Method, system and equipment for predicting movement of crowd in disaster, and medium

Abstract

The invention discloses a disaster crowd movement prediction method, system and device and a medium, and the method comprises the steps: firstly obtaining the disaster information of a sensing adjacent range of an intelligent agent, carrying out the sensing operation according to the disaster information of the sensing adjacent range, and obtaining a first disaster sensing index, a second disaster sensing index and a third disaster sensing index; wherein the first disaster sensing index is used for representing the rejection behavior of the agent in the face of the disaster, the second disaster sensing index is used for representing the attraction behavior generated by the agent, and the third disaster sensing index is used for representing the convergence behavior of the agent in the face of the disaster. Then, according to the three disaster sensing indexes, the moving direction of the intelligent agent is determined; and then, according to the moving direction, controlling the intelligent agent to move in the sensing adjacent range until the intelligent agent moves to the target grid unit, thereby obtaining a moving prediction result. According to the embodiment of the invention, the decision-making process of an individual in a disaster environment can be accurately reflected, so that the accuracy of crowd movement behavior prediction during a disaster period is remarkably improved.

Description

Technical Field

[0001] The embodiments of the present application relate to but are not limited to the field of disaster modeling and simulation technology, and in particular to a method, system, device and medium for predicting crowd movement during disasters. Background Art

[0002] In the relevant technical field, the simulation and prediction of crowd movement behavior during disasters generally have the problem of insufficient adaptability to disaster-specific situations. Specifically, these technologies often find it difficult to accurately predict the dynamic changes of crowds under extreme weather or emergencies. Especially when a disaster actually occurs, the behavior patterns of the crowd will be deeply affected by multiple complex factors such as the disaster situation (such as flooding), the shelters provided in the disaster area (such as buildings), and the psychological effects of the group (such as herd behavior). The current technology is still insufficient in comprehensively considering these key factors. In other words, existing technical means cannot fully simulate the complex decision-making process and its dynamic changes faced by individuals in a disaster environment. Summary of the invention

[0003] The following is a summary of the subject matter described in detail herein. This summary is not intended to limit the scope of the claims.

[0004] The embodiments of the present application provide a method for predicting crowd movement during disasters, which can more accurately reflect the complex decision-making process of individuals in a disaster environment and significantly improve the accuracy of predicting crowd movement behavior during disasters.

[0005] In a first aspect, an embodiment of the present application provides a method for predicting crowd movement during a disaster, comprising: obtaining disaster information within a perceptual proximity of an intelligent agent; performing a perceptual operation based on the disaster information within the perceptual proximity to obtain a first disaster perception index, a second disaster perception index, and a third disaster perception index of the intelligent agent, wherein the first disaster perception index is used to characterize the repulsive behavior of the intelligent agent in the face of a disaster, the second disaster perception index is used to characterize the attractive behavior generated by the intelligent agent, and the third disaster perception index is used to characterize the convergent behavior of the intelligent agent in the face of a disaster; determining a moving direction of the intelligent agent based on the first disaster perception index, the second disaster perception index, and the third disaster perception index; controlling the intelligent agent to move within the perceptual proximity according to the moving direction, and obtaining a movement prediction result when the intelligent agent moves to a target grid unit within the perceptual proximity.

[0006] In combination with the first aspect, in an embodiment of the present application, after the intelligent agent moves to the target grid unit within the perceived proximity range, the method further includes: when the capacity of the target grid unit exceeds a preset capacity threshold, controlling the intelligent agent to move to a secondary target grid unit within the perceived proximity range; the number of safe positions within the secondary target grid unit is less than the number of safe positions within the target grid unit.

[0007] In combination with the first aspect, in an embodiment of the present application, the moving directions of the multiple intelligent bodies are determined according to the first disaster perception index, the second disaster perception index and the third disaster perception index, including: assigning a first perception weight to the first disaster perception index, assigning a second perception weight to the second disaster perception index, and assigning a third perception weight to the third disaster perception index; according to the first perception weight, the second perception weight and the third perception weight, respectively adjusting the contribution of the first disaster perception index, the second disaster perception index and the third disaster perception index to obtain an overall disaster perception index; and determining the moving direction of the intelligent body according to the overall disaster perception index and a preset perception factor.

[0008] In combination with the first aspect, in an embodiment of the present application, the disaster information of the perceived proximity includes the disaster situation and the number of safe locations of at least one grid unit within the perceived proximity; the first disaster perception index, the second disaster perception index and the third disaster perception index of the intelligent agent are obtained based on the disaster information, including: obtaining the first disaster perception index based on the disaster situation of the grid unit where the intelligent agent is located and a preset distance coefficient; obtaining the second disaster perception index based on the number of safe locations of the grid unit where the intelligent agent is located; obtaining the third disaster perception index based on the first disaster perception index, the second disaster perception index, the grid unit where the intelligent agent is located and the target grid unit.

