Disaster prevention and refuge green land system layout optimization method and system based on multi-agent

Abstract

The application discloses a multi-agent-based disaster prevention and safety green land system layout optimization method and system, and relates to the technical field of city planning.The application comprises the following steps: receiving original geographic data in a region to be optimized, and preprocessing the original geographic data to form effective geographic data, wherein the original geographic data comprises original residential area data, original road network data and original disaster prevention and safety green land data.The application introduces a multi-agent evacuation simulation model in the planning and design of disaster prevention and safety green land, and optimizes multiple targets such as road congestion and green land overload during simulation, so that the real-time escape situation can be accurately simulated, real-time disaster risks of roads, crowd congestion risks during escape and capacity overload risks of disaster prevention and safety green land during disasters can be avoided, an optimal escape scheme can be formulated, and the evacuation path and strategy can be dynamically adjusted according to different emergency demands, so that an optimal scheme under multiple targets can be formulated.

Description

Technical Field

[0001] This invention relates to the field of urban planning technology, specifically to a method and system for optimizing the layout of disaster prevention and mitigation green space systems based on multi-agent intelligence. Background Technology

[0002] With the acceleration of urbanization, urban population density is constantly increasing, and the risk of urban disasters is also rising. In the event of sudden disasters such as earthquakes, fires, and floods, how to efficiently organize the evacuation of people to safe refuge places, such as disaster prevention and mitigation green spaces, has become an important issue in urban emergency management.

[0003] Existing evacuation simulation methods mainly suffer from the following problems: During the evacuation process, road congestion will affect the movement speed and route selection of personnel in real time, and traditional methods mostly use macroscopic models that fail to provide real-time feedback and adjustment; at the same time, in actual evacuation, there is an uneven utilization of some disaster prevention and mitigation green spaces, with some areas having extremely high utilization efficiency and others having extremely low utilization efficiency. To address these issues, we propose a method and system for optimizing the layout of disaster prevention and mitigation green spaces based on multi-agent systems. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for optimizing the layout of disaster prevention and mitigation green space based on multi-agent systems. By introducing multi-agent simulation, path planning and dynamic adjustment mechanisms, and combining multi-objective optimization layout methods, it provides more scientific and accurate planning decision support for urban disaster management.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for optimizing the layout of a disaster prevention and mitigation green space system based on multi-agent systems, comprising the following steps:

[0006] Receive raw geographic data from the area to be optimized, preprocess the raw geographic data to form effective geographic data, including raw residential area data, raw road network data and raw disaster prevention and mitigation green space data.

[0007] Set evacuation targets, select specific evacuation targets and simulate with a multi-agent evacuation model, and the multi-agent evacuation model outputs the optimal evacuation plan;

[0008] Among them, the multi-agent evacuation model uses effective geographic data as the data source to simulate and record the real-time evacuation status. The real-time evacuation status includes the number of people entering each disaster prevention and mitigation green space and the usage efficiency, the evacuation routes of each community, and the number of road users and the degree of congestion of each section of the road network within each time step.

[0009] The optimal evacuation plan is compared with the original road network data and the original disaster prevention and mitigation green space data to optimize the layout of roads and disaster prevention and mitigation green spaces in the area to be optimized.

[0010] Furthermore, the geographic data within the area to be optimized is obtained based on an internet map open platform and a planning map, as detailed below:

[0011] Residential area data includes boundary SHP data and population of each residential community;

[0012] Road network data includes road geographic location information and grade attributes, as well as road width data;

[0013] Data on disaster prevention and mitigation green spaces includes geographical location data and maximum capacity data, where maximum capacity data = total effective disaster mitigation area of ​​disaster prevention and mitigation green spaces / per capita effective disaster mitigation area;

[0014] The preprocessing of raw geographic data specifically involves:

[0015] The above-mentioned geographic data are integrated and processed using GIS software, a unified geographic coordinate system is established, and missing and outlier values ​​are handled.

[0016] Furthermore, the pre-built multi-agent evacuation model includes an agent module, an environment module, and a communication module, wherein:

[0017] The intelligent agent module includes residential area intelligent agents, evacuation intelligent agents, and green space intelligent agents;

[0018] The environment module includes basic GIS environmental data and global real-time information, specifically including road network data, residential area boundary data, green space location and maximum capacity data, as well as dynamically updated road weights and green space status data;

[0019] The communication module is used to support information exchange and resource sharing among residential intelligent agents, evacuation intelligent agents and green space intelligent agents, and to ensure collaboration and synchronization between residential intelligent agents, evacuation intelligent agents and environmental modules.

[0020] Furthermore, taking the evacuation target as the objective, the optimal evacuation plan after simulation is output, which specifically includes: evacuation route planning, disaster prevention and mitigation green space capacity update, road congestion feedback update, iterative control and evacuation simulation result output, and modification of weight ratio according to achieving different evacuation targets to generate the optimal evacuation plan.

[0021] Furthermore, the evacuation route planning is as follows:

[0022] Based on the Mesa framework, we define residential area agents, evacuation agents, and green space agents and their attributes to simulate the dynamic decision-making process of individual residents in different evacuation scenarios.

[0023] Among them, the residential area agent is used to manage residential area data, road network data, and disaster prevention and mitigation green space data in the GIS basic data information database; the evacuation agent is used to store the initial geographic data of all residential areas, roads, and green spaces, and record the real-time evacuation status of each residential area during the simulation process; the green space agent is used to comprehensively evaluate the weights of roads and green spaces, sort them by distance, select the top 10 nearest disaster prevention and mitigation green spaces as the target destination candidate green space list, calculate the optimal evacuation path through the A* algorithm, and track and record the evacuation status at different time steps in real time.

