A resource management and control method based on a smart city

Abstract

This invention relates to the field of smart city scheduling technology and discloses a resource management method based on smart cities, including: acquiring real-time resource pool capacity vectors, task node demand vectors, and topological relationships; comparing demand and available capacity vectors to determine resource occupancy saturation; mapping the topology to a multi-dimensional temporal constraint space and defining task nodes as parameterized task packages, whose boundaries are determined by saturation; monitoring the overlap of task package boundaries in the multi-dimensional space to identify conflicts and calculate weights; and calling a constraint solving algorithm to adjust the resource occupancy window and generate an allocation strategy. This invention converts conflicts into multi-dimensional temporal topology offsets, breaks the limitations of one-dimensional queuing, perceives conflict intensity based on saturation, decouples resource scheduling deadlock while satisfying temporal dependencies, and improves the integrity and configuration efficiency of smart city business scheduling.

Description

Technical Field

[0001] This invention belongs to the field of smart city scheduling technology, and in particular relates to a resource management method based on smart cities. Background Technology

[0002] Currently, cross-departmental business processes are orchestrated using standardized workflow engines to schedule physical and digital resource tokens, ensuring the orderly operation of traffic guidance, power distribution, and emergency security systems. As urban workloads increase, heterogeneous business processes generate high-frequency concurrent resource requests during peak hours. Existing technologies employ linear queuing based on preset priorities, relying on static weights to determine the task execution order. This single-dimensional scheduling method struggles to perceive the temporal dependencies and resource exclusivity relationships between different workflow nodes when handling cross-domain tasks with strong coupling characteristics. When resource constraints occur at local nodes, conflicts rapidly spread globally, resulting in scheduling deadlocks and process fragmentation.

[0003] To resolve resource conflicts, the industry has adopted methods such as increasing infrastructure redundancy or adding scheduling priority levels. However, in city-level collaborative scenarios, these methods only alleviate surface load pressure and cannot logically address the systemic vulnerability caused by task topology interference. If scheduling rules are overly refined, it can lead to resource idleness and system oscillations. In addition to the physical redundancy limitations of hardware facilities, the control methods also have shortcomings. For example, Chinese invention patent application CN121585664A discloses a heterogeneous workflow task scheduling method, device, and medium. Although it introduces entropy weighting and analytic hierarchy process to quantify task characteristics, the core logic still anchors the fitting of multi-dimensional attributes into a one-dimensional scalar total score weighted calculation mode. The discrete scoring scheduling mechanism cannot logically restore the complex spatiotemporal topological dependencies between concurrent tasks, making it difficult for the system to perceive the dynamic expansion and interference of task envelopes in the resource dimension. In city-level high-load environments, this solution lacks the ability to decouple resource conflict topology, faces cross-domain coupled tasks, cannot resolve deadlocks through phase shift or offset adjustment, and cannot guarantee the structural robustness of the scheduling sequence.

[0004] Therefore, how to construct a scheduling mechanism that analyzes concurrent workflow topology characteristics and eliminates resource interference based on multidimensional parameter space constraints is the technical problem to be solved by this invention. Summary of the Invention

[0005] This invention provides a resource management method based on smart cities, comprising the following steps: Step S1: Obtain the real-time available capacity vector of the smart city multi-dimensional resource library, the resource specification requirement vector of the task nodes to be scheduled, and the initial task topology relationship composed of the logical dependencies between the task nodes to be scheduled. Step S2: By comparing the numerical mapping relationship between the resource specification demand vector and the real-time available capacity vector, calculate the occupancy saturation, which is used to characterize the intensity of the request for various bottleneck resources by the task node to be scheduled, and use the occupancy saturation as a feature parameter to control the degree of spatiotemporal expansion of the task node to be scheduled in the virtual space. Step S3: Map the initial task topology relationship to a multi-dimensional temporal constraint space composed of resource dimension and time axis, and define the task node to be scheduled as a parameterized task package. The spatial occupancy envelope and spatiotemporal boundary threshold of the parameterized task package in the multi-dimensional temporal constraint space are calculated by the corresponding occupancy saturation. Step S4: Monitor the spatiotemporal boundary overlap status between different parameterized task packages in the multidimensional temporal constraint space in real time to identify resource preemption conflicts at the business level, and calculate the overlap weight of resource preemption conflicts in the spatiotemporal dimension based on the boundary overlap volume and resource scarcity index. Step S5: Invoke the multi-objective constraint solving algorithm. Under the logical constraint of satisfying the temporal dependency relationship of the task nodes to be scheduled, adjust the topology offset of the resource occupation window or safe time interval of different parameterized task packages in the multi-dimensional temporal constraint space according to the overlap weight, and generate a conflict-free global resource allocation strategy.

