calculating the combined value of the propagation coefficient product and the path length of each path in the propagation path table, and screening the propagation path with the maximum combined value as the to-be-cut path;

[0122] Removing the synergistic enhancement relationship edges on the to-be-cut path in the coupling relationship graph to form an optimized graph, counting the number of connection edges of each public controllable factor in the optimized graph to obtain a node connection value, calculating the sum of the propagation coefficients of the connection edges to obtain a propagation impact value, and multiplying the node connection value and the propagation impact value to obtain a control evaluation value;

[0123] Selecting the public controllable factor with the maximum control evaluation value as the master control factor, calculating the adjustment reference amount based on the control evaluation value of the master control factor, distributing the adjustment reference amount to the public controllable factors according to the control evaluation value proportion through the synergistic enhancement relationship edges in the optimized graph, and generating a final adjustment scheme.

[0124] In this embodiment, the adjustment parameters of the public controllable factors in the preliminary adjustment scheme are substituted into the corresponding positions of the coupling relationship graph to construct an initial state table. The preliminary adjustment scheme includes improving the light source type from level 65 to 67, reducing the lighting power density from 12 watts per square meter to 11 watts per square meter, improving the illuminance uniformity from 0.65 to 0.75, and improving the luminaire layout from level 6 to level 7. These adjustment parameters are filled into the corresponding node positions of the coupling relationship graph to construct an initial state table, which includes the parameter values before adjustment, the parameter values after adjustment, and the change amount. Based on the propagation rules of the synergistic enhancement relationship edges, multiple rounds of propagation operations are performed. The propagation rules of the synergistic enhancement relationship edges define how the parameter changes propagate along the edges and affect the connected indicators. The propagation rules include a propagation coefficient and a propagation direction. The propagation coefficient determines the amplification or attenuation degree of the change amount, and the propagation direction determines the positivity or negativity of the change. The propagation coefficient of the synergistic enhancement relationship edge between the light source type and the light quality evaluation indicator is 0.8, and the direction is positive. The propagation coefficient of the synergistic enhancement relationship edge between the lighting power density and the energy efficiency evaluation indicator is 1.2, and the direction is negative. In the multiple rounds of propagation operations, the first round calculates the change amount of the indicators to be optimized that are directly connected to the public controllable factors. The light source type is improved by 2 level units, which is converted to an improvement of 1.6 percentage points in the light quality evaluation indicator through the propagation coefficient of 0.8. The lighting power density is reduced by 1 watt per square meter, which is converted to an improvement of 1.2 percentage points in the energy efficiency evaluation indicator through the propagation coefficient of 1.2. The second round calculates the change amount of the indicators to be optimized that affect other indicators through the synergistic enhancement relationship edges. The light quality evaluation indicator is improved by 1.6 percentage points, which is converted to an improvement of 1.44 percentage points in the user satisfaction indicator through the propagation coefficient of 0.9 between the light quality evaluation indicator and the user satisfaction indicator. The numerical changes and propagation path information of the indicators to be optimized after each round of operation are recorded to form a predicted change amount table. The predicted change amount table includes the indicator name, the change amount, and the influence path. For example, the change amount of the light quality evaluation indicator is +1.6, and the influence path is "light source type → light quality evaluation indicator". The change amount of the user satisfaction indicator is +1.44, and the influence path is "light source type → light quality evaluation indicator → user satisfaction indicator".

[0125] From the predicted change amount table, the indicators to be optimized with negative numerical changes are selected as the deteriorating indicators. After multiple rounds of propagation operations, it is found that the change amount of the maintenance cost evaluation indicator is -2.1 percentage points, and the change amount of the environmental impact evaluation indicator is -0.8 percentage points, both of which are negative, and are identified as deteriorating indicators. The propagation path table is constructed by extracting the propagation paths corresponding to the deteriorating indicators. By tracing the influence path information in the predicted change amount table, it is found that the path causing the deterioration of the maintenance cost evaluation indicator is "light source type → light quality evaluation indicator → energy efficiency evaluation indicator → maintenance cost evaluation indicator", and the path causing the deterioration of the environmental impact evaluation indicator is "illuminance uniformity → light quality evaluation indicator → user satisfaction indicator → environmental impact evaluation indicator". These paths are recorded in the propagation path table, including the path number, the path content, and the deteriorating indicator name.

