Path optimization and intelligent resource scheduling method and system based on digital twinning

Abstract

The invention discloses a path optimization and resource intelligent scheduling method and system based on digital twinning. The method comprises the following steps: constructing a digital twinning three-dimensional model of disaster area infrastructure to realize three-dimensional scene reconstruction; constructing a first-aid repair path optimization model on the basis of a multi-objective function to optimize paths among the sites in the digital twinborn three-dimensional model; a job resource multi-target scheduling model is established, and intelligent resource scheduling of the digital twin three-dimensional model is realized; setting constraint conditions of scheduling and execution; updating the multi-target path optimization model and the job resource multi-target scheduling model once every other scheduling cycle by adopting a model prediction control idea; according to the method, collaborative optimization of the first-aid repair path and resource scheduling can be realized, an optimal scheme is generated by combining terrain and risk constraints, and the first-aid repair operation efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of infrastructure repair and emergency dispatch, in particular to a path optimization and resource intelligent scheduling method and system based on digital twinning. BACKGROUND

[0002] Currently, after natural disasters occur, infrastructure is often damaged to varying degrees, and traditional manual inspection and path planning methods are slow to react, lack global perception and dynamic decision-making capabilities, and cannot quickly respond and adjust path planning in post-disaster environments. These methods usually rely on static data and lack dynamic perception of real-time changes after disasters, making it difficult to respond to changing disaster conditions. In addition, traditional systems lack precise modeling and three-dimensional visualization functions for post-disaster infrastructure, making it difficult to accurately perceive damage conditions and resource states in disaster areas, affecting the optimization of repair paths and the rational scheduling of resources. Although some systems have introduced information technology for task management, these systems often cannot update post-disaster scenarios in real time, lack sufficient intelligent optimization capabilities, and cannot closely integrate path planning and resource scheduling, resulting in low scheduling efficiency and the inability to make quick decisions and adjustments when disaster conditions change.

[0003] Digital twinning technology has the ability to map physical and virtual, and can build accurate digital models of post-disaster infrastructure. The combination of path planning algorithms and resource scheduling strategies can improve repair efficiency and scheduling scientificity. Therefore, it is an urgent need for technology development to build a solution that integrates "digital twinning + path optimization + resource scheduling". SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a path optimization and resource intelligent scheduling method and system based on digital twinning, which aims to solve the problems in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solution: a path optimization and resource intelligent scheduling method based on digital twinning, comprising the following steps: Step S1: constructing a digital twinning three-dimensional model of disaster area infrastructure to realize three-dimensional scene reconstruction; Step S2: constructing a repair path optimization model based on a multi-objective function to optimize the paths between each location in the digital twinning three-dimensional model; Step S3: establishing a multi-objective scheduling model for work resources to realize intelligent scheduling of resources in the digital twinning three-dimensional model; Step S4: setting constraints for scheduling and execution, including personnel / device attendance constraints, path accessibility constraints, task completion time constraints, personnel capability matching constraints, and dynamic post-disaster obstacle constraints; Step S5: Using the model predictive control concept, the multi-objective path optimization model and the job resource multi-objective scheduling model are updated every scheduling cycle.

[0006] Furthermore, the specific process of step S2 is: Step S2.1: Based on the graph model, the disaster area roads / scenes in the digital twin 3D model are treated as a weighted directed graph. ,in, Represents a set of vertices, representing locations in the digital twin 3D model, Represents an edge set, representing the paths between locations in the digital twin 3D model. represents the weight function, which represents the original cost of the path; Step S2.2: Use the improved Dijkstra algorithm to find the starting point The minimum cost path to all vertices is the shortest path; for any vertex , defined from the starting point arrive The total cost of the minimum cost path for: ; Where, Indicates all starting points arrive Path; Indicates the path An edge in ; express The weight of The shortest path satisfies the Bellman optimality principle: ; Where, Indicates that from the vertex arrive 's directed edges; Indicates from arrive The total cost of the minimum cost path; Represents an edge The weight of Step S2.3: Determine the multi-objective attributes of path optimization, including length, time consumption, resource handling / transportation cost, and path risk index; quantify each edge in the weighted directed graph The multi-objective attributes of The multi-dimensional comprehensive cost : ; Where, Represents an edge The actual length of denotes the length of edge denotes the expected time consumption of edge denotes the resource carrying / transporting cost of edge denotes the risk index of edge , , , denote the weighted coefficients of length, time consumption, resource and risk, respectively; Step S2.4: Construct a target function of path optimization aiming at minimizing the path comprehensive cost: ; wherein, denotes the multi-dimensional comprehensive cost of path ; denotes the length cost sum of path , which is obtained by accumulating of all edges in the path; denotes the time consumption cost sum of path , which is obtained by accumulating of all edges in the path; denotes the resource carrying / transporting cost sum of path , which is obtained by accumulating of all edges in the path; denotes the risk cost sum of path , which is obtained by accumulating of all edges in the path; By minimizing , the path with the optimal comprehensive cost under the disaster scene is obtained.

