A power transmission line multi-objective route selection method and system

Abstract

The application discloses a power transmission line multi-target line selection method and system, which firstly acquires a target region space environment and line selection task information to construct an environment model, generates a candidate line set and extracts multi-target quantitative evaluation results; secondly, representative lines are screened through a preference function and human-computer interaction feedback is acquired; subsequently, structured preference information is generated according to the feedback and the evaluation results to update the preference function, and then the candidate line set is iteratively updated; finally, the above-mentioned interaction and optimization process is repeated until a termination condition is met, and a target power transmission line is determined. The method of the application can effectively fuse engineering preferences and realize multi-target intelligent line selection of the power transmission line facing actual engineering requirements by dynamically updating the preference function through human-computer interaction feedback and iteratively optimizing the candidate scheme.

Description

Technical Field

[0001] This invention relates to the field of power system planning technology, and in particular to a multi-objective route selection method and system for transmission lines. Background Technology

[0002] Transmission line route planning is a core step in the early stages of power grid construction, and its route selection directly affects the project's investment costs, construction difficulty, operational safety, and ecological impact. As power grid construction continues to extend into complex terrains and ecologically sensitive areas, the route selection process needs to comprehensively consider multiple conflicting optimization objectives, such as construction costs, line length, safety, and smoothness.

[0003] Currently, the industry typically uses Geographic Information Systems (GIS) to construct spatial environment models and employs multi-objective evolutionary algorithms to automatically search in a vast three-dimensional solution space to generate a Pareto optimal solution set containing multiple candidate transmission lines, thereby assisting planners in comparing and selecting alternatives and making decisions.

[0004] However, existing multi-objective route selection methods lack effective human-computer interaction and preference learning mechanisms. They cannot dynamically capture expert feedback during algorithm iteration and transform it into structured preference information to guide the search direction. This leads to a disconnect between the algorithm optimization process and the true intentions of humans. When faced with a massive number of candidate solutions, planners are prone to "decision paralysis" and find it difficult to efficiently and accurately determine the target transmission line that meets the engineering expectations. Summary of the Invention

[0005] This invention provides a multi-target route selection method and system for power transmission lines, so as to improve the scientificity and accuracy of power transmission line route selection.

[0006] To address the aforementioned technical problems, this invention provides a multi-target line selection method for power transmission lines, comprising: Acquire spatial environment information and route selection task information of the target area, and construct an environmental model for transmission line route selection based on the spatial environment information; Based on the environmental model and the route selection task information, a candidate line set containing multiple candidate transmission lines is generated, and the quantitative evaluation results of each candidate transmission line under multiple preset optimization objectives are extracted. The candidate route set is filtered based on the preference function to obtain at least two candidate routes of different categories. Structured preference information is generated based on the human-computer interaction feedback information of the candidate routes and the quantitative evaluation results. The preference function is updated based on the structured preference information. This step is repeated until the preset termination condition is met. The preset termination condition includes the inability to filter at least two candidate lines of different categories; If the preset termination condition is met, the target transmission line for the target area is determined based on the finally updated preference function.

[0007] As one preferred embodiment, the step of generating a candidate line set containing multiple candidate transmission lines based on the environmental model and the line selection task information includes: The starting point and ending point of the candidate route are determined based on the route selection task information; The candidate line is represented as a candidate transmission line consisting of the starting point, the ending point, and several variable intermediate tower positions, wherein the number and spatial position of the intermediate tower positions are variable coding parameters, and adjacent starting points, intermediate tower positions and ending points are connected in sequence to form a broken line transmission line path. Based on the environmental model, the feasibility of the candidate transmission lines is assessed by constraint, and the candidate transmission lines that meet the preset constraint conditions are included in the candidate line set.

[0008] As one preferred embodiment, the extraction of quantitative evaluation results for each candidate transmission line under multiple preset optimization objectives includes: Based on the described polygonal transmission line path, determine the line sampling points or the area through which the line passes; Based on the spatial environment information corresponding to the line sampling points or the areas through which the line passes in the environmental model, the quantitative evaluation results of each candidate transmission line under the multiple preset optimization objectives are calculated. Among them, the multiple preset optimization objectives include at least two of the following: construction cost, line length, safety index, and smoothness index.

[0009] As one preferred embodiment, the step of filtering the candidate route set based on a preference function to obtain at least two candidate routes of different categories includes: The candidate transmission line with the best preference evaluation is selected from the candidate line set and used as the preference anchor line; Select the candidate transmission line that differs most from the preferred anchor line from the candidate line set as the comparison line; Based on the degree of difference between candidate transmission lines, supplementary lines are selected from the remaining candidate transmission lines in the candidate line set; The candidate routes include the preferred anchor route, the comparison route, and the supplementary route.

[0010] As one preferred embodiment, the step of generating structured preference information based on the human-computer interaction feedback information of the candidate routes and the quantitative evaluation results includes: A structured context is generated based on the human-computer interaction feedback information, the quantitative evaluation results, the target metadata, and the historical interaction information. The structured context is input into a pre-built preference parsing model to obtain the structured preference information; The structured preference information includes: target priority, trade-off relationships and / or coupling relationships between targets, and preference confidence.

