Emergency evacuation planning method based on improved ant colony algorithm and genetic fusion

Abstract

The invention discloses an emergency evacuation planning method based on improved ant colony algorithm and genetic fusion, and belongs to the technical field of emergency management and public safety. The method comprises the following steps: constructing a multi-objective optimization model taking minimization of total evacuation time T, total road section risk R and total personnel panic index P as objectives, and solving the multi-objective optimization model based on fusion of an improved ant colony algorithm and a genetic algorithm to obtain an optimal evacuation path scheme. The evacuation efficiency, safety and personnel psychological factors are comprehensively considered, the state transition probability and the pheromone updating strategy of the ant colony algorithm are improved, the global search mechanism of the genetic algorithm is introduced for deep fusion, and finally, the evacuation efficiency can be improved in a complex evacuation scene. And a global optimal evacuation path giving consideration to timeliness, safety and personnel psychological factors is quickly planned, and the evacuation efficiency is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of emergency management and public safety, more specifically, relates to a personnel emergency evacuation planning method based on fusion of improved ant colony algorithm and genetic algorithm. The present application can be used in emergency scenarios such as fire, earthquake, terrorist attack, etc., to plan and support decision-making for the rapid, efficient and safe evacuation of trapped personnel. BACKGROUND

[0002] Personnel evacuation path planning is a core link in emergency safety management, mainly studying how to use operational and decision optimization theory and computer technology to quickly generate an efficient personnel evacuation plan after a sudden event occurs, so that trapped personnel can safely and orderly evacuate from the dangerous area to the safe area, and minimize the casualties caused by the sudden event. The on-site environment is often complex and variable, with dynamic risks such as reduced path traffic capacity, exit congestion, and even local area failure. Therefore, it is necessary to quickly and accurately assess the evacuation situation and develop a scientific and reasonable evacuation path planning scheme to avoid delays in evacuation time, exacerbation of channel congestion, and even secondary disasters such as stampedes.

[0003] At the level of personnel emergency evacuation path planning, existing research usually takes total evacuation time, path length, and other objective indicators as the target function, builds a model, and selects a suitable algorithm for solution, while ignoring the influence of individual panic, herd effect, and other factors in the evacuation process, which may deviate from the psychological needs of trapped personnel in the event of a sudden event. When a sudden event occurs, trapped personnel are affected by various factors such as evacuation environment, accident scale, and traffic volume, and are prone to strong negative emotions such as panic and anxiety. This emotion often leads to crowd congestion, exit congestion, prolongs evacuation time, and affects subjective judgment of path safety; as the source of danger spreads, the dynamic deterioration of the evacuation environment, such as channel congestion and exit availability changes, will exacerbate the panic of personnel. In this case, if the evacuation plan does not take into account the panic psychological characteristics of personnel, and only relies on static optimal path calculation, it is easy to lead to loss of control of personnel emotions, increasing personnel casualties and property losses.

[0004] According to the search, the patent application with the application number CN202511255162.4 discloses a large-scale crowd evacuation simulation method considering congestion state real-time decision-making. The implementation steps of the method are as follows: first, obtaining the crowd to be evacuated, the shelter, and the road network data in the region; second, constructing a road travel time calculation model; third, constructing a full shelter warning model; fourth, constructing a shortest time flow model considering congestion state real-time decision-making; fifth, obtaining the crowd evacuation movement process by using crowd simulation and visualizing the process. The deficiency of this method is that it relies on a macro model and ignores the psychological factors and behavior characteristics of individuals. Individuals may lose rational judgment under emergency situations. Secondly, the full shelter warning in this method is based on the number of people flowing in at the future time step, and its accuracy depends heavily on the accuracy of the model itself. Once the crowd flow prediction deviates due to psychological factors, the warning may fail.