[0126] Calculate the combined value of the product of propagation coefficients and the path length for each path in the propagation path table. The product of propagation coefficients is the product of the propagation coefficients of all cooperative reinforcement edges on the path, and the path length is the number of cooperative reinforcement edges contained in the path. The combined value is calculated by dividing the product of propagation coefficients by the square of the path length. This design aims to prioritize cutting short paths while ensuring propagation efficiency. The product of propagation coefficients for the deteriorated path of the maintenance cost assessment index is 0.8 × 0.7 × 0.9 = 0.504, the path length is 3, and the combined value is 0.504 ÷ 9 = 0.056; the product of propagation coefficients for the deteriorated path of the environmental impact assessment index is 0.75 × 0.9 × 0.8 = 0.54, the path length is 3, and the combined value is 0.54 ÷ 9 = 0.06. The propagation path with the largest combined value is selected as the path to be cut. In this embodiment, the combined value of the deteriorated path of the environmental impact assessment index is 0.06, which is greater than the 0.056 of the deteriorated path of the maintenance cost assessment index. Therefore, the deteriorated path of the environmental impact assessment index is selected as the path to be cut.

[0127] In the coupling relationship graph, synergistic enhancement edges on the paths to be cut are removed to form an optimized graph. Specifically, key links are identified in the path "illuminance uniformity → light quality assessment index → ​​user satisfaction index → ​​environmental impact assessment index," and synergistic enhancement edges between the user satisfaction index and the environmental impact assessment index are removed, resulting in an optimized coupling relationship graph. The number of connecting edges for each common adjustable factor in the optimized graph is counted to obtain the node connection value. Light source type is connected by 3 synergistic enhancement edges, with a node connection value of 3; lighting power density is connected by 2 synergistic enhancement edges, with a node connection value of 2; illuminance uniformity is connected by 2 synergistic enhancement edges, with a node connection value of 2; and luminaire layout is connected by 3 synergistic enhancement edges, with a node connection value of 3. The sum of the propagation coefficients of the connecting edges is calculated to obtain the propagation impact value. The propagation coefficients of the synergistic enhancement edges connected by light source type are 0.8, 0.7, and 0.9, respectively, with a propagation impact value of 2.4; the propagation impact value of lighting power density is 1.9; the propagation impact value of illuminance uniformity is 1.6; and the propagation impact value of luminaire layout is 2.2. The control evaluation value is obtained by multiplying the node connection value by the propagation effect value. The control evaluation value for light source type is 3×2.4=7.2; the control evaluation value for lighting power density is 2×1.9=3.8; the control evaluation value for illuminance uniformity is 2×1.6=3.2; and the control evaluation value for luminaire layout is 3×2.2=6.6.

[0128] The public controllable factor with the largest regulation evaluation value is selected as the main control factor. In this embodiment, the regulation evaluation value of the light source type is 7.2, which is the largest, so it is selected as the main control factor. The adjustment reference quantity is calculated based on the regulation evaluation value of the main control factor. The adjustment reference quantity calculation considers the relationship between the regulation evaluation value and the adjustment range. In this embodiment, the adjustment range of the light source type is 0 to 100, the current value is 65, the regulation evaluation value is 7.2, and the adjustment reference quantity is (100-65) x 7.2 ÷ 10 = 2.52, rounded up to 3 level units. The adjustment reference quantity is distributed to the public controllable factors in proportion to the regulation evaluation value through the synergistic enhancement relationship edge in the optimization graph. The regulation evaluation value proportion of each public controllable factor is light source type: lighting power density: illumination uniformity: luminaire layout = 7.2:3.8:3.2:6.6.

[0129] The adjustment amount of the lighting power density is 3 x 3.8 ÷ 7.2 = 1.58, rounded up to 2 watts per square meter, from 12 watts per square meter to 10 watts per square meter; the adjustment amount of the illumination uniformity is 3 x 3.2 ÷ 7.2 = 1.33, rounded up to 0.15, from 0.65 to 0.8; the adjustment amount of the luminaire layout is 3 x 6.6 ÷ 7.2 = 2.75, rounded up to 3 level units, from level 6 to level 9. The final adjustment scheme is generated: the light source type is raised from level 65 to level 68, the lighting power density is reduced from 12 watts per square meter to 10 watts per square meter, the illumination uniformity is improved from 0.65 to 0.8, and the luminaire layout is raised from level 6 to level 9.