[0011] Another aspect of the present invention provides a multi-target line selection system for transmission lines, comprising: The acquisition module is used to acquire spatial environmental information and route selection task information of the target area, and to construct an environmental model for transmission line route selection based on the spatial environmental information. The generation module is used to generate a set of candidate transmission lines containing multiple candidate transmission lines based on the environmental model and the route selection task information, and to extract the quantitative evaluation results of each candidate transmission line under multiple preset optimization objectives. The update module is used to filter the candidate route set based on the preference function to obtain at least two candidate routes of different categories. It generates structured preference information based on the interaction feedback information and quantitative evaluation results of the candidate routes, and updates the preference function based on the structured preference information. This step is repeated until the preset termination condition is met. Among them, the preset termination conditions include the inability to filter out at least two candidate routes of different categories; The output module is used to determine the target transmission line in the target area based on the finally updated preference function, provided that the preset termination conditions are met.

[0012] As one preferred embodiment, the generation module is specifically used for: The starting point and ending point of the candidate route are determined based on the route selection task information; The candidate line is represented as a candidate transmission line consisting of the starting point, the ending point, and several variable intermediate tower positions, wherein the number and spatial position of the intermediate tower positions are variable coding parameters, and adjacent starting points, intermediate tower positions and ending points are connected in sequence to form a broken line transmission line path. Based on the environmental model, the feasibility of the candidate transmission lines is assessed by constraint, and the candidate transmission lines that meet the preset constraint conditions are included in the candidate line set.

[0013] As one preferred embodiment, the generation module is further configured to: Based on the described polygonal transmission line path, determine the line sampling points or the area through which the line passes; Based on the spatial environment information corresponding to the line sampling points or the areas through which the line passes in the environmental model, the quantitative evaluation results of each candidate transmission line under the multiple preset optimization objectives are calculated. Among them, the multiple preset optimization objectives include at least two of the following: construction cost, line length, safety index, and smoothness index.

[0014] As one preferred embodiment, the update module is specifically used for: The candidate transmission line with the best preference evaluation is selected from the candidate line set and used as the preference anchor line; Select the candidate transmission line that differs most from the preferred anchor line from the candidate line set as the comparison line; Based on the degree of difference between candidate transmission lines, supplementary lines are selected from the remaining candidate transmission lines in the candidate line set; The candidate routes include the preferred anchor route, the comparison route, and the supplementary route.

[0015] As one preferred embodiment, the update module is specifically used for: A structured context is generated based on the human-computer interaction feedback information, the quantitative evaluation results, the target metadata, and the historical interaction information. The structured context is input into a pre-built preference parsing model to obtain the structured preference information; The structured preference information includes: target priority, trade-off relationships and / or coupling relationships between targets, and preference confidence.

[0016] Compared with the prior art, the beneficial effects of the present invention are at least one of the following: 1) The multi-objective route selection method for transmission lines provided by this invention effectively solves the problems of blind searching and disconnection from human intent in traditional multi-objective route selection algorithms by introducing an interactive preference learning mechanism based on "human-in-the-loop" principles during the iterative optimization process. Specifically, when preset interaction conditions are met, the method selects candidate lines from the candidate line set and obtains interactive feedback information. Then, it combines the quantitative evaluation results to generate structured preference information to dynamically update the preference function. This approach makes the implicit engineering experience and decision-making logic of planning experts explicit and mathematical, and guides the algorithm's search direction in real time, causing the candidate line set to continuously converge towards the optimization region that meets human expectations. This significantly reduces invalid searches and the generation of redundant candidate schemes, and significantly improves the computational efficiency and resource utilization of the route selection process.

[0017] 2) This invention digitally maps spatial environmental information by constructing an environmental model and quantitatively evaluates candidate transmission lines by combining multiple preset optimization objectives, achieving refined modeling of complex geographical environments and multidimensional engineering constraints. Simultaneously, it iteratively updates the candidate line set based on the updated preference function, and determines the final target transmission line by integrating the preference function and the candidate set after the termination condition is met. This ensures that the output scheme not only meets the constraints of physical environment feasibility but also highly aligns with the actual preferences of planning experts in terms of multidimensional objective trade-offs. This method effectively avoids the "decision paralysis" problem faced by planners when confronted with a massive number of Pareto optimal solutions in traditional methods, improving the scientific rigor, accuracy, and engineering practicality of transmission line selection. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a high-level comprehensive optimization method in one embodiment of the present invention; Figure 2 This is a structural block diagram of a high-level integrated optimization system in one embodiment of the present invention.

[0019] Figure label: Among them, 11 is the acquisition module; 12 is the generation module; 13 is the update module; and 14 is the output module. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0021] In the description of this invention, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0022] In the description of this invention, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0023] One embodiment of the present invention provides a multi-target line selection method for power transmission lines. For details, please refer to [link to documentation]. Figure 1 , Figure 1 The diagram shown is a flowchart of a multi-target line selection method for transmission lines according to one embodiment of the present invention, which includes steps S1-S4: S1: Obtain spatial environment information and route selection task information of the target area, and construct an environmental model for power transmission line route selection based on the spatial environment information.

