Disclosure of Invention
Aiming at the defects, the invention provides the dam-break flood emergency evacuation path priority planning method based on the improved NSGA-III, solves the problem that the existing evacuation path planning method is difficult to consider multi-objective optimization, evacuation priority grading and refuge point capacity constraint under the dam-break flood situation with strong burst property and wide influence range, realizes the differential evacuation path optimization of high, medium and low priority groups, improves the overall evacuation efficiency and emergency response capability, and maximally ensures the life safety of personnel.
In order to solve the technical problems, the invention adopts the following technical scheme:
the dam break flood emergency evacuation path priority planning method based on the improved NSGA-III comprises the following steps:
step S1, collecting relevant data and parameter information of a research area, and determining a digital elevation model diagram, a satellite image diagram, a river distribution diagram, a road network diagram and dam structure parameter information of the research area;
s2, setting parameters of the HEC-RAS model to perform dam break simulation;
Setting extreme flood scene and breach parameters in an HEC-RAS model based on the acquired digital elevation model diagram, satellite image diagram, river distribution diagram, road network diagram and dam structure parameters, carrying out two-dimensional unsteady flow dam break simulation in the HEC-RAS module, outputting a result file of a maximum flood submerged depth diagram, and importing a simulation result to an ArcGIS Pro platform;
Step S3, selecting evacuation points and difficulty avoidance points by combining flood inundation and grading;
In the ArcGIS Pro platform, combining a satellite image map, a road network map, a flood submerged depth map, a high-resolution population density map and a building block vector map, and comprehensively analyzing to select evacuation points and refuge points;
s4, constructing topology, calculating all shortest paths and caching;
Constructing a road network topological structure by utilizing a Python NetworkX tool, calculating the shortest path distance from each evacuation point to all avoidance points based on Dijkstra algorithm, and caching calculation results;
step S5, based on the space position of the evacuation points and the priority weights thereof, performing weighted K-means cluster initialization to construct an initial population solution of an NSGA-III algorithm;
step S6, the initial population solution of the NSGA-III algorithm is updated through multi-objective optimization iteration of the NSGA-III algorithm;
Under the NSGA-III multi-objective optimization algorithm framework, taking a minimized multi-objective function as an objective, and combining with refuge point capacity limitation, iteratively updating an initial population solution to obtain an initial Pareto solution set;
step S7, fine-tuning a Pareto solution set by a heuristic strategy;
Introducing a heuristic fine tuning strategy to the initially obtained Pareto solution set, namely carrying out local exchange optimization on the distribution relation between evacuation points and refuge points based on path length comparison results among the evacuation points with different priorities, and executing load exchange optimization aiming at the evacuation points with path lengths exceeding a preset improvement threshold;
s8, obtaining and visualizing an optimal evacuation path structure;
The heuristic fine tuning strategy is optimized to obtain an optimal Pareto solution set, an optimal refuge point corresponding to each evacuation point and a path allocation scheme thereof are output, and the final evacuation path result is imported into an ArcGIS Pro platform for visual display to generate results such as a distribution map, an evacuation pressure thermodynamic diagram and the like of each priority evacuation path for emergency dispatch reference.
Further, the step S2 of performing the two-dimensional unsteady flow dam break simulation in the HEC-RAS module comprises the following steps:
Drawing a range of a reservoir water Storage Area by using a Storage Area tool, defining a flood simulation Area boundary by using a 2D Flow Area tool, setting dam axis and breach parameters by using an SA/2D Connection tool, setting simulation working conditions including an initial water level, a dam break water level, simulation time length and time step, and starting unsteady Flow simulation after constructing a two-dimensional calculation grid.
Further, in the step S3, a process of selecting evacuation points and avoiding difficulties in the ArcGIS Pro platform includes the following steps:
based on the satellite image map, the building block vector map and the high-resolution population density map, intersections and open places of a high population gathering area of a flood inundation area are selected as evacuation points, and difficulty avoidance points are arranged in resident areas and large public facility areas outside the flood influence range.