[0130] In this embodiment, by introducing the propagation algorithm and path cutting mechanism, intelligent optimization and adjustment of green lighting scheme are realized. This method innovatively performs global propagation analysis on the impact of the preliminary adjustment scheme, accurately predicts the trend of changes in each indicator, and can actively identify and eliminate potential risk of indicator degradation. Through the combination evaluation model of the product of the propagation coefficient and the path length, the scientific and quantitative ordering of the degradation propagation path is realized, ensuring the accuracy and effectiveness of the cutting intervention. The construction of the optimization graph and the calculation method of the regulation evaluation value effectively balance the breadth and depth of indicator improvement, and realize the optimal configuration of multi-factor synergistic regulation.

[0131] In an alternative embodiment, based on the final adjustment scheme, a regulation instruction for a plurality of sub-regions is generated, and the regulation instruction is sent to the lighting device of the corresponding sub-region, comprising:

[0132] The adjustment parameters of the public controllable factors in the final adjustment scheme and the historical adjustment data of the sub-region lighting device are obtained, the trend of changes in the lighting parameters in the historical adjustment data is calculated, and a frequency domain response matrix between the lighting parameters is constructed;

[0133] The frequency components of the illumination parameter changes are analyzed based on the frequency domain response matrix, the resonant frequency and the resonance frequency in the illumination parameter are extracted, and the frequency adjustment instructions for the multiple sub-regions are generated;

[0134] The output spectrum data of the multiple sub-region lighting devices when the frequency adjustment instructions are executed are collected, the energy transfer coefficients between adjacent wavelength segments are calculated, the energy transfer coefficients are coupled with the frequency adjustment instructions to obtain the spectrum compensation amount, the execution period of the frequency adjustment instructions is adjusted based on the spectrum compensation amount, and the energy balance instructions for the multiple sub-regions are generated;

[0135] The driving circuit temperature and the light emitting element temperature of the multiple sub-region lighting devices when the energy balance instructions are executed are monitored, and the gradient field distribution of the temperature changes is constructed;

[0136] The heat accumulation area and the heat diffusion rate in the gradient field are calculated, the energy balance instructions are modified, the regulation and control instructions for the multiple sub-regions are generated, and the regulation and control instructions are sent to the lighting devices of the corresponding sub-regions.

[0137] In the embodiment, the adjustment parameters of the common controllable factors in the final adjustment scheme and the historical adjustment data of the sub-region lighting devices are obtained. The final adjustment scheme includes that the light source type is improved from level 65 to 68, the lighting power density is reduced from 12 watts per square meter to 10 watts per square meter, the illuminance uniformity is improved from 0.65 to 0.8, and the lamp layout is improved from 6 levels to 9 levels. At the same time, the historical adjustment data of the sub-region lighting devices in the past 30 days is extracted from the lighting control system database, including the lighting parameter adjustment records of the three sub-regions of the work area, the meeting area and the rest area. The historical adjustment data records the time point, the change amount and the influence on the light environment of the change of the lighting parameters. The change trend of the lighting parameters in the historical adjustment data is calculated, and the specific method is to arrange each lighting parameter in time sequence, and calculate the change rate through sliding window. In the work area sub-region, the average change rate of the lighting power density is 0.2 watts per square meter per day, and the average change rate of the illuminance uniformity is 0.015 per day. The frequency domain response matrix between the lighting parameters is constructed, the lighting parameter changes in the time domain are converted to the frequency domain by using discrete Fourier transform, and the frequency characteristics of different lighting parameter changes are analyzed. The row of the frequency domain response matrix represents different lighting parameters, the column represents different frequency components, and the matrix element value represents the amplitude of the parameter change at a specific frequency. In the work area sub-region, the response amplitude of the lighting power density at 0.5 hertz frequency is 2.3, and the response amplitude of the light source color temperature at 0.8 hertz frequency is 1.8.