[0024] Furthermore, the capacity update of the aforementioned disaster prevention and mitigation green space is detailed as follows:

[0025] Based on the number of agents entering the disaster prevention and mitigation green space, the remaining capacity of the green space is updated. When the capacity of the green space is insufficient, the agent module re-evaluates the shelter selection strategy, selects other nearby disaster prevention and mitigation green spaces with available capacity based on real-time environmental data, and regenerates the list for evacuation.

[0026] Furthermore, the road congestion feedback update follows these steps:

[0027] (71) The processed road network data is converted into graph structure data, where nodes represent road intersections and edges represent road segments, and the initial weights of the edges are set as follows:

[0028] The selection weight R for each road n Due to its risk factor D n And escape distance L n The decision was made jointly, and the specific calculation formula is as follows:

[0029] R n =D n ×L n

[0030] Escape distance L n The path length is calculated based on the A* optimal path algorithm, and the risk factor D is... n The road safety influencing factors were obtained after calculation and normalization using the objective entropy weight method.

[0031] (72) The intelligent agent module has planned and generated simulated evacuation routes, calculated the road congestion degree of each road segment at different time steps, and updated the weights of each road segment as follows:

[0032] The formula for calculating congestion level C is:

[0033]

[0034] In the formula, N represents the number of people escaping on the road segment. maxThe maximum number of passengers allowed on the road section;

[0035] The weights of road segments are adjusted based on the congestion level C. The new weights are as follows:

[0036] R n '=D n *L n *C

[0037] In the formula, R n 'For the new weights, D n Risk factor, L n Escape distance;

[0038] (73) The agent module performs the next iteration based on the updated road weights and recalculates the optimal evacuation path.

[0039] Furthermore, the iterative control and evacuation simulation results output are implemented through the following specific steps:

[0040] Set the evacuation target or maximum number of iterations as preset conditions. After each simulation, determine whether the preset conditions have been met and decide whether to continue iterating. Then perform multiple calculations and road congestion calculations to carry out the iterative simulation process until the preset conditions are met and the evacuation simulation results are output.

[0041] The evacuation simulation results include the evacuation time distribution of the intelligent agent module, the congestion data of each road section, the remaining capacity data of disaster prevention and mitigation green space, and the spatial distribution of the simulated evacuation routes.

[0042] The weighting ratios are adjusted based on achieving different evacuation objectives to generate the optimal evacuation plan, as detailed below:

[0043] The specific evacuation objectives include: the closest green space target, aiming for the shortest straight-line distance to the target green space; the total evacuation time target, aiming for the shortest total time; and the disaster prevention and mitigation green space utilization rate target, aiming for the highest comprehensive utilization rate of each disaster prevention and mitigation green space.

[0044] The various evacuation simulation schemes include: the closest green space scheme, which uses the straight-line distance to the residential area as the basis for ranking the target destination green space list, and uses the sum of the congestion coefficient and the danger coefficient as the road selection weight; the shortest evacuation time scheme, which uses the actual distance to the residential area as the basis for ranking the target destination green space list, and uses the sum of the road distance, danger coefficient, and congestion coefficient as the road selection weight; and the disaster prevention and mitigation green space personnel distribution balanced scheme, which uses the utilization efficiency of each green space as the basis for ranking the target destination green space list, prioritizes the use of the green space with the lowest utilization efficiency, and uses the sum of the danger coefficient and the congestion coefficient as the road selection weight.

[0045] Furthermore, the optimal evacuation plan is compared with the original road network data and the original disaster prevention and mitigation green space data to optimize the layout of roads and disaster prevention and mitigation green spaces in the area to be optimized, specifically including three aspects:

[0046] Optimization of new site selection for disaster prevention and mitigation green spaces: Based on areas with high residential density but insufficient coverage of existing green space services, an optimization plan for the selection of new green spaces is proposed, selecting plots with convenient transportation and low risk as candidate green spaces;

[0047] Existing disaster prevention and mitigation green spaces will be expanded and optimized: For existing green spaces that are overloaded with services, the area of ​​the green space will be expanded or evacuation facilities will be added.

[0048] Disaster prevention and escape route optimization: Based on the evacuation routes of the optimal evacuation plan, the main roads leading to the green space are modified or widened, and the road network structure is adjusted to optimize the connectivity of the routes.

[0049] According to a second aspect of the present invention, the present invention provides a multi-agent-based disaster prevention and mitigation green space system layout optimization system, used to implement the above-mentioned multi-agent-based disaster prevention and mitigation green space system layout optimization method, comprising:

[0050] The data processing module is used to receive the raw geographic data of the area to be optimized, and to preprocess the raw geographic data to form effective geographic data. The raw geographic data includes raw residential area data, raw road network data, and raw disaster prevention and mitigation green space data.

[0051] The scheme output module is used to set evacuation targets, select specified evacuation targets and simulate with a multi-agent evacuation model, and the multi-agent evacuation model outputs the optimal evacuation scheme.

[0052] Among them, the multi-agent evacuation model uses effective geographical data as the data source to simulate and record the real-time evacuation status. ， Real-time evacuation status includes the number of people entering each disaster prevention and mitigation green space and its usage efficiency within each time step, evacuation routes of each community, and the number of road users and congestion of each section of the road network.

[0053] The layout optimization scheme generation module compares the optimal evacuation plan with the original road network data and the original disaster prevention and mitigation green space data to optimize the layout of roads and disaster prevention and mitigation green spaces within the area to be optimized.

[0054] The solution output module includes:

[0055] The agent model building unit is used to construct a multi-agent evacuation model, which includes an agent module, an environment module, and a communication module. The agent module includes residential area agents, evacuation agents, and green space agents.