Further, when the NetworkX tool of Python is used to construct the road network topology in the step S4, the method includes the following steps:
Firstly, the preprocessed road data are converted into graph structures, wherein each road section corresponds to a side with weight in the graph structure, the weight of the side is the actual road distance, and the nodes represent road intersections or key geographic positions. Then, the evacuation area and the position of the shelter are projected to the nearest node in the network map through a space mapping method, so that the corresponding relation between the evacuation points and the shelter points in the road network is established;
The method comprises the steps of ensuring that the topological structure of a network accords with the actual road passing logic, avoiding unreasonable cross connection, checking the connectivity of the network, carrying out connectivity correction if a non-connected segment exists, confirming that all evacuation points and evacuation points are located on nodes, and adjusting if the evacuation points and the evacuation points are not in accord.
Further, in the step S4, the Dijkstra algorithm calculates a shortest path distance process from each evacuation point to all avoidance points, including the following steps:
Sequentially calculating the path lengths from all evacuation points to each refuge point by using a Dijkstra algorithm, and storing a binary group consisting of the evacuation points and the refuge points and the path lengths thereof into a cache structure such as a hash table and the like so as to support subsequent quick inquiry and repeated use and avoid repeated calculation paths of each model iteration;
The two groups of evacuation points and refuge points are ordered pairing of one evacuation point and one refuge point, and are used for uniquely identifying the corresponding relation between the two positions, and in the path cache, each of the two groups represents a specific evacuation path, and the shortest path length corresponding to the specific evacuation path is stored in the hash table as a numerical value.
Further, the method for initializing the weighted K-means cluster in the step S5 comprises the following steps:
setting weight according to the priority of the evacuation points, executing weighted clustering to generate initial grouping, matching the evacuation points closest to each clustering center as default evacuation targets, completing initial evacuation point distribution, and constructing an initial population solution of NSGA-III algorithm.
Further, the objective functions adopted by the NSGA-III multi-objective optimization in step S6 include:
the total length of all evacuation paths;
Consider the total path length of the priority weighting;
the total length of the high priority evacuation path.
Further, the objective function calculation process adopted by the NSGA-III multi-objective optimization is as follows:
;
wherein, theDecision variables, N is the total number of evacuation points 192, reviewed inAn evacuation point index which represents the allocation of the ith evacuation point, S is the total number of the evacuation points 32; Evacuation pointTo the refuge pointIs the shortest path length of (a); representing flood priority weightCorresponding to the weight [10, 5, 1 ]); a high priority evacuation point set; penalty terms are constrained for capacity.
Further, the heuristic fine-tuning strategy is divided into three stages:
step one, after the NSGA-III calculates the preliminary Pareto solution set in step S7, if the path length of a certain high-priority evacuation point is found to exceed a certain low or medium-priority point, performing the evacuation point allocation exchange operation, and when the path length of the medium-priority evacuation point is significantly greater than that of the low-priority evacuation point, performing the allocation exchange to improve the high-priority evacuation efficiency;
Step two, after the operation of the step one is finished, if the path length of a certain high-priority evacuation point is found to exceed a certain low-priority point, continuing to execute the exchange operation of the evacuation point until no more exchange pairs can be obviously improved or the preset iteration times are reached, and obtaining an optimized Pareto solution set;
And thirdly, screening out extreme evacuation points with path length exceeding a preset threshold value on the optimized Pareto solution set obtained in the second stage, traversing all the avoidance points, screening out candidate avoidance points which do not exceed capacity limit and can shorten the path to the maximum extent, replacing the original allocation avoidance points with the candidate avoidance points, and updating allocation information of the corresponding paths and the avoidance points.