[0005] For example, the application with the application number CN202510932969.0 discloses a dynamic gathering and distributing guidance method for emergency evacuation of urban rail transit passenger flow. The implementation steps of the method include: first, calculating the neighborhood radius that meets the service requirements based on the building internal structure, guide staff information, passenger type, and real-time location data. At the same time, set the threshold values of neighborhood radius and density; second, clustering analysis is performed on the passengers in emergency evacuation based on the threshold values of the previous step, identifying the aggregated clusters and determining the specific location, number, and passenger density within each cluster; third, calculating the center aggregation point of each cluster based on the generated clusters; passengers are guided to the aggregation point through a mobile phone software, and the guide staff organizes all passengers near the point to evacuate in an orderly manner. The deficiency of this method is that: first, this method requires real-time acquisition of the accurate position and type of all passengers, and the positioning accuracy is affected by factors such as building structure and signal interference, which is difficult to meet in real emergency scenarios; second, this method assumes that the space is free of traffic, but in reality, dynamic hazards such as fire, smoke, and collapsed objects can block the passage. At the same time, path planning and time sequence control are not performed from the global perspective of the station, and the evacuation paths of multiple clusters may intersect or converge at the same bottleneck point, such as a narrow exit, which can easily lead to new congestion and conflicts. SUMMARY

[0006] In view of the deficiencies of the above-mentioned prior art, the present application provides a personnel emergency evacuation path planning method based on the fusion of improved ant colony algorithm and genetic algorithm, so as to solve the problem of unreasonable emergency evacuation path planning due to the influence of ignoring the panic psychology of personnel in the event. The present application comprehensively considers the evacuation time, evacuation road risk and the influence of personnel panic, so as to improve the rationality of personnel emergency evacuation path planning in the evacuation process, reduce the casualties and property losses caused by the emergency, and provide a reliable method reference for personnel emergency evacuation path planning and decision-making.

[0007] To achieve the above-mentioned purpose, the technical scheme provided by the present application is as follows: The present application provides an improved ant colony algorithm and genetic fusion emergency evacuation planning method, comprising: A multi-objective optimization model is constructed, with the minimum total evacuation time T, total road risk R and total personnel panic index P as the target, and the total objective function of the model is as follows: ; Among them, f 1、 f 2、 f 3 are the shortest total evacuation time, the minimum total road risk and the minimum personnel panic index, ω 1、 ω 2、 ω 3 are the weight coefficients corresponding to the three targets; The fusion of improved ant colony algorithm and genetic algorithm is used to solve the multi-objective optimization model, that is, to obtain the optimal evacuation path scheme.

[0008] Further, the fusion of improved ant colony algorithm and genetic algorithm is used to solve the multi-objective optimization model, comprising: Initialize the algorithm parameters, including the population size of genetic algorithm N , the maximum number of iterations t max , the pheromone evaporation coefficient ρ , the heuristic factor α 1、 α 2、 α 3, the expected heuristic factor β , the pheromone constant Q , the multi-objective weight coefficient ω 1、 ω 2、 ω 3; Initialize the population of genetic algorithm, and perform pre-search iteration solution on the multi-objective optimization model based on genetic algorithm to obtain an initial optimization solution set; The initial high-quality solution set obtained based on the genetic algorithm is used to initialize the multi-dimensional pheromone distribution of the ant colony algorithm, the multi-objective optimization model is iteratively solved again based on the improved ant colony algorithm, and periodic genetic operations are performed in the iterative solving process of the ant colony algorithm until the final iteration termination condition is reached, that is, the iteration is stopped, and the optimal evacuation path scheme is output.

[0009] Further, when the improved ant colony algorithm is used to iteratively solve the multi-objective optimization model, the state transition probability formula is calculated as follows: ; In the above formula, denotes the pheromone concentration on the path (i, j) with respect to time, denotes the pheromone concentration on the path (i, j) with respect to risk, denotes the pheromone concentration on the path (i, j) with respect to panic; α 1、 α 2、 α 3 are the corresponding heuristic factors, β is the expected heuristic factor; is the adaptive heuristic function.

[0010] Further, the adaptive heuristic function is calculated as follows: ; wherein, ω d 、 ω r 、 ω p are the distance, risk, and panic weights, respectively, and ω d + ω r + ω p = 1; θ is the panic influence coefficient.