[0138] The frequency components of the lighting parameter changes are analyzed based on the frequency domain response matrix, and the resonance frequency and the sympathetic frequency in the lighting parameter are extracted. The resonance frequency refers to the main frequency component of the lighting parameter itself, and the sympathetic frequency refers to the frequency point of the enhanced effect generated by the similar frequency of multiple lighting parameters. The resonance frequency is identified by scanning the amplitude peak in the frequency domain response matrix, and the sympathetic frequency is identified by comparing the amplitude changes of different parameters at similar frequencies. The resonance frequency of the lighting power density is 0.5 Hz, the resonance frequency of the color rendering index is 0.8 Hz, and the resonance frequency of the two appears at 0.65 Hz. The resonance frequencies of the lighting power density and the color rendering index of the conference area appear at 0.6 Hz and 0.9 Hz respectively, and the resonance frequency appears at 0.75 Hz. The resonance frequencies of the lighting power density and the color rendering index of the rest area appear at 0.4 Hz and 0.7 Hz respectively, and the resonance frequency appears at 0.55 Hz.

[0139] The frequency adjustment instructions for the multiple sub-areas are generated, and the instruction content includes adjustment frequency, adjustment amplitude and duration. The frequency adjustment instruction of the working area is that the lighting power density is reduced by 2 W / m2 at a frequency of 0.65 Hz, and the duration is 10 minutes; the frequency adjustment instruction of the conference area is that the lighting power density is reduced by 1.5 W / m2 at a frequency of 0.75 Hz, and the duration is 8 minutes; the frequency adjustment instruction of the rest area is that the lighting power density is reduced by 1 W / m2 at a frequency of 0.55 Hz, and the duration is 6 minutes.

[0140] The output spectrum data of the multiple sub-region lighting devices when executing the frequency adjustment instruction is collected, and the spectral energy distribution of the lighting device in the wavelength range of 380-780 nm is monitored in real time using a spectral analyzer. The energy transfer coefficient between adjacent wavelength segments is calculated, and the wavelength range is evenly divided into 10 wavelength segments. For any two adjacent wavelength segments, the proportional relationship between the energy change of one wavelength segment and the energy change of the other wavelength segment is calculated as the energy transfer coefficient. In the working area, the energy transfer coefficient between the 450-480 nm wavelength segment and the 480-510 nm wavelength segment is 0.85, indicating that when the energy of the former wavelength segment decreases, 85% of the energy is transferred to the latter wavelength segment. The spectral compensation amount is obtained by coupling the energy transfer coefficient with the frequency adjustment instruction. The specific method is to multiply the frequency of the frequency adjustment instruction by the energy transfer coefficient, and then multiply it by the energy proportion of the wavelength segment. The spectral compensation amount of the 450-480 nm wavelength segment in the working area is 0.65×0.85×0.12=0.066, indicating that 6.6% of the energy needs to be compensated. The execution period of the frequency adjustment instruction is adjusted based on the spectral compensation amount. When the compensation amount is greater than the preset threshold 0.05, the execution period is extended by 1.5 times. The execution period of the working area is adjusted from 10 minutes to 15 minutes, and the execution period of the conference area is adjusted from 8 minutes to 12 minutes. Energy balancing instructions for multiple sub-regions are generated, and the instruction content includes adjustment frequency, adjustment amplitude, execution period and wavelength segment compensation parameter. The energy balancing instruction of the working area is: the lighting power density is reduced by 2 W / m2 at a frequency of 0.65 Hz, the execution period is 15 minutes, and the 450-480 nm wavelength segment compensation coefficient is 0.066.

[0141] The driving circuit temperature and the light emitting element temperature of the multiple sub-region lighting devices when executing the energy balancing instruction are monitored, and the temperature data is collected in real time using an embedded temperature sensor. A gradient field distribution of temperature change is constructed, and the temperature data of the lighting device is mapped to a two-dimensional plane to generate a continuous temperature gradient field through an interpolation algorithm. The driving circuit temperature of the working area lighting device is 45 degrees Celsius, the light emitting element temperature is 65 degrees Celsius, and the temperature gradient is 0.5 degrees Celsius / cm. The heat accumulation area in the gradient field and the heat diffusion rate are calculated. The heat accumulation area refers to the area with a large temperature gradient and a continuous temperature rise, and the heat diffusion rate represents the speed of heat transfer. The heat accumulation area of the working area lighting device is concentrated within 3 cm around the light emitting element, and the heat diffusion rate is 0.3 degrees Celsius / min.