[0006] Preferably, generating a conflict-free global resource allocation strategy includes: step S21, obtaining the boundary change trend of parameterized task packages; step S22, constructing a multidimensional spatial correction vector in the multidimensional temporal constraint space to characterize the spatiotemporal displacement correction amount based on the overlap weight and the boundary change trend; step S23, using the multidimensional spatial correction vector to correct the spatiotemporal coordinates of the parameterized task packages, decoupling the resource scheduling deadlock of concurrent services, and outputting the corrected global resource scheduling instruction.

[0007] Preferably, it also includes a topology simplification step based on load decay: step S41, setting an occupancy saturation threshold; step S42, identifying connected branches in the initial task topology relationship whose occupancy saturation is lower than the occupancy saturation threshold; step S43, folding the logical nodes of the connected branches and converting the connected branches into a constant resource consumption model to reduce the computational load of the multi-objective constraint solving algorithm in subsequent steps.

[0008] Preferably, in step S5, the topology offset adjustment of the resource occupancy window of the parameterized task package includes: step S51, calculating the conflict repulsion force based on the overlap weight; step S52, driving the path reset direction of the resource scheduling channel of the parameterized task package in the multi-dimensional temporal constraint space according to the conflict repulsion force; step S53, when the path reset direction is detected to be restricted, performing asynchronous phase translation on the conflicting nodes to eliminate physical resource competition.

[0009] Preferably, in step S3, the parameterized task package is represented by a rectangular envelope box, a spherical envelope box, or a polygonal envelope box, and different types of task nodes to be scheduled correspond to envelopes of different geometric shapes in the multidimensional temporal constraint space.

[0010] Preferably, the length of the major axis of the rectangular envelope box is determined by the expected duration of the task node to be scheduled in the time dimension, and the width of the minor axis of the rectangular envelope box is determined by the space occupancy envelope width of the parameterized task package determined by the occupancy saturation.

[0011] Preferably, in step S1, the task nodes to be scheduled include traffic flow guidance tasks, power grid load peak shaving tasks, and security monitoring linkage tasks, and different types of task nodes have preset task priorities.

[0012] Preferably, in step S5, the multi-objective constraint solving algorithm is based on a directed acyclic graph search strategy to find the collision-free scheduling path with the minimum comprehensive cost consisting of resource consumption cost and time delay cost in the multi-dimensional temporal constraint space.

[0013] Preferably, after generating a conflict-free global resource allocation strategy, the method further includes: step S101, collecting resource consumption feedback data of each subsystem of the smart city in real time; step S102, correcting the resource channel load weight in the multi-dimensional time-series constraint space based on the resource consumption feedback data, and feeding it back to step S1 as a correction parameter of the real-time available capacity vector to optimize the subsequently generated global resource allocation strategy.

[0014] Compared with existing technologies, the resource management method of this invention based on smart cities has the following advantages: 1. In the resource management of smart cities, by constructing a global directed acyclic graph and mapping task nodes to parameterized geometric primitives, this invention transforms the resource preemption conflict at the logical level into a geometric interference problem in the auxiliary design space. This multi-dimensional topology mapping mechanism breaks the limitations of traditional one-dimensional time axis queuing, enabling the system to pre-decouple potential execution deadlocks by introducing displacement vectors in the design space under the rigid constraints of workflow logical connectivity, thus ensuring the structural integrity of cross-department concurrent business under extreme conditions.

[0015] 2. Based on the real-time capacity vector of the global standard resource library and the resource demand vector of the task node, the comprehensive resource occupancy tension is calculated, providing the system with a physically meaningful basis for judging the intensity of conflict. This tension index directly drives the envelope expansion of parameterized geometric primitives, enabling the system to dynamically perceive the changes in the occupancy tension of each subsystem on bottleneck resources, thereby guiding the geometric constraint solver to generate a collision-free resource configuration blueprint in a resource-constrained environment.