[0011] Further, the initial high-quality solution set obtained based on the genetic algorithm is used to initialize the multi-dimensional pheromone distribution of the ant colony algorithm, the multi-objective optimization model is iteratively solved again based on the improved ant colony algorithm, and periodic genetic operations are performed in the iterative solving process of the ant colony algorithm, and the specific steps include: The initial high-quality solution set obtained based on the genetic algorithm is used to initialize the multi-dimensional pheromone distribution of the ant colony algorithm, and the parameter k is set to 1; Ant k constructs a path based on the improved multi-objective state transition probability; The three objective values of all the paths constructed by the ants are evaluated, and the single-objective weighted value is calculated; Based on single target fitness value, elite selection strategy is performed on the ant path to select high-quality individuals directly into the next generation, and the multi-dimensional pheromone distribution is updated; According to the set periodic trigger interval N GA Periodic genetic operation is performed, that is, the solution constructed by the current ant is taken as the initial population of the genetic algorithm, the selection, crossover and mutation operations of the genetic algorithm are performed once to generate a new offspring population, and the individuals with the highest fitness in the offspring population are used to replace the same number of individuals with the lowest fitness in the current ant colony, and the iteration operation is continued; It is judged whether the iteration termination condition is reached, if yes, the iteration is stopped and the optimal solution is output; otherwise, the fusion iteration solving of the genetic algorithm-improved ant colony algorithm is continued.

[0012] Further, the calculation formulas of the three targets: total evacuation time, total road risk and total personnel panic index are as follows: ; ; ; Wherein, W is the total number of nodes in the to-be-planned evacuation network environment, ; E is the total number of channels (or road sections), ; T is the evacuation time, d ij is the distance from node i to node j ; V ij ( t ) is the speed of node i to node j at time t; μ is the inhibition coefficient of panic emotion to speed, and 0<μ<1; S i ( t ), I i ( t ), R i ( t ) are the proportions of susceptible, panic and recovered individuals at node i at time t; is the decision variable, which represents whether the pedestrian n is on the road section (i, j); is the inherent risk from node i to node j ; λ is the amplification coefficient of panic emotion to risk; x ij ( t) is the number of people moving from node i to node j at time t; β i ( t ) is the panic emotion infection rate of node i at time t; N i ( t ) represents the total number of people at node i at time t.

[0013] Further, in the iterative solution process of the ant colony algorithm, the multi-dimensional pheromone distribution is updated by the following formula: ; represents the pheromone concentration on the path (i, j) with respect to time, represents the pheromone concentration on the path (i, j) with respect to risk, represents the pheromone concentration on the path (i, j) with respect to panic.

[0014] Further, the initialization of the genetic algorithm population is based on the genetic algorithm to pre-search and iteratively solve the multi-objective optimization model, comprising: using the genetic algorithm to pre-search, node i represents the evacuation starting point of the personnel, node j represents the next target node; first, initialize the genetic algorithm population, the population size is N , set the iteration number of the genetic algorithm, and start iteration; evaluate the three objective values of each individual in the initial population, and calculate the weighted single-objective fitness; perform genetic operations on the population to generate a new generation of population, the genetic operations include selection, crossover and mutation, until the pre-search iteration number is reached, that is, the initial optimization solution set of the genetic algorithm is obtained.

[0015] Further, the genetic operation on the population is specifically: selecting parent individuals by roulette selection method according to the fitness; performing crossover operation on the selected parent individuals to generate offspring individuals; performing node replacement mutation operation on the offspring individuals with a probability; repeating until the pre-search iteration number is reached, that is, the initial optimization solution set of the genetic algorithm is obtained.

[0016] Compared with the prior art, the technical scheme provided by the present application has the following beneficial effects: 1. The emergency evacuation path planning method of the present application has the advantages that: the multi-objective optimization model is constructed to minimize the total evacuation time, the total road risk and the total personnel panic index, and the weighted summation method is used to convert it into a single objective problem for solving, so that the performance of the emergency evacuation path planning method is better; the evacuation efficiency, safety and personnel psychological factors are comprehensively considered, the state transition probability and pheromone update strategy of the improved ant colony algorithm are combined with the global search mechanism of the genetic algorithm, and finally the globally optimal evacuation path is quickly planned under the complex evacuation scene, which takes into account the timeliness, safety and personnel psychological factors, and the evacuation efficiency is effectively improved.