[0142] When the heat diffusion rate is less than the preset safety threshold of 0.5 degrees Celsius per minute, the energy balance instruction does not need to be corrected; when the heat diffusion rate is greater than the preset safety threshold, the adjustment amplitude needs to be reduced or the execution period needs to be prolonged. The energy balance instruction is corrected to generate a regulation and control instruction for the multiple sub-regions. The regulation and control instruction of the working area is that the lighting power density is reduced by 2 watts per square meter at a frequency of 0.65 hertz, the execution period is 15 minutes, the compensation coefficient of the wavelength segment of 450-480 nanometers is 0.066, and the heat protection parameter is 0.3. The regulation and control instruction is sent to the lighting devices of the corresponding sub-regions through the lighting control network, and the lighting devices adjust the lighting power, color temperature and spectral distribution according to the instruction parameters after receiving the instruction, so as to realize fine execution of the green lighting scheme.

[0143] In the embodiment, by introducing the frequency domain analysis and the spectral energy management mechanism, high-precision execution and dynamic optimization of the green lighting regulation and control are realized. The time domain change of the lighting parameters is converted to the frequency domain for analysis, and the resonance characteristics and resonance points of the parameter change are accurately identified, breaking through the limitation of the traditional method which only considers static adjustment. Through the spectral energy migration analysis and compensation technology, the spectral imbalance problem that may be caused in the lighting parameter adjustment process is solved, and the visual comfort and color rendering performance are ensured. The introduction of the temperature gradient field monitoring and the heat protection mechanism effectively prevents the overheating risk of the devices that may be caused by high-intensity regulation and control, and prolongs the service life of the lighting system.

[0144] In a second aspect of the embodiment of the present application, an AI-assisted green lighting scheme evaluation and intelligent decision system is provided, and the system comprises:

[0145] A first unit is configured to identify the lighting demand differences of different spatial positions in the to-be-evaluated region, divide the to-be-evaluated region into multiple sub-regions, determine an evaluation weight for each sub-region, and generate a multi-dimensional evaluation index based on the evaluation weight;

[0146] A second unit is configured to construct a coupling relationship graph containing the constraint relationship and the synergistic enhancement relationship between the indexes based on the multi-dimensional evaluation index, determine an index optimization sequence according to the constraint relationship, perform influence source reverse tracing on the to-be-optimized indexes based on the index optimization sequence, and generate a set of controllable factors;

[0147] A third unit is configured to identify a public controllable factor capable of simultaneously affecting multiple to-be-optimized indexes from the set of controllable factors, trigger the synergistic enhancement relationship in the coupling relationship graph by adjusting the public controllable factor, and generate a preliminary adjustment scheme;

[0148] A fourth unit is configured to substitute the preliminary adjustment scheme into the coupling relationship graph to perform propagation calculation and obtain a predicted change amount, when the predicted change amount shows that there is index degradation, cut off the degradation propagation path in the coupling relationship graph, re-identify the public controllable factor, and generate a final adjustment scheme;

[0149] a fifth unit configured to generate control instructions for the plurality of sub-regions based on the final adjustment scheme, and send the control instructions to the lighting devices of the corresponding sub-regions.

[0150] In a third aspect, the present application provides an electronic device, comprising:

[0151] a processor;

[0152] a memory for storing processor-executable instructions;

[0153] The processor is configured to invoke the instructions stored in the memory to perform the method described above.

[0154] In a fourth aspect, the present application provides a computer-readable storage medium having stored thereon computer program instructions, which when executed by a processor, implement the method described above.

[0155] The present application can be a method, apparatus, system and / or computer program product. The computer program product can include a computer-readable storage medium having stored thereon computer-readable program instructions that, when executed by a computer, cause the computer to carry out various aspects of the present application.