[0009] In combination with the first aspect, in one embodiment of the present application, the perceived proximity range is obtained according to the following steps: obtaining a disaster simulation area, dividing the disaster simulation area into a plurality of grid units; and combining the plurality of grid units according to the type of the intelligent agent to obtain the perceived proximity range.

[0010] In combination with the first aspect, in an embodiment of the present application, the types of the intelligent agents include first-type intelligent agents and second-type intelligent agents, and the perception proximity range includes a first perception proximity range and a second perception proximity range; when the intelligent agent belongs to the first-type intelligent agent, the perception proximity range is the first perception proximity range; when the intelligent agent belongs to the second-type intelligent agent, the perception proximity range is the second perception proximity range; the number of grid units of the first perception proximity range is greater than the number of grid units of the second perception proximity range.

[0011] In a second aspect, an embodiment of the present application provides a disaster crowd movement prediction system, including: an information acquisition module, used to obtain disaster information within the perception proximity of an intelligent agent; a perception action module, used to perform perception operations based on the disaster information within the perception proximity, to obtain a first disaster perception index, a second disaster perception index, and a third disaster perception index of the intelligent agent, wherein the first disaster perception index is used to characterize the rejection behavior of the intelligent agent in the face of disasters; the second disaster perception index is used to characterize the attraction behavior generated by the intelligent agent, and the third disaster perception index is used to characterize the convergence behavior of the intelligent agent in the face of disasters; a movement decision module, used to determine the movement direction of the intelligent agent based on the first disaster perception index, the second disaster perception index, and the third disaster perception index; and a result prediction module, used to control the movement of the intelligent agent within the perception proximity according to the movement direction, and obtain a movement prediction result when the intelligent agent moves to a target grid unit within the perception proximity.

[0012] In combination with the second aspect, in one embodiment of the present application, the perception function module includes a repulsion layer, an attraction layer and a convergence layer. The repulsion layer is used to perform perception operations based on the disaster information in the perceived neighborhood to obtain a first disaster perception index. The attraction layer is used to perform perception operations based on the disaster information in the perceived neighborhood to obtain a second disaster perception index. The convergence layer is used to perform perception operations based on the disaster information in the perceived neighborhood to obtain a third disaster perception index.

[0013] On the other hand, an embodiment of the present application provides a mobility prediction device, comprising: at least one processor; at least one memory for storing at least one program; at least one of the programs is run by at least one of the processors to execute the method for predicting crowd mobility during disasters as described above.

[0014] On the other hand, an embodiment of the present application provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are used to execute the method for predicting crowd movement during disasters as described above.

[0015] A method for predicting the movement of people in times of disaster provided by an embodiment of the present application, first, obtains the disaster information of the perceived proximity range of the intelligent body, and obtains the first disaster perception index, the second disaster perception index and the third disaster perception index of the intelligent body according to the disaster information of the perceived proximity range. Among them, the first disaster perception index is used to characterize the repulsive behavior of the intelligent body in the face of disasters, the second disaster perception index is used to characterize the attractive behavior generated by the intelligent body, and the third disaster perception index is used to characterize the convergent behavior of the intelligent body in the face of disasters. By refining the disaster perception index (including the three dimensions of repulsive behavior, attractive behavior and convergent behavior), it is possible to more accurately describe the behavioral characteristics of the intelligent body in the disaster environment, capture the subtle changes in the behavior, and thus improve the accuracy and reliability of the prediction. Then, according to the first disaster perception index, the second disaster perception index and the third disaster perception index, the moving direction of the intelligent body can be determined. Then, according to the moving direction, the intelligent body is controlled to move within the perceived proximity range, and when the intelligent body moves to the target grid unit within the perceived proximity range, the movement prediction result is obtained. The embodiments of the present application simulate the behavior of people in disasters by introducing intelligent agents (which can represent individuals or group members). The intelligent agents can dynamically adjust their disaster perception index based on the real-time disaster information within their perceived proximity, and can more accurately reflect the complex decision-making process of individuals in disaster environments, thereby significantly improving the accuracy of predicting the mobility behavior of people during disasters. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flow chart of a method for predicting crowd movement during disasters provided in an embodiment of the present application;

[0017] Figure 2 This is a schematic diagram of the perceived proximity range of an intelligent agent provided by a specific example of the present application;

[0018] Figure 3 It is a schematic diagram of the perception process of an intelligent agent provided by a specific example of this application;

[0019] Figure 4 is a schematic diagram of the perception decision process provided by a specific example of this application;

[0020] Figure 5 The embodiment of this application provides Figure 1 Specific flow chart of step 120;

[0021] Figure 6 is a schematic diagram of the movement process of an intelligent agent provided by a specific example of the present application;

[0022] Figure 7 is a schematic diagram of the movement process of an intelligent agent provided by another specific example of the present application;