[0024] In the planning and construction of power transmission lines, planners typically need to determine a reasonable route and the placement of intermediate towers between a given starting point and ending point. Since the transmission line scheme directly affects construction costs, line length, operational safety, corridor coordination, and the impact on the surrounding environment, comprehensively and accurately acquiring spatial environmental information of the target area and constructing an environmental model is a crucial foundation for achieving multi-objective intelligent route selection.

[0025] In this embodiment, spatial environmental information of the target area is acquired. Preferably, the spatial environmental information may include at least one of the following: remote sensing image data, semantic data of ground features, topographic data, building area data, road area data, water body area data, farmland area data, mountain area data, existing road data, restricted area data, and security-sensitive area data. It should be noted that the spatial environmental information may come from remote sensing image processing results, geographic information layers, planning and control data, manually labeled data, or other spatial data sources, and this invention does not limit this.

[0026] In one optional implementation, remote sensing imagery of the target study area can be acquired, and image segmentation models can be used to extract semantic features from the remote sensing imagery to obtain at least one feature category from buildings, roads, water bodies, farmland, mountains, and other areas. Further, these feature categories can be converted into a feature semantic layer, an engineering cost layer, a safety-sensitive area layer, a restricted area layer, or an existing route constraint layer. Through this layering process, unstructured image data can be transformed into structured spatial constraint information, facilitating subsequent candidate route generation, constraint judgment, and multi-objective evaluation.

[0027] In another alternative implementation, existing geographic information layers, planning and control layers, restricted area layers, existing route layers, or security-sensitive area layers can be directly acquired and used as part of the environmental model. By spatially registering and overlaying these multi-source layers, an environmental model containing multi-dimensional spatial constraints can be constructed, thereby providing complete underlying data support for subsequent route selection.

[0028] In this embodiment, corresponding route selection task information also needs to be obtained. The route selection task information may include at least one of the following: the starting point, ending point, tower placement requirements, line length constraints, turning angle constraints, obstacle avoidance constraints, and existing line coordination constraints for the candidate route. The starting point and ending point can be a substation, access point, line endpoint, or a planned designated location; tower placement requirements may include a range of intermediate tower positions, spacing requirements between adjacent tower positions, and requirements that tower positions must not fall into restricted areas; existing line coordination constraints may include requirements such as maintaining a preset distance from existing lines, avoiding intersections, or satisfying parallel channel relationships.

[0029] Based on the spatial environment information and route selection task information obtained above, an environmental model for transmission line route selection can be constructed. By constructing an environmental model that integrates multi-source spatial data and engineering constraints, the accuracy and engineering applicability of subsequent route search and multi-objective evaluation can be effectively improved.

[0030] S2: Based on the environmental model and route selection task information, generate a candidate line set containing multiple candidate transmission lines, and extract the quantitative evaluation results of each candidate transmission line under multiple preset optimization objectives.

[0031] In the process of selecting transmission lines, the generation method of candidate lines and the construction of the evaluation index system directly affect the search efficiency and scheme quality of subsequent multi-objective optimization. Traditional methods usually use fixed grids or fixed nodes for path search, which is difficult to flexibly represent line schemes under different numbers of towers and spatial layouts. Therefore, in this embodiment, a coding method based on variable intermediate towers is used to generate candidate transmission lines, and multi-dimensional optimization objectives are combined for quantitative evaluation to support subsequent multi-objective iterative optimization and interactive preference guidance.

[0032] In this embodiment, multiple initial candidate transmission lines can be generated based on the environmental model and route selection task information. These initial candidate transmission lines can be generated using at least one of the following methods: path search, random tree expansion, heuristic generation, existing line adjustment, and random initialization.

[0033] Preferably, the starting point and ending point of the candidate lines are first determined based on the route selection task information. The starting point and ending point can be substations, access points, line endpoints, or planned designated locations. Then, the candidate lines are represented as candidate transmission lines consisting of a starting point, an ending point, and several variable intermediate tower locations. The number and spatial location of the intermediate tower locations are variable coding parameters. Preferably, the candidate transmission lines can be represented in the following form: in, Indicates the starting point of the candidate route. Indicates the destination of the candidate route. Indicates the middle tower position. This indicates the number of intermediate tower positions, and It falls within the preset quantity range.

[0034] Adjacent starting points, intermediate tower locations, and ending points are connected sequentially to form a broken-line transmission line path. It should be noted that encoding candidate transmission lines as a variable number of intermediate tower location sequences allows for simultaneous searching of line path directions and tower location layout schemes within the same optimization framework, avoiding suboptimal problems that may arise from phased optimization.

[0035] In another optional embodiment of the present invention, a number of low-cost feasible routes can be generated first using a path search method as basic candidate routes, and then candidate routes with different geometric structures can be generated using a random tree expansion method or a random initialization method to improve the diversity of the initial candidate route set. For scenarios involving the modification of existing routes or channel coordination, local adjustments can be made based on existing routes to generate initial candidate routes that meet the requirements of new routes.