[0056] The path planning unit, based on the Mesa framework, defines residential area agents, evacuation agents, and green space agents and their attributes to simulate the dynamic decision-making process of individual residents in different evacuation scenarios. The residential area agent manages residential area data, road network data, and disaster prevention and mitigation green space data in the GIS basic data database; the evacuation agent stores the initial geographic data of all residential areas, roads, and green spaces, and records the real-time evacuation status of each residential area during the simulation; the green space agent dynamically calculates the optimal evacuation path by comprehensively evaluating the weights of roads and green spaces, and tracks and records the evacuation status at different time steps in real time.

[0057] The disaster prevention and mitigation green space capacity update unit is used to dynamically adjust the remaining capacity of the green space based on the real-time recorded evacuation situation. When the green space is overloaded, the intelligent agent module is triggered to select other nearby disaster prevention and mitigation green spaces with available capacity and replan the simulated evacuation route.

[0058] The road congestion feedback update unit is used to convert the processed road network data into graph structure data, where nodes represent road intersections and edges represent road segments. It sets the initial weights of the edges, and then calculates the road congestion of each road segment at different time steps based on the simulated evacuation path planned and generated by the agent module. It also updates the weights of the edges of each road segment. Then, the agent module performs the next iteration based on the updated road weights and recalculates the optimal evacuation path.

[0059] The iterative control unit is used to set the evacuation target or the maximum number of iterations as preset conditions, control the iterative process of the evacuation simulation of the intelligent agent module, and output the evacuation simulation results after the preset conditions are met.

[0060] The evacuation simulation scheme generation unit is used to set multiple evacuation objectives and control the intelligent agent module to generate multiple evacuation simulation schemes according to the weight ratio of achieving different evacuation objectives. Each evacuation simulation scheme is the optimization result under different weight ratios, and the evacuation simulation scheme is used to achieve a comprehensive objective or a single objective.

[0061] This invention has at least the following beneficial effects:

[0062] This invention introduces a multi-agent evacuation simulation model into the planning and design of disaster prevention and mitigation green spaces, and optimizes multiple objectives such as road congestion and green space overload during simulation. This not only accurately simulates real-time escape situations and mitigates real-time road disaster risks, evacuation crowd congestion risks, and disaster prevention and mitigation green space capacity overload risks during disasters, thus helping to formulate optimal escape plans, but also dynamically adjusts evacuation routes and strategies according to different emergency needs, formulating optimal solutions under multi-objective conditions. This provides more scientific and accurate planning decision support for future urban pre-disaster emergency plan formulation and post-disaster response plans.

[0063] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0064] Figure 1 This is a flowchart illustrating the optimization method described in this invention;

[0065] Figure 2 This is a schematic diagram of the structure of the optimization system described in this invention;

[0066] Figure 3 This is a schematic diagram of the evacuation simulation results of the closest green space to a certain area in an embodiment of the present invention, taken 30 minutes.

[0067] Figure 4 This is a schematic diagram of the evacuation simulation results of the shortest evacuation time scheme of 30 minutes in a certain area in an embodiment of the present invention;

[0068] Figure 5 This is a schematic diagram of the evacuation simulation results of a 30-minute simulation of a disaster prevention and mitigation green space population equalization scheme in a certain region, as described in an embodiment of the present invention. Detailed Implementation

[0069] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0070] Example 1:

[0071] Please see Figure 1 and Figure 2 This invention provides a technical solution: a method for optimizing the layout of a disaster prevention and mitigation green space system based on multi-agent systems, comprising the following steps:

[0072] S1. Data Reception and Processing: Receive geographic data from the area to be optimized. The geographic data includes residential area data, road network data, and disaster prevention and mitigation green space data. Based on GIS software, integrate and process the above geographic data, unify the geographic coordinate system, handle missing values ​​and outliers, and construct a complete GIS basic data information database.

[0073] The geographic data was obtained based on the Internet map open platform and planning map. The residential area data includes boundary shape data and population of each residential community. The population was calculated by substituting the population data of the study area streets in the seventh national population census into the ratio of the building area of ​​the residential community to the building area of ​​the street.

[0074] Road network data includes extracted road data within the target area, including road geographic location information and grade attributes, and road width data;

[0075] The data on disaster prevention and mitigation green spaces includes the geographical location data of the green spaces. The maximum capacity data of the green spaces is obtained by converting the effective mitigation area of ​​disaster prevention and mitigation green spaces in the "Design Guidelines for Disaster Prevention and Mitigation of Urban Green Spaces" into per capita area. Specifically: Maximum capacity data of green spaces = Total effective mitigation area of ​​disaster prevention and mitigation green spaces / Effective mitigation area per capita.

[0076] S2. Road Network Construction: The processed road network data is converted into graph structure data, where nodes represent road intersections and edges represent road segments. Initial weights for the edges are set, as follows:

[0077] The selection weight R for each road n Due to its risk factor D n And escape distance L n The decision was made jointly, and the specific calculation formula is as follows:

[0078] R n =D n ×L n

[0079] Escape distance L n The path length is calculated based on the A* optimal path algorithm, and the risk factor D is... n The road safety influencing factors are obtained after calculation and normalization using the objective entropy weight method. Specific influencing factors may include the number of medical facility POIs, police station POIs, gas station POIs, industrial zones, building collapse risk, and hospitals within a 500m buffer zone of each road.

[0080] It should be noted that when performing Geographic Information System (GIS) analysis, it is often necessary to convert map data into a graph structure in order to perform operations such as path planning and network analysis. For the technical solution of this embodiment, Python is used to convert map data into a graph structure.

[0081] S3. Construction of the agent evacuation model: Construct a multi-agent evacuation model, which includes an agent module, an environment module, and a communication module. The agent module includes residential area agents, evacuation agents, and green space agents.