[0017] 2. The initial population is generated by the genetic algorithm, and the pheromone distribution of the ant colony algorithm is initialized, so that the quality and diversity of the initial solution group are ensured, the search direction of the ant colony algorithm is provided with a high starting point, the defects of the traditional ant colony algorithm, such as blind initial search and slow convergence speed, are fundamentally overcome, and the initialization performance and overall convergence efficiency of the algorithm are significantly improved. Meanwhile, through the adaptive weight mechanism and multi-dimensional pheromone cooperative updating, the dynamic adjustment of each target in the dynamic evacuation environment is realized, so that the ants can consider the time, risk and psychological factors when selecting the path, and the actual emergency decision-making scene is more in line with the actual situation.

[0018] 3. The pheromone update strategy of the elite cooperation is adopted, and the new elite solution generated by the genetic algorithm and the elite ant path in the original ant colony are used to update the environmental pheromone. This strategy balances the global exploration and local development process, ensures that the excellent path experience can be efficiently accumulated, cooperatively plays the core advantages of the two algorithms, and speeds up the convergence process of the algorithm to the global optimal solution.

[0019] 4. Through comparison and verification in simulation experiments at different dangerous source development stages, the results show that, compared with the basic ant colony algorithm and the sparrow search algorithm, the fusion and improvement method has faster convergence speed and more stable solution result, and the planned path can dynamically adapt to environmental changes, which fully proves the effectiveness and superiority of the present application in dealing with dynamic and multi-objective emergency evacuation problems, and provides reliable technical support for actual emergency evacuation decision-making. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figures 1-3 is an emergency evacuation scene diagram of an embodiment of the present application; Figure 4 is an implementation flowchart of the emergency evacuation path planning method of an embodiment of the present application; Figure 5 is a personnel evacuation path at the initial appearance stage of a dangerous source obtained by different algorithms; Figure 6 is a convergence diagram of each path of four algorithms at the first layer in the evacuation scene at the initial appearance stage of a dangerous source; Figure 7 is the second layer of four algorithms in the evacuation scene in the initial stage of the hazard source, and the convergence diagram of each path; Figure 8 is the evacuation path of the hazard source in the initial growth stage solved by different algorithms; Figure 9 is the evacuation path of the hazard source in the diffusion stage solved by different algorithms. DETAILED DESCRIPTION

[0021] In view of the deficiency that the shortest equivalent length of a single path is taken as the target in traditional evacuation path planning, the application comprehensively considers evacuation time, road section risk and personnel panic, takes the minimum total evacuation time, the minimum total road section risk and the minimum panic index as the objective function, establishes a new decision principle, and constructs an emergency evacuation path planning model, so that the personnel evacuation process under the sudden event can be more truly and accurately described, which has great theoretical value and practical significance for improving the personnel evacuation efficiency under the sudden event and reducing personnel casualties and property losses. Meanwhile, the traditional ant colony algorithm is improved, the improved ant colony algorithm is combined with the genetic algorithm, the shortest path search problem is converted into a comprehensive path optimal search problem, so as to be beneficial to improving the rationality of the evacuation path planning.

[0022] In the improved ant colony algorithm, pheromone is set for the three targets of the shortest total evacuation time, the minimum total road section risk and the minimum personnel panic index respectively, and the state transition probability is optimized, the initial population is generated by global search of the genetic algorithm, and is used as the initial pheromone of the ant colony algorithm, so as to realize the combination of the ant colony and the genetic algorithm, and cooperatively play the advantages of global and local search, and finally the global evacuation efficiency can be improved and the system safety can be enhanced. The model provides a theoretical basis and a dynamic and reasonable path planning scheme for emergency evacuation in the scene of a subway station.

[0023] To further understand the content of the present application, the present application will be described in detail in conjunction with specific embodiments. However, it should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present disclosure. In the following detailed description, many specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments can be implemented without these specific details. In addition, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessary confusion of the concepts of the present disclosure.

