USER-INTERACTIVE, MULTI-CRITERIA AND PARAMETER-CONTROLLED GEOGRAPHIC ZONING SYSTEM AND METHOD
Patent Information
- Application Number
- TR202514693
- Authority / Receiving Office
- TR · TR
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-10-08
- Publication Date
- 2026-06-22
Smart Images
Figure 00000015_0000
Abstract
Description
1 TARIFF USER-INTERACTIVE, MULTI-CRITERIA, AND PARAMETER-CONTROLLED. GEOGRAPHICAL ZONING SYSTEM AND METHOD Technical Area The invention relates to user-defined numerical data within the scope of geographic information systems. By processing parameters and geographic location data together, balanced and multi-criteria regions can be identified. It relates to a method and system for generating 10 different data from various sources. The resulting tables, columns, date filters, and value inputs are displayed via an interface. It enables the planning of geographical areas through their processing. In this context The method and system are based not only on spatial proximity criteria, but also on value. It also allows for zoning based on certain criteria. Thus, The acquired areas are directly applicable in field operations, balanced and repeatable. 15 It is made possible for it to be of a viable nature. The invention has a parameter-controlled and user-interactive structure for geographic zoning. in the process numerical values, spatial proximity criteria and balancing It offers a solution that considers the mechanisms together. Method and system; data 20 Filtering interface, date filter control mechanism, number of regions input field, clustering. This algorithm includes elements such as a balancing parameter and an optimization module. Thanks to the integrated work of the elements, geographical areas can be analyzed more accurately, balanced, and The operationally feasible division of the system is ensured. As a result, From logistics planning to sales organization, from the design of distribution networks to 25 It is applicable in a wide range of fields, including geographic data analysis. State of the Art The current technique involves clustering and regionalizing geographic data using various 30 methods. Methods have been developed. These methods are mostly distance or density based. It is working and only considers the geographical proximity of the data. However, this Because these approaches are based on a single criterion, many problems arise in real-world field applications. It is insufficient to meet the diverse needs, especially the unbalanced sales regions. the creation, efficient structuring of distribution networks and logistics operations 35 2 User-specific numerical values are also included in the process in areas such as efficient management. This needs to be done. Known methods do not take these value criteria into account. This leads to imbalances in practice. Among clustering algorithms, the K-means method is widely used in this field. 5 It is used. However, due to the random selection of starting centers, the clusters This leads to an imbalance. The K-means++ algorithm (enhanced clustering) (the algorithm) partially reduces this problem and makes the starting centers more balanced. It chooses K- for parallel processing of geographic data in patent number US10803096B2. A method based on the means++ algorithm is described. In this method, data is converted into a "tile" structure. It is divided and clustered in parallel processing units. This is how big data is generated. Speed and scalability are provided in processing the sets. However, the solution in question... It focuses solely on spatial proximity. User-specific value criteria, for example... because sales volume or order quantity was not integrated into the clustering process Operationally stable regions cannot be established. 15 Similarly, in patent US10437863B2, hierarchical clustering of geographic data is described. A solution has been presented regarding processing using this method. In this method, different "zoom" settings are used. Tile-based projections are used at these levels, with large clusters being divided into subsets. The data is being separated and more precisely classified. This approach allows data 20 It is useful in terms of scaling and managing border regions. However, only Since spatial location is taken into account, a balancing based on value criteria is possible. It is not. Also, border overlaps, gaps, or actual operational conditions The emergence of unsuitable areas cannot be resolved with this method. One of the known methods for determining the boundaries of geographical regions. It is the ConvexHull algorithm (convex hull method). This algorithm sorts points. It defines the convex boundary and forms a polygon around the clusters. However Due to its convex structure, it often covers unrealistically large areas. Thus, areas that are not used operationally are also included within the borders. 