[0082] Multi-agent systems define communication protocols and interaction rules between agents to simulate complex interaction processes, thereby achieving a comprehensive consideration of multi-objective optimization problems. Specifically, it can simultaneously optimize multiple dimensions such as evacuation time, route selection, and balanced utilization of refuge sites, ensuring that in emergency situations, it can not only effectively shorten the overall evacuation time and reduce casualties, but also avoid the problem of some refuge sites being overcrowded while other sites are idle by intelligently allocating refuge resources.

[0083] In the Jupyter Notebook environment, Python is used to model and simulate agent-based multi-agent evacuation models. Jupyter Notebook provides a write-and-run development approach, which facilitates rapid adjustment of code logic and parameters. Python's visualization library can quickly and dynamically display the evacuation process and results during subsequent scheme comparison. At the same time, open-source Python libraries (such as mesa, networkx, and geopandas) are used to build agent behavior logic and simulation processes, quickly establishing a multi-agent simulation framework.

[0084] The multi-agent evacuation model consists of three main parts: agents, environment module, and communication module. The model environment module includes GIS basic environmental data and global real-time information, specifically road network, residential area boundaries, green space location and capacity, as well as dynamically updated road weights and green space status data.

[0085] The communication module supports information interaction and resource sharing between intelligent agents, ensuring effective collaboration and synchronization between residential intelligent agents, evacuation intelligent agents and environmental data, thereby achieving efficient operation and dynamic adjustment of the entire evacuation simulation process;

[0086] S4. Evacuation Route Planning: Based on the Mesa framework, define residential area agents, evacuation agents, and green space agents and their attributes, and simulate the dynamic decision-making process of individual residents in different evacuation scenarios:

[0087] Among them, the residential area intelligent agent is used to manage residential area data, road network data and disaster prevention and mitigation green space data in the GIS basic data information database, specifically the geographical boundary information, population size, basic capacity information of disaster prevention and mitigation green space and basic road network data of each residential community;

[0088] The evacuation agent is used to store the initial geographic data of all residential areas, roads and green spaces, and to record the real-time evacuation status of each residential area during the simulation. During the simulation, the residential area agent records the core information of the evacuation, including the number of residents who have been evacuated, the number of residents who have arrived at each green space, and the evacuation route information of different green spaces.

[0089] The green space agent is used to comprehensively evaluate the weights of roads and green spaces, sort the top 10 nearest disaster prevention and evacuation green spaces by distance as the target destination green space candidate list, calculate the optimal evacuation route through the A* algorithm, and track and record the evacuation status at different time steps in real time.

[0090] It should be noted that during the evacuation simulation process of the intelligent agent module, each disaster prevention and mitigation green space is prioritized according to distance to form a list of green spaces to be evacuated. Each intelligent agent uses the A* optimal path algorithm based on the road network weight and the capacity of the current disaster prevention and mitigation green space list, and uses the road hazard coefficient as the weight to calculate the optimal route, simulating the process of the intelligent agent traveling along the path.

[0091] Among them, the A* algorithm is a heuristic search algorithm that finds the shortest path from the starting point to the target by combining the actual cost (the path cost from the starting point to the current point) and the estimated cost (the predicted cost from the current point to the target point) to minimize the total cost. Compared with methods such as Dijkstra, the A* algorithm can ensure that the optimal path solution is calculated. At the same time, it improves search efficiency and reduces the amount of computation through heuristic functions. In addition, the A* algorithm is more flexible and can adjust the heuristic function according to different environmental characteristics and needs. Therefore, it is suitable for large-scale, dynamic or complex path planning tasks.

[0092] S5. Disaster Prevention and Mitigation Green Space Capacity Update: Based on the real-time recorded evacuation situation, dynamically adjust the remaining capacity of the green space. When the green space is overloaded, trigger the intelligent agent module to select other nearby disaster prevention and mitigation green spaces with available capacity, and replan the simulated evacuation route.

[0093] Specifically, based on the number of agents entering the disaster prevention and mitigation green space, the remaining capacity of the green space is updated. When the capacity of the green space is insufficient, the agents re-evaluate the shelter selection strategy, select other nearby disaster prevention and mitigation green spaces with available capacity based on real-time environmental data, and regenerate the list for evacuation.

[0094] S6. Road Congestion Feedback Update: The evacuation process is simulated multiple times. In each simulation, the intelligent agent module plans a simulated evacuation path. Based on the planned simulated evacuation path, the road congestion level of each road segment at different time steps is calculated, and the weights of each road segment are updated, as follows:

[0095] The formula for calculating congestion level C is:

[0096]

[0097] In the formula, N represents the number of people escaping on the road segment. max The maximum number of passengers allowed on the road section;

[0098] The weights of road segments are adjusted based on the congestion level C. The new weights are as follows:

[0099] R n '=D n *L n *C

[0100] In the formula, R n 'For the new weights, D n Risk factor, L n Escape distance;

[0101] Then the agent module performs the next iteration based on the updated road weights and recalculates the optimal evacuation path.

[0102] S7. Iterative Control and Evacuation Simulation Result Output: Set the evacuation target as a preset condition, control the iterative process of the evacuation simulation of the intelligent agent module, and output the evacuation simulation result after the preset condition is met;

[0103] Specifically, for the evacuation plan, evacuation efficiency targets are set (such as total evacuation time and arrival rate). Multiple evacuation simulations and road congestion calculations are performed within the model, and an iterative simulation process is conducted until the number of iterations reaches a preset upper limit. The final output is the evacuation simulation results, which may include: the evacuation time distribution of the agents, the congestion situation of each road segment, the usage of disaster prevention and mitigation green spaces, and the spatial distribution of evacuation routes.