[0036] After generating candidate transmission lines, a constraint feasibility assessment is performed on the candidate transmission lines based on an environmental model. The constraint feasibility assessment may include at least one of the following: line length constraint assessment, corner constraint assessment, tower location constraint assessment, restricted area avoidance assessment, and existing line coordination assessment. Preferably, the line length constraint can be expressed as: in, This represents the maximum permissible length of the line. Angle constraints can be expressed as: in, This represents the maximum allowed turning angle. The constraint on the number of intermediate tower positions can be expressed as: in, and These represent the lower and upper limits of the number of intermediate tower positions, respectively. The tower position placement constraint can be expressed as: in, This represents raster or layer information in the environment model. This represents the set of land feature categories where tower placement is prohibited. Configuration feasibility constraints can be expressed as: Accordingly, the set of feasible candidate transmission lines can be represented as: The constraints described above are merely examples. In other implementations, constraints such as restricted area avoidance, existing line coordination, tower spacing, boundary constraints, or construction accessibility constraints may be further added.

[0037] For each candidate transmission line included in the candidate line set, it is necessary to extract its quantitative evaluation results under multiple preset optimization objectives. In this embodiment, line sampling points or line-traversed areas are determined along the broken-line transmission line path for constraint judgment and multi-objective evaluation. The line sampling points can be points or grid cells sampled along the line segment at preset intervals; the line-traversed areas can be the spatial areas traversed or influenced by the candidate line. Through the line sampling points or line-traversed areas, the corresponding land feature categories, engineering costs, safety distances, and constraint information can be read, thereby calculating the quantitative evaluation results of each candidate transmission line under multiple preset optimization objectives.

[0038] In this embodiment, the multi-objective evaluation indicators include at least construction cost, line length, safety indicators, and smoothness indicators. Each indicator can be uniformly set to be as small as possible, or it can be normalized according to actual needs. The multi-objective optimization problem for candidate transmission lines can be expressed as: in, Indicates construction cost, Indicates the length of the line. Indicates safety indicators, Indicates the smoothness index, This represents the set of candidate transmission lines that satisfy the basic constraints.

[0039] Specifically, the construction cost can be determined based on at least one of the following: tower site layout cost of the candidate transmission line, land occupation cost, and cost of crossing sensitive features. Preferably, the construction cost can be expressed as: in, Indicates the type of land feature or terrain label corresponding to a spatial location; Indicates candidate transmission line The corresponding set of line sampling points or the set of sampling units within the area traversed by the line; This represents the cost function for tower site layout; A function representing the cost of land use by type; This represents the cost function for traversing sensitive features.

[0040] The line length can be determined by summing the distances between adjacent starting points, intermediate towers, and ending points in the candidate transmission line, and can be expressed as: in, This represents the Euclidean distance. The route length metric is used to encourage candidate routes to reduce unnecessary detours while meeting constraints.

[0041] Safety indicators can be determined based on the distance relationship or exposure level between the candidate transmission line and at least one of the following: building areas, road areas, water bodies, existing lines, restricted areas, and safety-sensitive areas. In one optional embodiment, it is set that... Indicates the line sampling point The distance to the nearest safety-sensitive or risky area can be taken as the lower quantile of the safety distance from the line sampling points as the critical safety distance: in, For example, the preset quantiles Furthermore, the normalized safety clearance mass can be defined: in, This serves as a reference safe distance. Risk exposure ratios can also be defined: in, This is an indicator function. The line safety score can be expressed as: in, and Let be the weighting coefficient. To incorporate the safety metric into a multi-objective optimization framework where smaller values ​​are better, we can set: Therefore, minimize This is equivalent to maximizing the line safety score.

[0042] The smoothness index can be determined based on the angle, bend, or tortuosity of the candidate transmission line at intermediate tower locations. For intermediate tower locations... Its turning angle can be expressed as: in, To prevent constants with a denominator of zero, the smoothness index can be expressed as: in, This indicates the allowed free-turn threshold. Indicates the threshold for sharp bends. and This corresponds to the penalty weight. This indicator is used to reduce sharp bends and unnecessary turns in candidate routes, making the route shape more conducive to engineering implementation.

[0043] S3: Filter the candidate route set based on the preference function to obtain at least two candidate routes of different categories. Generate structured preference information based on the human-computer interaction feedback information and quantitative evaluation results of the candidate routes, and update the preference function based on the structured preference information. Repeat this step until the preset termination condition is met. The preset termination condition includes the inability to filter at least two candidate routes of different categories.

[0044] In the multi-objective route selection process for transmission lines, the candidate route set typically contains a large number of schemes with different trade-offs among multiple optimization objectives. Presenting all candidate routes to planners at once would not only increase the cognitive burden of interaction but also potentially lead to redundant feedback information and reduced preference analysis efficiency. Therefore, this embodiment employs a candidate route selection strategy based on a preference function, combining human-computer interaction feedback and a preference analysis model. The preference function is updated progressively within a limited number of interaction rounds, enabling subsequent search processes to converge towards route areas that align with the planners' engineering preferences.