[0024] It should be noted that the terms used herein are only used to describe specific embodiments and are not intended to limit the exemplary embodiments according to the present application. Unless otherwise explicitly stated, all technical terms and scientific terms used in the present application have the same meaning as generally understood by those skilled in the art to which the present application belongs.

[0025] The application can also be implemented in other ways different from the following description, can be similarly extended without departing from the spirit and essence of the application, but based on the embodiments of the application, other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application.

[0026] Specifically, the emergency evacuation planning method based on the fusion of improved ant colony algorithm and genetic algorithm of the application comprises: Step one, a multi-objective optimization model is constructed, which takes the minimization of total evacuation time T, total road risk R and total personnel panic index P as the target, and the total objective function of the model is as follows: ; Among them, f 1、 f 2、 f 3 are the shortest total evacuation time, the minimum total road risk and the minimum personnel panic index respectively, ω 1、 ω 2、 ω 3 are the weight coefficients corresponding to the three targets respectively; that is, the multi-objective is weighted into a single objective optimization function by weight coefficients ω 1、 ω 2、 ω 3.

[0027] Step two, the multi-objective optimization model is solved based on the fusion of improved ant colony algorithm and genetic algorithm, that is, the optimal evacuation path scheme is obtained.

[0028] Further, as shown in Figure 4 , the multi-objective optimization model is solved based on the fusion of improved ant colony algorithm and genetic algorithm, specifically comprising: S21, initialize algorithm parameters, including genetic algorithm population size N , maximum iteration number t max , pheromone evaporation coefficient ρ , heuristic factor α 1、 α 2、 α 3, expected heuristic factor β , pheromone constant Q , multi-objective weight coefficient ω 1、 ω 2、 ω 3; S22, initialize genetic algorithm population, iteratively solve the multi-objective optimization model based on genetic algorithm, and obtain an initial optimization solution set; S23, based on the initial optimization solution set obtained by solving the genetic algorithm, the multi-dimensional pheromone of the ant colony algorithm (including the time dimension pheromone matrixτ time , risk dimension pheromone matrix τ risk and panic dimension pheromone matrix τ panic ) distribution is initialized, the multi-objective optimization model is iteratively solved again based on the improved ant colony algorithm, and periodic genetic operation is performed in the iterative solving process of the ant colony algorithm; S24, when the final iteration termination condition is reached, such as reaching the maximum number of iterations of the ant colony algorithm t max , that is, stop iteration, and output the optimal evacuation path scheme.

[0029] Further, in the S22, the multi-objective optimization model is iteratively solved based on the genetic algorithm, comprising: S221, pre-searching is performed using the genetic algorithm, and nodes i represent the evacuation starting point of the personnel, and nodes j represent the next target node; first, the genetic algorithm population is initialized, the population size is N , the number of iterations of the genetic algorithm is set, and iteration is started; S222, each individual in the initial population is evaluated for its three objective values, and its weighted single-objective fitness (i.e. ) is calculated; the specific calculation formulas of the three objective values (total evacuation time, total road risk, and total personnel panic index) are as follows: ; ; ; Wherein, W is the total number of nodes in the evacuation network environment to be planned, ; E is the total number of channels (or road sections), ; T is the evacuation time, d ij is the distance from node i to node j ; V ij ( t ) is the speed of node i to node j at time t; μ is the inhibition coefficient of panic emotion on speed, and 0<μ<1; S i ( t ), I i ( t ), R i ( t) represent the proportions of susceptible individuals, panicked individuals, and recovered individuals at node i at time t; Let be the decision variable, representing pedestrians. n Is it on road segment (i, j)? For the node i To the node j The inherent risks; λ This represents the amplification factor of panic on risk. x ij ( t () represents the time from node t. i Move to node j Number of personnel; β i ( t () represents the panic infection rate at node i at time t; N i ( t () represents the total number of people at node i at time t. S i ( t )· I i ( t )· N i ( t This reflects the potential scale of panic spread.

[0030] S222. Perform genetic operations on the population to generate a new generation of population, wherein the genetic operations include selection, crossover, and mutation.