30 Alternatively, the ConcaveHull algorithm (concave enclosure method) is used. It can draw more realistic boundaries. However, the parametric nature of this method allows for gaps. This leads to the formation or overlapping of boundaries, especially in the field. Such overlaps in operations occur when the same area is included in multiple zones, or This leads to serious problems, such as some areas not being covered at all. 35 3 DBSCAN algorithm (Density-Based Spatial), one of the density-based methods (Clustering of Applications with Noise) is used for clustering geographic data. This method is used to create clusters based on density threshold values, and It excludes noise points. However, patent number US20150095336A1 states 5 As can be seen, these types of algorithms do not allow the user to predetermine the number of clusters. This does not result in clusters. Therefore, in low-density areas, clusters remain scattered. In high-density areas, excessive clustering occurs. The user's inability to intervene in the process, the regions are balanced and operationally feasible This prevents it from being created in this way. 10 There are other density-based approaches as well. For example, "Adaptive Spatial". In the patent application titled "Density Based Clustering," user movement data Clustering is performed by conducting density analysis. However, this type The solutions are more focused on behavioral analysis and 15 sales or distribution regions. It is not directly related to operational balance. User-specific value criteria. The failure to consider these methods in the process makes them inadequate for field operations. This causes it to remain there. Another shortcoming of known techniques is the absence of a parameter control mechanism. 20 The systems described in the existing documentation work even when data entries are incomplete or incorrect. It runs the algorithm and produces a result. For example, when no date filter is entered or The algorithm still works when table columns are left incomplete, but the resulting regions... It is not feasible under real operational conditions. Incorrect areas are being assigned, The workload is unevenly distributed and resources are used inefficiently. 25 In conclusion, when the methods and systems described in current techniques are examined, Geographic data is processed only according to a single criterion and in a limited way, user-specific value. where criteria are not taken into account, and problems such as border overlaps or gaps cannot be resolved. and it appears that there is no parameter validation mechanism. Therefore, 30 a new balanced, parametric, user-interactive design with a boundary improvement algorithm A method and system are needed. 4 Purpose of the Invention The aim of this invention is to improve the process of clustering and regionalizing geographic data. to address the shortcomings seen in current techniques and to provide the user with a balanced, reliable and A new method and system that provides operationally viable areas 5 It is about developing. One of the aims of the invention is to classify clusters not only according to the distance criterion, but also also balanced according to user-defined value criteria. The aim is to ensure its creation. In the structure developed to achieve this goal, there are 10 improved... The clustering algorithm (K-means++ based approach) more accurately identifies starting centers. In balanced selection, the balancing parameter (diffFactor) is the magnitude of the difference between the clusters. or reduces differences in value. Thus, both geometric proximity and functionality are improved. Regions are categorized simultaneously based on criteria such as cargo volume, order volume, or sales quantity. is obtained in a balanced manner. 15 Another objective is to address the unrealistic aspects that arise in known border demarcation methods. It is the elimination of areas and overlapping regions. The border used in this context Improvement algorithm (ImproveConvex) for convex or concave contouring. The shortcomings of the methods are being addressed, the overlapping of boundaries is prevented, and the gaps are filled. This prevents its occurrence. Thus, it can be applied directly in the field, without intersections, and Consistent geographical regions are being created. The invention also addresses potential omissions or errors that may occur in data entry. It aims to prevent the clustering process from being negatively affected. For this purpose, 25 The system uses data filtering interface and date filter validation steps. The algorithm is only allowed to work with accurate and complete datasets. The algorithm doesn't work if missing or incorrect parameters are provided, which is incorrect. This prevents the formation of regional groups and increases the reliability of the system. Another objective is to give the user direct control over the process. The user Thanks to its interface, parameters such as table, column, filter and region count can be adjusted step by step. These can be defined, so the process can be customized according to user needs. The structure enables the user to make operational decisions quickly, accurately, and repeatably. It provides an opportunity. 35 Finally, the aim of the invention is to bring all these features together in a single integrated system. by bringing together solutions that are scattered across known techniques, it is sufficient on its own. The goal is to eliminate the elements that are not separate. In this way, the invention is not merely the sum of its parts. No, it's a parameter-controlled, user-interactive, multi-criteria, and boundary-improvement method. 