[0023] Figure 8is a structural diagram of a disaster-time crowd movement prediction system provided by an embodiment of the present application;

[0024] Fig. 9 It is a structural diagram of a mobile prediction device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0026] It should be noted that, although the logical order is shown in the flowchart, in some cases, the steps shown or described can be performed in a different order from that in the flowchart. The terms "first", "second", etc. in the specification, claims and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions that can be implemented in this application, so they have no technical substantive significance. Any modification of the structure, change of the proportional relationship or adjustment of the size, without affecting the effects that can be produced by this application and the purposes that can be achieved, should still fall within the scope of the technical content disclosed in this application. At the same time, the terms such as "upper", "lower", "left", "right", "middle" and "one" quoted in this specification are only for the convenience of narration, and are not used to limit the scope of implementation of this application. The change or adjustment of the relative relationship should also be regarded as the scope of implementation of this application without substantial change of the technical content.

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0028] In the relevant technical field, the simulation and prediction of crowd movement behavior during disasters generally have the problem of insufficient adaptability to disaster-specific situations. Specifically, these technologies often find it difficult to accurately predict the dynamic changes of crowds under extreme weather or emergencies. Especially when a disaster actually occurs, the behavior patterns of crowds will be profoundly affected by multiple complex factors such as flooding conditions, shelters provided by buildings, and group psychological effects (such as herd behavior). The current technical framework is still insufficient in comprehensively considering these key factors. In other words, existing technical means cannot fully simulate the complex decision-making process and its dynamic changes faced by individuals in a disaster environment.

[0029] In view of this, the embodiment of the present application provides a method for predicting the movement of people in a disaster, a system for predicting the movement of people in a disaster, a mobile prediction device and a computer-readable storage medium. First, the disaster information of the perceived proximity range of the intelligent agent is obtained, and the first disaster perception index, the second disaster perception index and the third disaster perception index of the intelligent agent are obtained according to the disaster information of the perceived proximity range. Among them, the first disaster perception index is used to characterize the repulsive behavior of the intelligent agent in the face of disasters, the second disaster perception index is used to characterize the attractive behavior generated by the intelligent agent, and the third disaster perception index is used to characterize the convergent behavior of the intelligent agent in the face of disasters. By refining the disaster perception index (including the three dimensions of repulsive behavior, attractive behavior and convergent behavior), the behavioral characteristics of the intelligent agent in the disaster environment can be more accurately described, and the subtle changes in the behavior can be captured, thereby improving the accuracy and reliability of the prediction. Then, according to the first disaster perception index, the second disaster perception index and the third disaster perception index, the moving direction of the intelligent agent can be determined. Then, according to the moving direction, the intelligent agent is controlled to move within the perceived proximity range, and when the intelligent agent moves to the target grid unit within the perceived proximity range, the movement prediction result is obtained. The embodiments of the present application simulate the behavior of people in disasters by introducing intelligent agents (which can represent individuals or group members). The intelligent agents can dynamically adjust their disaster perception index based on the real-time disaster information within their perceived proximity, and can more accurately reflect the complex decision-making process of individuals in disaster environments, thereby significantly improving the accuracy of predicting the mobility behavior of people during disasters.

[0030] The embodiments of the present application are further described below in conjunction with the accompanying drawings.

[0031] Reference Figure 1 , Figure 1 1 is a flow chart of a method for predicting the movement of people in times of disaster provided by an embodiment of the present application. The method includes but is not limited to steps 110 to 140.

[0032] Step 110: Obtain disaster information of the agent's perceived neighborhood;

[0033] Step 120: Perform a perception operation based on the disaster information of the perceived neighboring range to obtain a first disaster perception index, a second disaster perception index, and a third disaster perception index of the intelligent agent, wherein the first disaster perception index is used to characterize the repulsive behavior of the intelligent agent in the face of disasters; the second disaster perception index is used to characterize the attractive behavior generated by the intelligent agent; and the third disaster perception index is used to characterize the convergent behavior of the intelligent agent in the face of disasters;

[0034] Step 130: determining a moving direction of the intelligent agent according to the first disaster perception index, the second disaster perception index, and the third disaster perception index;

[0035] Step 140: Control the agent to move within the perceived proximity range according to the moving direction, and obtain a movement prediction result when the agent moves to the target grid unit within the perceived proximity range.

[0036] Steps 110 to 140 are described in detail below.

[0037] In a feasible embodiment, an agent refers to an entity that can perceive the environment, make decisions and take actions. In simulation and prediction systems, agents can be regarded as software entities representing individuals or group members. They can make decisions based on the set rules or algorithms and the received information (such as environmental data, the behavior of other agents, etc.), and perform corresponding actions accordingly. In disaster simulation and prediction scenarios, agents can be used to simulate the behavior of the crowd. Each agent represents an individual or a small group. They can perceive the surrounding disaster conditions (such as flooding, building collapse, etc.) and decide their own direction of action based on this information and their own decision-making rules (such as finding safe shelters, avoiding dangerous areas, etc.). In this way, the agent can simulate the dynamic behavior of the crowd in a disaster and provide valuable reference information for disaster emergency response and management.