Dam break flood emergency evacuation path priority planning system based on improved NSGA-III, comprising:
The parameter determining unit is used for acquiring digital elevation model diagrams, satellite image diagrams, river distribution diagrams, road network diagrams and dam structure parameter information of the research area;
the flood simulation unit is used for performing two-dimensional unsteady flow dam break simulation in the HEC-RAS and deriving a simulation result;
The space point selection unit is used for comprehensively analyzing the flood influence range, the building distribution and the population density in the ArcGIS Pro platform, scientifically selecting evacuation points and refuge points and dividing the priority class of the evacuation points;
The path pre-calculation unit is used for constructing a road network topological structure and caching the shortest paths between all evacuation points and refuge points through a Dijkstra algorithm;
the cluster initialization unit is used for generating an initial group based on a weighted K-means method and constructing an NSGA-III initial population solution;
The multi-objective optimization unit is used for executing path optimization under an NSGA-III multi-objective optimization framework to obtain an initial Pareto solution set;
The heuristic fine tuning unit is used for carrying out high-priority path tuning and extreme path load exchange optimization on the initial solution set;
the result visualization unit is used for exporting the optimal Pareto solution set distribution result between the evacuation points and the refuge points to form a file, and displaying the distribution map of each priority evacuation path and the evacuation pressure thermodynamic diagram in the ArcGIS Pro platform.
Compared with the prior art, the invention has the following technical effects:
According to the dam-break flood emergency evacuation path planning method and system based on the improved NSGA-III, dam-break flood simulation, evacuation point and refuge point selection, multi-objective optimization and heuristic strategies are fused, efficient evacuation path optimization for people with different priorities can be achieved under the extreme dam-break scene, the overall evacuation efficiency and the priority are improved, and an evacuation optimization model adapting to the spatial distribution characteristics of flood risks is constructed by carrying out two-dimensional unsteady flow dam-break simulation in HEC-RAS and combining with scientific selection and priority division of the evacuation point and the refuge point by an ArcGIS Pro platform.
The method introduces a path caching mechanism and weighted K-means initialization under an NSGA-III multi-objective optimization framework, remarkably improves the known convergence speed and feasibility, and simultaneously, carries out local fine adjustment on the preliminary solution set through heuristic priority exchange and load adjustment strategies to further optimize the evacuation efficiency of high-priority groups. Compared with the traditional evacuation method based on shortest path or single objective optimization, the method can more reasonably consider the shortest path, the priority and the capacity limit of the refuge points, and provide high-timeliness and operability decision support for dam-break flood disaster emergency response.
Detailed Description
The present invention is further described below with reference to examples, but it should not be construed that the scope of the above subject matter of the present invention is limited to the following examples. Various substitutions and alterations are made according to the ordinary skill and familiar means of the art without departing from the technical spirit of the invention, and all such substitutions and alterations are intended to be included in the scope of the invention.
An embodiment, as shown in fig. 1, of a dam break flood emergency evacuation path priority planning method based on an improved NSGA-III, includes the following steps:
Step S1, collecting relevant data and parameter information of a research area;
digital Elevation Model (DEM) map, satellite image map, river profile map, road network map and dam structure parameter information of the study area are determined.
The dam structure parameter information comprises dam elevation, dam length, dam width, reservoir capacity-water level curve, design flood level and water storage capacity.
S2, setting parameters of the HEC-RAS model to perform dam break simulation;
Based on the acquired digital elevation model diagram, satellite image diagram, river distribution diagram, road network diagram and dam structure parameters, setting extreme flood scene and breach parameters in an HEC-RAS model, carrying out two-dimensional unsteady flow dam break simulation on the HEC-RAS module, outputting a result file of a maximum flood submerged depth diagram, and importing a simulation result to an ArcGIS Pro platform.
The crumple parameters comprise crumple mode, crumple position, crumple shape, crumple weir flow coefficient and crumple forming time.
The method comprises the steps of drawing a reservoir water Storage Area range by using a Storage Area tool, defining a flood simulation Area boundary by using a 2D Flow Area tool, setting dam axis and breach parameters by using an SA/2D Connection tool, setting simulation working conditions including initial water level, dam break water level, simulation duration, time step and the like, constructing a two-dimensional calculation grid, and then starting up the non-constant Flow simulation.