5 And as a system, it presents a new technical impact. The invention is designed to fulfill the above objectives by processing geographic datasets, user data... Parameters obtained through an interactive structure, geographical proximity and user-specific processing by considering the value criteria together, thus ensuring a balanced, non-conflicting and 10 It is a system that enables the creation of operationally viable areas, The data is presented in the form of geographic coordinates, value columns, and operational parameters. a database where user selections such as tables, columns, and dates are stored to be done and the data narrowed down according to the criteria determined before clustering a data filtering interface that allows 15 selected through the data filtering interface By performing integrity, type, and required field checks on the parameters, identifies errors or omissions. a parameter that prevents the algorithm from working if the inputs are present The verification module selects starting centers in a balanced way, taking into account geographical proximity and A clustering system that forms groups by considering user-specific value criteria together. The algorithm calculates the total workload or value criteria for each cluster, classifying them into 20 clusters. to reduce the differences between them to the tolerance range by redefining them in border regions A balancing algorithm that makes assignments to convexhull boundary polygons of clusters. or extracts using concavehull-based methods and eliminates gaps or overlaps a boundary improvement algorithm that makes it operationally feasible, 25 for clustering algorithm, balancing algorithm and boundary refinement algorithm centralized tolerance ranges, convergence criteria, and configuration options. a parameter management system that manages data in this way and provides the user with template-based selection options. and the configuration module enables the execution of algorithms during runtime. By recording the parameters used, tolerance values, and initial conditions, the same 30 that allows the same results to be obtained when run again under the conditions a data processing and computing unit requires the user to specify the number of tables, columns, filters, and regions, etc. enabling the parameters to be collected step by step and the process to be interactive. a user interface that allows management, mapping of clustering results visualizing the workload and value distributions for each region in tabular form. It includes a reporting interface that presents and reports the outputs in file formats. 35 6 Explanation of the Figures Figure 1 shows the user-interactive, multi-criteria and parameter-controlled geographical design that is the subject of the invention. This is a representative block diagram of the zoning system. Explanation of Part References 10. Database 20. Data Filtering Interface 30. Parameter Validation Module 10 40. Parameter Management and Configuration Module 50. Clustering Algorithm 60. Balancing Algorithm 70. Boundary Improvement Algorithm 80. Data Processing and Computing Unit 15 90. User Interface 100. Reporting Interface Detailed Description of the Invention 20 The invention involves combining geographic data points with user-defined numerical values. by processing and clustering them into balanced and operationally viable regions It provides a method and system that enables data acquisition, filtering, and validation, parametric configuration, clustering, balancing, boundary creation and 25 It integrates improvement and reporting steps in an unified manner. The system is started via the user interface (90) and the user can access the table, column, date, It allows the user to specify parameters such as the number of areas and regions. 30 documents with functional value such as delivery notes, delivery history, invoice history, sales quantity or sales amount marking the weight columns and indicating that clustering is based not only on positional proximity, It also ensures that it is done in a balanced way according to value criteria. The parameters are received by the parameter management and configuration module (40), Defaults, scaling coefficients, and iteration limits are defined, and the system... is being made consistent across the board. 35 7 The raw data to be processed is stored and accessed on the database (10). Data base (10); geographical coordinates, date information, weight / score columns, customer or It is structured to include fields such as delivery ID and geographic coordinates. preferably using a global reference system such as WGS84 or application-specific projection 5 It is stored in the system in a normalized form. However, the database (10) Access includes not only row-based reading, but also spatial indexing and range queries. It operates within a supportive structure. Thus, it can handle even large volumes of datasets. The effective preparation of the sub-clusters necessary for clustering is ensured. User-defined filters are set via the data filtering interface (20) is implemented. Data filtering interface (20); date, region, product group, channel type or In line