[0016] 3. By introducing a graph pruning mechanism based on tension decay, the system can identify and logically fold connected components with resource occupancy tension below the threshold, reducing redundant topological branches to a constant resource consumption model. This mechanism ensures that when a massive influx of micro-workflows is triggered by a city-level emergency, the computational scale of the global directed acyclic graph is always kept within the convergence limit of the control engine, guaranteeing the deterministic response of the scheduling center within a short control cycle. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the process of the smart city resource management method with multidimensional time-series constraints of the present invention; Figure 2 This is a logical mapping diagram of the feature elements of the global conflict-free resource allocation strategy of this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0019] It should be noted that all directional and positional terms used in this invention, such as: up, down, left, right, front, back, vertical, horizontal, inner, outer, top, bottom, transverse, longitudinal, center, etc., are only used to explain the relative positional relationship and connection between components in a specific state (as shown in the accompanying drawings). They are only for the convenience of describing this invention and do not require that this invention be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention. In addition, the descriptions of "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated.

[0020] In the description of this invention, unless otherwise explicitly specified and limited, the terms installation, connection, and linking should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections; they can refer to direct connections or indirect connections through an intermediate medium; they can refer to the internal connection of two components. For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances.

[0021] In the description of this specification, references to the terms "an embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example, and the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0022] A resource management method based on smart cities includes the following steps: Step S1: Obtain the real-time available capacity vector of the smart city multi-dimensional resource library, the resource specification requirement vector of the task nodes to be scheduled, and the initial task topology relationship composed of the logical dependencies between the task nodes to be scheduled. Step S2: By comparing the numerical mapping relationship between the resource specification demand vector and the real-time available capacity vector, calculate the occupancy saturation, which is used to characterize the intensity of the request for various bottleneck resources by the task node to be scheduled, and use the occupancy saturation as a feature parameter to control the degree of spatiotemporal expansion of the task node to be scheduled in the virtual space. Step S3: Map the initial task topology relationship to a multi-dimensional temporal constraint space composed of resource dimension and time axis, and define the task node to be scheduled as a parameterized task package. The spatial occupancy envelope and spatiotemporal boundary threshold of the parameterized task package in the multi-dimensional temporal constraint space are calculated by the corresponding occupancy saturation. Step S4: Monitor the spatiotemporal boundary overlap status between different parameterized task packages in the multidimensional temporal constraint space in real time to identify resource preemption conflicts at the business level, and calculate the overlap weight of resource preemption conflicts in the spatiotemporal dimension based on the boundary overlap volume and resource scarcity index. Step S5: Invoke the multi-objective constraint solving algorithm. Under the logical constraint of satisfying the temporal dependency relationship of the task nodes to be scheduled, adjust the topology offset of the resource occupation window or safe time interval of different parameterized task packages in the multi-dimensional temporal constraint space according to the overlap weight, and generate a conflict-free global resource allocation strategy.

[0023] Preferably, generating a conflict-free global resource allocation strategy includes: step S21, obtaining the boundary change trend of parameterized task packages; step S22, constructing a multidimensional spatial correction vector in the multidimensional temporal constraint space to characterize the spatiotemporal displacement correction amount based on the overlap weight and the boundary change trend; step S23, using the multidimensional spatial correction vector to correct the spatiotemporal coordinates of the parameterized task packages, decoupling the resource scheduling deadlock of concurrent services, and outputting the corrected global resource scheduling instruction.

[0024] Preferably, the occupancy saturation is determined by the following formula: Where S is the occupancy saturation; n is the total number of bottleneck resource types; Let be the resource requirement of the task node to be scheduled for the i-th bottleneck resource; Let be the real-time capacity value of the i-th bottleneck resource; This is the preset weighting factor corresponding to the i-th type of bottleneck resource.

[0025] Preferably, it also includes a topology simplification step based on load decay: step S41, setting an occupancy saturation threshold; step S42, identifying connected branches in the initial task topology relationship whose occupancy saturation is lower than the occupancy saturation threshold; step S43, folding the logical nodes of the connected branches and converting the connected branches into a constant resource consumption model to reduce the computational load of the multi-objective constraint solving algorithm in subsequent steps.

[0026] Preferably, in step S5, the topology offset adjustment of the resource occupancy window of the parameterized task package includes: step S51, calculating the conflict repulsion force based on the overlap weight; step S52, driving the path reset direction of the resource scheduling channel of the parameterized task package in the multi-dimensional temporal constraint space according to the conflict repulsion force; step S53, when the path reset direction is detected to be restricted, performing asynchronous phase translation on the conflicting nodes to eliminate physical resource competition.