[0104] S8. Optimal evacuation simulation scheme selection: Set multiple evacuation objectives, modify the weight ratio according to achieve different evacuation objectives, control the intelligent agent module to generate multiple evacuation simulation schemes, and each evacuation simulation scheme is the optimization result under different weight ratios;

[0105] Specifically, to compare the merits of different evacuation simulation schemes, the following evacuation indicators can be used for quantitative evaluation:

[0106] The green space is the closest target, and the dynamic green space is the closest for escape;

[0107] The total evacuation time target aims to minimize the straight-line distance to the target green space.

[0108] The target for the utilization rate of disaster prevention and mitigation green spaces is to achieve the highest comprehensive utilization rate of each disaster prevention and mitigation green space.

[0109] The various evacuation simulation schemes include: the shortest green space distance scheme, which uses the straight-line distance from the residential area as the basis for ranking the target destination green space list, and uses the sum of the congestion coefficient and the danger coefficient as the road selection weight; the shortest evacuation time scheme, which uses the actual distance from the residential area as the basis for ranking the target destination green space list, and uses the sum of the road distance, the danger coefficient, and the congestion coefficient as the road selection weight; and the disaster prevention and mitigation green space personnel distribution balanced scheme, which uses the utilization efficiency of each green space as the basis for ranking the target destination green space list, prioritizes the use of the green space with the lowest utilization efficiency, and uses the sum of the danger coefficient and the congestion coefficient as the road selection weight.

[0110] The constraints for achieving these goals include: each residential area must have at least one disaster prevention and mitigation green space; the number of users of each disaster prevention and mitigation green space must not exceed its capacity; disaster prevention and mitigation green spaces that are fully occupied cannot be candidates; and roads with the highest level of congestion cannot be candidates.

[0111] By comparing the advantages and disadvantages of each generated scheme through these indicators and constraints, the most suitable evacuation scheme for the current emergency needs is finally selected, such as the shortest evacuation time scheme, the least congestion scheme, and the scheme for balanced distribution of people in disaster prevention and mitigation green spaces.

[0112] It should be noted that after each plan is generated, dynamic evacuation route maps and road congestion heat maps are drawn using Python's matplotlib. Green space and residential area distribution maps are overlaid on the GIS platform to present the layout results intuitively. Through different visualization methods, the results are comprehensively displayed and compared to help decision-makers make quick judgments in emergency situations.

[0113] S9. Generate layout optimization scheme: Compare the optimal evacuation scheme for the overall goal or a single goal with the current status of disaster prevention and mitigation green space to derive specific optimization schemes, which may include the following four suggested directions:

[0114] Recommendations for the selection of new green spaces for disaster prevention and mitigation: Based on the areas with high residential density but insufficient coverage of existing green spaces, recommendations are made for the selection of new green spaces, prioritizing sites with convenient transportation and low risk as candidate green spaces;

[0115] Existing strategies for expanding disaster prevention and mitigation green spaces: For situations where existing green spaces are overloaded, it is recommended to improve service capacity by expanding the green space area or adding evacuation facilities;

[0116] Disaster prevention and escape route optimization: Based on the selected optimal evacuation route plan, the main roads leading to the green space are modified or widened to reduce road obstructions and improve road capacity. At the same time, by adjusting the road network structure, the connectivity of the routes is optimized to reduce detour time.

[0117] The technical solution of the present invention will now be described in detail with reference to the embodiments:

[0118] like Figures 3 to 5 As shown, the embodiment takes the main urban area of ​​a city as the research object. The main urban area of ​​the city is a typical high-density urban area with a dense population in residential areas. At the same time, the types of green spaces are rich and complex, mainly scattered and fragmented, which cannot guarantee the efficiency of active escape and evacuation. Therefore, there is an urgent need for reasonable and effective pre-disaster evacuation planning to ensure the orderly escape of disaster personnel and maximize the personal safety of residents.

[0119] Figure 3-5 The simulation results of a disaster prevention and mitigation green space model in the main urban area of ​​a city under three target categories were presented. Figure 3 , 5 The blue lines represent escape routes for each residential area; the red blocks represent residential areas, with darker colors indicating unsuccessful escapes and lighter colors indicating successful escapes; the green blocks represent green spaces, with darker colors indicating overloaded capacity and lighter colors indicating unloaded capacity. Figure 4 Green indicates unobstructed traffic (smooth flow), while red indicates overloaded traffic (congestion).

[0120] As shown in the figure, among the three models, Model 3 (the scheme for balanced population distribution in disaster prevention and mitigation green spaces) has the best overall performance. Figure 5 As shown, the simulation can complete the emergency evacuation of most communities within 30 minutes. At the same time, the occupancy rate of disaster prevention and mitigation green spaces is relatively low, and the population distribution is more reasonable and balanced. This indicates that the green space selection strategy with balanced population distribution can not only improve the overall evacuation efficiency, but also effectively prevent congestion and resource depletion in individual green spaces due to excessive population flow, thereby improving the overall disaster prevention and mitigation capabilities.

[0121] like Figure 4 As shown in the simulation results of Model 2, the main roads in the central area of ​​the main urban area are in a state of severe congestion for 30 minutes. This phenomenon reflects that the traffic bottleneck problem in the central area is particularly prominent during peak evacuation periods, which seriously affects evacuation efficiency. Therefore, in the future, the main roads and some branch roads in the central, western and southwestern parts of the main urban area need to be widened to increase traffic flow and pedestrian flow, so as to ensure that evacuation traffic can maintain efficient operation and actual evacuation efficiency during disasters.