with similar business practices, we narrow the data set and the scope of the analysis. It defines filters as reducing data volume and noise before clustering. It appears that this facilitates the algorithm reaching convergence faster. Filter 15 Data consistency after implementation, parameter validation module (30) is checked. Parameter validation module (30); presence of mandatory fields, value whether the time ranges are reasonable, and whether there are any cases where the date range is not defined. Checking if it is missing and whether the selected columns are compatible with the data types. The process is stopped when verification fails, and the user is charged 20 The algorithm is asked to correct the deficiencies. This stage involves correcting the errors or missing data. It prevents the formation of faulty zones by preventing their operation. The system processes and calculates structured and verified data using a data processing and computing unit (80) It is preparing to cluster on it. Data processing and computing unit (80); numerical 25 an optimized system in terms of stability, memory usage, and iteration performance It provides the execution layer. This layer supports multi-core processing and preferably By working with vectorized mathematical routines, even on large datasets It offers scalability. Data processing and computing unit (80), one for each harness. It generates a "study descriptor" and includes the parameters used, randomness seeds, and 30 By recording interim results, it ensures that experiments are reproducible. The clustering algorithm (50) is run during the clustering phase. Clustering algorithm (50) is based on k-means++ logic for the selection of starting centers using a launch method and increasing the spatial diversity of the starting centers 35 8 This approach determines the imbalance of the clusters from the very beginning. It prevents getting stuck and shortens the convergence time. Cluster assignment and In the central update steps, the metric includes not only geographical distance but also user data. will also include a balance term derived from the weight / value columns marked by It is defined in this way. For this purpose, a total cost function is defined for each data point. 5 This function is calculated and includes: (i) geographical distance component, (ii) value difference component and preferably (iii) a penalty component weighted by normalization coefficients This is how appointments are made, not just based on the "nearest center" logic, but also on the same This is done in a way that is consistent with the principle of "targeted balance" over time. After obtaining the clustering results, the balancing algorithm (60) is activated. The balancing algorithm (60) includes the total load and / or workload for each cluster. in calculating load indicators, inter-cluster differences with the target tolerance band comparing and local reassignments to reduce deviations outside of tolerance is carrying out these reassignments. These reassignments are located in border regions and involve more than one 15 including data points close to the cluster, meaningful in the total cost function. It is accepted when an improvement is achieved. The balancing algorithm (60), iteration updating cluster centers throughout or making corrections only at the assignment level. It does so. Convergence occurs either when the tolerance band is satisfied or when the greatest improvement is achieved. The step is completed when the value falls below the predetermined threshold. Thus, 20 Clusters are categorized not only based on geometric proximity but also in terms of workload and value criteria. It is becoming balanced. Boundary improvement algorithm for generating boundary polygons for balanced clusters (70) This is carried out by [company name]. In this stage, a basic polygon is first created for each cluster. 25 The inference is made as follows: concave enclosure, convex enclosure or according to the data distribution. A hybrid approach is chosen. Boundary improvement algorithm (70), gaps and narrow It smooths out the sharp curves that form the throats, and smooths out very small protrusions, and It eliminates potential overlaps between neighboring clusters. The risk of intersection... The boundary is set at 30°, taking into account the load and proximity information of neighboring clusters along its edges. The line is being repositioned. Conflict resolutions are carried out according to a specific priority rule. The managed and preferred solution is one that reduces the total cost function and clusters. It is an alternative that does not disrupt the balance between them. Ultimately, it is applicable in the field. Non-intersecting and visually consistent polygons are obtained. 