[0027] Preferably, in step S3, the parameterized task package is represented by a rectangular envelope box, a spherical envelope box, or a polygonal envelope box, and different types of task nodes to be scheduled correspond to envelopes of different geometric shapes in the multidimensional temporal constraint space.

[0028] Preferably, the length of the major axis of the rectangular envelope box is determined by the expected duration of the task node to be scheduled in the time dimension, and the width of the minor axis of the rectangular envelope box is determined by the space occupancy envelope width of the parameterized task package determined by the occupancy saturation.

[0029] Preferably, in step S1, the task nodes to be scheduled include traffic flow guidance tasks, power grid load peak shaving tasks, and security monitoring linkage tasks, and different types of task nodes have preset task priorities.

[0030] Preferably, in step S5, the multi-objective constraint solving algorithm is based on a directed acyclic graph search strategy to find the collision-free scheduling path with the minimum comprehensive cost consisting of resource consumption cost and time delay cost in the multi-dimensional temporal constraint space.

[0031] Preferably, after generating a conflict-free global resource allocation strategy, the method further includes: step S101, collecting resource consumption feedback data of each subsystem of the smart city in real time; step S102, correcting the resource channel load weight in the multi-dimensional time-series constraint space based on the resource consumption feedback data, and feeding it back to step S1 as a correction parameter of the real-time available capacity vector to optimize the subsequently generated global resource allocation strategy.

[0032] Example 1: Under concurrent operational conditions integrating traffic guidance in a large-scale urban core area, local power grid load shaving, and emergency rescue for fires in high-rise buildings, heterogeneous subsystems experience concurrent preemption pressure on communication bandwidth, right-of-way sensing nodes, and public computing resources within the same control cycle. Due to overlapping physical resource requirements of standardized workflow engines from different departments, and multi-dimensional topological dependencies between traffic guidance nodes and medical rescue route planning nodes, rescue tasks stagnate due to resource locking, resulting in a technical state of coexisting local resource exhaustion and global resource idleness. The system obtains the real-time available capacity vector of the smart city's multi-dimensional resource library, the resource specification requirement vector of the task nodes to be scheduled, and the initial task topology relationship composed of the logical dependencies between the task nodes to be scheduled. By comparing the numerical mapping relationship between the resource specification requirement vector and the real-time available capacity vector, the system calculates the saturation S, which characterizes the intensity of the task nodes' requests for various bottleneck resources. The saturation S follows the formula: Where S is the occupancy saturation; n is the total number of bottleneck resource types; Let be the resource requirement of the task node to be scheduled for the i-th bottleneck resource; Let be the real-time capacity value of the i-th bottleneck resource; This is the preset weighting factor corresponding to the i-th type of bottleneck resource.

[0033] The system maps the initial task topology to a multi-dimensional temporal constraint space composed of resource dimensions and a time axis, and defines the task nodes to be scheduled as parameterized task packages. Occupancy saturation S is used as a feature parameter to control the spatiotemporal expansion of the task nodes to be scheduled within the virtual space, determining the spatial occupancy envelope and spatiotemporal boundary threshold of the parameterized task packages within the multi-dimensional temporal constraint space. The system monitors the spatiotemporal boundary overlap status between different parameterized task packages within the multi-dimensional temporal constraint space in real time to identify resource preemption conflicts at the business level, and calculates the overlap weight of resource preemption conflicts in the spatiotemporal dimension based on the boundary overlap volume and resource scarcity index. A multi-objective constraint solving algorithm is then invoked to satisfy the requirements of the task nodes to be scheduled. Under the logical constraints of node temporal dependencies, topological offset adjustments are made to the resource occupancy windows or safe time gaps of different parameterized task packages based on overlap weights. By introducing multi-dimensional spatial misalignment vectors between interfering primitives, resource scheduling deadlocks in concurrent businesses are decoupled, resulting in a conflict-free global resource allocation strategy. After the system issues resource scheduling instructions, the medical rescue workflow obtains a priority execution sequence without interference, the average waiting delay of cross-departmental workflows is reduced, and the overall turnover rate of global standard resources is improved. By converting the high-concurrency logical resource preemption state into a computable geometric interference problem within the auxiliary design space, the system transforms the business scheduling process into an active constraint blueprint design process.