[0122] like Figure 3 As shown, Model 1, due to its fixed target green space and lack of dynamic adjustment capability, does not achieve ideal evacuation results when faced with complex and ever-changing road congestion. It may fail to make full use of other green space resources, leading to overload or waste of some green space capacity. However, Model 1 still demonstrates a relatively reasonable evacuation route and green space utilization. As a basic evacuation model based on straight-line distance, decision-makers and urban planners can more easily understand and apply its results. It can provide a baseline reference for other more complex models and be used to formulate and adjust disaster prevention and mitigation policies.

[0123] By summarizing the simulation results of the three target models at each time step, we can also gain in-depth evaluation and suggestions on the congestion level of other relevant escape routes in the main urban area of ​​a city, as well as effective suggestions and planning guidance for future escape and disaster avoidance routes.

[0124] In summary, this invention, by introducing a multi-agent evacuation simulation model into the planning and design of disaster prevention and mitigation green spaces and optimizing multiple objectives such as road congestion and green space overload during simulation, can not only accurately simulate real-time escape situations and mitigate real-time road disaster risks, evacuation crowd congestion risks, and disaster prevention and mitigation green space capacity overload risks during disasters, thus helping to formulate optimal escape plans, but also dynamically adjust evacuation routes and strategies according to different emergency needs, formulating optimal solutions under multi-objective conditions. This provides more scientific and accurate planning decision support for future urban pre-disaster emergency plan formulation and post-disaster response plans.

[0125] Example 2:

[0126] like Figure 2 As shown, this embodiment provides a multi-agent-based disaster prevention and mitigation green space system layout optimization system, used to implement the multi-agent-based disaster prevention and mitigation green space system layout optimization method described in Embodiment 1, including:

[0127] The data processing module receives geographic data from the area to be optimized, including residential area data, road network data, and disaster prevention and mitigation green space data. Based on GIS software, it integrates and processes these geographic data, unifies the geographic coordinate system, handles missing and outlier values, and constructs a complete GIS basic data information database. The data processing module also receives raw geographic data from the area to be optimized and preprocesses it to form effective geographic data, including raw residential area data, raw road network data, and raw disaster prevention and mitigation green space data.

[0128] The scheme output module is used to set evacuation targets, select specified evacuation targets and simulate with a multi-agent evacuation model, and the multi-agent evacuation model outputs the optimal evacuation scheme.

[0129] Among them, the multi-agent evacuation model uses effective geographical data as the data source to simulate and record the real-time evacuation status. ， Real-time evacuation status includes the number of people entering each disaster prevention and mitigation green space and its utilization efficiency, evacuation routes of each community, and the number of road users and congestion of each road segment in the road network within each time step; outputs the optimal evacuation plan after simulation; real-time evacuation status includes the real-time green space overload status and the real-time road congestion status of each road segment in the evacuation route.

[0130] The layout optimization scheme generation module compares the optimal evacuation plan with the original road network data and the original disaster prevention and mitigation green space data to optimize the layout of roads and disaster prevention and mitigation green spaces within the area to be optimized.

[0131] The solution output module includes:

[0132] The agent model building unit is used to construct a multi-agent evacuation model, which includes an agent module, an environment module, and a communication module. The agent module includes residential area agents, evacuation agents, and green space agents.

[0133] The path planning unit, based on the Mesa framework, defines residential area agents, evacuation agents, and green space agents and their attributes to simulate the dynamic decision-making process of individual residents in different evacuation scenarios. The residential area agent manages residential area data, road network data, and disaster prevention and mitigation green space data in the GIS basic data database; the evacuation agent stores the initial geographic data of all residential areas, roads, and green spaces, and records the real-time evacuation status of each residential area during the simulation; the green space agent dynamically calculates the optimal evacuation path by comprehensively evaluating the weights of roads and green spaces, and tracks and records the evacuation status at different time steps in real time.

[0134] The disaster prevention and mitigation green space capacity update unit is used to dynamically adjust the remaining capacity of the green space based on the real-time recorded evacuation situation. When the green space is overloaded, the intelligent agent module is triggered to select other nearby disaster prevention and mitigation green spaces with available capacity and replan the simulated evacuation route.

[0135] The road congestion feedback update unit is used to convert the processed road network data into graph structure data, where nodes represent road intersections and edges represent road segments. It sets the initial weights of the edges, and then calculates the road congestion of each road segment at different time steps based on the simulated evacuation path planned and generated by the agent module. It also updates the weights of the edges of each road segment. Then, the agent module performs the next iteration based on the updated road weights and recalculates the optimal evacuation path.

[0136] The iterative control unit is used to set the evacuation target or the maximum number of iterations as preset conditions, control the iterative process of the evacuation simulation of the intelligent agent module, and output the evacuation simulation results after the preset conditions are met.

[0137] The evacuation simulation scheme generation unit is used to set multiple evacuation objectives and control the intelligent agent module to generate multiple evacuation simulation schemes according to the weight ratio of achieving different evacuation objectives. Each evacuation simulation scheme is the optimization result under different weight ratios, and the evacuation simulation scheme is used to achieve a comprehensive objective or a single objective.

[0138] Specifically, the aforementioned data processing module, scheme output module, and layout optimization scheme generation module can be embedded into a computer processing system. The computer, based on the aforementioned multi-agent-based disaster prevention and mitigation green space system layout optimization method, calls upon the aforementioned modules to complete the task of optimizing the disaster prevention and mitigation green space system plan. The aforementioned data processing module, scheme output module, and layout optimization scheme generation module can execute operations according to the specific steps given by the aforementioned multi-agent-based disaster prevention and mitigation green space system layout optimization method.