[0156] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. AI-based assisted green lighting scheme evaluation and intelligent decision making method, characterized in that, The method comprises the following steps: Identify the differences in lighting requirements at different spatial locations in the area to be evaluated, divide the area to be evaluated into multiple sub-areas, determine the evaluation weight for each sub-area respectively, and generate a multi-dimensional evaluation index based on the evaluation weight; wherein the lighting requirement parameters include the required illuminance value, color temperature requirement, color rendering index requirement and glare limit value; Based on the multi-dimensional evaluation index, a coupling relationship graph containing the constraint relationship and synergistic enhancement relationship between the indexes is constructed, and the index optimization sequence is determined according to the constraint relationship. The influence source of the to-be-optimized index is traced back in reverse based on the index optimization sequence, and a set of controllable factors is generated; From the set of controllable factors, identify the public controllable factors that can simultaneously affect multiple to-be-optimized indexes, trigger the synergistic enhancement relationship in the coupling relationship graph by adjusting the public controllable factors, and generate a preliminary adjustment scheme; Substitute the adjustment parameters of the public controllable factors in the preliminary adjustment scheme into the corresponding positions in the coupling relationship graph to construct an initial state table, perform multiple rounds of propagation operations based on the propagation rules of the synergistic enhancement relationship edges, record the numerical changes and propagation path information of the to-be-optimized indexes after each round of operation, and generate a predicted change table; From the predicted change table, select the to-be-optimized indexes with negative numerical changes as the degradation indexes, extract the propagation path corresponding to the degradation indexes to construct a propagation path table; Calculate the combination value of the propagation coefficient product and the path length of each path in the propagation path table, and select the propagation path with the maximum combination value as the to-be-cut path; Remove the synergistic enhancement relationship edges on the to-be-cut path in the coupling relationship graph to form an optimization graph, count the number of connection edges of each public controllable factor in the optimization graph to obtain a node connection value, calculate the sum of the propagation coefficients of the connection edges to obtain a propagation influence value, and multiply the node connection value by the propagation influence value to obtain a control evaluation value; Select the public controllable factor with the maximum control evaluation value as the master control factor, calculate the adjustment reference amount based on the control evaluation value of the master control factor, and distribute the adjustment reference amount to the public controllable factors in proportion to the control evaluation value through the synergistic enhancement relationship edges in the optimization graph to generate a final adjustment scheme; Based on the final adjustment scheme, generate control instructions for multiple sub-areas, and send the control instructions to the lighting devices in the corresponding sub-areas.

2. The method of claim 1, wherein, Identifying the differences in lighting requirements at different spatial locations in the area to be evaluated, dividing the area to be evaluated into multiple sub-areas, determining the evaluation weight for each sub-area respectively, and generating a multi-dimensional evaluation index based on the evaluation weight comprises: Obtaining lighting requirement parameters of multiple spatial position points in the area to be evaluated; Based on the lighting requirement parameters, a demand difference matrix between the spatial position points is constructed, and spatial position points with a deviation exceeding a preset deviation threshold are identified as boundary position points by traversing the demand difference matrix; The area to be evaluated is segmented into multiple sub-areas with the boundary position points as boundaries, and the lighting requirement parameters of the spatial position points in each sub-area are aggregated to obtain sub-area characteristic parameters; Statistical the deviation distribution of the lighting requirement parameters of the spatial position points in the sub-area relative to the sub-area characteristic parameters, calculate the demand fluctuation index based on the deviation distribution, and take the ratio of the demand fluctuation index to the preset reference fluctuation index as the fluctuation coefficient; The evaluation weight is obtained by performing a weighted combination operation on the fluctuation coefficient and the sub-region characteristic parameters, and the energy efficiency evaluation value, the light quality evaluation value and the environmental impact evaluation value are calculated based on the evaluation weight for each sub-region to form the multi-dimensional evaluation index.