[0031] Specifically, in this application, parent individuals are selected by roulette wheel selection based on fitness; crossover operation is performed on the selected parent individuals to generate offspring individuals; node replacement mutation operation is performed on the offspring individuals with probability; this process is repeated until the pre-search iteration number is reached, thus obtaining the initial optimized solution set of the genetic algorithm.

[0032] In the first execution of the genetic algorithm, an initial population of size N is randomly generated using real number encoding, with each chromosome representing a complete evacuation path.

[0033] Furthermore, in step S23, the multi-objective optimization model is iteratively solved again based on the improved ant colony algorithm, including: S231. The initial high-quality solution set obtained by the genetic algorithm is used to initialize the multidimensional pheromone distribution of the ant colony algorithm. That is, the path information represented by the best individual obtained by the genetic algorithm pre-search is converted into the initial pheromone concentration of the ant colony algorithm, and the parameter setting is initialized to k=1. S232, Ant k constructs a path based on improved multi-objective state transition probability; In each main loop iteration, ant k constructs a complete evacuation path scheme based on pheromone concentration and heuristic information. Preferably, in this application, ant k starts from the starting point and selects the next node according to the following improved state transition probability formula: ; In the above formula, This represents the pheromone concentration over time along the path (i, j). This represents the pheromone concentration related to risk along the path (i, j). This represents the pheromone concentration related to panic along the path (i, j); α 1. α 2. α 3 represent the corresponding heuristic factors. β As a desired heuristic factor; This is an adaptive heuristic function.

[0034] Traditional ant colony algorithms rely solely on pheromone concentration and heuristic distance for state transition probabilities, neglecting the influence of multiple objectives. This invention proposes a multi-objective adaptive state transition probability formula. The improved rule enables ants to consider time, risk, and psychological factors simultaneously when choosing a path, and to comprehensively evaluate the path during path selection, thus better aligning with real-world emergency decision-making scenarios.

[0035] Traditional heuristics typically only provide the reciprocal of the path length. This invention proposes an improved multi-objective adaptive heuristic function to collaboratively optimize three objectives: time, risk, and panic. The specific formula is as follows: ; in, ω d , ω r , ω p These are weighted by distance, risk, and panic, respectively, which can dynamically balance the importance of each objective. ω d + ω r + ω p =1; θ This is the panic impact coefficient. The larger the value, the more sensitive the ants are to panic and the more they tend to avoid panic events.

[0036] In traditional algorithms, ants are guided by the single objective of "shortest distance" when choosing the next node. However, the improved algorithm is guided by a heuristic function that integrates efficiency, safety, and psychological factors, which enables the algorithm to directly respond to multi-objective requirements in the path construction stage.

[0037] S233. Evaluate the three objective values ​​of all paths constructed by ants and calculate the single objective weighted value; based on the single objective fitness value, implement an elite selection strategy for ant paths, select high-quality individuals to directly enter the next generation, and update the multidimensional pheromone distribution; Furthermore, the elite selection strategy applied to the ant paths, selecting high-quality individuals to directly enter the next generation, specifically involves calculating the fitness value of all ant-constructed schemes during the iteration process according to the following formula, and updating the global historical optimal solution: .

[0038] Traditional ant colony optimization algorithms use a single pheromone matrix, typically set to a uniform constant. This invention utilizes a high-quality solution set obtained through pre-search using a genetic algorithm to initialize a pheromone distribution with a clear search orientation. It proposes a pheromone matrix with three dimensions: time, risk, and panic. The multidimensional pheromone distribution on each edge (i, j) is updated according to the following formula. When ants select the next node, they are no longer guided solely by the single objective of "shortest distance," but can simultaneously optimize multiple objectives, adapting to dynamic environmental changes.

[0039] .

[0040] In this application, when updating the multidimensional pheromone distribution, the pheromone evaporation coefficient... ρ The range is between 0 and 1; This represents the pheromone increment caused by ant k along path (i, j) in the target dimension z: .

[0041] S234. During the iterative solution process of the ant colony algorithm, according to the set periodic trigger interval... N GA Perform periodic genetic operations, that is, use the solution constructed by the current ants as the initial population of the genetic algorithm, and repeat S22 to S23.