[0139] It should be noted that the division of the various modules in the above system is merely a division of logical functions. In actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated. These modules can be implemented entirely in software through processing element calls; they can be fully implemented in hardware; or some modules can be implemented by processing element calls to software, while others are implemented in hardware. For example, the data processing module can be a separate processing element or integrated into a chip in the aforementioned device. Alternatively, it can be stored as program code in the memory of the aforementioned device, and its functions can be called and executed by a processing element of the device. The implementation of other modules is similar. Furthermore, these modules can be fully or partially integrated together or implemented independently. The processing element mentioned here can be an integrated circuit with signal processing capabilities. During implementation, each step of the above method or each of the above modules can be completed through integrated logic circuits in the hardware of the processor element or through software instructions.

[0140] For example, these modules can be one or more integrated circuits configured to implement the above methods, such as one or more Application Specific Integrated Circuits (ASICs), one or more Digital Signal Processors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs). As another example, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together to form a system-on-a-chip (SOC).

[0141] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0142] For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances. When an element is referred to as being "assembled on," "mounted on," "fixed to," or "set on" another element, it may be directly on the other element or there may be an intermediate element present. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be an intermediate element present. The terms "vertical," "horizontal," "upper," "lower," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only possible embodiments.

[0143] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

[0144] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," 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 this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

Claims

1. A multi-agent based layout optimization method for disaster prevention and refuge green space system, applied to the features of, The method comprises the following steps: Receiving original geographic data in a region to be optimized, and preprocessing the original geographic data to form effective geographic data, wherein the original geographic data comprises original residential area data, original road network data and original disaster prevention and refuge green land data; Setting a evacuation target, selecting the designated evacuation target and a multi-agent evacuation model for simulation, and outputting an optimal evacuation scheme from the multi-agent evacuation model; The multi-agent evacuation model takes the effective geographic data as a data source, simulates and records real-time evacuation states, and the real-time evacuation states comprise the number of people entering each disaster prevention and refuge green land and the use efficiency of each disaster prevention and refuge green land, the evacuation path of each residential area, the number of people using each road segment of the road network and the congestion degree of each road segment of the road network at each time step; The optimal evacuation scheme is compared with the original road network data and the original disaster prevention and refuge green land data, and the layout of the roads and the disaster prevention and refuge green lands in the region to be optimized is optimized; The pre-constructed multi-agent evacuation model comprises an agent module, an environment module and a communication module, wherein: The agent module comprises residential area agents, evacuation agents and green land agents; The environment module comprises GIS basic environment data and global real-time information, specifically comprising road network data, residential area boundary data, green land position and maximum capacity data and dynamically updated road weight and green land state data; The communication module is used for supporting information interaction and resource sharing among the residential area agents, the evacuation agents and the green land agents, and ensuring cooperation and synchronization among the residential area agents, the evacuation agents and the environment module; The road congestion degree is updated, and the specific steps are as follows: (71) converting the processed road network data into graph structure data, wherein a node represents a road intersection, an edge represents a road segment, and the initial weight of the edge is set, specifically as follows: The selection weight R of each road n The dangerous coefficient D of each road n And the escape distance L of each road n Together determine, the specific calculation formula is: Escape distance L n For based on A The path length calculated by the optimal path algorithm, the danger coefficient D n For road safety influence factor obtained after objective entropy weight method calculation and normalization processing; (72) calculating the road congestion degree of each road segment at different time steps according to the simulation evacuation path generated by the agent module, and updating the weight of each road segment edge, specifically as follows: The calculation formula of the congestion degree C is as follows: In the formula, N is the number of escapees on the road segment, N max is the maximum carrying capacity of the road segment. According to the congestion degree C, the weight of the road segment is adjusted, and the new weight is as follows: wherein is a new weight, D n is a danger coefficient, L n is an escape distance; (73) the agent module re-calculates the optimal evacuation path according to the updated road weight for the next iteration; The iteration control and the evacuation simulation result output, and the specific steps are as follows: Setting the evacuation target or the maximum number of iterations as a preset condition, judging whether the preset condition is met after each simulation, deciding whether to continue iteration, then performing multiple calculations and road congestion calculations, and performing the iteration simulation process until the evacuation simulation result is output after the preset condition is met; The evacuation simulation result comprises the evacuation time distribution of the agent module, the congestion condition data of each road segment, the remaining capacity data of the disaster prevention and refuge green land and the spatial distribution of the simulation evacuation path; According to the implementation of different evacuation targets, the weight ratio is modified and the optimal evacuation scheme is generated, specifically as follows: The evacuation target specifically comprises a green land distance nearest target, a total evacuation time target and a disaster prevention and refuge green land utilization rate target. The multiple evacuation simulation schemes specifically include: a green land distance closest scheme, taking the straight line distance from the residential community as the sorting basis of the target end green land list, and taking the sum of the congestion coefficient and the danger coefficient as the road selection weight; a shortest evacuation time scheme, taking the actual distance from the residential community as the sorting basis of the target end green land list, and taking the sum of the road distance, the danger coefficient and the congestion coefficient as the road selection weight; a disaster prevention and safety green land personnel distribution equalization scheme, taking the use efficiency of each green land as the sorting basis of the target end green land list, and preferentially using the green land with the lowest use efficiency, and taking the sum of the danger coefficient and the congestion coefficient as the road selection weight; The optimal evacuation scheme is compared with the original road network data and the original disaster prevention and safety green land data, and the layout of the roads and the disaster prevention and safety green lands in the region to be optimized is optimized, specifically including three aspects: Disaster prevention and safety green land site selection optimization: according to the region with high residential density but insufficient service coverage of the existing green land, an optimization scheme for adding new green land site selection is proposed, and a block with convenient transportation and low risk is selected as a candidate green land; Existing disaster prevention and safety green land expansion optimization: for the green land with overloaded service, the green land area is expanded or the evacuation function facilities are increased; Disaster prevention and safety escape path optimization: combined with the evacuation path of the optimal evacuation scheme, the main roads leading to the green land are reconstructed or widened, and the road network structure is adjusted to optimize the path connectivity. 2.The multi-agent based layout optimization method of disaster prevention and shelter greenfield system according to claim 1, wherein: The geographic data in the region to be optimized is obtained based on an Internet map open platform and a planning plan, specifically as follows: The residential area data includes the boundary SHP data of each residential community and the population; The road network data includes the geographic position information and grade attribute of the road and the road width data; The disaster prevention and safety green land data includes the green land geographic position data and the green land maximum capacity data, wherein the green land maximum capacity data = disaster prevention and safety green land effective safety total area / personnel effective safety area; The preprocessing of the original geographic data specifically includes: Based on the GIS software, the above-mentioned geographic data is integrated and processed, the geographic coordinate system is unified, and the missing values and abnormal values are processed. 3.The multi-agent based disaster prevention and shelter green space system layout optimization method of claim 1, wherein, The multi-agent evacuation model takes effective geographic data as the data source, simulates and records the real-time evacuation state, takes the evacuation target as the target, outputs the optimal evacuation scheme after simulation, and specifically includes: evacuation path planning, disaster prevention and safety green land capacity updating, road congestion degree feedback updating, iteration control and evacuation simulation result output, and modification of the weight ratio according to the realization of different evacuation targets and generation of the optimal evacuation scheme. 4.The multi-agent based disaster prevention and refuge green space system layout optimization method according to claim 3, characterized in that, The evacuation path planning specifically includes: Based on the mesa framework, the residential community agent, the evacuation agent and the green land agent and their attributes are defined, and the dynamic decision-making process of the individual residents in different evacuation scenarios is simulated; The residential area agent is used for managing the residential area data, road network data and disaster prevention and safety green land data in the GIS basic data information base; the evacuation agent is used for storing the initial geographic data of all residential areas, roads and green lands, and recording the real-time evacuation state of each residential area in the simulation process; the green land agent is used for taking the top 10 distance closest disaster prevention and safety evacuation green lands as the target end point standby green land list through comprehensive evaluation of the weights of roads and green lands, using A Algorithm to calculate the optimal evacuation path, and real-time tracking and recording of the evacuation situation at different time steps. 5.The multi-agent based disaster prevention and shelter green space system layout optimization method according to claim 4, characterized in that, The disaster prevention and safety green land capacity updating specifically includes: According to the number of agents entering the disaster prevention and safety green land, the remaining capacity of the green land is updated, when the green land capacity is insufficient, the agent module reevaluates the refuge selection strategy, selects other disaster prevention and safety green lands with the closest distance and available capacity according to the surrounding real-time environmental data, and generates a list for evacuation again.