[0007] Further, the feasible path is optimized based on genetic algorithm; the multi-dimensional comprehensive cost constructed in step S2.3 is taken as the fitness function, and the following process is executed: Initial feasible path generation: run step S2.2 and step S2.3 to obtain an initial feasible path ; Take as an individual, and start the genetic algorithm to optimize the feasible path: Generate an initial population by k-shortest perturbation in the neighborhood of ; Perform cross operation, mutation operation and selection operation on the initial population in turn; When the preset maximum evolution number is reached or the fitness values of two consecutive generations are less than the preset threshold, stop, and select the optimal individual in the current population as the optimized feasible path.

[0008] ​​Further, when the digital twin model detects new obstacles or new repair tasks, the following are performed: Edge weight temporary adjustment: Set the risk cost of the corresponding edge to a maximum value in the affected road section; Local path repair: Re-run the improved Dijkstra algorithm only on the affected subgraph in the weighted directed graph to obtain a local repair path; Population update and re-evolution: Inject the local repair path into the population of the genetic algorithm as a new individual and restart the optimization process.

[0009] Further, the specific process of step S3 is as follows: Step S3.1: Convert the resource scheduling optimization into a "task-resource" matching problem; define: Resource set: , represents the class resource, represents the total number of resources; Task set: , represents the task, represents the total number of tasks; Decision variable: , when 0, the class resource is not assigned to the task , when 1, assigned to , , ; Step S3.2: Define the single-item cost of resource and task matching from multiple core dimensions respectively: Response cost: , represents the response time / presence time of the resource to the task ; Distance cost: , represents the spatial distance between the resource and the task ; Skill matching penalty: , represents the "degree of mismatch" between the ability of the resource and the requirements of the task , represents the skill set possessed by , represents the required skill set of , represents the penalty coefficient; Resource running cost term: , representing the running cost of the resource executing the task; Step S3.3: The multi-dimensional cost of matching the resource with the task in step S3.2 is fused through a weight coefficient to form an optimization objective function of the total cost: ; In the formula, in the formula, , , all represent scheduling preference weights.

[0010] Further, in step S4, the personnel / equipment attendance constraint is represented as: , representing a resource availability function, representing time; the path accessibility constraint is represented as: , representing a path accessibility function, all represent nodes of the path; the task completion time limit constraint is represented as: , representing a task time limit function; the personnel capability matching constraint is represented as: ; the dynamic post-disaster obstacle constraint is represented as: , representing an edge obstacle function.

[0011] Further, in step S5, the specific process of updating is: inputting the latest field image, equipment state and traffic information, re-optimizing and generating a scheduling instruction: Path planning update: considering newly appeared obstacles or new tasks, adjusting the path planning; Resource dynamic rescheduling: updating the real-time location and work progress of the resource, and accordingly performing dynamic rescheduling; Instruction and work order issuance: generating new path instructions and work orders, and issuing them to mobile terminals; Manual interaction support: allowing manual intervention in the scheduling process to support task merging or splitting operations.

[0012] A path optimization and resource intelligent scheduling system based on digital twinning, comprising: A scene modeling reconstruction module for constructing a digital twin three-dimensional model of the infrastructure in the disaster area to realize three-dimensional scene reconstruction; A path optimization module for constructing a repair path optimization model based on a multi-objective function to optimize the paths between each location in the digital twin three-dimensional model; A resource scheduling module for establishing a multi-objective scheduling model of work resources to realize intelligent scheduling of resources in the digital twin three-dimensional model; A constraint module is configured to set constraints for scheduling and execution, including personnel / device attendance constraints, path accessibility constraints, task completion time constraints, personnel capability matching constraints, and dynamic post-disaster obstacle constraints. A real-time updating module is configured to update the multi-objective path optimization model and the job resource multi-objective scheduling model every other scheduling period using model predictive control.