[0038] In a feasible embodiment, the perceptual proximity of an agent generally refers to the surrounding environment area where the agent can perceive and obtain information. Within this range, the agent can collect data about the environment, such as disaster conditions, topography, the location and behavior of other agents, etc. Specifically, the perceptual proximity may include the following aspects: Physical space range: the size of the physical space that the agent can perceive. For example, in a disaster simulation, the perceptual range of the agent may be limited to its "field of vision" or "hearing range", that is, the surrounding environment area that the agent can "see" or "hear". Information acquisition range: In addition to the physical space, the perceptual proximity may also involve the type and content of information that the agent can obtain. For example, the agent may be able to obtain specific information about the intensity of the disaster, the structure of the building, the location of the shelter, etc., which is crucial for the decision-making and action of the agent. It should be noted that in some cases, the perceptual proximity of the agent may not be fixed, but can be dynamically adjusted according to environmental changes or the needs of the agent. For example, when a disaster occurs, the agent can expand its perceptual range to obtain more information about the disaster situation, or narrow the range to perceive the changes in the surrounding environment more finely.

[0039] In a feasible embodiment, the step of determining the proximity range that the intelligent agent can perceive generally includes the following steps: first, clarify the specific area of ​​the disaster simulation and divide the area into multiple grid units; then, according to the type of the intelligent agent, combine the divided grid units to define the proximity range that the intelligent agent can effectively perceive.

[0040] In a feasible embodiment, when the divided grid cells are combined, they can be combined based on indirect neighborhood. Specifically, the neighborhood refers to a set of multiple (e.g., 8) grid cells adjacent to a grid cell. These adjacent grid cells share vertices or edges with the grid cell. On a two-dimensional plane, the neighborhood of a grid cell generally includes adjacent grid cells above, below, on the left, on the right, and in four diagonal directions. For example, the specific area of ​​the disaster simulation can be divided into multiple 1km×1km grid cells. According to the type of the agent, the current grid cell where the agent is located can be combined with multiple 1km×1km adjacent grid cells located above, below, on the left, on the right, and in four diagonal directions of the grid cell to obtain the perceived proximity of the agent. It is worth noting that during the movement of the agent, the current grid cell where the agent is located will change accordingly. This dynamicity directly causes the perceived proximity of the agent to change accordingly. In other words, no matter how the current grid cell changes, the agent's perception proximity can always be flexibly adjusted by combining the current grid cell of the agent with multiple grid cells that are immediately above, below, to the left, to the right, and in the four diagonal directions. This combination ensures that the agent can always build a comprehensive and dynamic perception environment based on its latest position.

[0041] In a feasible embodiment, a certain number of people may live in a grid unit, and these people may be represented by other agents. In order to be closer to reality, the initial population of each grid unit can be scientifically estimated and initialized based on the WorldPop data. WorldPop data can provide detailed and accurate population distribution information, making the prediction results more in line with the actual situation, and helping to more deeply understand and predict the population flow when a disaster occurs.

[0042] In a feasible embodiment, the types of intelligent agents include first-type intelligent agents and second-type intelligent agents, and the perception proximity range includes a first perception proximity range and a second perception proximity range; when the intelligent agent belongs to the first-type intelligent agent, the perception proximity range is the first perception proximity range; when the intelligent agent belongs to the second-type intelligent agent, the perception proximity range is the second perception proximity range; the number of grid units of the first perception proximity range is greater than the number of grid units of the second perception proximity range.

[0043] In a feasible embodiment, the first type of intelligent agent may be a local resident of the disaster area, and the second type of intelligent agent may be a non-local resident of the disaster area. Because the local residents have a deep understanding of the disaster area environment, established social network connections, frequent contact with local media, proficiency in culture and language, and rich historical experience, they show a stronger ability to obtain information than non-local residents. Therefore, compared with non-local residents, local residents have a wider range of perceived proximity. In other words, the first type of intelligent agent has a wider range of perceived proximity than the second type of intelligent agent. Figure 2 As shown, the perceived proximity range of local residents (first type of agents) can cover 5×5 adjacent grid cells, while the perceived proximity range of non-local residents (second type of agents) is relatively smaller, covering 3×3 adjacent grid cells.