[0045] In this embodiment, the candidate route set is filtered based on a preference function to obtain at least two candidate routes of different categories. Let the current set of candidate routes that need to be truncated or filtered be . The current preference function is The number of candidate routes to be displayed is First, the candidate transmission line with the best current preference evaluation is selected as the preference anchor line: The smaller the value of the preference function, the more the candidate transmission line conforms to the current planning preference. When the current preference function is unavailable, the preference anchor line can be determined according to a preset equilibrium weight or a preset initial preference function.

[0046] Secondly, the candidate transmission line with the greatest difference from the preferred anchor line in the normalized multi-objective evaluation space is selected from the candidate line set as the comparison line, which can be expressed as: in, Indicates candidate transmission lines The normalized multi-objective evaluation vector.

[0047] Then, supplement the remaining candidate routes according to the maximum and minimum distance criterion: in, This indicates the set of candidate routes selected for display. Through this method, the candidate routes presented to planners are both relevant to the current preference estimate and possess a certain degree of differentiation, which is beneficial for extracting more effective preference information from a limited number of selections. It should be noted that candidate routes include preference anchor routes, comparison routes, and supplementary routes; this invention does not limit the number of routes displayed or impose specific selection criteria.

[0048] After selecting candidate routes, these routes are displayed to planners, and their interactive feedback is obtained. Planners can choose a relatively satisfactory route from the candidate route set, or provide sorting, pairwise comparison, or other forms of preference feedback. It should be noted that this application does not limit the specific interactive interface and input format for the selection feedback.

[0049] Structured preference information is generated using a preference parsing model based on interactive feedback information, multi-objective evaluation metrics of candidate routes, and historical interaction information. The preference parsing model may include at least one of a large language model, a natural language understanding model, a rule parsing model, a machine learning model, and a human-computer interaction preference parsing model.

[0050] In one alternative implementation, the input to the preference parsing model is organized as a structured context, including target metadata (target name, number of targets, and target optimization direction), historical interaction information (historical selection feedback, historical candidate route evaluation information, and historical preference estimation information), candidate profile information of the current candidate route (original multi-objective evaluation indicators and normalized multi-objective evaluation indicators), and the planner's selection feedback in the current round. Based on this structured context, the preference parsing model outputs structured preference information, which includes target priority, trade-off relationships and / or coupling relationships between targets, and preference confidence.

[0051] Specifically, in one embodiment of the present invention, in the first Each interaction round organizes the input of the preference parsing model into a structured context: in, This represents the target metadata, including the target name, the number of targets, and the target optimization direction; It represents historical interaction information, including at least one of historical selection feedback, historical candidate route evaluation information, and historical preference estimation information; The candidate profile information representing the current candidate route includes at least one of the original multi-objective evaluation index and the normalized multi-objective evaluation index. This indicates the planners' feedback on their choices in the current round.

[0052] The preference parsing model outputs structured preference information based on the structured context: in, This represents a preference analysis model; Indicates target priority; This indicates the trade-off relationship and / or coupling relationship between objectives; This indicates the confidence level of the preference.

[0053] In one alternative implementation, target priority can be represented as a set of pairwise priority relationships between targets: in, Represents the target index set. This indicates that planners have a target The level of importance attached to it is higher than the target. .

[0054] The trade-offs and / or coupling relationships between objectives can be represented as: in, Indicate target and target There exists a strength of joint consideration, mutual trade-offs, or coupling relationships. Preference confidence can be expressed as: in, The larger the value, the higher the reliability estimate of the current structured preference information by the preference parsing model; The smaller the value, the less evidence there is of the current interaction or the greater the uncertainty.

[0055] After generating structured preference information, the preference function is updated based on the structured preference information and cumulative selection feedback. The cumulative selection feedback includes candidate transmission lines selected by planners in one or more interaction rounds, as well as candidate transmission lines that were not selected. Candidate transmission lines selected by planners can be used as samples with higher preferences, and candidate transmission lines that were not selected can be used as samples with lower preferences, constructing a pair of preference data. Preferably, the preference function may include at least one of a first-order target weight term, a target nonlinear term, and an inter-target interaction term to simultaneously characterize the first-order importance of the target, the nonlinear influence of the target, and the inter-target interaction influence.

[0056] Specifically, in one embodiment of the present invention, the multi-objective evaluation index of candidate transmission lines is normalized into a unified representation where smaller is better. Let the objective vector of the candidate transmission line be... , its first The normalized value of each objective can be expressed as: in, and They represent the first The lower and upper bounds of the reference for each target. To prevent constants with zero denominators, let the normalization objective vector be: In one alternative implementation, a structure-enhanced preference function can be constructed. First, a nonlinear feature transformation is performed on the normalized target vector: in, and These can be preset or fixed random parameters. Further, construct second-order target features: And construct combined features: The preference function can be expressed as: in, Let these be the parameters to be learned. Furthermore, the parameters can be written as: in, This represents the first-order weight term of the objective. The parameters represent the target nonlinear term and the interaction term between targets. The parameter represents the nonlinear characteristic. Therefore, the preference function can simultaneously characterize the first-order importance of the target, the nonlinear influence of the target, and the interaction between targets.