[0013] An electronic device includes a processor, a memory, and a bus, the processor and the memory being connected through the bus, wherein the memory is configured to store a set of program codes, and the processor is configured to invoke the program codes stored in the memory to execute a path optimization and resource intelligent scheduling method based on digital twinning.

[0014] A non-volatile computer storage medium stores computer executable instructions, and the computer executable instructions execute a path optimization and resource intelligent scheduling method based on digital twinning.

[0015] Compared with the prior art, the present application has the following beneficial effects:

[0016] (1) In the post-disaster environment where disaster conditions change dynamically, the digital twinning model can be used to map the infrastructure state in real time, and the fault can be accurately located and the global disaster situation can be perceived. This innovation can reflect the actual situation of post-disaster infrastructure in real time, while traditional methods usually rely on static data or manual inspection, which is slow and difficult to respond quickly to sudden changes.

[0017] (2) The collaborative optimization of repair path and resource scheduling is realized, and the optimal scheme is generated combined with the terrain and risk constraints to improve the efficiency of repair operation. The existing technology often handles path planning and resource scheduling separately, lacking systematicness and real-time optimization. The present application combines path planning and resource scheduling organically to ensure dynamic optimization of path and resource in complex post-disaster environment, thereby greatly improving the efficiency of repair.

[0018] (3) While considering the terrain restrictions and resource availability, the repair task automatic dispatching and resource dynamic allocation are realized to ensure the safety and efficiency of the repair process. Traditional resource scheduling relies on manual experience and lacks accurate matching and real-time performance, while the present application intelligently schedules and intelligently matches and allocates tasks and resources based on real-time data, avoiding resource waste and scheduling delay, and improving overall efficiency and safety.

[0019] (4) Through the dynamic updating and rolling optimization mechanism, the scheduling scheme can be continuously adjusted according to the changes on the scene, and quickly adapt to the environment with disaster changes. This function greatly improves the adaptability and robustness of the system in a complex and dynamic environment. Compared with the existing static optimization method, the present application has strong self-adaptive ability, can adjust in time according to the latest post-disaster image, equipment state and traffic information, and avoids the lag of the traditional scheme in the changing environment.

[0020] (5) Based on the multi-dimensional cost optimization objective function, compared with the traditional single-target path planning, the present application can comprehensively consider multiple factors such as path length, time consumption, resource consumption and safety, and realize all-round balance of path optimization and resource scheduling. This multi-objective optimization method can meet the needs of efficiency, safety and cost in the post-disaster repair process, far beyond the limitations of traditional path planning methods relying on a single target. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 The method flowchart of the present application. DETAILED DESCRIPTION

[0022] As shown in Figure 1 , the present application provides a technical solution: a path optimization and resource intelligent scheduling method based on digital twinning, comprising the following steps: Step S1: constructing a digital twinning three-dimensional model of disaster area infrastructure to realize three-dimensional scene reconstruction; In the post-disaster environment, various data sources such as unmanned aerial vehicle inspection images and GIS geographic information data are collected and utilized, point cloud reconstruction and image recognition technology are applied, and the BIM model is docked and fused to generate a digital twinning three-dimensional model of disaster area infrastructure.

[0023] The digital twinning three-dimensional model is used to provide an accurate virtual mapping, so that the specific state of the post-disaster infrastructure such as "broken pole", "disconnected line", "foundation displacement" and the like can be understood in real time, and accurate data support is ensured for subsequent path optimization and resource scheduling. This step generates a digital model by point cloud reconstruction, image recognition and BIM model docking, which provides a basis for subsequent path planning, task scheduling and resource allocation. The real-time updating capability of the digital twinning model ensures that the model is consistent with the actual situation on the scene.

[0024] Step S2: constructing a repair path optimization model based on a multi-objective function to optimize the paths between each place in the digital twinning three-dimensional model.