[0044] In a feasible embodiment, the perception ability of the intelligent agent is reflected in that they can actively collect and analyze environmental information from the nearby perception range, such as Figure 3 As shown. Within this range, the agent can identify key clues to disaster information, such as disaster risk areas, safe areas (such as safe buildings, etc.), and the dynamic changes in the behavior of surrounding agents (such as surrounding people). It is worth noting that some grid cells can be designated as uninhabited areas. These areas may be temporarily or permanently unsuitable for human habitation due to disaster damage, safety hazards or other special reasons. When identifying this information, the agent can comprehensively consider various factors, including but not limited to the severity of the disaster, the distribution of safe areas, the activity patterns of surrounding agents, and the specific conditions of uninhabited areas, so as to make more intelligent and efficient decisions.

[0045] In a feasible embodiment, after the intelligent agent perceives the disaster information in its surrounding environment, it can make corresponding response decisions based on these perception results. Figure 4As shown in Figure 1, the agent's reaction stage can be abstracted into three different levels: the repulsion layer, the attraction layer, and the convergence layer, which together embody the basic dynamics of group behavior commonly seen in disaster scenarios. Specifically, the repulsion layer symbolizes the direction of repulsion in the group, which reflects the instinct to escape from danger and seek safety. When a disaster occurs, people will instinctively avoid areas that are considered dangerous, such as in remote areas or flood-prone areas. This repulsion effect usually prevents people from moving to grids containing uninhabited spaces or high-risk flood areas, which is out of deep concern for personal safety and security. The attraction layer represents the direction of the group's attraction, which points to safe locations such as elevated buildings or shelters. These locations can provide protection from direct dangers such as drowning and water debris, and provide temporary safe shelter for victims until emergency rescue services arrive or the situation improves (such as floods subside). The convergence layer reveals people's behavioral patterns in disasters that rely on social information, that is, by observing the behavior of individuals around them to assess the current situation and make decisions accordingly. This behavior pattern reflects the complex psychological and behavioral responses of humans in the face of disasters. People tend to follow the actions of others to find safe shelters or escape from dangerous areas. After superimposing the effects of these three levels, the direction of the combined force in the result layer will guide the movement direction of the intelligent agent. This process not only reflects the adaptability and decision-making ability of the intelligent agent to the disaster environment, but also reveals the behavioral laws and psychological characteristics of humans in disasters.

[0046] See also Figure 5 In step 120, a specific process of performing a perception operation based on the disaster information in the perceived neighborhood to obtain the first disaster perception index, the second disaster perception index and the third disaster perception index of the intelligent agent may at least include steps 510 to 530.

[0047] Step 510: Obtaining a first disaster perception index according to the disaster situation of the current grid unit of the agent and a preset distance coefficient;

[0048] Step 520: Obtain a second disaster perception index according to the number of safe locations of the current grid unit of the agent;

[0049] Step 530: Obtain a third disaster perception index according to the first disaster perception index, the second disaster perception index, the current grid unit of the agent, and the target grid unit.

[0050] In a feasible embodiment, the repulsive behavior exhibited by the intelligent agent in the face of a disaster can be quantitatively represented by the first disaster perception index; and its attractive behavior caused by the disaster (i.e., the tendency to move to a safe area) can be reflected by the second disaster perception index; as for the convergent behavior of the intelligent agent in a disaster situation, that is, the tendency to act synchronously with other individuals, it can be described by the third disaster perception index.

[0051] It is worth noting that these three disaster perception indexes can be expressed in vector form. Specifically, the first disaster perception index represents the repulsive force acting on the intelligent agent, the second disaster perception index reflects the attractive force acting on the intelligent agent, and the third disaster perception index reflects the convergent force acting on the intelligent agent. In the process of interaction, these three forces can form a combined force. This combined force, as the result of the intelligent agent's comprehensive consideration of disaster information, can determine the direction and path of the intelligent agent's subsequent movement.

[0052] In a feasible embodiment, the disaster information perceived within the vicinity covers the disaster conditions and the number of safe locations of at least one grid unit. Specifically, the disaster conditions of the grid unit involve areas that may become uninhabitable or difficult to observe due to flooding, reduction of human activities or other unfavorable factors, such as flooded areas or uninhabited areas. The number of safe locations refers to the total number of safe buildings available for refuge in the grid unit. In addition, the preset distance coefficient can quantitatively express the trend that the agent's repulsive behavior gradually decreases with increasing distance.

[0053] In a feasible embodiment, as shown in formula (1), Ω_(m×m) represents the sensing proximity range of the agent, which includes m×m grid units. represents the vector from grid cell i containing disaster risk points or uninhabited areas. Distance coefficient First Disaster Perception Index F rep represents the sum of the distance decay direction vectors of the grids that exclude the observing agent from the flooded or uninhabited area, that is, the sum of the direction vectors generated by the agent due to the exclusion of the flooded or uninhabited area.

[0054]

[0055] In a feasible embodiment, as shown in formula (2), the second disaster perception index F attr is the direction vector from the grid cell with the largest number of safe locations (e.g., the largest roof area) to the agent, where v imax Represents the vector from the grid i with the largest building roof area within Ω_(m×m) to the intelligent agent.