35 9 Throughout this process, the parameter management and configuration module (40); tolerance bands, iteration numbers, convergence criteria, metric weights, and boundary refinement It centrally manages settings such as parameters. Parameter management and configuration module (40), different configuration templates according to user profile can define and quickly select the preferred template via the user interface (90) 5 It can be implemented. At the same time, the results obtained with the selected parameter sets... Summary statistics and quality measures per harness for comparison of results. It is produced. Evaluation and distribution of results reporting interface (100) out of 10 is being implemented. Reporting interface (100); clusters on the map layer and Visualizing boundary polygons, the total workload, total value, and points for each region. It presents performance indicators such as number of trips, average distance, and similar data in tabular form. The reporting interface (100) allows exporting outputs in various formats; Vector boundary files, summary report PDFs, and machine-readable JSON / CSV 15 It produces the outputs. The reporting interface (100) logs the parameter log of the run and It also includes the study descriptor in the reports, so that the results can be tracked and This ensures that it is repeatable. In an alternative application form, service only 20 without user interface (90) An application programming interface that operates through calls can be used. This In this case, the parameters are managed by the parameter management and configuration module (40). The validations are sent via the provided endpoints and parameter validation is performed. This is done automatically by module (30). Similarly, the reporting interface (100) 25 Integration only with data services or file outputs without It is possible. In such tangible representations, visual output production is not mandatory, It is sufficient to write the boundary polygons and area summaries to the file system. The system includes protective measures for extreme cases related to data quality. Outlier When the value density is high, the distance metric in the clustering algorithm (50) is 30 It applies flexible scaling in a way that limits the impact of outliers. Excessive Adjustment algorithm for unbalanced distributions (60), from border regions By making gradual reassignments, it ensures that the tolerance band is approached. Data leakage or incompatible data types parameter validation module (30) process to stop and report the error via the user interface (90) and suggest improvements It offers. Data processing and computation are required for the application to work on large-scale datasets. unit (80); fragmented execution based on region, time period or organizational hierarchy 5 It supports this. In this way, data is processed in parallel in the form of independent work units. The results are processed and integrated on the final results reporting interface (100). Boundary improvement algorithm (70) to ensure that the boundaries do not overlap during integration guaranteed by and potential overlays are automatically corrected by rules. is being eliminated. 10 In conclusion, the system consists of a database (10), a data filtering interface (20), and parameter validation. module (30), parameter management and configuration module (40), clustering algorithm (50), balancing algorithm (60), boundary improvement algorithm (70), data processing and 15 elements of calculation unit (80), user interface (90) and reporting interface (100) Its integrated operation is balanced, seamless, and operationally feasible. It produces zoning results. This integrated structure is based on data known only individually. It is not the sum of methods, but data quality assurances, multi-criteria balancing and limits It offers a new solution that creates a technical impact through improvement steps.
Claims
11 REQUESTS 1. Geographic datasets, with parameters obtained through a user-interactive structure, Processing by taking into account both geographical proximity and user-specific value criteria. and thus create balanced, non-overlapping and operationally viable zones. 5 It is a system that enables its creation, and its characteristic is; data includes geographic coordinates, value columns, and operational parameters a database in which it is stored (10), The user makes selections such as tables, columns, and dates, and the data is displayed in 10 a cluster that allows for narrowing down the clusters according to criteria determined beforehand data filtering interface (20), integrity, type and of the parameters selected via the data filtering interface (20) by performing mandatory field checks and finding erroneous or incomplete entries. a parameter validation module that prevents the algorithm from working (30), 15 By selecting starting points in a balanced way, considering geographical proximity and the user a clustering that forms clusters by considering specific value criteria together algorithm (50), Clusters by calculating the total workload or value criteria for each cluster. to reduce the differences between them to the tolerance range in border regions 20 a re-assignment algorithm (60), Boundary polygons of clusters can be determined using convexhull or concavehull-based methods by removing and eliminating gaps or conflicts, thus providing operational benefits. a boundary improvement algorithm that makes it feasible (70), clustering algorithm (50), balancing algorithm (60) and boundary improvement 25 tolerance ranges, convergence criteria and configuration for the algorithm (70) centrally manages options and provides the user with template-based selection possibilities. a