[0025] Step S2.1: abstracting the disaster area road / scene in the digital twinning three-dimensional model into a weighted directed graph based on a graph model , wherein, represents a set of vertices, representing locations such as repair task points, resource warehouses, etc. in the digital twin three-dimensional model, represents a set of edges, representing paths between locations in the digital twin three-dimensional model, represents a weight function, representing the original cost (such as time, distance, risk) of the path.

[0026] Step S2.2: Find the minimum cost path, i.e. the shortest path, from the starting point to all vertices using the improved Dijkstra algorithm; for any vertex , define the total cost of the minimum cost path from the starting point to as: ; wherein represents all paths from the starting point to ; represents an edge in the path ; represents the weight of .

[0027] The shortest path satisfies the Bellman Optimality principle: ; wherein represents the directed edge from vertex to ; represents the total cost of the minimum cost path from to ; represents the weight of edge .

[0028] Step S2.3: Determine the multi-objective attributes of path optimization, including length, time consumption, resource handling / transportation cost, and path risk indicators; Quantify the multi-objective attributes of each edge in the weighted directed graph, set the weighting coefficients, and construct the multi-dimensional comprehensive cost of edge : ; wherein represents the actual length of edge ; represents the expected time consumption of edge (taking into account traffic, waterlogging, etc.); represents the resource handling / transportation cost of edge .​​ denotes the edge ; , , , denote the weighted coefficients of length, time consumption, resource, and risk, respectively.

[0029] Step S2.4: Construct the objective function of path optimization with the goal of “minimum path synthesis cost”: ; wherein, denotes the multi-dimensional synthesis cost of the path ; denotes the length cost sum of the path , which is obtained by accumulating the of all edges in the path; denotes the time consumption cost sum of the path , which is obtained by accumulating the of all edges in the path; denotes the resource carrying / transportation cost sum of the path , which is obtained by accumulating the of all edges in the path; denotes the risk cost sum of the path , which is obtained by accumulating the of all edges in the path.

[0030] By minimizing , the path with the optimal synthesis cost under the disaster area scenario (while balancing multiple requirements such as length, time consumption, resource, and risk) can be obtained.

[0031] For the above post-disaster infrastructure repair path optimization objective function based on digital twinning, it needs to be pointed out that the function is constructed around the core requirements of post-disaster repair “efficiency, safety, and cost”, and the expression can be simplified as the total path cost equal to the weighted sum of distance cost, time cost, resource cost, and safety penalty cost. Among them, the length cost is quantified by the product of and , the purpose is to reduce invalid detours and avoid the “near but not accessible” problem caused by traditional distance as the only target under the premise of ensuring feasibility; the time cost is composed by considering the product of and , which can increase the weight when repairing in key load areas to ensure the fastest speed to reach the fault point and solve the disadvantages of traditional static path “only distance not actual time consumption”; the resource cost is composed by the product of and , The weight can be increased in resource shortage to prefer low-cost routes and reduce post-disaster resource waste; the safety penalty cost is the product of the safety risk and the safety weight and , which is significantly increased in high-risk scenarios such as earthquakes and floods, and is forced to avoid dangerous areas to avoid the problem of ignoring safety risks in traditional path planning. Compared with traditional artificial experience planning or single-target algorithms, the function can adjust the path strategy according to the disaster changes (such as new obstacles and risk area expansion) in each scheduling period through multi-dimensional weighting and dynamic weight adjustment combined with real-time data updating of the digital twin model, solves the problem that traditional methods cannot cope with dynamic post-disaster scenarios, and realizes the coordinated optimization of efficient, safe and economic repair paths, and improves the response speed and operation safety of post-disaster infrastructure repair.

[0032] wherein the feasible path is optimized based on a genetic algorithm (the path planning is combined with an improved Dijkstra algorithm and a genetic algorithm to balance global optimization and local feasibility, and supports path re-planning and path obstacle avoidance for sudden tasks).

[0033] The multi-dimensional comprehensive cost constructed in step S2.3 is used as the fitness function, and the following process is executed: 1. Initial feasible path generation: run steps S2.2 and S2.3 to obtain an initial feasible path .