[0057] For the The interaction feedback from previous rounds can be used to construct a pairwise preference dataset: in, Indicates the first The normalized target vector corresponding to the candidate transmission lines selected by the planners in the round. Indicates the first The one not selected in the round The normalized target vector corresponding to each candidate transmission line. Indicates the index of the selected candidate transmission line.

[0058] In one alternative implementation, the preference function parameters can be updated using the following objective function: in, Represents the set of parameters for the preference function. This represents the ranking loss based on pairwise preference data. These represent constraints based on target priority. This represents constraint terms based on trade-offs and / or coupling relationships between objectives. and These are the weighting coefficients.

[0059] The ranking loss can be expressed as: in, This represents the Sigmoid function. Since a smaller preference function value indicates better performance in this embodiment, minimizing this ranking loss can help selected candidate lines achieve better preference scores than unselected candidate lines.

[0060] The target priority constraint can be expressed as: in, This indicates the preset initial weights or equilibrium weights. , This is the priority interval parameter. This term is used to align the first-order objective weights in the preference function with the objective priorities in the structured preference information.

[0061] The constraints on the trade-offs and / or coupling relationships between objectives can be expressed as: in, This indicates the parameters corresponding to the interaction items between targets. This represents the single-objective nonlinear parameter in the second-order term of the objective. To represent the prior relationship between targets obtained from structured preference information, for example, we can let: in, This is the scaling factor. Preference confidence level. Used to adjust the strength of the influence of structured preference information on preference function updates. When At higher levels, the objective priority constraints and inter-objective relationship constraints have a stronger impact on the preference function update; when At lower levels, its impact weakens to avoid over-reliance on unstable preference analysis results.

[0062] After updating the preference function, the updated preference function is used for subsequent candidate route sampling, candidate route selection, and iterative updates of the candidate route set, so as to guide subsequent multi-objective iterative optimization to converge to the route area that conforms to the engineering preferences of the planners.

[0063] Specifically, in the sampling of candidate lines, candidate transmission lines with better preference scores and differences can be selected as candidates for the next round of interaction; in the screening of candidate lines, the preference score can be used as one of the bases for retaining, sorting or truncating candidate transmission lines; in the updating of the candidate line set, candidate transmission lines that are more compatible with the current preference function can be given a higher probability of retention.

[0064] The steps of screening candidate lines, obtaining interactive feedback information, generating structured preference information, and updating the preference function are repeated until a preset termination condition is met. The preset termination condition may include at least one of the following: reaching a preset number of iterations, reaching a preset number of interaction rounds, convergence of the candidate line set, convergence of the preference function, the candidate line quality meeting preset requirements, inability to select at least two candidate lines of different categories, and receiving a termination command. Through this iterative interaction process, the planner's engineering preferences can continuously influence the generation and search phases of candidate lines, effectively reducing the search consumption in irrelevant line areas and the burden of posterior screening, and improving the generation efficiency and engineering applicability of transmission line route selection schemes under complex spatial constraints and multi-objective trade-offs.

[0065] S4: If the preset termination conditions are met, determine the target transmission line in the target area based on the final updated preference function.

[0066] In the multi-objective route selection process for transmission lines, through multiple rounds of human-computer interaction and preference function updates, the candidate route set gradually converges towards the region that meets the engineering preferences of the planners. When the interaction process meets the preset termination conditions, the target transmission line that best meets the planners' preferences and satisfies all engineering constraints needs to be determined from the final candidate route set as the final route selection planning scheme.

[0067] In this embodiment, it is first determined whether the current iterative optimization and interaction process meets a preset termination condition. The preset termination condition may include at least one of the following: reaching a preset number of iterations, reaching a preset number of interaction rounds, convergence of the candidate line set, convergence of the preference function, the candidate line quality meeting preset requirements, inability to filter at least two candidate lines of different categories, and receiving a termination command input by the planner. It should be noted that when the distribution difference of the remaining candidate transmission lines in the candidate line set in the multi-objective evaluation space is less than a preset threshold, or when the planner's selection feedback in consecutive rounds of interaction does not cause a significant change in the preference function, the convergence condition can be considered met. This invention does not limit the scope of the invention; the specific judgment logic of the preset termination condition can be configured according to actual engineering needs.

[0068] If the preset termination conditions are met, the target transmission line is determined from the current candidate line set based on the finally updated preference function. Specifically, the final updated preference function is used to evaluate the preferences of each candidate transmission line in the current candidate line set, obtaining the final preference score for each candidate transmission line. Preferably, the candidate transmission line with the best final preference score (e.g., the smallest preference function value) can be selected as the target transmission line. It should be noted that the finally updated preference function has fully incorporated the engineering preferences expressed by planners in multiple rounds of interaction; therefore, the target transmission line determined based on this preference function can best meet the actual needs of planners.