[0034] Example 2: In a government affairs area resource scheduling test environment constructed using a discrete event simulation model, the system runs on a total capacity of 500 TFLOPS computing power. and a total capacity of 10Gbps bandwidth resources The test used a virtual cluster; the test data came from anonymized real-time logs from the city operation center, and Gaussian white noise with a signal-to-noise ratio of 20dB was superimposed on the collected task specification parameters; the test verified the resource allocation effect of multiple heterogeneous service flows during high-concurrency peak periods. The value of the sampling period T depends on the frequency of change of service flow characteristics; where T is the sampling period, and when the characteristic frequency f of the monitored service flow increases, in order to meet the Nyquist sampling criterion and reduce the risk of signal aliasing, the setting value of the sampling period T is adjusted to the lower limit of its value range; where f is the characteristic frequency, for the traffic flow pulse condition with a characteristic frequency f of 10Hz, the sampling period T is set to The set value is 50ms; this setting balances the real-time requirements of signal acquisition with the computational load of the processor; the experiment sets up a control group and an experimental group. The control group adopts a static priority queuing method, while the experimental group adopts this resource management method; under the condition that the initial system load rate is less than 0.3, the allocation success rate of both the control group and the experimental group remains above 98.2%; when the business concurrency increases, causing the calculated occupancy saturation S to enter the range of 0.75 to 0.9, the control group experiences scheduling deadlock due to its inability to resolve the topological dependency between the road administration task and the power monitoring task, and its task processing latency increases from 15.2ms to 420.5ms.

[0035] The experimental group generated a parameterized task package when the occupancy saturation S reached a threshold of 0.7. The boundary threshold of the parameterized task package expanded with the increase of occupancy saturation S. Measurement data showed that when S increased from 0.75 to 0.85, the expansion of the spatial occupancy envelope in the time axis dimension increased from 5.1ms to 12.8ms. In the interference region where the spatiotemporal boundaries overlapped, the experimental group implemented an 8.5ms topology offset adjustment on the traffic guidance task package by calculating a multidimensional spatial correction vector. The task processing latency of the experimental group stabilized at around 45.6ms, without the latency growth trend seen in the control group, and was consistent with the preset weighting factor. Boundary tests show that when When the value is 0.15, the system's protection index for core rescue missions decreases, while the critical mission abandonment rate increases; among them, This is a preset weighting factor corresponding to the i-th type of bottleneck resource; when When set to 1.6, the increased space occupancy envelope led to a decrease in resource utilization from 85.4% to 62.1%; within a numerical window of 0.2 to 1.5, the system achieved a balance between scheduling determinism and resource turnover efficiency; the processed task flow showed a 75% reduction in scheduling oscillation amplitude under 20dB noise interference; the experimental results confirmed the stability of the technical approach of mapping logical conflicts to spatiotemporal geometric interference and performing topology decoupling when handling concurrent business.

[0036] Example 3: In a sudden emergency situation integrating urban high-rise building fire alarm perception and traffic guidance linkage control, the fire rescue workflow generates concurrent requests for right-of-way perception nodes, backbone network emergency bandwidth, and auxiliary decision computing resources. Due to the evolution of the disaster, the resource specification demand vector of the video backhaul link undergoes nonlinear jumps, and the standardized signal control nodes on the rescue vehicle's travel path and the traffic guidance tasks of regular social vehicles experience resource locking within the same control cycle. The system experiences scheduling interference between the hard constraint of meeting the timeliness of fire rescue and the elastic constraint of ensuring the operation of the regional traffic network, causing the queuing method to induce global oscillation of the topological dependency chain in the critical area of ​​resource scarcity.