[0034] 2. Take the initial feasible path as an individual, start the genetic algorithm, and optimize the feasible path: In the neighborhood of the initial feasible path, an initial population is generated by k-shortest perturbation. The initial population is sequentially subjected to crossover, mutation and selection operations; wherein the crossover operation uses PMX crossover (partial match crossover) to ensure the legality of the vertex order after crossover; the mutation operation uses a combination of single-point mutation and local re-routing; and the selection operation combines tournament selection and elite retention strategy to avoid loss of optimal solution.

[0035] In the genetic algorithm, the fitness value is usually required to be "the larger the better". The optimization objective here is "minimization of multi-dimensional comprehensive cost", and the fitness value is a numerical value that measures the "goodness" of the path; the larger the value, the better the path.

[0036] When the maximum evolution generation is reached or the individual fitness value of two consecutive generations is less than the preset threshold, the process is stopped, and the optimal individual in the current population is selected as the optimized feasible path.

[0037] When the maximum evolution generation is reached or the individual fitness value of two consecutive generations is less than the preset threshold, the process is stopped, and the optimal individual in the current population is selected as the optimized feasible path.

[0038] ​When the digital twin model detects new obstacles (dynamic post-disaster obstacle constraint triggers) or new repair tasks appear: 1. Temporary adjustment of edge weights: Set the risk cost of the corresponding edge to a maximum value (prohibit path selection of this section) in the affected road section.

[0039] 2. Local path repair: Only re-run the improved Dijkstra algorithm in the affected subgraph of the weighted directed graph to quickly obtain a local repair path.

[0040] 3. Population update and re-evolution: Inject the local repair path into the population of genetic algorithm and use it as a new individual to restart the optimization process to ensure that the global optimality is not destroyed.

[0041] Step S3: Establish a multi-objective scheduling model of operation resources to realize intelligent scheduling of resources in the digital twin three-dimensional model.

[0042] Step S3.1: Convert resource scheduling optimization into a "task-resource" matching problem; define: Resource set: , (including various resources such as people, vehicles, and equipment), denotes the type of resource, denotes the total number of resources.

[0043] Task set: , (repair and operation tasks to be performed, etc.), denotes the task, denotes the total number of tasks.

[0044] Decision variable: , when the type of resource is not assigned to the task, , when , , .

[0045] Step S3.2: Define the single-item cost of resource and task matching from multiple core dimensions (time, space, skill fit, operation cost) to provide a basis for subsequent total cost fusion, as follows:

[0046] Response cost (time dimension) , represents the response time / arrival time of the resource to the task (reflects the scheduling response efficiency).

[0047] Distance cost (spatial dimension) , represents the spatial distance between resources and tasks , embodying the spatial transfer cost.

[0048] Skill matching penalty (ability matching dimension) , represents the "degree of mismatch" between the ability of resources and the requirements of tasks , represents the set of skills possessed by the resource , represents the set of skills required by the task , represents the penalty coefficient.

[0049] Resource operation cost term (economic cost dimension) , represents the operation cost of resources performing tasks (such as vehicle fuel cost, labor cost, etc.), the calculation formula is "unit use cost × task estimated time".

[0050] Step S3.3: The multi-dimensional cost of matching resources and tasks in step S3.2 is fused through weight coefficients to form the optimization objective function of total cost: ; In the formula, , , all represent scheduling preference weights, used to balance the relative importance of "spatial distance, response time, skill matching" three types of cost.

[0051] For the above resource scheduling optimization objective function, it needs to be pointed out that this function is constructed around the core demand of "task-resource" accurate matching, and the expression can be simplified as the total scheduling cost equal to the weighted sum of spatial distance cost, response time cost and ability matching penalty cost. Among them, and quantify the product, the purpose is to prioritize scheduling nearby resources, reduce round-trip time, and avoid the low efficiency caused by the traditional scheduling "seek far and abandon near"; the response time cost is composed of the product of and , which can increase the weight in hospital, communication base station and other key tasks to ensure fast response and solve the problem of traditional manual dispatch response lag; the ability matching penalty cost is the product of and , Significant improvement in professional tasks such as high-pressure operation, equipment replacement, etc. Force matching resources with corresponding qualifications to avoid rework or safety accidents caused by traditional scheduling "resource mismatch". Compared with the traditional resource allocation method relying on human experience, this function can adjust the allocation strategy according to the resource location, skill change in each scheduling cycle through multi-factor quantization and dynamic weight adjustment, combined with real-time updating of resource state of digital twin model, solve the problem of low resource utilization and poor matching accuracy of traditional method, realize the "fast, adaptive, efficient" collaborative scheduling of repair resources, and improve the resource utilization efficiency and task completion quality of post-disaster infrastructure repair.