6. A multi-agent based layout optimization system for disaster prevention and refuge greenbelt system, for implementing the multi-agent based layout optimization method for disaster prevention and refuge greenbelt system according to any one of claims 1 to 5, characterized in that, ​ The data processing module is used to receive the raw geographic data of the area to be optimized, and to preprocess the raw geographic data to form effective geographic data. The raw geographic data includes raw residential area data, raw road network data, and raw disaster prevention and mitigation green space data. The scheme output module is used to set evacuation targets, select specified evacuation targets and simulate with a multi-agent evacuation model, and the multi-agent evacuation model outputs the optimal evacuation scheme. Among them, the multi-agent evacuation model uses effective geographic data as the data source to simulate and record the real-time evacuation status. The real-time evacuation status includes the number of people entering each disaster prevention and mitigation green space and the usage efficiency, the evacuation routes of each community, and the number of road users and the degree of congestion of each section of the road network within each time step. The layout optimization scheme generation module compares the optimal evacuation plan with the original road network data and the original disaster prevention and mitigation green space data to optimize the layout of roads and disaster prevention and mitigation green spaces within the area to be optimized. The solution output module includes: The agent model building unit is used to construct a multi-agent evacuation model, which includes an agent module, an environment module, and a communication module. The agent module includes residential area agents, evacuation agents, and green space agents. The path planning unit, based on the Mesa framework, defines residential area agents, evacuation agents, and green space agents and their attributes to simulate the dynamic decision-making process of individual residents in different evacuation scenarios. The residential area agent manages residential area data, road network data, and disaster prevention and mitigation green space data in the GIS basic data database; the evacuation agent stores the initial geographic data of all residential areas, roads, and green spaces, and records the real-time evacuation status of each residential area during the simulation; the green space agent dynamically calculates the optimal evacuation path by comprehensively evaluating the weights of roads and green spaces, and tracks and records the evacuation status at different time steps in real time. The disaster prevention and mitigation green space capacity update unit is used to dynamically adjust the remaining capacity of the green space based on the real-time recorded evacuation situation. When the green space is overloaded, the intelligent agent module is triggered to select other nearby disaster prevention and mitigation green spaces with available capacity and replan the simulated evacuation route. The road congestion feedback update unit is used to convert the processed road network data into graph structure data, where nodes represent road intersections and edges represent road segments. It sets the initial weights of the edges, and then calculates the road congestion of each road segment at different time steps based on the simulated evacuation path planned and generated by the agent module. It also updates the weights of the edges of each road segment. Then, the agent module performs the next iteration based on the updated road weights and recalculates the optimal evacuation path. The iterative control unit is used to set the evacuation target or the maximum number of iterations as preset conditions, control the iterative process of the evacuation simulation of the intelligent agent module, and output the evacuation simulation results after the preset conditions are met. The evacuation simulation scheme generation unit is used to set multiple evacuation objectives and control the intelligent agent module to generate multiple evacuation simulation schemes according to the weight ratio of achieving different evacuation objectives. Each evacuation simulation scheme is the optimization result under different weight ratios, and the evacuation simulation scheme is used to achieve a comprehensive objective or a single objective.

Images