A method and system for collaborative optimization of interconnected discrete quantities based on spatial topological constraints

Abstract

This invention relates to the field of transportation information technology, and particularly to a method and system for coordinated traffic optimization of interchanges based on spatial topological constraints. The method includes the following steps: acquiring road segment connection data of the interchange, and interpreting the ring-ramp structure based on the road segment connection data to obtain closed-loop topology skeleton data; expanding the cyclic load-bearing chain based on the closed-loop topology skeleton data to obtain intra-ring flow path data; performing retention coupling processing based on the intra-ring flow path data to obtain cyclic retention potential field data; locating amplification sources based on the cyclic retention potential field data to obtain self-reinforcing core area data; generating topology discharge collaboratively based on the self-reinforcing core area data to obtain ring-ramp control data; and performing cyclic suppression verification based on the ring-ramp control data to obtain closed-loop optimization evaluation data. This invention achieves the identification and suppression of traffic cyclic retention in ring ramps, improving traffic efficiency and operational stability in interchange scenarios.

Description

Technical Field

[0001] This invention relates to the field of information technology in transportation, and in particular to a method and system for collaborative optimization of interconnected discrete quantities based on spatial topology constraints. Background Technology

[0002] As urban transportation networks continue to expand, interchanges, as crucial connecting nodes between urban expressways and highways, are becoming increasingly complex in structure. This is especially true for interchanges with multi-level ring ramps and interwoven multi-path structures, which are prone to problems such as difficulties in traffic flow organization, amplified local congestion, and decreased traffic efficiency during peak hours. Current technologies for analyzing and optimizing traffic flow at interchanges rely on segment-level or node-level flow statistics methods, adjusting through signal timing, entrance control, or local flow restrictions. While these methods can alleviate local congestion to some extent, they often lack the ability to model the overall complex topology of the interchange. In actual operation, the closed-loop nature of ring ramp structures easily leads to repeated vehicle back-and-forth and accumulated traffic flow. When multiple paths overlap in the same segment, the cumulative effect of traffic flow causes a rapid increase in local load and creates coupling effects between paths, causing congestion to propagate and even amplify along the closed-loop structure. Existing methods treat each path as a relatively independent traffic unit, making it difficult to represent the interaction between different paths and their impact on the overall traffic state, resulting in inaccurate identification of the congestion evolution process. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention proposes a method and system for collaborative optimization of interconnected discrete quantities based on spatial topological constraints, thereby resolving at least one of the aforementioned technical problems.

[0004] This application provides a method for collaborative optimization of interconnected discrete quantities based on spatial topological constraints, the method comprising: Obtain the road segment connection data of the interchange, and interpret the ring-ramp structure based on the road segment connection data of the interchange to obtain the closed-loop topology skeleton data; Based on the closed-loop topology skeleton data, the cyclic bearing chain is expanded to obtain the internal circulation path data; based on the internal circulation path data, the cyclic stagnant coupling processing is performed to obtain the cyclic stagnant potential field data. The amplification source is located based on the cyclic stagnant potential field data to obtain the self-reinforcing core region data; topology discharge is collaboratively generated based on the self-reinforcing core region data to obtain the ring-turn control data. Based on the loop control data, cyclic suppression verification is performed to obtain closed-loop optimization evaluation data.

[0005] This invention utilizes spatial topology modeling of interchange networks and employs cyclic load-bearing chain expansion and retention coupling to represent the multi-path interweaving relationships of traffic flow within ring ramps. Compared to traditional analysis methods based solely on single road segments or local nodes, this method jointly represents the cyclic retention and propagation characteristics of traffic flow in closed-loop structures at both the path and topology levels. By constructing a cyclic retention potential field, discrete traffic flow states are transformed into a continuous distribution, thereby improving the accuracy of identifying congestion evolution trends. The amplification source localization mechanism accurately identifies key areas leading to traffic self-amplification, and combined with topological discharge co-generation, it achieves coordinated control of injection, pressure-bearing, and discharge paths, effectively suppressing traffic accumulation and amplification effects in closed-loop structures.

[0006] Optionally, the step of interpreting the ring-ramp structure based on the interchange segment connection data to obtain closed-loop topology skeleton data includes: A directed road network is constructed based on the road segment connection data to obtain the initial topology map data; Closed-loop path data is obtained by performing closed-loop path retrieval on the initial topology map data; Based on the closed-loop path data, loop-turn structure constraint screening is performed to obtain valid closed-loop structure data; Node-edge merging is performed on the effective closed-loop structure data to obtain the closed-loop topology skeleton data.

[0007] This invention constructs a directed road network from road segment connection data and performs closed-loop path retrieval and structural constraint screening, achieving accurate identification and effective extraction of ring ramp structures in interchanges. Compared to methods that rely solely on geometric shapes or empirical rules for loop identification, this scheme utilizes directed topological representation to clarify the traffic relationships and directional constraints between road segments, preventing reverse or unreachable paths from being misjudged as closed-loop structures. Simultaneously, through ring ramp structure constraint screening, closed-loop paths are filtered for continuity and hierarchical accessibility consistency, effectively eliminating pseudo-closed-loop paths with unreasonable structures or actual impassability. Node-edge merging processing provides a skeletonized representation of the closed-loop structure, reducing the complexity of the original road network data and making the loop path unfolding and traffic modeling processes more efficient and stable.

[0008] Optionally, the step of expanding the cyclic bearer chain based on the closed-loop topology skeleton data to obtain the intra-loop flow path data includes: Based on the closed-loop topology skeleton data, the entry node inside the loop is identified to obtain the loop start node data; The data of the starting node of the loop is expanded by directed path recursion to obtain the flow path sequence data; By performing backtracking behavior identification on the flow path sequence data, loop path identification data is obtained; Based on the cyclic path identifier data, the path carrying capacity is associated and mapped to obtain the path carrying attribute data; The path carrying attribute data is compiled into a path set to obtain the internal loop flow path data.

[0009] This invention achieves a structured and path-level representation of traffic flow within a ring ramp by expanding the closed-loop topology skeleton into a cyclic carrying capacity chain. Compared to analysis methods based solely on single paths or local road segments, identifying the entry node clearly defines the starting position of traffic entering the closed-loop structure, providing clear source localization for flow analysis. Through directed path recursive expansion, it can comprehensively describe the multi-path propagation process of traffic within the closed loop, and effectively identify repeated flows and cyclic paths through return behavior discrimination, thereby revealing potential traffic stagnation and backflow characteristics within the closed-loop structure. By mapping path carrying capacity, path structure information is combined with capacity and flow efficiency, expanding path representation from a single topological relationship to a carrying capacity attribute with physical meaning. The compilation of path sets into a unified intra-ring flow path data structure reduces redundancy in path representation and enhances the comparability and correlation between different paths.

[0010] Optionally, the step of performing retention coupling processing based on the intra-loop flow path data to obtain cyclic retention potential field data includes: Based on the flow path data within the ring, the overlapping area is identified to obtain the overlapping area data; Traffic aggregation calculations are performed on the overlapping occupancy area data to obtain the segment load data; Perform time-series retention analysis on the section load data to obtain section retention data; Nonlinear interactive interferometry is performed based on the segment retention data to obtain coupled retention data; Spatially continuous mapping is performed based on coupled stagnant data to obtain cyclic stagnant potential field data.

[0011] This invention achieves a representation of segment load and stagnation status by identifying path overlap, superimposing flow, and performing temporal stagnation analysis on the flow path data within the loop. Through nonlinear interferometry, the interaction relationships between different paths within the same overlapping segment are modeled in a unified manner, enabling the coupling effects between multiple paths to go beyond simple superposition and reflect enhancement or suppression effects under complex interleaving conditions. By using spatial continuous mapping, the coupling stagnation results of discrete segments are transformed into a continuous potential field expression along the closed-loop topology, which not only improves the spatial representation ability of stagnation distribution but also enhances the accuracy of identifying local high-stagnation regions and their evolution trends.

[0012] Optionally, the step of performing nonlinear interferometric measurement based on the segment retention data to obtain coupled retention data includes: Nonlinear mapping processing is performed on the segment retention data to obtain nonlinear feature space data; Interferometric construction is performed on nonlinear feature space data to obtain path interference data; Interference measurement is performed based on path interference data to obtain coupled stagnant data.

[0013] This invention transforms the original stagnation information from a linear superposition model to a nonlinear expression in a multidimensional feature space by performing nonlinear mapping processing on the segment stagnation data and constructing interactive interference relationships. This makes the potential complex relationships between different paths explicit. Compared with traditional analysis methods based on simple correlation or statistical superposition, the interactive interference construction can represent the interaction strength and changing trend between paths under different stagnation states, thus more accurately reflecting the coupling effect under multi-path interleaving conditions. By measuring the interference of path interference data, a quantitative characterization of the degree of interaction between paths is achieved, making the coupled stagnation data not only structurally correlated but also comparable and computable.

[0014] Optionally, the step of constructing path interference data by cross-interference of the nonlinear feature space data includes: Path feature phase alignment is performed based on nonlinear feature space data to obtain aligned feature data; Interactive perturbation injection is performed on the alignment feature data to obtain perturbation injection data; Nonlinear expansion is performed on the injected disturbance data to obtain nonlinear interferometric data; Interference stability screening is performed on the nonlinear interferometric data to obtain effective interferometric data; Based on the effective interferometric data, the interferometric relationships are organized to obtain path interferometric data.

[0015] In this invention, phase alignment, perturbation injection, and nonlinear expansion in a nonlinear feature space are used to represent the interaction relationships between different path dwell states. Phase alignment of path features effectively eliminates the misalignment effects of different paths in the time dimension. Interactive perturbation injection introduces the state changes of one path into the response analysis of another path, transforming the passive observation of potential influence relationships between paths into active construction, thereby improving the sensitivity and interpretability of interaction relationship identification. Nonlinear expansion analyzes the feature changes before and after the perturbation in a high-dimensional space, helping to reveal nonlinear interaction effects under complex conditions. Interference stability screening eliminates spurious interference relationships caused by randomness or short-term fluctuations, retaining effective interference with continuity and consistency. Path interference data is formed by organizing the interference relationships.