[0052] The objective function comprehensively considers response speed, path cost and resource fit degree, solves the resource allocation sequence through genetic algorithm, realizes optimal resource matching and automatic generation of scheduling work order.

[0053] Step S4: Set the constraints of scheduling and execution, including personnel / equipment attendance constraints, path accessibility constraints, task completion time limit constraints, personnel ability matching constraints and dynamic post-disaster obstacle constraints.

[0054] The personnel / equipment attendance constraint is represented as: , represents the resource availability function, represents time; the constraint means that only "currently available" personnel or equipment can be scheduled (ensure that the resource can be actually called).

[0055] The path accessibility constraint is represented as: , represents the path accessibility function, both represent the nodes of the path; the constraint means that the nodes (or points) of the path must be located in the "post-disaster marked accessible area" (ensure that the physical path is feasible).

[0056] The task completion time limit constraint is represented as: , represents the task time limit function; the constraint means that critical repair tasks need to be completed within a "specified time window (deadline )" (ensure the timeliness of the task).

[0057] The personnel ability matching constraint is represented as: ; the constraint means that the allocated resources (such as personnel) must have "all the abilities required to complete the task" (ensure that the task can be effectively executed).

[0058] The dynamic post-disaster obstacle constraint is represented as: , represents the edge obstacle function; this constraint means that in the post-disaster environment, obstacles may appear at any time (such as landslides, collapsed buildings, water accumulation, etc.). In the path planning process, traditional methods often only consider static obstacle data, and cannot cope with the rapidly changing environment after the disaster. The dynamic post-disaster obstacle constraint requires the system to be able to identify and update obstacle information in real time, ensuring that the latest obstacles are avoided during path planning.

[0059] For the constraint conditions in the post-disaster infrastructure repair path and resource scheduling system based on digital twinning, it needs to be pointed out that these constraints are key restrictions to ensure the feasibility, safety and efficiency of the repair scheme, mainly including personnel / equipment attendance constraints, path accessibility constraints, task completion time constraints, personnel capability matching constraints and dynamic post-disaster obstacle constraints. Among them, the attendance constraint ensures that only available resources are dispatched to avoid repeated dispatching; the accessibility constraint ensures that the path is physically passable based on real-time terrain data from digital twinning, and prevents "feasible on paper but blocked on site"; the time limit constraint sets a time window for critical tasks to prioritize core load recovery; the capability matching constraint requires resources to have corresponding qualifications to avoid accidents caused by "mismatching"; the dynamic post-disaster obstacle constraint ensures that the path planning always avoids the latest risk areas by real-time identification and updating of sudden obstacles such as landslides, water accumulation and collapsed buildings. Compared with the traditional problem of missing or single constraints, these multi-dimensional constraints are integrated into the objective function penalty term and take effect simultaneously, ensuring the safety and feasibility bottom line while retaining dynamic adjustment flexibility, solving the contradiction between traditional "constraint rigidity or no constraint out of control".

[0060] Step S5: Using the model predictive control (MPC) idea, update the multi-objective path optimization model and the multi-objective operation resource scheduling model every other scheduling period.

[0061] The specific process of updating is: input the latest field image, equipment status and traffic information, re-optimize and generate scheduling instructions:

[0062] Path planning update: considering newly appeared obstacles or new tasks, adjust the path planning. Resource dynamic rescheduling: update the real-time location and operation progress of resources, and accordingly perform dynamic rescheduling.

[0063] Instruction and work order issuance: generate new path instructions and work orders and issue them to mobile terminals (such as terminals of repair personnel and equipment).

[0064] Artificial interaction support: allow human intervention in the scheduling process to support task merging or splitting operations.

[0065] This rolling optimization strategy can realize the adaptive evolution of the scheduling scheme and improve the resilience and adaptability of the system in the dynamic post-disaster environment.