[0056]

[0057] In a feasible embodiment, the third disaster perception index F eerd The calculation process of F follows formula (3), which is composed of the sum of the expected movement direction vectors of the agent around each observed grid cell. Specifically, F herdis a vector that reflects the collective movement trend of agents under the influence of convergent behavior. In this framework, represents the expected movement vector of agent k, which is based on the independent decision-making process of the agent under the influence of repulsion and attraction only. herd The agent can decide its movement direction based on the combined effect of these forces (in the case of the repulsive force itself). The repulsive force encourages the agent to move away from potentially dangerous areas, while the attractive force guides the agent to move towards safe areas. The process of the agent calculating its movement vector is as follows: First, the basic movement vector is determined based on the vector difference between the current grid coordinates and the target grid cell coordinates. This vector points to where the agent expects to go and is based on the agent's assessment of the current environment (including the repulsive and attractive forces). Then, the movement vectors of all agents in the same grid are averaged to obtain

[0058]

[0059] In a feasible embodiment, after obtaining the first disaster perception index, the second disaster perception index and the third disaster perception index, the moving direction of the intelligent agent can be further determined based on these indexes. Specifically, first, a first perception weight is assigned to the first disaster perception index, a second perception weight is assigned to the second disaster perception index, and a third perception weight is assigned to the third disaster perception index. The setting of these weights is intended to reflect the relative importance of each perception index in the decision-making process of the intelligent agent. Then, according to the first perception weight, the second perception weight and the third perception weight, the contribution of the first disaster perception index, the second disaster perception index and the third disaster perception index are adjusted respectively, and an overall disaster perception index that comprehensively reflects the impact of the disaster environment can be obtained. Then, according to the overall disaster perception index and the preset perception factor, the moving direction of the intelligent agent can be determined. As shown in formula (4), after calculating the first disaster perception index F rep , Second Disaster Perception Index F attr and the third disaster perception index F herd After that, you can rep Assign weight w rep (For example, the setting value is 1), attr Assign weight w attr (For example, the setting value is 0.8), which is F herd Assign weight w herd(For example, the set value is 0.2). The actual application of these weights is to adjust the contribution of each perception index to the overall disaster perception index, thereby optimizing the agent's decision-making process for the direction of movement. In addition, in order to more comprehensively capture the inherent uncertainty in population mobility, a perception factor ε is introduced to analyze this variable nature. The dual sources of this uncertainty are: on the one hand, the movement of the agent itself is probabilistic, which means that even when facing the same environmental conditions, the agent may choose different movement paths; on the other hand, if the agent repeatedly moves to the same grid unit, it indicates that it has entered a stable mode without a strong directional preference. At this time, the agent is in a state of equilibrium. After calculation by formula (4), a resultant force F can be finally obtained, which can drive the agent to move to the grid unit with the highest safety.

[0060] F=F rep × rep +F attr × attr +F herd × herd +ε formula (4).

[0061] In a feasible embodiment, according to the guidance of the moving direction, the agent can select and move to a specific grid cell among multiple grid cells within its perceived proximity. However, in some situations, this selection is not absolutely certain, but is affected by the probability associated with the angle between the result vector (i.e., the predetermined moving direction) and the potential candidate direction vector. Figure 6 As shown in the figure, the probability distribution of the agent among multiple movement options is used to illustrate the possibility of its moving to any grid. Each grid is equipped with a unique ID, which is marked with numbers in the lower right corner of the respective grid cell. Within the agent's neighborhood, each grid corresponds to a specific angle based on polar coordinates: 0° corresponds to ID5, 45° corresponds to ID2, 90° corresponds to ID1, 135° corresponds to ID0, 180° corresponds to ID3, 225° corresponds to ID6, 270° corresponds to ID7, and 315° corresponds to ID8. In such a layout, if the result vector points to the area between ID5 and ID2, the agent has a probability of moving to any of the two grids. In this case, P2 and P5 can be calculated based on the angle difference θ between the result vector and 0° (ID5 direction), Figure 6 In the above figure, P5 represents the probability of the agent moving to the ID5 grid cell, and P2 represents the probability of the agent moving to the ID2 grid cell.