[0016] Optionally, the step of locating the amplified source based on the cyclic stagnant potential field data to obtain the self-reinforcing core region data includes: Local high-value regions are extracted based on the cyclic stagnant potential field data to obtain preliminary stagnant region data; The preliminary data on the remaining areas were continuously screened to obtain data on the stable remaining areas. Back-coupling analysis was performed on the data from the stable retention zone to obtain the data from the feedback enhancement zone. The amplification effect is determined based on the feedback enhancement region data, and the self-enhancing region data is obtained. The self-enhancing region data is filtered for core regions to obtain the self-enhancing core region data.

[0017] In this invention, through data analysis and layer-by-layer screening based on the cyclic stagnation potential field, this step achieves accurate identification and hierarchical positioning of traffic amplification sources in a closed-loop structure. Local high-value area extraction quickly identifies regions with significant stagnation intensity; continuous screening effectively eliminates unstable regions caused by short-term fluctuations or instantaneous congestion, making the identification results more reliable; back-coupling analysis models the correlation between stagnation regions and path back-feedback behavior, revealing the intrinsic connection between stagnation accumulation and cyclic feedback at the structural level; amplification effect discrimination identifies key regions with a continuously increasing trend, thus distinguishing between ordinary stagnation regions and self-amplifying regions with amplification potential; and core area screening extracts the set of nodes with the greatest impact on overall traffic efficiency.

[0018] Optionally, the step of generating ring-turn control data by topology discharge coordination based on self-enhancing core region data includes: Based on the self-enhancing core area data, the associated path backtracking is performed to obtain the core associated path data; The core associated path data is partitioned into path functions to obtain topology control partition data; Based on the topology control partition data, the partition control amount is generated to obtain local control parameter data; The local control parameter data are topologically coordinated to obtain the ring-turn control data.

[0019] This invention achieves an effective transformation from "problem identification" to "structured control" by using a self-reinforcing core area as the center for path backtracking and functional partitioning. Path backtracking clarifies the topological relationships between the core area and upstream injection paths, intra-ring pressure-bearing paths, and downstream discharge paths, giving the control range a clear structural boundary. Functional path partitioning differentiates paths with different function types, effectively avoiding efficiency losses caused by a uniform control strategy. Partition control quantity generation transforms structural information into specific parameter expressions, making the control process executable and quantifiable. Topological location collaborative orchestration combines the control parameters of each partition according to closed-loop topological relationships, ensuring the control process conforms to the flow propagation direction and structural dependencies.

[0020] Optionally, the step of performing cyclic suppression verification based on loop control data to obtain closed-loop optimization evaluation data includes: Based on the control data of the ring-turn control, control mapping is applied to obtain the verification input state data; The input status data for verification is cyclically processed and simulated to obtain the verification flow data; The cycle strength data is obtained by evaluating the return intensity of the verification and circulation results data. Based on the cyclic intensity data, the retention change is judged to obtain closed-loop optimization evaluation data.

[0021] This invention achieves closed-loop verification and quantitative evaluation of the control strategy's effectiveness by mapping loop control data to corresponding paths and performing cyclic flow simulation. Through control mapping loading, various control parameters can form calculable input states within the topology; through cyclic flow simulation, the evolution of flow under control within the closed-loop structure can be dynamically simulated, thus realistically reflecting the flow state after control; through return intensity evaluation, the repeated flow behavior in the closed loop is quantitatively represented, effectively identifying the degree of change in the cyclic effect; and through retention change discrimination, the retention distribution and duration before and after control are compared and analyzed, enabling an intuitive judgment of the cyclic suppression effect.

[0022] Optionally, this application also provides a spatial topology-constrained interconnection variable collaborative optimization system for executing the spatial topology-constrained interconnection variable collaborative optimization method described above, wherein the spatial topology-constrained interconnection variable collaborative optimization system includes: The closed-loop topology construction module is used to acquire the road segment connection data of the interchange and interpret the ring-ramp structure based on the road segment connection data of the interchange to obtain the closed-loop topology skeleton data. The cyclic flow modeling module is used to expand the cyclic bearing chain based on the closed-loop topology skeleton data to obtain the flow path data within the loop; and to perform stagnation coupling processing based on the flow path data within the loop to obtain the cyclic stagnation potential field data. The amplification source analysis and collaborative control module is used to locate the amplification source based on the cyclic stagnant potential field data to obtain the self-reinforcing core region data; and to perform topology discharge collaborative generation based on the self-reinforcing core region data to obtain the annular control data. The cyclic suppression evaluation module is used to perform cyclic suppression verification based on the loop control data to obtain closed-loop optimization evaluation data.

[0023] The purpose of this invention is to use interchange segment connection data as a basis, construct a closed-loop topological skeleton through ring-turn structure interpretation, and realize a structured expression of traffic relationships and path connectivity in interchanges; through the expansion of cyclic bearing chains, the traffic flow in the closed-loop structure is elevated from a single segment perspective to a holistic representation of the multi-path flow process; and combined with the stagnation coupling processing mechanism, the superposition of traffic flow, temporal stagnation and nonlinear interaction relationships of different paths in overlapping segments are uniformly modeled to construct a continuously distributed cyclic stagnation potential field, thereby realizing a spatial expression of the stagnation evolution process. By amplifying the source localization step, the high retention region in the potential field is correlated with the path return characteristics to identify the self-reinforcing core region with a continuously enhancing trend. Through topological discharge collaborative generation, the associated path of the core region is divided into different action areas of injection, pressure bearing and discharge, and corresponding control parameters are generated and topological-level collaborative orchestration is performed to realize the linkage control of flow input, retention diffusion and discharge release. Through cyclic suppression verification, the control results are deduced and the return intensity is evaluated to realize the quantitative verification of the degree of cyclic effect suppression in the closed-loop structure. Attached Figure Description

[0024] Other features, objects, and advantages of this application will become more apparent from the following detailed description of the non-limiting embodiments, taken with reference to the accompanying drawings: Figure 1 A flowchart illustrating the steps of an embodiment of a collaborative optimization method for interconnected discrete quantities based on spatial topology constraints is shown. Figure 2 A flowchart illustrating the steps of a method for interpreting a ring-turn structure according to an embodiment is shown; Figure 3 A flowchart illustrating the steps of a cyclic support chain unfolding method according to an embodiment is shown; Figure 4 A flowchart illustrating the steps of an embodiment of a method for locating an amplification source is shown. Figure 5 A flowchart illustrating the steps of a cycle suppression verification method according to one embodiment is shown. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0025] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0026] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. Functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0027] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0028] Please see Figures 1 to 5 This application provides a method for collaborative optimization of interconnected discrete quantities based on spatial topological constraints, the method comprising: S1. Obtain the road segment connection data of the interchange, and interpret the ring-ramp structure based on the road segment connection data of the interchange to obtain the closed-loop topology skeleton data. In one embodiment, the system acquires Building Information Model (BIM) data and Geographic Information Network (GIS) data corresponding to the interchange. The BIM data characterizes the three-dimensional structural features of the ramps, including height information, longitudinal slope, and hierarchical relationships between different ramps. The GIS data characterizes the connectivity between roads and their spatial distribution. The system performs spatial registration processing on these two types of data under a unified coordinate system, enabling each ramp component in the BIM to be mapped to nodes and connections in the road network structure, thereby constructing a unified topology containing directional attributes. Based on the constructed directed topology, the system traverses and analyzes paths that satisfy driving direction constraints, identifies path combinations that can form closed-loop connections, and treats these as loop structures. Combining the hierarchical and structural constraint information in the BIM, the system verifies the drivability of candidate paths, eliminating paths with cross-layer breaks or those that are practically impassable, thus obtaining closed-loop topology skeleton data reflecting the internal loop-ramp connections of the interchange.

[0029] S2. Expand the cyclic bearing chain based on the closed-loop topology skeleton data to obtain the internal circulation path data; perform retention coupling processing based on the internal circulation path data to obtain the cyclic retention potential field data. In one embodiment, the system identifies entrance nodes connected to external roads in the geographic information road network based on closed-loop topology skeleton data. Using each entrance node as a starting point, the system expands the inner-loop path segment by segment along directed connections. During path expansion, the system adjusts the capacity of each path segment by incorporating ramp length, curvature characteristics, and speed limit information provided in the building information model, ensuring that each path forms a carrying capacity chain containing both structural and carrying capacity attributes. The system also performs repeated access detection on nodes within the path. When the path passes the same node again during expansion, the corresponding segment is marked as a return segment, and its location and range information are recorded, thus forming inner-loop flow path data containing path sequence, return identifier, and carrying capacity information.