[0066] A path optimization and resource intelligent scheduling system based on digital twinning, comprising: A scene modeling reconstruction module for constructing a digital twin three-dimensional model of disaster area infrastructure to realize three-dimensional scene reconstruction; A path optimization module for constructing a path optimization model based on a multi-objective function to optimize paths between locations in the digital twin three-dimensional model; A resource scheduling module for establishing a multi-objective scheduling model to realize intelligent scheduling of resources in the digital twin three-dimensional model; A constraint module for setting constraints for scheduling and execution, including personnel / device attendance constraints, path accessibility constraints, task completion time constraints, personnel capability matching constraints, and dynamic post-disaster obstacle constraints; A real-time update module for updating the multi-objective path optimization model and the multi-objective scheduling model of work resources once every scheduling period using model predictive control.

[0067] An electronic device comprising a processor, a memory, and a bus, wherein the processor and the memory are connected through the bus, wherein the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute a path optimization and resource intelligent scheduling method based on digital twinning.

[0068] A non-volatile computer storage medium, the computer storage medium storing computer executable instructions, the computer executable instructions executing a path optimization and resource intelligent scheduling method based on digital twinning.

[0069] Although embodiments of the present application have been shown and described, it is to be understood that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A path optimization and resource intelligent scheduling method based on digital twins, characterized in that: The steps include: Step S1: Build a digital twin 3D model of the disaster area infrastructure to achieve 3D scene reconstruction; Step S2: constructing a repair path optimization model based on a multi-objective function to optimize the paths between various locations in the digital twin 3D model; Step S3: Establish a multi-objective scheduling model for job resources to realize intelligent resource scheduling in the digital twin 3D model; Step S4: Set scheduling and execution constraints, including personnel / equipment attendance constraints, path accessibility constraints, task completion time constraints, personnel capability matching constraints, and dynamic post-disaster obstacle constraints; Step S5: Using the model predictive control concept, the multi-objective path optimization model and the job resource multi-objective scheduling model are updated every scheduling cycle.

2. The method for path optimization and intelligent resource scheduling based on digital twins according to claim 1, characterized in that: The specific process of step S2 is: Step S2.1: Based on the graph model, the disaster area roads / scenes in the digital twin 3D model are treated as a weighted directed graph. ,in, Represents a set of vertices, representing locations in the digital twin 3D model, Represents an edge set, representing the paths between locations in the digital twin 3D model. represents the weight function, which represents the original cost of the path; Step S2.2: Use the improved Dijkstra algorithm to find the starting point The minimum cost path to all vertices is the shortest path; for any vertex , defined from the starting point arrive The total cost of the minimum cost path for: ； Where, Indicates all starting points arrive Path; Indicates the path An edge in ; express The weight of The shortest path satisfies the Bellman optimality principle: ； Where, Indicates that from the vertex arrive ; Indicates from arrive The total cost of the minimum cost path; Represents an edge The weight of Step S2.3: Determine the multi-objective attributes of path optimization, including length, time consumption, resource handling / transportation cost, and path risk index; quantify each edge in the weighted directed graph The multi-objective attributes of The multi-dimensional comprehensive cost : ； Where, Represents an edge The actual length of Represents an edge Estimated time required; Represents an edge The cost of resource handling / transportation; Represents an edge risk indicators; 、 、 、 Represent the weighting coefficients of length, time, resources and risk respectively; Step S2.4: With the goal of minimizing the overall cost of the path, construct the objective function for path optimization: ； Where, Indicates the path The multi-dimensional comprehensive cost; Indicates the path The sum of the length costs of all edges on the path Accumulate and obtain; Indicates the path The sum of the time cost of all edges on the path Accumulate and obtain; Indicates the path The sum of the resource handling / transportation costs is determined by the Accumulate and obtain; Indicates the path The sum of the risk costs of all edges on the path Accumulate and obtain; By minimizing , and obtain the path with the optimal comprehensive cost in the disaster area scenario.

3. The method for path optimization and intelligent resource scheduling based on digital twins according to claim 2, characterized in that: Optimize feasible paths based on genetic algorithms; Using the multi-dimensional comprehensive cost constructed in step S2.3 as the fitness function, execute the following process: Initial feasible path generation: Run steps S2.2 and S2.3 to obtain an initial feasible path ; by For individuals, start the genetic algorithm to optimize the feasible path: exist In the neighborhood of , the initial population is generated by k-shortest perturbation; Perform crossover, mutation, and selection operations on the initial population in sequence; When the preset maximum number of evolutionary generations is reached or the fitness value of two consecutive generations is less than the preset threshold, the process stops and the best individual in the current population is selected as the optimized feasible path.