Step S3, selecting evacuation points and difficulty avoidance points by combining flood inundation and grading;
In the ArcGIS Pro platform, a satellite image map, a road network map, a flood submerging depth map, a high-resolution population density map and a building block vector map are combined, evacuation points and refuge points are selected after comprehensive analysis, and meanwhile, the evacuation points are divided into three priority levels of high, medium and low according to the flood submerging depth and influence degree.
Specifically, 80 high-priority evacuation points, 56 medium-priority evacuation points with the water depth of 5-10 meters, 56 low-priority evacuation points with the water depth of 5 meters and less than 192 evacuation points and 32 difficult avoidance points are screened according to the influence of more than 10 meters of water depth.
After the HEC-RAS finishes dam break flood evolution simulation, grid results such as maximum water Depth (Max Depth) and the like can be exported through RAS MAPPER, and then the grid results are imported into an ArcGIS Pro platform to manufacture and display the flood Depth map. The high resolution population density map is from a publicly published chinese seven-population grid dataset (resolution 100 m) that has been publicly shared at the Figshare platform. The building block vector diagram is generated by processing a building boundary extraction algorithm based on Google Earth remote sensing images (with the spatial resolution of 0.5 m) acquired in 2020-2022, and can effectively represent the spatial distribution and morphological characteristics of buildings in a research area. The data set is publicly available and can be downloaded and acquired from Zenodo databases and is used for supporting the space references selected by evacuation points and refuge points.
The process for selecting evacuation points and avoiding difficulties in the ArcGIS Pro platform comprises the following steps:
based on the satellite image map, the building block vector map and the high-resolution population density map, intersections and open places of a high population gathering area of a flood inundation area are selected as evacuation points, and difficulty avoidance points are arranged in resident areas and large public facility areas outside the flood influence range.
In order to scientifically and reasonably select evacuation points and refuge points, researching and combining a road network diagram and a satellite image diagram, in a high-density building area within the influence range of flood, visually interpreting satellite images to preferentially lay the evacuation points at road intersections and key nodes, and meanwhile, setting up the refuge points in a resident concentrated area outside the influence range of flood. In order to further verify the rationality of evacuation points and refuge points, 100 m resolution population density map and building block vector data are introduced for research, and although the data have certain limitations in spatial resolution and precision, the data can be used as auxiliary references. Based on the superposition analysis of the data, the preliminarily set evacuation points and the refuge points are properly adjusted so as to enhance the scientificity and rationality of the spatial distribution of the evacuation points and the refuge points.
S4, constructing topology, calculating all shortest paths and caching;
and constructing a road network topological structure by utilizing a Python NetworkX tool, calculating the shortest path distance from each evacuation point to all the avoidance points based on a Dijkstra algorithm, and caching the calculation result.
The construction of the road network topology by using the NetworkX tool of Python comprises the following steps:
Firstly, the preprocessed road data (including the information of the start point, the end point, the distance and the like of the road) are converted into graph structures, wherein each road section corresponds to a weighted side in the graph structure, the weight of the side is the actual road distance, and the node represents the road intersection or the key geographic position. And then, projecting the positions of the evacuation area and the shelter to the nearest node in the network map by a space mapping method, so that the corresponding relation between the evacuation points and the shelter points in the road network is established.
The method comprises the steps of ensuring that the topological structure of a network accords with the actual road passing logic, avoiding unreasonable cross connection, checking network connectivity, carrying out connectivity correction if a non-connected segment exists, confirming that all evacuation points and evacuation points are located on network nodes, and adjusting if the evacuation points and the evacuation points are not in accord.
The Dijkstra algorithm calculates the shortest path distance process from each evacuation point to all avoidance points, including:
and sequentially calculating the path lengths from all evacuation points to each evacuation point by using a Dijkstra algorithm, and storing the (evacuation points, difficulty avoidance points) binary groups and the path lengths thereof into a cache structure such as a hash table and the like so as to support subsequent quick inquiry and repeated use and avoid repeated calculation paths of each model iteration.