3. The method of claim 1, wherein, A coupling relationship graph including a constraint relationship and a synergistic enhancement relationship between indexes is constructed based on the multi-dimensional evaluation index, and an index optimization sequence is determined according to the constraint relationship. The influence source is traced in reverse based on the index optimization sequence to generate a controllable factor set including: The change history data of the multi-dimensional evaluation index is collected, and the response time delay and the change synchronization between indexes are calculated based on the historical data to generate a constraint relationship matrix and a synergistic enhancement relationship matrix respectively; The multi-dimensional evaluation index is constructed as an index node, the response time delay value is extracted from the constraint relationship matrix to form a constraint relationship edge between the index nodes, and the change synchronization value is extracted from the synergistic enhancement relationship matrix to form a synergistic enhancement relationship edge between the index nodes. The constraint relationship edge and the synergistic enhancement relationship edge are superimposed between the index nodes to form the coupling relationship graph; The constraint relationship edge is extracted from the coupling relationship graph to construct a constraint relationship link, and the constraint relationship strength is obtained by counting the number of constraint relationship edges connected by the index nodes in the constraint relationship link. The constraint relationship strength is sorted to construct the index optimization sequence; The index node with the maximum constraint relationship strength is selected as the to-be-optimized index from the index optimization sequence, and the influence source node is identified in reverse along the constraint relationship link starting from the to-be-optimized index to construct an influence factor tree. The influence propagation coefficient is obtained by counting the number of constraint relationship edges and synergistic enhancement relationship edges of the end node of the influence factor tree in the coupling relationship graph, and the execution cost coefficient is obtained by counting the execution conditions of the end node. The influence propagation coefficient and the execution cost coefficient are multiplied to obtain the regulation and control cost coefficient. The end nodes with a regulation and control cost coefficient less than a preset cost threshold are selected to form the controllable factor set.

4. The method of claim 3, wherein, The constraint relationship edge is extracted from the coupling relationship graph to construct a constraint relationship link, and the constraint relationship strength is obtained by counting the number of constraint relationship edges connected by the index nodes in the constraint relationship link. The constraint relationship strength is sorted to construct the index optimization sequence including: An index node pair in the coupling relationship graph is traversed to obtain the constraint relationship edge weight value between the index node pair, and an adjacency matrix of the index node is constructed; A weight value dense distribution area is identified in the adjacency matrix to form an index node group, and a node group difference value is obtained by calculating the variance and mean of the weight value in the index node group; The constraint relationship edges are grouped and aggregated based on the node group difference value to form a constraint relationship edge group, and a link complexity value of the constraint relationship link is constructed by calculating the transmission path length value and the branch number value of the constraint relationship edges in the constraint relationship edge group; The distribution probability value of the link complexity value at different levels of the constraint relationship link is calculated, and a cumulative density function is constructed based on the distribution probability value. The index optimization sequence is generated by sorting the constraint relationship links in the constraint relationship link according to the constraint relationship strength value from high to low in the segmentation interval and extracting the inflection point position value from the cumulative density function to divide the segmentation interval.

5. The method of claim 1, wherein, Identify the common controllable factors that can simultaneously affect multiple to-be-optimized indicators from the set of controllable factors, trigger the synergistic enhancement relationship in the coupling relationship graph by adjusting the common controllable factors, and generate a preliminary adjustment scheme including: Calculate the number of to-be-optimized indicators connected in the coupling relationship graph for each controllable factor in the set of controllable factors to obtain the indicator coverage value, calculate the number of synergistic enhancement relationship edges connected for each controllable factor to obtain the synergy value, and multiply the indicator coverage value and the synergy value to obtain the comprehensive evaluation value; Filter the controllable factors whose comprehensive evaluation value exceeds the preset evaluation threshold as common controllable factors, and calculate the ratio of the comprehensive evaluation value of each common controllable factor to the sum of the comprehensive evaluation values of all common controllable factors to obtain the adjustment weight value; Multiply the adjustment weight value of each common controllable factor by the propagation coefficient of the corresponding synergistic enhancement relationship edge to obtain the adjustment gain value, and filter the common controllable factor with the largest adjustment gain value as the dominant adjustment factor; Generate a reference adjustment amount based on the adjustment weight value of the dominant adjustment factor, distribute the reference adjustment amount to the remaining common controllable factors according to the proportion of the adjustment weight value to obtain a linkage adjustment amount, adjust the parameters of the common controllable factors according to the linkage adjustment amount, and generate a preliminary adjustment scheme.