[0062] In a feasible embodiment, when the agent moves to the target grid unit within its perceived proximity, a movement prediction result can be obtained. This target grid unit is the grid unit with the largest number of safe locations within the perception range of the agent. In addition, when the agent is in a balanced state, that is, when the agent has no clear preference for the direction of movement, its final movement result can be predicted based on the grid unit where the agent finally stays. Further, considering the availability of resources and the limitations of infrastructure, when the capacity of the target grid unit to which the agent is about to move exceeds the preset capacity threshold, the agent can turn to the secondary target grid unit within the perceived proximity. Assume that the infrastructure and resource allocation of the target grid unit are only sufficient to support the operation of 10 agents, and the grid unit is currently fully loaded with 10 agents. In this case, if another agent plans to move to the grid unit, in order to avoid resource overload or overload of infrastructure, the agent can turn to detect other grid units within its proximity, namely the secondary target grid units. As alternatives, these secondary target grid units should be able to meet the basic needs of the agent without exceeding its capacity limit or the upper limit of resource allocation. During this process, the agent can restart a three-stage process (including information acquisition, perception, and movement decision) to find and move to a new target grid cell (sub-target grid cell). Although the number of safe locations in this sub-target grid cell is less than that of the target grid cell, it is still a feasible option. Figure 7 As shown in Figure 2, the process aims to ensure that the agent can efficiently utilize limited resources while ensuring safety and avoiding overcrowding of grid cells.

[0063] It should be noted that the embodiment of the present application aims to guide each agent to reach the largest number of safe locations within its perceived proximity (such as the Ω_(m×m) area), for example, the grid cell with the largest roof area of ​​the building. In this process, unless the agent finds that the optimal grid cell is overloaded when it arrives at the optimal grid cell, backtracking is generally not allowed, because compared with the previous step of movement, arriving at the current location is considered to be a better choice for disaster avoidance, which reflects the principle of gradual optimization. If the agent frequently revisits the same grid cell, it indicates that it has entered a stable movement mode and lacks a clear direction preference. This state can be called a "balanced" state. In this case, the agent can be allowed to stop moving after revisiting the same grid cell three times in a row. This rule is intended to reduce unnecessary continuous movement and computational costs, while still giving the agent the flexibility to moderately explore its surrounding environment before finally settling (i.e., terminating movement). This method can simulate the behavioral decision-making process of individuals in the real world under disaster conditions as much as possible while maintaining simulation efficiency. It is worth noting that in real life, the movement of individuals is often constrained by a complex road network and does not present a discrete movement mode as in the grid cell simulation. Therefore, in order to more continuously and realistically reflect the movement status of individuals in a certain area, Gaussian filtering processing technology can be applied to the individual movement prediction results in the grid cells.

[0064] See also Figure 8 , Figure 88 is a structural diagram of a disaster crowd movement prediction system provided by an embodiment of the present application. The prediction system highly integrates an information acquisition module 810, a perception module 820, a mobile decision module 830 and a result prediction module 840 to achieve accurate movement prediction of an intelligent agent in a disaster environment. First, the information acquisition module 810 can collect disaster information within the perceived proximity of the intelligent agent to provide a basis for subsequent decision-making. The perception module 820 further processes this information and calculates the first disaster perception index, the second disaster perception index and the third disaster perception index of the intelligent agent through its three internal sublayers - the repulsion layer 821, the attraction layer 822 and the convergence layer 823. Each of these three indexes carries different meanings: the first disaster perception index reflects the repulsion behavior of the intelligent agent in the face of disasters, that is, it tends to stay away from dangerous areas; the second disaster perception index reflects the attraction behavior of the intelligent agent due to disasters, which may point to safer or more favorable areas; the third disaster perception index reveals the convergence behavior of the intelligent agent in the face of disasters, that is, it tends to gather towards other intelligent agents or safe areas. Subsequently, the movement decision module 830 can comprehensively judge and determine the movement direction of the agent based on the three disaster perception indexes. Finally, the result prediction module 840 controls the agent to move within the perceived proximity according to the direction given by the movement decision module. When the agent successfully moves to the target grid area, the movement prediction result of the agent can be obtained, thereby completing the comprehensive prediction and simulation of the agent's movement behavior.

[0065] It should be noted that since the disaster crowd movement prediction system of this embodiment can implement the disaster crowd movement prediction method of the previous embodiment, the disaster crowd movement prediction system of this embodiment and the disaster crowd movement prediction method of the previous embodiment have the same technical principles and the same beneficial effects. In order to avoid repetition, they will not be repeated here.

[0066] Reference Fig. 9 The embodiment of the present application also discloses a mobility prediction device, which includes: at least one processor 910; at least one memory 920, used to store at least one program; when the at least one program is executed by the at least one processor 910, the method for predicting crowd mobility during disasters as described above can be implemented.

[0067] The embodiment of the present application also discloses a computer-readable storage medium, which stores a computer program executable by a processor. When the computer program executable by the processor is executed by the processor, the method for predicting crowd movement during disasters as described above can be implemented.