[0030] The system acquires real-time or historical traffic flow information for corresponding road segments from geographic information data and maps this flow information to corresponding segments of each loop's flow path. Alternatively, the system selects structurally similar constructed road segments from a historical database as reference samples based on road grade (e.g., arterial roads, ramps), design speed, and number of lanes, and extracts their time-period average flow as a benchmark value. Simultaneously, it adjusts this benchmark flow by combining planned traffic demand data or traffic prediction model results. For example, it uses 50% to 80% of the design capacity as the initial flow range, taking the upper limit during peak hours (e.g., 1500 to 2000 vehicles per lane per hour) and the median value during off-peak hours (e.g., 600 to 1200 vehicles per lane per hour). Traffic input is generated when no actual measurement data is available. For segments shared by multiple paths, the system superimposes their flow to obtain the segment's load level and, combined with the segment's capacity, determines whether traffic congestion exists. When the load on a section exceeds its capacity for a continuous period of time, the system records the congestion intensity and duration characteristics of that section. For example, the system compares the load and capacity of each section in fixed time steps (e.g., 30 seconds or 60 seconds). When the load exceeds the capacity, the system records the overload amount for that time step as the baseline value for congestion intensity. The system then accumulates or averages the overload amounts across consecutive time steps to obtain the congestion intensity characterization value for that section. The system counts the number of consecutive time steps in which the section is in an overloaded state. When the continuous overload time reaches a preset threshold (e.g., 3 consecutive time steps), it is determined to be a valid congestion, and the duration is recorded as the congestion duration. If the load recovers below capacity even once, it is considered a congestion interruption and the count restarts, thus forming a clear record of congestion time segments. The system performs uniform-scale feature mapping on stagnation characteristics, converting discrete stagnation information into comparable feature representations. It also constructs inter-path influence relationships based on shared segment relationships. For example, when two or more paths contain the same segment, the system treats that segment as a coupling node and establishes connections between paths participating in that segment. Influence weights are determined based on the traffic share in the shared segment. For instance, if a path's traffic share in a segment exceeds 50%, it is considered to have a dominant influence on that segment and other paths; if the share is between 20% and 50%, it is considered to have a moderate influence; and if it is below 20%, it is considered to have a weak influence. The system spatially integrates the stagnation characteristics of each segment along the closed-loop topology, transforming discretely distributed stagnation information into continuously distributed cyclic stagnation potential field data within the loop.

[0031] S3. Based on the cyclic stagnant potential field data, the amplification source is located to obtain the self-reinforcing core region data; based on the self-reinforcing core region data, topology discharge is collaboratively generated to obtain the ring-turn control data; In one embodiment, the system analyzes the retention distribution of each segment based on cyclic retention potential field data. By comparing and identifying local retention intensities, it extracts peak regions where the retention level is significantly higher than that of surrounding segments. The system then tracks the retention changes in these regions over multiple consecutive time windows to screen out stable retention areas with continuous retention. Combining the return marker information in the loop flow path, the system identifies segments located at the front of the return path whose retention changes can trigger increased retention in subsequent segments, and determines these segments as reinforcement zones with feedback effects. When such segments show a gradually increasing retention level over a continuous time range, they are identified as self-reinforcing regions. The system filters self-reinforcing regions based on their impact range and duration. Impact range is characterized by the number of consecutive segments covered and the number of associated paths. For example, a region covering at least four consecutive segments and having three or more impact paths is considered a large-scale impact region. Duration is measured by the length of time the region maintains an amplifying trend within a continuous time window. For example, a region lasting five or more time windows is considered a long-lasting region. The system retains only regions that simultaneously meet the following criteria: the number of impact segments reaches a preset lower limit (e.g., ≥4) and the duration reaches a preset threshold (e.g., ≥5 time windows) as self-reinforcing core regions. Regions with small impact ranges or short durations are eliminated or downgraded. The system obtains self-reinforcing core region data based on the above.

[0032] Centered on the self-enhancing core area, the system performs backtracking analysis on its related paths based on the topology of the geographic information road network, determines the upstream and downstream paths associated with it, and judges the adjustability of each path by combining the structural information in the building information model, including the range of traffic capacity changes and diversion conditions. Based on the path's position in the topology, the system divides relevant paths into injection zones, pressure zones, and discharge zones. (The system identifies all path segments leading to the self-reinforcing core area. A path segment is classified as an injection zone when its downstream nodes can directly reach the core area or through no more than two segments, and the flow direction is towards the core area. A segment is classified as a pressure zone when it is within the core area's coverage area or its topological distance from a core area segment is no more than one segment, and the segment exhibits continuous stagnation (e.g., load exceeding capacity for three consecutive time windows). A segment is classified as a discharge zone when its upstream nodes are located in the core area or pressure zone, and the flow direction is towards the outside of the closed loop or a low-load area.) Corresponding control strategy parameters are generated for each zone. For the injection zone, inbound flow is limited; for the pressure zone, the flow rhythm is adjusted to reduce load; and for the discharge zone, the flow priority is increased to accelerate drainage. Specifically, for the injection zone, the system determines the flow based on the difference between the current inbound flow and the segment's capacity. The system generates inlet flow restriction parameters. For example, when the inbound flow exceeds the capacity by more than 10%, the inbound flow will be restricted to within 90% of the capacity. For pressure zones, the system generates pressure release rhythm parameters based on the duration of congestion and the length of the section. For example, when the congestion duration exceeds four time windows, the release interval for that section will be set to be once every 1 to 2 time steps. For release zones, the system generates priority parameters based on the remaining capacity of the downstream section and the connection strength with the core area. For example, when the downstream load is less than 70% of its capacity, the release priority of that section will be increased, allowing it to receive traffic from upstream first. The injection zone refers to the upstream section that is directly connected to the self-reinforcing core area and whose traffic flow direction is towards the core area. The pressure zone refers to the section located inside or near the core area that bears the function of traffic congestion and transmission and is the main area bearing the cyclic backlog. The release zone refers to the downstream section located after the core area that is connected to the external road network or low-load area and bears the function of traffic discharge and diversion. The system coordinates and arranges various control parameters according to the order of the path in the closed-loop structure, so that the overall control strategy is consistent with the traffic flow direction, thereby forming loop control data.

[0033] S4. Based on the loop control data, perform loop suppression verification to obtain closed-loop optimization evaluation data.

[0034] In one embodiment, the system loads various control parameters from the ring-loop control data into a geographic information traffic simulation model to simulate and analyze the traffic flow evolution process after the control is implemented. During the simulation, the system continuously records and statistically analyzes the traffic flow changes, vehicle dwell time, and path return status of each segment, forming post-control status data. The system compares and analyzes key indicators before and after control, including features such as return intensity and dwell time. The return intensity is defined as the ratio of the number of path returns to the total number of turns within a unit of time, combined with the repeated occupation of the corresponding segment, and is corrected by adding the degree of repeated occupation of the segment. For example, if a path has a total of 100 turns in an assessment period, and 25 of them involve return behavior, its basic return intensity can be expressed as 25%. If the number of times the key segment corresponding to the path is repeatedly occupied in the same period exceeds the average level (e.g., higher than 20% of the network average), the return intensity is adjusted upward (e.g., increased by 5% to 10%). When the return intensity after regulation is lower than the pre-regulation level, and the range and persistence of high-stagnation zones both decrease, the current regulation strategy is deemed to have an inhibitory effect on cyclical stagnation. The system assesses the overall operational status by combining changes in traffic efficiency. For example, the system calculates the average travel time of major paths before and after regulation; when the average travel time decreases by more than 10%, traffic efficiency is considered improved. It also calculates the average stagnation time of sections; when the stagnation time decreases by more than 15%, the operational status is considered improved. Furthermore, it assesses the number of vehicles passing through per unit time (e.g., vehicles per minute); when this indicator increases by more than 5% to 10%, overall traffic capacity is enhanced. Based on these indicators, the system determines the overall operational status as optimized when at least two indicators simultaneously reach a preset improvement threshold; if only some indicators improve, it is considered partially optimized; if the indicators show no significant change or deteriorate, it is considered unoptimized. The output includes closed-loop optimization assessment data, including the decrease in cycle intensity and the degree of improvement in traffic efficiency.

[0035] Optionally, the step of interpreting the ring-ramp structure based on the interchange segment connection data to obtain closed-loop topology skeleton data includes: S11. Construct a directed road network based on the road segment connection data to obtain the initial topology map data; In one embodiment, the system acquires road segment connection data, which includes at least the road segment number, starting node identifier, ending node identifier, traffic direction, and hierarchical information. The hierarchical information may include the structural layer number of the road segment, the relative elevation range, and the overpass / underpass relationship identifier. The structural layer number distinguishes whether the road segment is located at ground level, overpass level, or underpass level; for example, the ground mainline can be marked as the first layer, the overpass ramp as the second layer, and the underpass as the negative first layer. The relative elevation range indicates the spatial height range of the road segment, such as recording the elevation values ​​corresponding to the starting and ending points of the road segment. The overpass / underpass relationship identifier indicates whether the road segment has an overpass, underpass, or same-layer connection with intersecting road segments. The system constructs a node set based on each node identifier and an edge set based on the road segment connection relationships, thereby forming a road network topology with directional attributes. For bidirectional road segments, the system splits them into two directed connections with opposite directions. During the construction process, the system adds corresponding attribute information to each connection relationship, including road segment length, design speed, and hierarchical identifier. Simultaneously, the system performs unified numbering and deduplication on all nodes, eliminating duplicate nodes, and establishes an adjacency list structure based on nodes and connection relationships. For situations where road segments intersect spatially but are not actually connected, such as grade-separated interchanges at different levels, the system uses hierarchical identifiers to constrain connection relationships, avoiding incorrectly connecting road segments that do not have actual traffic relationships, thereby forming initial topology data that conforms to real traffic flow logic.

[0036] S12. Perform closed-loop path retrieval on the initial topology map data to obtain closed-loop path data; In one embodiment, the system uses initial topology data to sequentially select each node in the road network as a starting node for path retrieval. The system progressively expands the path along directed connections, recording the node sequence of the current path in real time during the expansion process. When a path returns to the starting node during traversal and the number of nodes contained in the path reaches a preset minimum path length requirement, the path is determined to be a closed path that meets the conditions. The system standardizes the identified paths, uniquely identifying them using a unified path representation method and removing duplicate paths. Simultaneously, for paths with duplicate nodes that do not return to the starting node, the system determines them as non-closed paths and removes them. To control the scale of path search, the system sets a maximum path length limit during path expansion; when the path length exceeds this limit, the expansion of the current path is terminated. The system aggregates all paths that meet the closure criteria to form closed-loop path data.