6. The method of claim 1, wherein, Based on the final adjustment scheme, generate control instructions for multiple sub-regions, and send the control instructions to the lighting devices of the corresponding sub-regions including: Obtain the adjustment parameters of the common controllable factors in the final adjustment scheme and the historical adjustment data of the sub-region lighting devices, calculate the change trend of the lighting parameters in the historical adjustment data, and construct a frequency domain response matrix between the lighting parameters; Based on the frequency domain response matrix, analyze the frequency components of the lighting parameter changes, extract the resonant frequency and the resonant frequency in the lighting parameters, and generate frequency adjustment instructions for multiple sub-regions; Collect the output spectrum data of the multiple sub-region lighting devices when executing the frequency adjustment instructions, calculate the energy transfer coefficient between adjacent wavelength segments, and perform coupling operation on the energy transfer coefficient and the frequency adjustment instructions to obtain a spectrum compensation amount, adjust the execution period of the frequency adjustment instructions based on the spectrum compensation amount, and generate energy balance instructions for multiple sub-regions; Monitor the driving circuit temperature and light emitting element temperature of the multiple sub-region lighting devices when executing the energy balance instructions, and construct a gradient field distribution of temperature change; Calculate the heat accumulation area and heat diffusion rate in the gradient field, modify the energy balance instructions, generate control instructions for multiple sub-regions, and send the control instructions to the lighting devices of the corresponding sub-regions.

7. AI-assisted green lighting scheme evaluation and intelligent decision system for implementing the method according to any one of the preceding claims 1-6, characterized in that, Including: The first unit is configured to identify the lighting demand differences of different spatial positions in the to-be-evaluated region, divide the to-be-evaluated region into multiple sub-regions, determine an evaluation weight for each sub-region, and generate multi-dimensional evaluation indicators based on the evaluation weight; wherein the lighting demand parameters include lighting required illuminance value, color temperature requirement, color rendering index requirement, and glare limit value; The second unit is configured to construct a coupling relationship graph including the constraint relationship and synergistic enhancement relationship between the indicators based on the multi-dimensional evaluation indicators, determine an indicator optimization sequence according to the constraint relationship, perform influence source reverse tracing on the to-be-optimized indicators based on the indicator optimization sequence, and generate a set of controllable factors. The third unit is configured to identify a common controllable factor capable of simultaneously affecting multiple to-be-optimized indexes from the set of controllable factors, trigger a synergistic enhancement relationship in the coupling relationship graph by adjusting the common controllable factor, and generate a preliminary adjustment scheme. The fourth unit is configured to substitute adjustment parameters of the common controllable factor in the preliminary adjustment scheme into an initial state table at a corresponding position in the coupling relationship graph, perform multiple rounds of propagation operation based on a propagation rule of the synergistic enhancement relationship edge, record numerical changes and propagation path information of the to-be-optimized indexes after each round of operation, and generate a predicted change amount table. The fourth unit is configured to substitute adjustment parameters of the common controllable factor in the preliminary adjustment scheme into an initial state table at a corresponding position in the coupling relationship graph, perform multiple rounds of propagation operation based on a propagation rule of the synergistic enhancement relationship edge, record numerical changes and propagation path information of the to-be-optimized indexes after each round of operation, and generate a predicted change amount table. The fourth unit is configured to substitute adjustment parameters of the common controllable factor in the preliminary adjustment scheme into an initial state table at a corresponding position in the coupling relationship graph, perform multiple rounds of propagation operation based on a propagation rule of the synergistic enhancement relationship edge, record numerical changes and propagation path information of the to-be-optimized indexes after each round of operation, and generate a predicted change amount table. The fourth unit is configured to substitute adjustment parameters of the common controllable factor in the preliminary adjustment scheme into an initial state table at a corresponding position in the coupling relationship graph, perform multiple rounds of propagation operation based on a propagation rule of the synergistic enhancement relationship edge, record numerical changes and propagation path information of the to-be-optimized indexes after each round of operation, and generate a predicted change amount table. The fourth unit is configured to substitute adjustment parameters of the common controllable factor in the preliminary adjustment scheme into an initial state table at a corresponding position in the coupling relationship graph, perform multiple rounds of propagation operation based on a propagation rule of the synergistic enhancement relationship edge, record numerical changes and propagation path information of the to-be-optimized indexes after each round of operation, and generate a predicted change amount table. The fifth unit is configured to generate, based on the final adjustment scheme, a control instruction for multiple sub-regions, and send the control instruction to a lighting device in a corresponding sub-region.

8. An electronic device, comprising: The computer program instructions are executed by the processor to implement the method of any one of claims 1-6. The computer program instructions are executed by the processor to implement the method of any one of claims 1-6. The computer program instructions are executed by the processor to implement the method of any one of claims 1-6. ​ 9. A computer-readable storage medium having stored thereon computer program instructions, wherein, ​

Images