4. The method for path optimization and intelligent resource scheduling based on digital twins according to claim 3 is characterized by: When the digital twin model detects a new obstacle or a new repair task, it executes: Temporary adjustment of edge weights: In the affected road section, the risk cost of the corresponding edge is set to the maximum value; Local path repair: rerun the improved Dijkstra algorithm only on the affected subgraph in the weighted directed graph to obtain the local repair path; Population update and re-evolution: Inject the local repair path into the population of the genetic algorithm and restart the optimization process as a new individual.

5. The method for path optimization and intelligent resource scheduling based on digital twins according to claim 4 is characterized in that: The specific process of step S3 is: Step S3.1: Convert resource scheduling optimization into a "task-resource" matching problem; definition: Resource Collection: , Indicates the Class resources, Indicates the total quantity of resources; Task collection: , Indicates the tasks, Indicates the total number of tasks; Decision variables: , when it is 0, the Class resources Not allocated to Task , when it is 1, Assigned to , , ; Step S3.2: Define the individual costs of matching resources with tasks from multiple core dimensions: Response cost: , indicating resources To the task Response time / arrival time; Distance cost: , indicating resources With the task The spatial distance between Skill Compatibility Penalty: , indicating resources Capabilities and tasks The required "degree of mismatch", express The skill set you have, express The required skill set, represents the penalty coefficient; Resource operating cost items: , represents the running cost of resources to perform tasks; Step S3.3: The multi-dimensional cost of matching resources and tasks in step S3.2 is integrated through weight coefficients to form the optimization objective function of the total cost: ； In the formula, in the formula, 、 、 Both represent scheduling preference weights.

6. The method for path optimization and intelligent resource scheduling based on digital twins according to claim 5, characterized in that: In step S4, the personnel / equipment attendance constraint is expressed as: , represents the resource availability function, represents time; the path accessibility constraint is expressed as: , represents the path accessibility function, Both represent nodes of the path; the task completion time constraint is expressed as: , Represents the task time limit function; Personnel capability matching constraints are expressed as: ; Dynamic post-disaster obstacle constraints, expressed as: , represents the edge obstacle function.

7. The method for path optimization and intelligent resource scheduling based on digital twins according to claim 6, characterized in that: In step S5, the specific process of updating is: inputting the latest on-site images, equipment status and traffic information, re-optimizing and generating dispatch instructions: Path planning update: taking into account new obstacles or new tasks and adjusting the path planning; Dynamic resource rescheduling: Update the real-time location of resources and job progress, and perform dynamic rescheduling accordingly; Instruction and work order issuance: Generate new route instructions and work orders and send them to mobile terminals; Manual interaction support: allows manual intervention in the scheduling process and supports task merging or splitting operations.

8. A path optimization and resource intelligent scheduling system based on digital twins, characterized by: include: The scene modeling and reconstruction module is used to build a digital twin 3D model of the disaster area infrastructure to achieve 3D scene reconstruction; The path optimization module is used to build an emergency repair path optimization model based on a multi-objective function to optimize the paths between various locations in the digital twin 3D model; The resource scheduling module is used to establish a multi-objective scheduling model for job resources and realize intelligent resource scheduling in the digital twin 3D model; The constraint module is used to set scheduling and execution constraints, including personnel / equipment attendance constraints, path accessibility constraints, task completion time constraints, personnel capability matching constraints, and dynamic post-disaster obstacle constraints; The real-time update module is used to update the multi-objective path optimization model and the multi-objective scheduling model of job resources every scheduling cycle using the model predictive control idea.

9. An electronic device, characterized in that: It includes a processor, a memory and a bus, wherein the processor and the memory are connected through the bus, wherein the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute a path optimization and resource intelligent scheduling method based on digital twins as described in any one of claims 1 to 7.

10. A non-volatile computer storage medium storing computer-executable instructions, characterized in that: The computer-executable instructions execute a path optimization and resource intelligent scheduling method based on digital twins as described in any one of claims 1-7.

Images