[0037] S13. Screen the loop-turn structure constraints based on the closed-loop path data to obtain valid closed-loop structure data; In one embodiment, the system performs loop-loop structure constraint screening on each candidate path based on closed-loop path data. The system verifies the continuity of the path, requiring adjacent connections to be directly connected topologically to ensure no breaks or jumps. It also judges the directional consistency of the path, requiring it to unfold continuously along the predetermined direction of travel, without reverse jumps or directional conflicts. Combining the hierarchical identification information of road segments, the system constrains the hierarchical differences between adjacent nodes within the path, requiring these differences to be within a preset allowable range to ensure the path conforms to the actual traffic rules of the interchange structure and avoids cross-level impassable connections. The system screens the overall scale of the path by summing the lengths of each segment to determine if the total length meets a preset minimum loop scale requirement, thus excluding paths that are too short or lack practical traffic significance. When a path simultaneously meets the constraints of continuity, directional consistency, hierarchical accessibility, and minimum scale, it is determined to be a valid closed-loop structure; otherwise, it is discarded. Through the above screening process, pseudo-closed-loop paths caused by data noise or topological misconnections can be effectively filtered out, thereby obtaining effective closed-loop structure data.

[0038] S14. Perform node-edge merging processing on the effective closed-loop structure data to obtain the closed-loop topology skeleton data.

[0039] In one embodiment, the system performs node and edge merging processing on each closed-loop path based on effective closed-loop structure data to achieve structural simplification and element retention. The system analyzes continuous edge segments in the path; when multiple adjacent edges are continuously connected in the topology, have no branch nodes in between, and maintain consistent traffic attributes, these continuous edge segments are merged into a single skeleton edge to reduce redundant representation and improve the clarity of the structural expression. During the merging process, the system identifies and retains key nodes in the path, including entrance nodes connected to the external road network, nodes with branching relationships, and merging nodes where multiple paths converge. These nodes serve as skeleton nodes to maintain the connection relationships of the closed-loop structure. After merging, the resulting skeleton structure still maintains the closure characteristics and traffic direction consistency of the original path. The system recalculates the attribute information of the merged skeleton edges, where the length of the skeleton edge is obtained by summing the lengths of its constituent original edge segments, and its traffic capacity and other attributes are updated based on the comprehensive characteristics of each edge segment. This results in closed-loop topology skeleton data with a reduced number of nodes and a clear structural hierarchy.

[0040] Optionally, the step of expanding the cyclic bearer chain based on the closed-loop topology skeleton data to obtain the intra-loop flow path data includes: S21. Identify the entry node within the loop based on the closed-loop topology skeleton data to obtain the loop start node data; In one embodiment, the system performs a traversal analysis of the nodes in the closed-loop topology skeleton data to identify entry nodes capable of entering the closed-loop structure. The system designates nodes that simultaneously meet the following conditions as entry nodes: first, the node has an input connection from outside the closed-loop structure; second, the node also has an output connection pointing to nodes inside the closed loop, thus enabling the import of external traffic flow into the closed-loop structure. The system analyzes the source of input traffic for entry nodes, filtering nodes by calculating the proportion of traffic from outside the closed loop to the total input traffic. When the proportion of external input traffic in the node's total input traffic reaches a preset threshold, the node is retained as a valid entry node; otherwise, it is discarded to avoid misidentifying local internal return flow nodes as entry nodes. The system aggregates all nodes that meet the conditions to form loop start node data.

[0041] S22. Perform directed path recursion expansion on the data of the starting node of the loop to obtain the flow path sequence data; In one embodiment, the system uses each entry node in the loop start node data as the starting point for path expansion, and recursively expands the path along the directed connections in the closed-loop topology skeleton. During the expansion process, the system adopts a layer-by-layer expansion approach, traversing all reachable output connections of the current node to generate the next layer of nodes, and continuously recording the sequence of nodes traversed by the path and the corresponding sequence of connections, thereby gradually forming a complete path representation. During the path expansion process, the system constrains and controls the path length, stopping expansion when the path length reaches a preset maximum range; simultaneously, it detects access patterns within the path, identifying repeated access to the same combination of connections as a loop path and terminating the expansion of the current path. The system caches the generated path results and prunes duplicate paths during the expansion process to avoid redundant computation. The system outputs a set of path sequences starting from each loop start node, where each path is represented in node order.

[0042] S23. Perform backtracking behavior judgment on the flow path sequence data to obtain the loop path identification data; In one embodiment, the system performs backtracking behavior discrimination processing on each path based on the flow path sequence data. The system performs repeated access detection on the node sequence in each path. When a node reappears at a subsequent position in the path sequence, it determines that the path has backtracking behavior. Specifically, when the starting and ending nodes of the path are the same, and the path forms a complete closure, it is marked as a closed-loop backtracking path; when a node repeatedly appears in the middle of the path, it is identified as a local backtracking, and the corresponding backtracking segment range is recorded. The system evaluates the scale of the backtracking segment by accumulating the lengths of each segment within the segment to obtain the total length of the backtracking segment, and simultaneously records the number of backtracking occurrences and their positional distribution within the path. To avoid misjudging short-distance disturbances or local fluctuations as valid backtracking behavior, the system sets a minimum backtracking length threshold; only when the cumulative length of the backtracking segment exceeds this threshold is the corresponding backtracking identifier retained. The system adds backtracking marking information to each path, including the backtracking segment range, the number of backtracking occurrences, and their positional characteristics, thereby forming cyclic path identification data.

[0043] S24. Perform path carrying capacity association mapping based on the circular path identifier data to obtain path carrying attribute data; In one embodiment, the system maps each path to its corresponding road segment connection based on cyclic path identification data, and extracts attribute information such as the capacity, length, and design speed of each road segment in the path. The system evaluates the overall carrying capacity of the path, where the path's capacity is determined by the minimum capacity of each road segment as a constraint to represent the path's bottleneck capacity level. For example, the system obtains the design capacity of each road segment in the path and uses the minimum capacity as the basic bottleneck capacity of the path. The system adjusts this based on the proportion of low-capacity road segments in the path; for instance, when the proportion of road segments with capacity below the path's average exceeds 30%, the overall carrying capacity is reduced (by 10% to 20%). The system accumulates the travel time of each segment in the path based on the length and corresponding design speed of each road segment to obtain the overall travel time characteristics of the path. The system further refines the path's carrying capacity attributes by incorporating return behavior information from the cyclic path identification data. When a route contains a return segment, the system reduces the route's effective capacity based on the size and frequency of the return segment, and adds corresponding additional costs to the travel time. For example, the system calculates the proportion of the return segment's length in the entire route. If the return segment's length accounts for 10% to 30% of the total route length, the effective capacity is reduced by approximately 10%; if the proportion exceeds 30%, the reduction can be increased to 20% to 30%. The system also adjusts the capacity based on the frequency of returns. For instance, if there are 1 to 2 returns within an evaluation period (e.g., 30 to 60 seconds), an additional 5% reduction is applied; if there are 3 or more returns, an additional 10% reduction is applied. Regarding travel time, the system treats return segments as repeated travel processes and adds extra calculations to the travel time. For example, each return segment adds 50% to 100% of the original travel time. When the return frequency is high (e.g., more than 3 times), the return segment time can be doubled as an additional cost. For instance, if the basic travel time for a route is 10 minutes, the return segment travel time is 2 minutes, and there are 2 returns, an additional 2 to 3 minutes can be added, adjusting the total travel time to 12 to 13 minutes. The system generates path carrying attribute data for each route, including information such as traffic capacity, travel time, and the degree of impact of returns.

[0044] S25. Compile the path carrying attribute data into a path set to obtain the internal loop flow path data.

[0045] In one embodiment, the system performs unified compilation processing on each path based on path carrying attribute data. The system classifies paths according to their start and end nodes and return characteristics (including return segment location, number of returns, and return intensity; the cumulative length of the return segment or the proportion of repeated occupancy is used as a measure of return intensity), grouping paths with the same entry point and similar return behavior characteristics into the same category. The system assigns a unified number to each path and analyzes the overlap between different paths, identifying shared road segments to establish a path association structure representing the spatial intersection and overlap of different paths. The system performs redundancy detection on the path set, removing or merging sub-paths completely contained within other paths. The system integrates path sequence information, return identification information, and carrying attribute information to construct intra-loop flow path data.

[0046] Optionally, the step of performing retention coupling processing based on the intra-loop flow path data to obtain cyclic retention potential field data includes: Based on the flow path data within the ring, the overlapping area is identified to obtain the overlapping area data; In one embodiment, the system identifies and processes the spatial overlap relationships between paths based on the flow path data within the loop. The system reads the node sequence and edge set corresponding to each path and compares and analyzes the edge sets of any two or more paths. When different paths have the same edge segments, or when there are segments with consistent start and end nodes and overlapping spatial positions, this part of the area is determined to be a path overlap area. After identifying the overlap segment, the system integrates multiple consecutive overlapping edge segments, merging them into a continuous overlap area, and records the path number set involved in the overlap area, the start and end nodes of the segment, and the spatial range and length information of the segment. At the same time, for multiple adjacent overlap areas, if their participating path sets are consistent and there are no branching or structural forking nodes between the segments, the system will merge them to form a continuous occupation area representation. The system generates overlap area data, which includes the unique identifier of the overlap area, the participating path set, the corresponding edge set, and the spatial position information of the segment.