parameter management and configuration module (40), the processes used during execution that enable the execution of algorithms By recording the parameters, tolerance values, and initial conditions, the same 30 When run again under the same conditions, it is possible to obtain the same results. a data processing and computing unit that recognizes (80), Step-by-step configuration of parameters such as the number of tables, columns, filters, and regions from the user. a system that enables the acquisition of data and allows the process to be managed interactively. User interface (90), 35 12 Visualizing clustering results on a map, workload for each region. and presenting the value distributions in tabular form and the outputs in file formats. a reporting interface (100), It includes. 5 2. According to claim 1, the system is characterized by the following features: located on the said database (10). Probability distribution of cluster centers from data points based on k-means++ method. by selecting clusters that are far apart from each other, and in this way, between the clusters It includes a clustering algorithm (50) that increases spatial diversity. 10 3. It is a system according to claim 1, and its characteristic is the workload or value generated between sets. their differences through the reassignment of data points located in border regions a balancing algorithm that enables the reduction to the specified tolerance range (60) It includes. 15 4. It is a system according to claim 1, and its characteristic is that the boundary polygons of the sets are convexhull or extraction of the polygons using concavehull methods, voids and eliminating conflicts and establishing operationally viable limits. It includes a boundary improvement algorithm (70) that enables its transformation. 20 5. The system, according to claim 1, has the following characteristic: the parameter used during operation. by recording the sets, tolerance values and initial conditions, the same conditions can be maintained. data that allows the same results to be obtained when run again under the same conditions. It contains a processing and calculation unit (80). 25 6. The system is defined in Claim 1 as follows: it allows the user to specify the number of tables, columns, filters, and regions. It allows parameters such as these to be collected step by step, manages the process, and It includes a user interface (90) that enables user interaction.
7. It is a system according to claim 1, and its characteristic is that the clustering results are shown on a map. visualization, tabular distribution of workload and value for each region. It enables the presentation of files and the export of their outputs in different file formats. It includes a reporting interface (100). 35 13 8. Balanced processing of geographic datasets by combining them with user-specific value criteria. It is a method that enables the creation of regions, and its characteristic feature is; Parameters are entered by the user via the user interface (90), The integrity of the entered parameters is checked by the parameter validation module (30), 5 Checking in terms of type and required field, by clustering algorithm (50) using verified parameters Selection of cluster starting points and geographical proximity of data points Assigning groups to clusters based on value criteria, Algorithm for balancing workload or value differences between clusters 10 (60) by reassigning data points in border regions elimination, Extraction of boundaries belonging to sets using convexhull or concavehull doing it with methods and the boundary improvement algorithm of the boundaries in question Correction of gaps and overlaps by (70), 15 The parameter sets used during the execution of algorithms, tolerance data processing and calculation unit (80) of values and initial conditions recorded by, through the reporting interface (100) of the obtained clustering results 20 visualization on the map and outputs in table or file format export, It includes the steps of the process.
9. The method according to claim 8, its characteristic is; clustering of cluster starting centers 25 probability distribution based on k-means++ method by algorithm (50) The selection process involves a step.
10. This method, according to claim 8, is characterized by its ability to account for differences in workload between clusters. Tolerance 30 by using diffFactor parameter by balancing algorithm (60) It involves the step of reducing it to the range.
11. This method, according to claim 8, is characterized by the fact that the boundaries of the sets are only concavehull. extraction by using the method and boundary improvement algorithm (70) This includes the process step of resolving the conflicts. 35 14 12. It is a method according to claim 8, and its characteristic is that it is used in the execution of algorithms. parameter sets, initial conditions and tolerance values data processing and to be recorded by the calculation unit (80) and re-recorded under the same conditions The process involves obtaining the same result when run. 5 13. This method, according to claim 8, is characterized by its reporting interface for clustering results. (100) includes the export process step only in table format.