[0037] The system acquires the real-time available capacity vector, the resource specification demand vector of the task nodes to be scheduled, and the initial task topology from the multi-dimensional resource database of the smart city. To establish a mapping between resource occupancy status and geometric space, the system defines the time series as the principal axis of the multi-dimensional temporal constraint space and abstracts the physical specifications of n bottleneck resources into mutually orthogonal dimensions. The system defines the task nodes to be scheduled as parameterized task packages with specific geometric attributes. The coordinate center of the parameterized task package in the multi-dimensional temporal constraint space is determined by the task start time and resource index, and its envelope span in the resource dimension depends on the element values ​​of the resource specification demand vector. By comparing the numerical mapping relationship between the resource specification demand vector and the real-time available capacity vector, the system calculates the occupancy saturation S, which characterizes the intensity of the bottleneck resource request by the task nodes to be scheduled. The occupancy saturation S follows the formula: Where S is the occupancy saturation; n is the total number of bottleneck resource types; The resource requirement of the task node to be scheduled for the i-th bottleneck resource is expressed in units of computing power or bandwidth rate. Let be the real-time capacity value of the i-th bottleneck resource; To determine the preset weighting factor for the i-th type of bottleneck resource, the system uses the occupancy saturation S as the spatial expansion coefficient of the geometric primitives, calculating the expansion boundaries of the parameterized task package along each spatial axis. When the system detects an intersection between the projection ranges of the road administration task package and the emergency rescue task package on the auxiliary computing power axis and time axis, a logical conflict occurs. The system calculates the overlap weight based on the Euclidean volume of the overlapping boundary region and the resource scarcity index. Based on the artificial potential field theory, the repulsive potential energy between spatial primitives is positively correlated with the overlapping volume and the environmental field strength. The system uses the following calculation formula... Determine the overlap weight, where W is a dimensionless overlap weight characterizing the repulsion strength, and its legal value is greater than zero. The overlapping volume of the three-dimensional orthogonal projection of the two parameterized task packages in the multidimensional temporal constraint space; The rated total physical capacity of the bottleneck resource corresponding to the conflict; This represents the real-time available capacity for the current sampling period.

[0038] The system initiates a multi-objective constraint solution procedure to decouple deadlock states and obtain the boundary variation trend of parameterized task packages. Based on the overlap weights and boundary variation trends, a multi-dimensional spatial correction vector is constructed in the multi-dimensional temporal constraint space to characterize the spatiotemporal displacement correction. The magnitude of this vector is positively correlated with the overlap weights, and its direction is the tangential direction to avoid interference regions. The spatiotemporal coordinates of the parameterized task packages are corrected using the multi-dimensional spatial correction vector. By generating displacement offsets on the time axis or switching equivalent paths in the resource dimension, a conflict-free global resource allocation strategy is generated. A transformation operation is issued to the physical control layer for the spatial geometric bias vector to obtain the atomic scheduling cycle of the underlying standardized workflow engine. Extract the projection magnitude Δt of the multidimensional space correction vector onto the time axis, and then apply the formula... Determine the discrete delay number of ticks, where N is a positive integer representing the number of ticks in the converted instruction; This represents the rounding up operation, shifting the trigger timestamp of the target task node in the directed acyclic graph by N atomic scheduling cycles. It triggers the token bucket suspension action of the underlying concurrency controller by overwriting the resource scheduling channel control register. When the latency tick number N is detected to be greater than the set business tolerance threshold, it triggers the alternative path allocation mechanism. It reads the multi-dimensional space correction vector and maps it directly to the physical media access control address index of the backup computing cluster using the resource dimension scalar bias. It then sends a routing update message to complete the underlying data flow switch and outputs the corrected global resource scheduling command. The system introduces an online compensation loop, which monitors the rate of change of the first derivative of the real-time available capacity vector V to correct the preset weight factor in real time. The system adjusts the corresponding weighting factor when the capacity of a resource continuously decreases due to physical link losses. By increasing the exclusive envelope radius of such task packages in space, suppressing scheduling stagnation caused by resource contention, and generating corrected global resource allocation instructions, after the system completes topology offset adjustment and issues scheduling instructions, the emergency rescue task flow obtains a definite low-latency execution path under the premise that the traffic load imbalance rate in the guaranteed area is less than 15%, and the comprehensive turnover rate of cross-departmental standard resources is maintained above 88.5%. By transforming the risk of resource preemption at the logical level into a quantifiable geometric interference and vector correction process in the spatiotemporal constrained space, the system enables the scheduling logic to achieve closed-loop self-healing under extreme conditions.