[0047] Traffic aggregation calculations are performed on the overlapping occupancy area data to obtain the segment load data; In one embodiment, the system statistically analyzes and overlays traffic flow data for each overlapping area within a preset time range based on overlapping area data. The system uses a set time window as the statistical unit, summarizing the vehicle flow entering the segment within each overlapping area from each participating path within that time range. This is achieved by acquiring road network structure data and connecting to traffic monitoring data sources (such as roadside detection equipment, checkpoint records, or historical statistical databases) to obtain the number of vehicles passing through each road segment within the corresponding time window, using this number as the basic data input for path flow. For example, the number of vehicle passes through each road segment is counted in 30-second or 60-second increments, serving as the vehicle flow value within that time window. This yields the total load level of the segment within the corresponding time window. During the statistical process, the system reads the traffic flow data of each path participating in the overlapping area one by one and accumulates the entry traffic of each path within the time window to represent the segment occupancy under the combined effect of multiple paths. Simultaneously, the system combines the design capacity and segment length information corresponding to the segment to provide a supplementary description of the load level, indicating the segment's carrying capacity under current traffic conditions. For overlapping areas with return paths, the system treats the traffic generated by the return paths as a separate component for statistical analysis and includes it in the total load of the segment. The system generates segment load data, which includes information such as segment identifier, corresponding time window, cumulative load level, and segment throughput capacity.

[0048] Perform time-series retention analysis on the section load data to obtain section retention data; In one embodiment, the system performs time-series analysis on load changes in each overlapping area over a continuous time range based on segment load data to identify the congestion status of the segment. The system compares the load levels of each overlapping area across multiple consecutive time windows. When the actual load of a segment consistently exceeds its corresponding traffic release capacity threshold across multiple consecutive time windows, the segment is determined to be in a congestion state. During the determination process, the system calculates the difference between the load and release capacity within each time window. When this difference is consistently positive and the duration reaches a preset congestion determination duration requirement, a stable congestion phenomenon is confirmed for that segment. The system records the congestion process, including the start time, end time, duration, and maximum load difference during the congestion process. For congestion phenomena occurring multiple times in the same segment across different time periods, the system divides them into multiple independent congestion events, which are then numbered and recorded separately. The system obtains segment congestion data to represent the degree of traffic congestion and its persistence characteristics in each overlapping area over time.

[0049] Nonlinear interactive interferometry is performed based on the segment retention data to obtain coupled retention data; In one embodiment, the system performs nonlinear interactive interference analysis on the mutual influence relationships between segments based on segment retention data. The system extracts retention characteristics for each segment, such as retention intensity (the degree of overload of the segment's actual load level relative to its passage release capacity within a certain time window), retention duration, and retention change trend. These characteristics are organized into segment retention characteristic descriptions to represent the changes in the retention state of a segment over different time ranges. For segments with shared nodes or adjacent connections in the topology, the system performs unified mapping and comparative analysis of their retention characteristics. By observing the changes in the corresponding characteristics of adjacent segments before and after a change in the retention state of a segment, the system identifies the correlation and change relationships between different segments. When the retention characteristics of adjacent segments show a synchronous strengthening or weakening trend with the change of the target segment, it determines that there is an interference relationship between them and distinguishes the direction and degree of influence of the interference. The system records the identified interference relationships, including the segment pairs involved in the interference, the direction of the interference, and the corresponding interference intensity / synchronization degree of retention characteristics, thereby forming coupled retention data.

[0050] Spatially continuous mapping is performed based on coupled stagnant data to obtain cyclic stagnant potential field data.

[0051] In one embodiment, the system performs spatial continuous mapping processing on the retention distribution in the closed-loop structure based on coupled retention data. The system establishes sequential coordinates along the loop for each segment according to the directed arrangement of nodes in the closed-loop topology skeleton, ensuring that each overlapping segment corresponds to a unique location identifier in the closed-loop structure. The retention intensity of this segment is then mapped to the corresponding location, forming an initial discrete distribution. For adjacent segments that are physically separated but topologically continuous, the system performs smoothing and completion processing on the retention changes between them, enabling a continuous transition in retention intensity between adjacent segments. This transforms the discrete segment retention results into a retention change curve continuously distributed along the closed loop. Simultaneously, for locations with diverging or merging nodes, the system retains the abrupt change characteristics of their retention intensity. The system constructs cyclic retention potential field data, including sequential coordinates along the closed-loop topology, retention intensity values ​​at corresponding locations, and relevant segment identifier information, representing the spatial distribution and continuous evolution characteristics of the retention state in the loop structure.

[0052] Optionally, the step of performing nonlinear interferometric measurement based on the segment retention data to obtain coupled retention data includes: Nonlinear mapping processing is performed on the segment retention data to obtain nonlinear feature space data; In one embodiment, the system extracts and maps the congestion change characteristics of each segment over a continuous time range based on segment congestion data. The system extracts features such as congestion intensity, congestion duration, congestion growth trend, and congestion recovery speed from the segment congestion data. The congestion growth trend can be determined by the congestion changes between adjacent time windows, and the congestion recovery speed characterizes how quickly a segment transitions from a congested state to a normal traffic state. The system organizes these features into the original feature description of the segment and performs uniform scaling on each dimension of the features. The system inputs the processed feature data into a preset nonlinear mapping rule (such as...). , This is the nonlinear feature representation vector obtained after nonlinear mapping. This is a nonlinear mapping function performed on the original feature vectors to transform the original features from a linear space to a high-discrimination feature space. This function can be a kernel mapping function or a multi-level nonlinear transformation rule. The feature vector of a segment within the current time window represents the segment's stagnation status information, including stagnation intensity, stagnation duration, stagnation change trend, and stagnation recovery speed. The function employs a radial basis mapping method, converting the distance relationship between the input features and several preset center points into new feature components. This makes similar stagnation patterns more concentrated after mapping, and patterns with significant differences easier to separate. Alternatively, a nonlinear mapping network with one or more activation transformation layers can be used to transform the input features layer by layer to obtain new feature representations. This structural transformation of the original features allows them to form a more discriminative expression in the new feature space, thereby expanding and reconstructing stagnation change patterns that were originally difficult to distinguish directly. Through this mapping process, the stagnation status of each segment within different time ranges can be represented in feature form. The system generates nonlinear feature space data, including segment identifiers, corresponding time windows, and the mapped feature representation results.

[0053] The nonlinear mapping network is constructed using a multi-layer feedforward structure. The system uses segmental retention feature vectors as input layer nodes, with input feature dimensions including retention intensity, retention duration, retention trend, and retention recovery speed. One or more hidden layers are set in the middle of the network, each containing several neurons. The number of nodes can be set according to the input feature dimensions, for example, 1 to 3 times the input dimension. The layers are connected via fully connected connections, and a nonlinear activation transformation is performed after each layer's output. This activation transformation uses a function with nonlinear mapping capabilities, such as a monotonic function that compresses or amplifies the input. During network construction, the system initializes the connection relationships of each layer and iteratively adjusts the network parameters using preset training data or historical retention samples to improve the discriminative power of the output features across different retention modes. After multi-layered successive transformations, the nonlinear feature representation is obtained.

[0054] Interferometric construction is performed on nonlinear feature space data to obtain path interference data; In one embodiment, the system constructs interference relationships between segments based on nonlinear feature space data. The system performs pairwise analysis on segments located on adjacent paths or shared node paths according to the path to which the segment belongs and its adjacency relationship in the topology, and extracts the feature representation of the corresponding segment in the nonlinear feature space. The system combines the features of each pair of segments by splicing, comparing differences, or associating the features of two segments to construct an interactive feature expression that reflects the mutual influence between segments. Based on the above interactive features, the system determines the interference relationship between segments. When a change in the state of a segment causes synchronous changes, increased change amplitude, or delayed recovery in associated segments within a continuous time range, it is identified as having an interference relationship and classified as either enhanced interference or suppressed interference. The system records valid interference segment pairs and their corresponding interference types and feature expressions, forming path interference data.

[0055] Interference measurement is performed based on path interference data to obtain coupled stagnant data.

[0056] In one embodiment, the system performs quantitative analysis on the interaction between pairs of interference segments based on path interference data. The system statistically analyzes the frequency and duration of each pair of interference segments within a continuous time range; it compares and analyzes the characteristic changes between the perturbation segment and the response segment to characterize the strength and degree of interference. The perturbation segment refers to the segment selected as the source of change during the path interference construction process, whose characteristic changes participate in the interference analysis as external inputs or triggering factors. The response segment refers to the segment whose characteristic changes are affected by the perturbation segment in the same interference relationship, located in a position adjacent to the perturbation segment in the topology or reachable through path connections. The system distinguishes between interference relationships between segments within the same path and interference relationships between segments of different paths, obtaining the coupling type and recording its direction and range of influence. When the influence of one segment on another related segment remains consistent across multiple continuous time windows, and the stagnation change of the corresponding segment exhibits obvious synchronous changes or stable delayed response characteristics, the system determines that a stable coupling relationship exists between the segment pairs to identify the set of coupled segments. The system integrates features such as the duration of interference, the number of involved segments, and the magnitude of change to form a result on the coupling relationship. It generates coupling persistence data, including information such as the set of coupling segments, coupling type, coupling duration, and the range of coupling influence.

[0057] Optionally, the step of constructing path interference data by cross-interference of the nonlinear feature space data includes: Path feature phase alignment is performed based on nonlinear feature space data to obtain aligned feature data; In one embodiment, the system performs phase alignment processing on feature changes between different paths based on nonlinear feature space data. The system extracts segment feature representations of each path within a continuous time range according to the path number, and constructs corresponding path feature sequences in chronological order to characterize the stagnation features of each path over time. For stagnation changes between different paths exhibiting temporal lag or advance, the system employs a sliding comparison method, attempting multiple alignments of the feature sequences of two paths within a preset time offset range. By gradually adjusting the time position of one path, the system compares the degree of difference between the features of the two paths under different offsets and selects the offset with the smallest difference as the phase compensation result. The system performs time alignment processing on the path feature sequences according to this compensation result, ensuring that the feature changes of the two paths correspond in the time dimension. For path pairs with required offsets exceeding the preset range, the system determines that their changes are asynchronous and does not participate in the subsequent construction of interference relationships. The system generates aligned feature data, including path pair identifiers, corresponding phase compensation information, and aligned feature sequences.