[0068] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting crowd movement during a disaster, characterized in that: include: Obtain disaster information in the agent's perceived neighborhood; Performing a perception operation according to the disaster information of the perceived proximity range, obtaining a first disaster perception index, a second disaster perception index and a third disaster perception index of the intelligent agent, wherein the first disaster perception index is used to characterize the rejection behavior of the intelligent agent in the face of disasters, the second disaster perception index is used to characterize the attraction behavior generated by the intelligent agent, and the third disaster perception index is used to characterize the convergence behavior of the intelligent agent in the face of disasters; determining a moving direction of the agent according to the first disaster perception index, the second disaster perception index, and the third disaster perception index; The agent is controlled to move within the perception proximity range according to the moving direction, and a movement prediction result is obtained when the agent moves to a target grid unit within the perception proximity range.

2. The method for predicting crowd movement during a disaster according to claim 1, characterized in that: After the agent moves to the target grid unit within the sensing proximity, the method further includes: When the capacity of the target grid unit exceeds a preset capacity threshold, the agent is controlled to move to a secondary target grid unit within the perception proximity; the number of safe positions in the secondary target grid unit is less than the number of safe positions in the target grid unit.

3. The method for predicting crowd movement during a disaster according to claim 1, characterized in that: The step of determining the moving directions of the plurality of intelligent agents according to the first disaster perception index, the second disaster perception index, and the third disaster perception index comprises: Assigning a first perception weight to the first disaster perception index, assigning a second perception weight to the second disaster perception index, and assigning a third perception weight to the third disaster perception index; According to the first perception weight, the second perception weight and the third perception weight, respectively, the first disaster perception index, the second disaster perception index and the third disaster perception index are adjusted for contribution to obtain an overall disaster perception index; The moving direction of the intelligent agent is determined according to the overall disaster perception index and a preset perception factor.

4. The method for predicting crowd movement during a disaster according to claim 1, characterized in that: The disaster information of the perceived proximity range includes the disaster situation and the number of safe locations of at least one grid unit within the perceived proximity range; The step of obtaining a first disaster perception index, a second disaster perception index, and a third disaster perception index of the intelligent agent according to the disaster information includes: Obtaining a first disaster perception index according to the disaster situation of the grid unit where the agent is located and a preset distance coefficient; Obtaining a second disaster perception index according to the number of safe locations of the grid unit where the agent is located; A third disaster perception index is obtained according to the first disaster perception index, the second disaster perception index, the grid unit where the agent is located, and the target grid unit.

5. The method for predicting crowd movement during a disaster according to claim 1, characterized in that: The perceived proximity range is obtained according to the following steps: Acquire a disaster simulation area, and divide the disaster simulation area into a plurality of grid units; A plurality of the grid units are combined according to the type of the agent to obtain a perceived proximity range.

6. The method for predicting crowd movement during a disaster according to claim 5, characterized in that: The types of the intelligent agents include first-type intelligent agents and second-type intelligent agents, and the perception proximity range includes a first-type perception proximity range and a second-type perception proximity range; when the intelligent agent belongs to the first-type intelligent agent, the perception proximity range is the first-type perception proximity range; when the intelligent agent belongs to the second-type intelligent agent, the perception proximity range is the second-type perception proximity range; the number of the grid units of the first-type perception proximity range is greater than the number of the grid units of the second-type perception proximity range.

7. A disaster-time crowd movement prediction system, characterized in that: include: The information acquisition module is used to obtain disaster information in the vicinity of the agent's perception; A perception module, used to perform a perception operation according to the disaster information of the perceived proximity range, and obtain a first disaster perception index, a second disaster perception index and a third disaster perception index of the intelligent agent, wherein the first disaster perception index is used to characterize the repulsive behavior of the intelligent agent in the face of disasters; the second disaster perception index is used to characterize the attractive behavior generated by the intelligent agent; and the third disaster perception index is used to characterize the convergent behavior of the intelligent agent in the face of disasters; A movement decision module, used for determining a movement direction of the intelligent agent according to the first disaster perception index, the second disaster perception index and the third disaster perception index; The result prediction module is used to control the intelligent agent to move within the perception proximity range according to the movement direction, and obtain a movement prediction result when the intelligent agent moves to a target grid unit within the perception proximity range.

8. The disaster crowd movement prediction system according to claim 7, characterized in that: The perception module includes a repulsion layer, an attraction layer and a convergence layer. The repulsion layer is used to perform perception operations based on the disaster information in the perception neighborhood to obtain a first disaster perception index. The attraction layer is used to perform perception operations based on the disaster information in the perception neighborhood to obtain a second disaster perception index. The convergence layer is used to perform perception operations based on the disaster information in the perception neighborhood to obtain a third disaster perception index.

9. A mobile prediction device, characterized in that: include: at least one processor; at least one memory for storing at least one program; At least one of the programs is executed by at least one of the processors to execute the method for predicting crowd movement during disasters as described in any one of claims 1 to 6.

10. A computer-readable storage medium storing computer-executable instructions, characterized in that: The computer executable instructions are used to execute the method for predicting crowd movement during disasters as described in any one of claims 1 to 6.

Images