[0039] Example 4: When deploying standardized resource management logic to new business nodes that include heterogeneous communication protocols and dynamic topology logic, the system calls an offline stress calibration procedure based on bottleneck tension before initiating the resource allocation process. By injecting simulated resource specification demand vectors of different magnitudes into the task nodes to be scheduled and monitoring physical response latency, the system determines the sensitivity coefficient of bottleneck resources under specific business loads and uses it as a preset weighting factor. The initial assignment is based on simulating at least 50 sets of traffic surge pulses with occupancy intensity, recording the resource lock-in time under each set of traffic surge pulses, and fitting the spatial expansion coefficient of the parameterized task package using the rate of change of the second derivative of the resource lock-in time and the resource demand. The offline pressure calibration procedure is executed in an isolated test network segment containing the traffic generator and nanosecond-level hardware timestamp probes. The specific calibration steps are as follows: the simulated load is continuously increased from 0% to 100% with a 5% resource occupancy increment, and 50 sets of response delay values ​​are continuously recorded at each increment node; the response delay difference between two adjacent increment nodes is calculated and divided by the 5% increment to obtain the sensitivity slope of the resource under the current load; this slope is mapped to a numerical range of 0.2 to 1.5 as a preset weighting factor. Initially, a baseline resource specification demand vector is injected into the tested node to record the basic response delay. The flow generator injects the resource demand in increments according to a set constant step size. The hardware timestamp probe synchronously collects the corresponding resource locking time. The first derivative is extracted by calculating the ratio of the resource locking time increment to the demand increment between adjacent probe points. The second derivative is calculated by differentiating adjacent first derivatives. The coordinate point where the second derivative is greater than zero and maintained for three consecutive probe cycles is determined as the physical saturation critical point. The demand scalar corresponding to the saturation critical point is read. The slope value of the preceding linear growth interval is extracted as the corresponding bottleneck resource sensitivity coefficient. The preset basic geometric envelope radius is multiplied by the sensitivity coefficient to calculate the characteristic space expansion coefficient corresponding to different concurrency intensities, generating a quantitative calibration lookup table that characterizes the mapping relationship between resource saturation state and geometric boundary.

[0040] When the system faces the business restructuring of cross-regional government collaboration and multi-level medical resource scheduling, it uses on-site pre-calibration procedures based on historical load fluctuations to redefine the basic range of the multi-dimensional time-series constraint space. After receiving the initial task topology, it analyzes the silent load records of the perception layer nodes to determine the static deviation of the real-time available capacity vector V, and uses the static deviation as the displacement offset value of the multi-dimensional coordinate system. Within a 200ms detection window, it uses the feedback scheduling deviation rate to correct the judgment threshold of the overlap weight, so that the geometric contour of the parameterized task package maintains the integrity of the logical dependency relationship under different physical topologies. After completing the on-site baseline calibration, the system enters a stable operation state of resource scheduling, the emergency rescue task flow obtains a determined low-latency transmission path, and the comprehensive turnover rate of cross-departmental standard resources is maintained above 88.5%.

[0041] Example 5: In situations involving the long-term operation of a massive urban digital infrastructure and the backlog of inactive business nodes, the system utilizes a topology simplification procedure based on load decay to reduce computational overhead within the multi-dimensional temporal constraint space. It acquires the historical access frequency and resource usage records of each task node to be scheduled, calculates a decay factor E to characterize node failure trends, and shows a negative correlation between the decay factor E and the node's inactivity duration ΔT. When the inactivity duration ΔT exceeds a preset active period, the system automatically reduces the logical weight of the node in the initial task topology relationship and removes the corresponding parameterized task package from the current geometric interference calculation sequence. This shrinks the complex task topology relationship into a core business skeleton containing only high-frequency interactive actions, generating simplified resource allocation instructions. This time-sensitive dynamic topology pruning method reduces the memory usage of the multi-objective constraint solving algorithm by approximately 28.5% when handling concurrent emergency medical rescue services, and maintains the scheduling latency of cross-departmental standard resources within a stable range of 45ms.

[0042] During the system's processing of heterogeneous service scheduling requests, the spatiotemporal boundary overlap between parameterized task packages is monitored. The gradient vector field of the interference region is calculated to determine the direction and magnitude of the multidimensional spatial correction vector D, where D is the multidimensional spatial correction vector. The magnitude of the multidimensional spatial correction vector D is determined based on the product of the overlap volume and the resource scarcity index, and it points to the backward extension of the line connecting the centroids of the interference primitives. The system applies the multidimensional spatial correction vector D to correct the spatiotemporal coordinates of the parameterized task packages, and iterates through 5 control cycles to approximate the distribution point that minimizes global constraint conflicts until the total interference volume in the multidimensional temporal constraint space is lower than the convergence threshold. This generates a conflict-free global resource allocation strategy, and the medical rescue workflow and traffic guidance logic are topologically decoupled on shared computing power nodes, reducing the conflict rate of cross-departmental standard resource allocation to 3.2%.

[0043] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.