[0069] In one optional implementation, if multiple candidate transmission lines with similar or identical preference scores exist in the final candidate line set, the final target transmission line can be determined by further combining preset secondary evaluation indicators or secondary confirmation by planners. The secondary evaluation indicators may include environmental impact indicators, construction accessibility indicators, or corridor coordination indicators, etc. Furthermore, multiple candidate transmission lines with the highest preference scores can be compiled into a recommended scheme set for planners to conduct final manual review and selection. This invention is not limited thereto; the method for determining the target transmission line can be flexibly adjusted according to the actual application scenario.

[0070] After determining the target transmission line, detailed information about the target transmission line is output. This detailed information may include the line path information, spatial location information of intermediate tower sites, quantitative evaluation results of each preset optimization objective, and corresponding preference scores. Preferably, the line path and tower site layout scheme of the target transmission line can be overlaid and displayed in a geographic information system or 3D visualization environment to facilitate intuitive engineering review and subsequent detailed design by planners. Through this method, the results of multi-objective optimization can be effectively combined with the subjective engineering preferences of planners, resulting in a high-quality transmission line selection scheme that satisfies both objective engineering constraints and subjective preferences.

[0071] Another embodiment of the present invention provides a multi-target line selection system for transmission lines. For details, please refer to [link to relevant documentation]. Figure 2 , Figure 2 The diagram shown illustrates a structural block diagram of a multi-target line selection system for power transmission lines according to one embodiment of the present invention, comprising: The acquisition module is used to acquire spatial environmental information and route selection task information of the target area, and to construct an environmental model for transmission line route selection based on the spatial environmental information. The generation module is used to generate a set of candidate transmission lines containing multiple candidate transmission lines based on the environmental model and the route selection task information, and to extract the quantitative evaluation results of each candidate transmission line under multiple preset optimization objectives. The update module is used to filter the candidate route set based on the preference function to obtain at least two candidate routes of different categories. It generates structured preference information based on the interaction feedback information and quantitative evaluation results of the candidate routes, and updates the preference function based on the structured preference information. This step is repeated until the preset termination condition is met. Among them, the preset termination conditions include the inability to filter out at least two candidate routes of different categories; The output module is used to determine the target transmission line in the target area based on the finally updated preference function, provided that the preset termination conditions are met.

[0072] Preferably, in one embodiment of the present invention, the generation module is specifically used for: Determine the starting and ending points of the candidate routes based on the route selection task information; The candidate line is represented as a candidate transmission line consisting of a starting point, an ending point, and several variable intermediate tower positions. The number and spatial location of the intermediate tower positions are variable coding parameters. Adjacent starting points, intermediate tower positions, and ending points are connected in sequence to form a broken line transmission line path. Based on the environmental model, the feasibility of candidate transmission lines is assessed under constraints, and candidate transmission lines that meet the preset constraints are included in the candidate line set.

[0073] Preferably, in one embodiment of the present invention, the generation module is further configured to: Determine the sampling points or areas traversed by the line based on the broken-line transmission line path; Based on the spatial environment information corresponding to the line sampling points or the areas through which the line passes in the environmental model, the quantitative evaluation results of each candidate transmission line under multiple preset optimization objectives are calculated. Among them, several preset optimization objectives include at least two of the following: construction cost, line length, safety indicators, and smoothness indicators.

[0074] Preferably, in one embodiment of the present invention, the updating module is specifically used for: The candidate transmission line with the best preference evaluation is selected from the candidate line set and used as the preference anchor line; The candidate transmission line with the greatest difference from the preferred anchor line is selected from the candidate line set as the comparison line; Based on the degree of difference between candidate transmission lines, supplementary lines are selected from the remaining candidate transmission lines in the candidate line set; The candidate routes include preferred anchor route, comparison route, and supplementary route.

[0075] Preferably, in one embodiment of the present invention, the updating module is specifically used for: Structured context is generated based on human-computer interaction feedback information, quantitative evaluation results, target metadata, and historical interaction information. The structured context is input into a pre-built preference parsing model to obtain structured preference information; The structured preference information includes: target priority, trade-off relationships and / or coupling relationships between targets, and preference confidence.

[0076] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A multi-target route selection method for power transmission lines, characterized in that, include: Acquire spatial environment information and route selection task information of the target area, and construct an environmental model for transmission line route selection based on the spatial environment information; Based on the environmental model and the route selection task information, a candidate line set containing multiple candidate transmission lines is generated, and the quantitative evaluation results of each candidate transmission line under multiple preset optimization objectives are extracted. The candidate route set is filtered based on the preference function to obtain at least two candidate routes of different categories. Structured preference information is generated based on the human-computer interaction feedback information of the candidate routes and the quantitative evaluation results. The preference function is updated based on the structured preference information. This step is repeated until the preset termination condition is met. The preset termination condition includes the inability to filter at least two candidate lines of different categories; If the preset termination condition is met, the target transmission line for the target area is determined based on the finally updated preference function.