[0058] Interactive perturbation injection is performed on the alignment feature data to obtain perturbation injection data; In one embodiment, the system performs perturbation injection processing on the interaction relationship between paths based on alignment feature data. The system selects the alignment feature sequence of one path as the perturbation source and introduces its feature changes over a continuous time range into the corresponding feature sequence of another path to simulate the mutual influence process between paths in a coupled state. For the feature representation of the target path in each time window, the system obtains the change components from the corresponding time position of the source path in feature dimensions related to retention intensity, retention change trend, or recovery speed, thereby forming a target feature representation that includes the influence of external perturbations. This method demonstrates the linkage effect on the target path when the state of the source path changes. The system only performs the above processing on path pairs that have adjacency, shared nodes, or overlapping segments in the topological structure; path pairs without actual association are not perturbated. The system generates perturbation injection data, including the source path identifier, target path identifier, corresponding time range, and the perturbed feature sequence.

[0059] Nonlinear expansion is performed on the injected disturbance data to obtain nonlinear interferometric data; In one embodiment, the system performs nonlinear expansion processing on the perturbation-injected path feature sequence based on perturbation injection data to identify the state change characteristics caused by the interaction between paths. The system compares the feature representations of the target path before and after the perturbation, analyzes the differences between the two over a continuous time range, and tracks the evolution trend of the differences over time. During the analysis, when the features after the disturbance show continuous enhancement in some dimensions, prolonged recovery process, or significant change in feature pattern, the system determines that there is a nonlinear interference response. For example, the system compares the difference of the same feature dimension before and after the disturbance in continuous time windows. When the difference shows a monotonically increasing or overall increasing trend in no less than 3 consecutive time windows, and the cumulative increase reaches more than 15% of the baseline value of that dimension, it is determined to be a continuous enhancement. The prolonged recovery process can be determined by comparing the number of time windows required for the segment before and after the disturbance to recover from a high stagnation state to a normal state. When the recovery time after the disturbance increases by at least 2 time windows compared to before the disturbance, or the extension exceeds 20% of the original recovery time, it is determined to be a prolonged recovery process. Significant changes in feature pattern can be judged by changes in the direction or structure of feature change, such as from single-peak change to multi-peak change, from stable fluctuation to continuous oscillation, or the direction of change of the feature sequence reverses within a continuous time window and remains for at least 2 time windows, while its change amplitude exceeds 10% of the baseline fluctuation range, then it is determined to be a significant change in pattern. The system locates the aforementioned differential changes according to path segments, mapping the interference effects to specific segment locations to clarify the spatial range of the interference. For anomalous changes occurring only within a single time window, the system first marks them as interference events to be determined and then verifies their stability over subsequent time ranges. Based on the specific forms of interaction between different paths, the system can classify them into three categories: enhancing interference, suppressing interference, and delayed interference, thus obtaining interference behavior types. Enhancing interference refers to the situation where the relevant characteristics of the target path continuously increase or the degree of retention significantly intensifies after the introduction of disturbance; suppressing interference refers to the situation where the characteristics of the target path decrease or the degree of retention weakens after the introduction of disturbance, such as a decrease in retention intensity or an increase in recovery speed; delayed interference refers to the situation where the target path does not immediately change significantly after the disturbance, but only shows a significant response after several time windows, and this delay time reaches a preset threshold (e.g., not less than 2 time windows). The system generates nonlinear interference data, including interference path pairs, interference occurrence segments, differential change sequences, and interference behavior types.

[0060] Interference stability screening is performed on the nonlinear interferometric data to obtain effective interferometric data; In one embodiment, the system performs stability screening on each interference relationship within a continuous time range based on nonlinear interferometric data to distinguish between genuine and valid interference relationships and pseudo-interference phenomena caused by occasional fluctuations. The system performs consistency analysis on the interference performance of the same path pair within multiple continuous time windows. When the path pair exhibits the same type of interference effect within at least a preset number of continuous time windows, and the interference segments are continuously distributed or adjacent to each other in the topological structure, the interference relationship is determined to be a stable interference relationship. For interference phenomena that only occur in individual time windows, or interference segments that frequently change spatially and lack continuity (the system records the position of the interference segment corresponding to each time step within a continuous time window and statistically analyzes the changes in segment position between adjacent time windows. When the number of position jumps of the interference segment exceeds a preset threshold (e.g., more than 2 times) within a preset continuous time window range (e.g., 5 time windows), and the interference segments corresponding to adjacent time windows are not adjacent in the topological structure (e.g., the topological distance is greater than 1 segment), the interference relationship is determined to be spatially discontinuous), the system determines it to be an unstable interference and removes it. The system performs constraint analysis on the range of interference intensity variation. When the fluctuation of interference differences within multiple time windows is within a preset range, the interference relationship is considered to have good stability and is retained. After screening, the system generates valid interference data, including information such as the path pairs of stable interferences, the corresponding interference segment range, duration, and interference type.

[0061] Based on the effective interferometric data, the interferometric relationships are organized to obtain path interferometric data.

[0062] In one embodiment, the system categorizes and organizes interference relationships based on valid interference data. The system classifies all identified stable interference relationships according to paths and, combined with the spatial relationships within the closed-loop topology framework, constructs an interference relationship table between paths, representing the interaction and influence between different paths. During the organization process, for cases where the same pair of paths exhibits stable interference in multiple segments, the system merges the relevant segments to form path-level interference relationships and records the corresponding segment range and interference duration. Simultaneously, for cases where a path interferes with multiple paths simultaneously, the system distinguishes the interference direction based on its positional relationship within the topology, labeling it as upstream or downstream interference relationships to represent the directional characteristics of interference propagation. The system adds corresponding information to each path interference relationship, including the frequency of interference occurrence, the number of affected segments, and the duration, to represent the stability and influence range of the interference relationship. The system obtains path interference data.

[0063] Optionally, the step of locating the amplified source based on the cyclic stagnant potential field data to obtain the self-reinforcing core region data includes: S31. Extract local high-value regions based on cyclic stagnant potential field data to obtain preliminary stagnant region data; In one embodiment, the system extracts local high-value regions from the retention distribution in a closed-loop structure based on cyclic retention potential field data. The system divides the retention potential field into several continuous segments according to the order of the closed-loop topological coordinates and reads the retention intensity value of each segment within the current time range. The system compares the retention intensity of each segment with its adjacent segments before and after it. When the retention intensity of a segment is simultaneously higher than that of its adjacent segments before and after it, and its value exceeds a preset retention intensity threshold, the segment is identified as a local high-value point. If multiple adjacent segments are consecutively identified as high-value points, the system merges them to form a continuous local high-value region and records the starting segment, ending segment, and the location and corresponding value of the segment with the highest retention intensity within this region. When the number of interval segments between two local high-value regions is less than a preset connection threshold, the system merges them into the same high-value region. The system generates preliminary retention region data, including region identifiers, corresponding segment ranges, peak segment locations, and their retention intensity information.

[0064] S32. Continuously screen the preliminary retention area data to obtain stable retention area data; In one embodiment, the system continuously screens the changes of each retention area over a continuous time range based on preliminary retention area data. The system tracks the spatial location and corresponding retention intensity of each retention area within multiple continuous time windows and establishes a temporal series correlation between the areas to describe the evolution of the same area at different times. During the screening process, for retention areas that repeatedly appear at the same location or adjacent topological locations, the system determines whether they have persistent characteristics. When an area persists for at least a preset number of consecutive time windows, and the fluctuation range of its peak retention intensity within that time range is within a preset range, the area is identified as a stable retention area. Conversely, for high-value areas that only appear briefly in individual time windows or a small number of discrete time ranges, the system considers them transient accumulation phenomena and removes them. For high-value areas that move slowly in space but remain within the same topological neighborhood, the system considers them as a continuation of the same retention area in the time dimension and uniformly identifies them. The system outputs stable retention area data, including information such as the area duration, average peak retention intensity, and corresponding topological coverage.

[0065] S33. Perform back-coupling analysis on the stable retention area data to obtain the feedback enhancement area data; In one embodiment, the system performs coupling analysis on the return relationships between each retention area and the loop flow path based on stable retention area data. The system maps the segments corresponding to stable retention areas to the loop flow path data, identifies all paths passing through the area, and focuses on screening loop paths with return identifiers. The system analyzes the temporal correlation between stable retention areas and return segments. When a stable retention area is located upstream of a return path, and within a subsequent time range after the retention intensity of the area increases, the retention intensity of the relevant segments traversed by the return path increases synchronously, it is determined that there is a return coupling relationship between the stable retention area and the corresponding return segment. The system records the topological distance, associated path identifiers, and affected segment range between the stagnation zone and the return zone, representing its impact characteristics. The topological distance is defined as the number of nodes or edges traversed along the closed-loop topological path towards the return zone, starting from the center node of the stable stagnation zone and using the skeleton nodes as the counting unit. The associated path identifiers are defined as the set of path numbers that pass through the stable stagnation zone and have return identifiers, which are pre-assigned unique numbers to each path in the loop flow path data and recorded during the mapping process. The affected segment range is defined as the system scanning the stagnation intensity changes along the path forward and backward with the return zone as the center. When the stagnation intensity of an adjacent segment continuously increases relative to the baseline level and the change exceeds a preset threshold (e.g., exceeding 15% of the baseline value), the segment is included in the affected range. When the stagnation change returns to below the threshold, the expansion stops. When a stable retention region significantly enhances retention for multiple return paths, the system marks this region as a feedback enhancement region. For stable retention regions that are not coupled with return paths, only their basic retention attributes are retained, and they are not included in the feedback enhancement region set. The system generates feedback enhancement region data.