The (evacuation points, escape points) doublets refer to ordered pairing consisting of one evacuation point and one escape point, and are used for uniquely identifying the corresponding relation between the two positions, and in the path cache, each such doublet represents a specific evacuation path, and the corresponding shortest path length is stored as a numerical value in the hash table. By the method, the path length from any evacuation point to the evacuation point can be rapidly inquired in the subsequent model iteration, and repeated calculation is avoided.
And S5, based on the spatial position of the evacuation points and the priority weights thereof, performing weighted K-means cluster initialization to construct an initial population solution of the NSGA-III algorithm.
The weighted K-means clustering initialization method comprises the steps of setting weights according to priorities of evacuation points, executing weighted clustering to generate initial groups, matching the evacuation points closest to each clustering center as default evacuation targets, completing initial evacuation point distribution, and constructing an initial population solution of an NSGA-III algorithm.
Step S6, the initial population solution of the NSGA-III algorithm is updated through multi-objective optimization iteration of the NSGA-III algorithm;
Under the NSGA-III multi-objective optimization algorithm framework, taking a minimized multi-objective function as an objective, and combining with refuge point capacity limitation, iteratively updating an initial population solution to obtain an initial Pareto solution set;
the objective functions adopted by the NSGA-III multi-objective optimization comprise:
;
wherein, theDecision variables, N is the total number of evacuation points 192, reviewed inAn evacuation point index which represents the allocation of the ith evacuation point, S is the total number of the evacuation points 32; Evacuation pointTo the refuge pointIs the shortest path length of (a); representing flood priority weightCorresponding to the weight [10, 5, 1 ]); a high priority evacuation point set; penalty terms are constrained for capacity.
Step S7, fine-tuning a Pareto solution set by a heuristic strategy;
And introducing a heuristic fine tuning strategy to the initially obtained Pareto solution set, namely carrying out local exchange optimization on the distribution relation between the evacuation points and the refuge points based on the path length comparison results among the evacuation points with different priorities, and executing load exchange optimization aiming at the evacuation points with the path length exceeding a preset improvement threshold.
The heuristic fine-tuning strategy is divided into three phases:
after the NSGA-III calculates the preliminary Pareto solution set, if the path length of a certain high-priority evacuation point is found to be greater than a certain low-priority or medium-priority point, performing the evacuation point allocation exchange operation, and when the path length of the medium-priority evacuation point is significantly greater than the low-priority evacuation point, performing the allocation exchange to improve the high-priority evacuation efficiency.
And step two, after the operation of the step one is finished, if the path length of a certain high-priority evacuation point is found to exceed a certain low-priority point, continuing to execute the exchange operation of the evacuation point until no more exchange pairs can be obviously improved or the preset iteration times are reached, and obtaining the optimized Pareto solution set.
And thirdly, screening out extreme evacuation points with path length exceeding a preset threshold value on the optimized Pareto solution set obtained in the second stage, traversing all the avoidance points, screening out candidate avoidance points which do not exceed capacity limit and can shorten the path to the maximum extent, replacing the original allocation avoidance points with the candidate avoidance points, and updating allocation information of the corresponding paths and the avoidance points.
The Pareto solution set is based on an original NSGA-III algorithm, and the optimal solution set obtained after three-stage local path adjustment and refuge point redistribution optimization has higher evacuation efficiency and more reasonable priority path distribution.
S8, obtaining and visualizing an optimal evacuation path structure;
The heuristic fine tuning strategy is optimized to obtain an optimal Pareto solution set, an optimal refuge point corresponding to each evacuation point and a path allocation scheme thereof are output, and the final evacuation path result is imported into an ArcGIS Pro platform for visual display to generate results such as a distribution map, an evacuation pressure thermodynamic diagram and the like of each priority evacuation path for emergency dispatch reference.