2. The multi-target line selection method for transmission lines as described in claim 1, characterized in that, Based on the environmental model and the route selection task information, a candidate line set containing multiple candidate transmission lines is generated, including: The starting point and ending point of the candidate route are determined based on the route selection task information; The candidate line is represented as a candidate transmission line consisting of the starting point, the ending point, and several variable intermediate tower positions, wherein the number and spatial position of the intermediate tower positions are variable coding parameters, and adjacent starting points, intermediate tower positions and ending points are connected in sequence to form a broken line transmission line path. Based on the environmental model, the feasibility of the candidate transmission lines is assessed by constraint, and the candidate transmission lines that meet the preset constraint conditions are included in the candidate line set.

3. The multi-target line selection method for transmission lines as described in claim 2, characterized in that, The extraction of quantitative evaluation results for each candidate transmission line under multiple preset optimization objectives includes: Based on the described polygonal transmission line path, determine the line sampling points or the area through which the line passes; Based on the spatial environment information corresponding to the line sampling points or the areas through which the line passes in the environmental model, the quantitative evaluation results of each candidate transmission line under the multiple preset optimization objectives are calculated. Among them, the multiple preset optimization objectives include at least two of the following: construction cost, line length, safety index, and smoothness index.

4. The multi-target line selection method for transmission lines as described in claim 1, characterized in that, The process of filtering the candidate route set based on a preference function to obtain at least two candidate routes of different categories includes: The candidate transmission line with the best preference evaluation is selected from the candidate line set and used as the preference anchor line; Select the candidate transmission line that differs most from the preferred anchor line from the candidate line set as the comparison line; Based on the degree of difference between candidate transmission lines, supplementary lines are selected from the remaining candidate transmission lines in the candidate line set; The candidate routes include the preferred anchor route, the comparison route, and the supplementary route.

5. The multi-target line selection method for transmission lines as described in claim 1, characterized in that, The step of generating structured preference information based on the human-computer interaction feedback information of the candidate routes and the quantitative evaluation results includes: A structured context is generated based on the human-computer interaction feedback information, the quantitative evaluation results, the target metadata, and the historical interaction information. The structured context is input into a pre-built preference parsing model to obtain the structured preference information; The structured preference information includes: target priority, trade-off relationships and / or coupling relationships between targets, and preference confidence.

6. A multi-target line selection system for transmission lines, characterized in that, include: The acquisition module is used to acquire spatial environmental information and route selection task information of the target area, and to construct an environmental model for transmission line route selection based on the spatial environmental information. The generation module is used to generate a set of candidate transmission lines containing multiple candidate transmission lines based on the environmental model and the route selection task information, and to extract the quantitative evaluation results of each candidate transmission line under multiple preset optimization objectives. The update module is used to filter the candidate route set based on the preference function to obtain at least two candidate routes of different categories. It generates structured preference information based on the interaction feedback information and quantitative evaluation results of the candidate routes, and updates the preference function based on the structured preference information. This step is repeated until the preset termination condition is met. Among them, the preset termination conditions include the inability to filter out at least two candidate routes of different categories; The output module is used to determine the target transmission line in the target area based on the finally updated preference function, provided that the preset termination conditions are met.

7. The multi-target line selection system for transmission lines as described in claim 6, characterized in that, The generation module is specifically used for: The starting point and ending point of the candidate route are determined based on the route selection task information; The candidate line is represented as a candidate transmission line consisting of the starting point, the ending point, and several variable intermediate tower positions, wherein the number and spatial position of the intermediate tower positions are variable coding parameters, and adjacent starting points, intermediate tower positions and ending points are connected in sequence to form a broken line transmission line path. Based on the environmental model, the feasibility of the candidate transmission lines is assessed by constraint, and the candidate transmission lines that meet the preset constraint conditions are included in the candidate line set.

8. The multi-target line selection system for transmission lines as described in claim 7, characterized in that, The generation module is further configured to: Based on the described polygonal transmission line path, determine the line sampling points or the area through which the line passes; Based on the spatial environment information corresponding to the line sampling points or the areas through which the line passes in the environmental model, the quantitative evaluation results of each candidate transmission line under the multiple preset optimization objectives are calculated. Among them, the multiple preset optimization objectives include at least two of the following: construction cost, line length, safety index, and smoothness index.

9. The multi-target line selection system for transmission lines as described in claim 6, characterized in that, The update module is specifically used for: The candidate transmission line with the best preference evaluation is selected from the candidate line set and used as the preference anchor line; Select the candidate transmission line that differs most from the preferred anchor line from the candidate line set as the comparison line; Based on the degree of difference between candidate transmission lines, supplementary lines are selected from the remaining candidate transmission lines in the candidate line set; The candidate routes include the preferred anchor route, the comparison route, and the supplementary route.

10. The multi-target line selection system for transmission lines as described in claim 6, characterized in that, The update module is specifically used for: A structured context is generated based on the human-computer interaction feedback information, the quantitative evaluation results, the target metadata, and the historical interaction information. The structured context is input into a pre-built preference parsing model to obtain the structured preference information; The structured preference information includes: target priority, trade-off relationships and / or coupling relationships between targets, and preference confidence.

Images