[0066] S34. Based on the feedback enhancement region data, the amplification effect is determined, and the self-enhancing region data is obtained; In one embodiment, the system analyzes the retention trend of each region within a continuous time range based on feedback enhancement region data to determine whether it has an amplification effect. The system tracks the retention intensity changes of each feedback enhancement region within multiple continuous time windows. When the retention intensity of a region shows a gradual upward trend within a continuous time range, and the associated return path segments also show synchronous enhancement characteristics within the same time range, the region is determined to have an amplification effect. The system compares the retention intensity changes between adjacent time windows; when the change shows a continuous increase within multiple consecutive time windows, it confirms a stable amplification trend. The system statistically analyzes the range of segments affected by the region; when the number of segments covered by its influence reaches a preset requirement, the region is identified as a self-enhancing region. For feedback enhancement regions that only exhibit high retention intensity but do not have a continuous amplification trend, the system retains them as general feedback regions and does not include them in the self-enhancing region results. The system generates self-enhancing region data, including amplification duration, affected segment range, and associated return path set information.

[0067] S35. Perform core region filtering on the self-enhancing region data to obtain the self-enhancing core region data.

[0068] In one embodiment, the system performs centralized filtering on self-reinforcement region data to determine the self-reinforcement core region. The system evaluates self-reinforcement regions from multiple aspects, including peak retention intensity, duration of amplification effect, number of affected return paths, and coverage area, prioritizing regions that simultaneously possess high retention intensity, long duration, and wide impact range. For multiple self-reinforcement regions that are spatially adjacent and whose affected path sets highly overlap, the system merges them to form a core region to avoid redundant representation. When multiple core regions exist, the system compares their coverage of return paths and their degree of obstruction to the overall closed-loop flow, selecting the region with the wider impact and more significant effect as the self-reinforcing core region. For example, the system sorts the peak retention intensity of each candidate region, classifying the top 20% as high-intensity regions; it statistically analyzes the duration of the amplification effect, classifying a region as a long-lasting region when it maintains an amplification trend for at least five consecutive time windows; it statistically analyzes the number of affected return paths, classifying a region as having a wide impact when it has at least three associated return paths; and it analyzes the coverage area, classifying a region as having a large coverage area when it continuously covers at least four segments. During the selection process, the system prioritizes regions that simultaneously meet at least three high-level conditions as self-reinforcing core regions. The system generates self-reinforcing core region data, including core region identifier, corresponding segment range, peak retention intensity, duration, and associated path set.

[0069] Optionally, the step of generating ring-turn control data by topology discharge coordination based on self-enhancing core region data includes: Based on the self-enhancing core area data, the associated path backtracking is performed to obtain the core associated path data; In one embodiment, the system performs backtracking analysis on paths related to the core area based on self-reinforcing core area data. Starting from the skeleton nodes and edges covered by the self-reinforcing core area, the system traces and extends paths along both the reverse and forward directions of traffic flow within the closed-loop topology skeleton. In the reverse direction, the system identifies all paths directly connected to the core area and inputting traffic flow into it, marking them as upstream injection paths. In the forward direction, the system identifies paths connected to the core area and responsible for traffic diversion and discharge, marking them as downstream discharge paths. Paths located within the closed loop, forming a continuous carrying relationship with the core area, and responsible for traffic transmission are marked as pressure-bearing paths. During path backtracking, the system only retains paths with actual traffic connections to the core area and whose distance within the topology does not exceed a preset range. When a path possesses both injection and discharge attributes, the system splits it according to its relative position to the core area, dividing it into sub-paths with different functions. The system generates core associated path data, which includes path identifier, path type, corresponding start and end nodes, and topological distance to the core area.

[0070] The core associated path data is partitioned into path functions to obtain topology control partition data; In one embodiment, the system performs functional zoning on each path based on core associated path data. The system divides different path segments into several control areas with clearly defined functional attributes based on the path's position within the closed-loop structure and its flow direction association with the self-reinforcing core area. For path segments whose endpoints or intermediate nodes are directly connected to the core area and whose traffic flow points towards the core area, the system classifies them as injection control areas / injection areas. For continuous segments located within the closed loop of the core area and responsible for traffic retention and transmission, the system classifies them as pressure-relieving areas / pressure-bearing areas. For path segments whose starting point is after the core area and which output traffic to the outside of the closed loop or low-load areas, the system classifies them as release and diversion areas / release areas. During the zoning process, the system prioritizes merging nodes, diverging nodes, and key backbone nodes as zoning boundaries. For cases where functional attributes change within the path, the system splits the path into multiple sub-segments with different functional attributes based on node locations. The system generates topology control zoning data, including zoning identifiers, zoning types, corresponding path sets, and segment ranges.

[0071] Based on the topology control partition data, the partition control amount is generated to obtain local control parameter data; In one embodiment, the system generates corresponding local control parameters for each functional zone based on topology control zoning data. The system generates differentiated control strategies for different types of zones to achieve zoned regulation of traffic flow within the closed-loop structure. For the injection control zone, the system determines the upper limit of the number of vehicles allowed to enter the closed loop per unit time based on the difference between the current inflow and the corresponding capacity of the zone (using the stable capacity of the zone as a benchmark, with a safety margin (e.g., 10%~20%); for example, when the zone capacity is 100 vehicles per minute, the upper limit can be set to 80~90 vehicles). If overload is detected for two consecutive time windows, the upper limit is further reduced by 10%; if normal conditions are restored for three consecutive time windows, the upper limit is increased each time. (5% until the baseline value is restored). When the incoming flow exceeds the section's carrying capacity, the incoming flow is reduced to form the entry restriction parameter. For example, if the current incoming flow is 120 vehicles / minute and the upper limit is 90 vehicles, then 30 vehicles will be reduced. When the excess is within 10%, a slight reduction is adopted (the release interval is reduced by 10%). When it is between 10% and 30%, a moderate reduction is adopted (entry is restricted to the upper limit). When it exceeds 30%, a strong restriction is adopted (only about 50% to 70% of the basic passage demand is retained), thus forming a graded restriction parameter. For the pressure-bearing slow-release zone, the system generates pressure release rhythm parameters based on the retention duration and length characteristics of the zone. For example, when the retention duration is ≥ 5 time windows and the zone length is ≤ 200 meters, it is set to high-frequency release (release once every 1 time step); when the duration is 3-5 time windows and the length is between 200-400 meters, it is set to medium-frequency release (release once every 2 time steps); and when the duration is < 3 time windows and the length is > 400 meters, it is set to low-frequency release (release once every 3 time steps). The system prioritizes releasing traffic from sections closest to the core area. For drainage and diversion zones, the system generates priority diversion parameters based on the connection strength between the section and the core area, as well as the available space in the downstream area. For example, when the load of a downstream section is below 70% of its capacity, its priority is raised to the highest level (100% priority passage); when the load is between 70% and 90%, it is set to the second priority (70% to 90%); and when it exceeds 90%, execution is delayed or the priority is reduced (e.g., below 50%). The system generates local control parameter data, including information such as the type of control parameter, the affected section, the corresponding parameter value, and the applicable time range.

[0072] The local control parameter data are topologically coordinated to obtain the ring-turn control data.

[0073] In one embodiment, the system performs coordinated orchestration processing on the execution relationship of each zone's control strategy within the closed-loop topology based on local control parameter data. The system determines the execution priority order of different control parameters according to the relative position of each control zone within the closed-loop structure, ensuring that the overall control process conforms to the traffic flow transmission patterns. During the scheduling process, the system prioritizes the injection control zone located upstream to implement relevant control measures; it also schedules the pressure-bearing and slow-release zones located within or adjacent to the core area to implement control measures; for the release and diversion zones located downstream, the system executes them synchronously or with appropriate delays based on the actual situation, thus forming a coordinated control process. For example, when the current load level of the downstream section is lower than a preset proportion of its capacity (e.g., not exceeding 70% of its maximum capacity) and there is no obvious stagnation in the downstream continuous sections, the system determines that it has sufficient diversion space, and at this time, the release and diversion parameters are set to be executed synchronously with the pressure-bearing and slow-release zones; when the load of the downstream section is close to or exceeds its capacity (e.g., reaching more than 80%), or when there is a stagnation trend within two or more consecutive time windows, the system determines that its diversion capacity is limited, and the release and diversion parameters are appropriately delayed, with the delay time set to one or more time steps (e.g., 1 to 3 time steps); when there are high stagnation zones or high return intensity in the downstream section, the system extends the execution delay and prioritizes reducing the upstream injection intensity to avoid forming new cyclic backlogs. For multiple control parameters located on the same path chain, the system constructs an execution sequence according to the topological connection order to avoid situations where upstream flow continues to enter before downstream flow has been completed. When there is a conflict in control direction between adjacent partitions, the system prioritizes retaining control parameters that are closer to the self-reinforcing core area to ensure priority adjustment of the core area. The system generates ring-turn control data, which includes information such as the control parameters, execution order, duration of action, and associated path identifiers for each segment.

[0074] Optionally, the step of performing cyclic suppression verification based on loop control data to obtain closed-loop optimization evaluation data includes: S41. Apply control mapping based on the ring-turn control data to obtain the verification input state data; In one embodiment, the system performs topology mapping loading on each control parameter based on loop-turn control data to construct the input state for verification analysis. The system reads the type, effective segment, execution order, and corresponding time range of each parameter in the control data and maps them to the closed-loop topology skeleton and the flow path structure within the loop. During the mapping process, for injection control parameters / parameters corresponding to injection control areas, the system loads them into the corresponding inlet segment to constrain the inflow of that segment; for pressure-releasing parameters / parameters corresponding to pressure-releasing areas, the system loads them into the corresponding pressure-relief segment to adjust the release rhythm or passage interval of the segment; for venting and guiding parameters / parameters corresponding to venting and guiding areas, the system loads them into the corresponding downstream segment to adjust its passage priority or output capacity. Simultaneously, the system reads the segment operation status information at the start of the verification, including segment flow, retention, and return path status, and constructs an initial state description accordingly. When multiple control parameters correspond to the same segment within the same time range, the system processes the parameters sequentially according to a preset execution order, which can be written using overwriting or merging methods. The system receives the verification input status data, including the segment identifier, initial flow, retention, control parameters after loading, and corresponding execution order information.