Claims

1.A resource management method based on a smart city, characterized in that, Includes the following steps: Step S1: Obtain the real-time available capacity vector of the smart city multi-dimensional resource library, the resource specification requirement vector of the task nodes to be scheduled, and the initial task topology relationship composed of the logical dependencies between the task nodes to be scheduled. Step S2: By comparing the numerical mapping relationship between the resource specification demand vector and the real-time available capacity vector, calculate the occupancy saturation, which is used to characterize the intensity of the request for various bottleneck resources by the task node to be scheduled, and use the occupancy saturation as a feature parameter to control the degree of spatiotemporal expansion of the task node to be scheduled in the virtual space. Step S3: Map the initial task topology relationship to a multi-dimensional temporal constraint space composed of resource dimension and time axis, and define the task node to be scheduled as a parameterized task package. The spatial occupancy envelope and spatiotemporal boundary threshold of the parameterized task package in the multi-dimensional temporal constraint space are calculated by the corresponding occupancy saturation. Step S4: Monitor the spatiotemporal boundary overlap status between different parameterized task packages in the multidimensional temporal constraint space in real time to identify resource preemption conflicts at the business level, and calculate the overlap weight of resource preemption conflicts in the spatiotemporal dimension based on the boundary overlap volume and resource scarcity index. Step S5: Invoke the multi-objective constraint solving algorithm. Under the logical constraint of satisfying the temporal dependency relationship of the task nodes to be scheduled, adjust the topology offset of the resource occupation window or safe time interval of different parameterized task packages in the multi-dimensional temporal constraint space according to the overlap weight, and generate a conflict-free global resource allocation strategy. 2.The resource management method based on smart city according to claim 1, wherein, The process of generating a conflict-free global resource allocation strategy includes: step S21, obtaining the boundary change trend of parameterized task packages; step S22, constructing a multidimensional spatial correction vector in the multidimensional temporal constraint space to characterize the spatiotemporal displacement correction amount based on the overlap weight and the boundary change trend; and step S23, using the multidimensional spatial correction vector to correct the spatiotemporal coordinates of the parameterized task packages, decoupling the resource scheduling deadlock of concurrent services, and outputting the corrected global resource scheduling instruction. 3.The resource management method based on smart city according to claim 1, wherein, It also includes a topology simplification step based on load decay: step S41, setting an occupancy saturation threshold; step S42, identifying connected branches in the initial task topology where the occupancy saturation is lower than the occupancy saturation threshold. Step S43: Fold the logical nodes of the connected components and convert the connected components into a constant resource consumption model to reduce the computational load of the multi-objective constraint solving algorithm in subsequent steps. 4.The resource management method based on smart city according to claim 1, wherein, In step S5, the topology offset adjustment of the resource occupancy window of the parameterized task package is performed, including: step S51, calculating the conflict repulsion force based on the overlap weight; step S52, driving the path reset direction of the resource scheduling channel of the parameterized task package in the multi-dimensional temporal constraint space according to the conflict repulsion force; step S53, when the path reset direction is detected to be restricted, performing asynchronous phase translation on the conflicting nodes to eliminate physical resource competition. 5.The resource management method based on smart city according to claim 1, wherein, In step S3, the parameterized task package is represented by a rectangular envelope box, a spherical envelope box, or a polygonal envelope box, and different types of task nodes to be scheduled correspond to envelopes of different geometric shapes in the multidimensional temporal constraint space. 6.The resource management method based on smart city according to claim 5, wherein, The length of the major axis of the rectangular envelope box is determined by the expected duration of the task node to be scheduled in the time dimension, and the width of the minor axis of the rectangular envelope box is determined by the space occupancy envelope width of the parameterized task package determined by the occupancy saturation. 7.The resource management method based on smart city according to claim 1, wherein, In step S1, the task nodes to be scheduled include traffic flow guidance tasks, power grid load peak shaving tasks, and security monitoring linkage tasks, and different types of task nodes have preset task priorities. 8.The resource management method based on smart city according to claim 1, wherein, In step S5, the multi-objective constraint solving algorithm is based on the directed acyclic graph search strategy to find the collision-free scheduling path with the minimum comprehensive cost consisting of resource consumption cost and time delay cost in the multi-dimensional temporal constraint space. 9.The resource management method based on smart city according to claim 1, wherein, After generating a conflict-free global resource allocation strategy, the process also includes: step S101, collecting resource consumption feedback data from each subsystem of the smart city in real time; step S102, correcting the resource channel load weights in the multi-dimensional time-series constraint space based on the resource consumption feedback data, and feeding them back to step S1 as correction parameters of the real-time available capacity vector to optimize the subsequently generated global resource allocation strategy.

Images