In order to further understand the technical solution of the present invention, the method of the present invention is described below by way of specific examples. In this example, a dam and a water outlet area of a reservoir are selected as research areas for implementation, and positions of escape points and evacuation points with three priority levels of high, medium and low are selected, as shown in fig. 2, and the specific implementation process is as follows:
1) Acquiring information such as a Digital Elevation Model (DEM) diagram, a satellite image diagram, a river distribution diagram, a road network diagram, dam structure parameters and the like of a research area;
2) Setting extreme flood situations and breach parameters in the HEC-RAS two-dimensional hydrodynamic model, carrying out two-dimensional unsteady flow dam break simulation, and outputting result files such as a maximum flood submerged depth map;
3) The simulation result is imported into an ArcGIS Pro platform, and the evacuation points and the refuge points are scientifically selected after comprehensive analysis by combining a satellite image map, a road network map, a flood submerging depth map, a population density map and a building block vector map, and meanwhile, the evacuation points are divided into three priority levels of high, medium and low according to the flood submerging depth and influence degree.
Figure 3 shows the evacuation path distribution of the priority avoidance points of the improved NSGA-III model. The points of different shapes in the figure represent evacuation points of different flood priorities, black squares are high priority (. Gtoreq.10m), grey dots are medium priority (5-10 m), and grey triangles are low priority (< 5 m). The modified NSGA-III model after heuristic fine tuning enables the path length of the evacuation points of three flood priorities to show obvious grading effect. It can be seen from the figure that the path lengths of the high priority evacuation points are mostly concentrated in relatively short intervals, medium priority ones, low priority ones are distributed relatively wider. Meanwhile, the extreme maximum value of the evacuation path is reduced to about 10000 meters, so that the occurrence of overlong paths is effectively avoided, and the rationality of the evacuation path planning is improved.
Fig. 4 shows an evacuation path plan based on evacuation points under the influence of a flood depth ∈10m, i.e. the evacuation path distribution of high priority evacuation points. The evacuation route is connected with the high-priority evacuation points and the refuge points at different positions, so that the evacuation planning layout of the high-priority evacuation points is intuitively presented, the evacuation route arrangement formulated for the high-priority area when the dam-break flood risk is dealt with is reflected, and the understanding of the evacuation direction of the high-priority evacuation points and the distribution situation of the accessible refuge points is facilitated.
Based on the same inventive concept, the dam-break flood emergency evacuation path planning method, system and device based on the improved NSGA-III according to the above embodiment of the present application, correspondingly, another embodiment of the present application further provides dam-break flood emergency evacuation path priority planning based on the improved NSGA-III, as shown in fig. 5, which is a schematic diagram of the system structure, the system includes:
A parameter determining unit 501, configured to obtain information such as a digital elevation model map, a satellite image map, a river distribution map, a road network map, and a dam structure parameter of a research area;
a flood simulation unit 502, configured to perform two-dimensional unsteady flow dam break simulation in the HEC-RAS, and derive a simulation result;
The space point selection unit 503 is configured to comprehensively analyze a flood influence range, building distribution and population density in the ArcGIS Pro platform, scientifically select evacuation points and refuge points, and divide priority classes of the evacuation points;
The path pre-calculation unit 504 is configured to construct a road network topology structure, and buffer the shortest paths between all evacuation points and evacuation points through Dijkstra algorithm;
the cluster initialization unit 505 is configured to generate an initial group based on a weighted K-means method, and construct an NSGA-III initial population solution;
A multi-objective optimization unit 506, configured to perform path optimization under an NSGA-III multi-objective optimization framework, and obtain an initial Pareto solution set;
the heuristic fine tuning unit 507 is configured to perform high-priority path tuning and extreme path load switching optimization on the initial solution set;
the result visualization unit 508 is configured to export a file of an optimal Pareto solution set allocation result between the evacuation points and the refuge points, and display an evacuation path distribution map and an evacuation pressure thermodynamic diagram of each priority in the arcgipro platform.
The description of the present invention has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the invention in the form disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, and to enable others of ordinary skill in the art to understand the invention for various embodiments with various modifications as are suited to the particular use contemplated.