[0075] S42. Perform a loop flow simulation on the verification input status data to obtain the verification flow data; In one embodiment, the system performs cyclic flow simulation processing on traffic flow in a closed-loop structure based on the verification input state data. The system updates the state of each segment step by step according to a preset time step, and sequentially executes processes such as flow input, segment release, and path transition within each time step. During the simulation, the system updates the new flow of the entrance segment according to the injection control parameters; the system adjusts the release rhythm, release sequence, and output capacity of each segment according to the pressure relief parameters and release guidance parameters, thereby determining the release amount of each segment in the current time step. The system updates the segment state of the next time step by combining the segment holding volume of the previous time step, the current inbound flow, and the current release flow, and simultaneously records the corresponding flow changes and congestion. For paths containing return segments, after a vehicle completes one release, the system re-injects it into the corresponding return node according to the path sequence to simulate the cyclic flow process. During the simulation, the system continuously records information such as traffic flow, congestion, release volume, and return frequency for each segment. For vehicles that cannot be released due to insufficient segment capacity, they are retained as congestion volume for the next time step and used in subsequent calculations. The system generates verification flow data, which includes segment status information, path flow results, and return behavior records for each time step.

[0076] S43. Evaluate the return intensity of the verification and circulation results data to obtain the cycle intensity data; In one embodiment, the system performs intensity assessment of backtracking behavior in the closed-loop structure based on the verification flow result data. The system filters all paths with backtracking identifiers from the flow results and performs statistical analysis on the repeated entry behavior of each path within a preset assessment time range, including the number of times the same segment is repeatedly entered, the number of segments involved, and the duration of the backtracking. The system statistically analyzes the frequency of each backtracking path re-entering the same segment within a unit time range and, combined with the path's occurrence in the overall flow process, assesses the proportion of its backtracking behavior in the overall flow. Simultaneously, for segments repeatedly occupied by the same path or multiple paths within multiple time steps, the system identifies them as high-backtracking segments, used to characterize locations with concentrated loop backlog. Based on the spatial order of the closed-loop topology, the system integrates and analyzes these high-backtracking segments, generating descriptions of loop intensity at both the path and segment levels. The system generates loop intensity data, including the number of path backtrackings, the number of times segments are repeatedly occupied, the duration of the backtracking, and the distribution of high-backtracking segments.

[0077] S44. Based on the cyclic intensity data, determine the retention change to obtain closed-loop optimization evaluation data.

[0078] In one embodiment, the system performs retention change discrimination processing on the control effect based on cycle intensity data. The system compares and analyzes the cycle intensity data after the control is implemented with the baseline cycle intensity data before the control, and simultaneously compares the changes in retention level, retention duration, and high retention area range of each segment before and after the control. During the discrimination process, when the number of high return segments decreases, the path return duration shortens, and the average retention level of the segment shows a downward trend after the control, the system determines that the current control strategy has an effective inhibitory effect on the closed-loop cycle. For segments that still maintain a high cycle intensity after the control, the system marks them as residual risk segments and records the corresponding path information and their position in the topology. The system generates closed-loop optimization evaluation data, including information such as changes in cycle intensity, retention reduction effect, set of residual risk segments, and overall optimization judgment results.

[0079] Optionally, this application also provides a spatial topology-constrained interconnection variable collaborative optimization system for executing the spatial topology-constrained interconnection variable collaborative optimization method described above, wherein the spatial topology-constrained interconnection variable collaborative optimization system includes: The closed-loop topology construction module is used to acquire the road segment connection data of the interchange and interpret the ring-ramp structure based on the road segment connection data of the interchange to obtain the closed-loop topology skeleton data. The cyclic flow modeling module is used to expand the cyclic bearing chain based on the closed-loop topology skeleton data to obtain the flow path data within the loop; and to perform stagnation coupling processing based on the flow path data within the loop to obtain the cyclic stagnation potential field data. The amplification source analysis and collaborative control module is used to locate the amplification source based on the cyclic stagnant potential field data to obtain the self-reinforcing core region data; and to perform topology discharge collaborative generation based on the self-reinforcing core region data to obtain the annular control data. The cyclic suppression evaluation module is used to perform cyclic suppression verification based on the loop control data to obtain closed-loop optimization evaluation data.

[0080] Therefore, the embodiments should be regarded as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended application documents rather than the foregoing description. Thus, it is intended that all changes falling within the meaning and scope of the equivalents of the application documents be incorporated into the invention.

[0081] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for collaborative optimization of interconnected discrete quantities based on spatial topological constraints, characterized in that, The method includes: Obtain the road segment connection data of the interchange, and interpret the ring-ramp structure based on the road segment connection data of the interchange to obtain the closed-loop topology skeleton data; Based on the closed-loop topology skeleton data, the entry node within the loop is identified to obtain the loop start node data; the loop start node data is expanded by directed path recursion to obtain the flow path sequence data; the flow path sequence data is judged by back-trip behavior to obtain the loop path identifier data; the loop path identifier data is mapped by path carrying capacity association to obtain path carrying attribute data; the path carrying attribute data is compiled into a path set to obtain the loop flow path data; the loop flow path data is identified by path overlap area to obtain overlap occupancy area data; the overlap occupancy area data is calculated by traffic superposition to obtain segment load data; the segment load data is analyzed by time-series retention to obtain segment retention data; the segment retention data is measured by nonlinear interactive interference to obtain coupled retention data; and the coupled retention data is mapped by spatial continuity to obtain the loop retention potential field data. Local high-value regions are extracted from the cyclic stagnation potential field data to obtain preliminary stagnation region data. Continuous screening of the preliminary stagnation region data yields stable stagnation region data. Back-coupling analysis is performed on the stable stagnation region data to obtain feedback enhancement region data. Amplification effect discrimination is performed on the feedback enhancement region data to obtain self-enhancing region data. Core region screening is performed on the self-enhancing region data to obtain self-enhancing core region data. Related path backtracking is performed on the self-enhancing core region data to obtain core related path data. Path function partitioning is performed on the core related path data to obtain topology control partition data. Partition control parameters are generated based on the topology control partition data to obtain local control parameter data. Topology position collaborative arrangement is performed on the local control parameter data to obtain ring-turn control data. Based on the loop control data, cyclic suppression verification is performed to obtain closed-loop optimization evaluation data.

2. The method according to claim 1, characterized in that, The process of interpreting the ring-ramp structure based on the interchange segment connection data to obtain closed-loop topology skeleton data includes: A directed road network is constructed based on the road segment connection data to obtain the initial topology map data; Closed-loop path data is obtained by performing closed-loop path retrieval on the initial topology map data; Based on the closed-loop path data, loop-turn structure constraint screening is performed to obtain valid closed-loop structure data; Node-edge merging is performed on the effective closed-loop structure data to obtain the closed-loop topology skeleton data.

3. The method according to claim 1, characterized in that, The process of performing nonlinear interferometric measurement based on segment retention data to obtain coupled retention data includes: Nonlinear mapping processing is performed on the segment retention data to obtain nonlinear feature space data; Interferometric construction is performed on nonlinear feature space data to obtain path interference data; Interference measurement is performed based on path interference data to obtain coupled stagnant data.

4. The method according to claim 3, characterized in that, The process of constructing path interference data by performing interactive interferometry on nonlinear feature space data includes: Path feature phase alignment is performed based on nonlinear feature space data to obtain aligned feature data; Interactive perturbation injection is performed on the alignment feature data to obtain perturbation injection data; Nonlinear expansion is performed on the injected disturbance data to obtain nonlinear interferometric data; Interference stability screening is performed on the nonlinear interferometric data to obtain effective interferometric data; Based on the effective interferometric data, the interferometric relationships are organized to obtain path interferometric data.

5. The method according to claim 1, characterized in that, The process of performing cyclic suppression verification based on loop-turn control data to obtain closed-loop optimization evaluation data includes: Based on the control data of the ring-turn control, control mapping is applied to obtain the verification input state data; The input status data for verification is cyclically processed and simulated to obtain the verification flow data; The cycle strength data is obtained by evaluating the return intensity of the verification and circulation results data. Based on the cyclic intensity data, the retention change is judged to obtain closed-loop optimization evaluation data.

6. A collaborative optimization system for interconnected discrete-time quantities based on spatial topological constraints, characterized in that, For executing the spatial topology-constrained interconnected variable collaborative optimization method as described in claim 1, the spatial topology-constrained interconnected variable collaborative optimization system comprises: The closed-loop topology construction module is used to acquire the road segment connection data of the interchange and interpret the ring-ramp structure based on the road segment connection data of the interchange to obtain the closed-loop topology skeleton data. The cyclic flow modeling module is used to expand the cyclic bearing chain based on the closed-loop topology skeleton data to obtain the flow path data within the loop; and to perform stagnation coupling processing based on the flow path data within the loop to obtain the cyclic stagnation potential field data. The amplification source analysis and collaborative control module is used to locate the amplification source based on the cyclic stagnant potential field data to obtain the self-reinforcing core region data; and to perform topology discharge collaborative generation based on the self-reinforcing core region data to obtain the annular control data. The cyclic suppression evaluation module is used to perform cyclic suppression verification based on the loop control data to obtain closed-loop optimization evaluation data.

Images