[0042] This application integrates an improved ant colony algorithm with a genetic algorithm, thereby leveraging the combined advantages of global exploration and local search. Simulation experiments have verified the algorithm's effectiveness. This invention can significantly improve the efficiency and quality of path planning, providing safe and efficient optimal evacuation routes for personnel in locations such as subway stations during emergencies.

[0043] Reference Figures 1-3This invention, using a two-story subway station as an example, further illustrates and verifies the effectiveness of the disclosed solution through simulation experiments, simplifying each floor into a 40×40 two-dimensional simulation map. To verify the actual effect of the algorithm, this invention constructs three typical scenarios based on the evolution logic of hazard sources: the initial appearance stage of hazard sources, the initial growth stage of hazard sources, and the full spread stage of hazard sources. For each of the above scenarios, dynamic adjustments and optimizations are implemented for personnel escape routes.

[0044] 1. Simulation conditions.

[0045] The software platform for the simulation experiment of this invention is: Windows 10 operating system and MATLAB 2022a.

[0046] The present invention further describes a specific scenario in the embodiments. The evacuation scenario is set as subway station A, which has two levels: a platform and a concourse. The layout of the subway station is simplified, with each level represented by a 40×40 two-dimensional simulation map. For example... Figure 1 , Figure 2 As shown, the first floor is the platform level, with four exits leading to the second floor. The red markings indicate the evacuation start point on the first floor, and the green markings indicate the evacuation end point on the first floor. The second floor is the concourse level. People evacuate from the exits on the first floor to the second floor to the final exit location, and the source of danger is located on the second floor.

[0047] In the simulation experiment, the population size N The number of ants is 150. K The value is 50; to balance the three objectives of shortest path, minimum risk, and minimum panic, this paper designs an independent heuristic factor. α 1. α 2. α The values ​​for 3 are 1.0, 1.2, and 0.8 respectively; the specific parameter settings are shown in Table 1 below.

[0048] Table 1 Algorithm Parameter Settings

[0049] 2. Simulation content and result analysis.

[0050] The fusion algorithm of this application is compared with the basic ACO algorithm (ant colony algorithm), the improved ACO algorithm, and the SSA algorithm (sponge search algorithm). In the initial stage of hazard source appearance, the optimal evacuation path lengths of the four algorithms in the first and second layers are shown in Tables 2 and 3.

[0051] Table 2 Comparison of Optimal Evacuation Path Lengths for People on the First Floor

[0052] Table 3 Comparison of Optimal Evacuation Path Lengths for Personnel on the Second Floor

[0053] According to Table 2, Table 3, Figure 5 , Figure 6 , Figure 7 It can be seen that the fusion improved ant colony algorithm-genetic algorithm of this application has improved the optimal path compared with the other three algorithms. The optimal path length of the first layer four-path fusion algorithm is reduced by an average of about 10.31% compared with the basic ant colony algorithm and by an average of about 14.07% compared with the sparrow search algorithm. The optimal path length of the second layer four-path fusion algorithm is reduced by an average of about 12.50% compared with the basic ant colony algorithm and by an average of about 18.25% compared with the sparrow search algorithm.

[0054] In this embodiment of the invention, the diffusion stage is set so that the two exits cannot pass due to the spread of the hazard source. The optimal path lengths for personnel evacuation of the four algorithms in the second layer are shown in Tables 4 and 5.

[0055] Table 4 Comparison of Optimal Evacuation Path Lengths for Second-Level Personnel During the Initial Growth Stage of Hazard Sources

[0056] Table 5 Comparison of optimal evacuation path lengths for personnel in the second stage of hazard diffusion.

[0057] Combining Table 4 and Table 5, Figure 8 and Figure 9 It can be seen that all four algorithms can obtain the optimal objective function value during the iteration process, and the overall convergence trend is similar to that in the hazard discovery stage. However, the path length increases in all four algorithms. This may be because the spread of the hazard leads to increased complexity in path planning, reduced traversable paths, and increased hazard level, thus increasing the difficulty of path planning. The improved ACO-GA algorithm proposed in this application generally outperforms the basic ACO algorithm, the improved ACO algorithm, and the SSA algorithm in terms of final convergence value, indicating that this algorithm can find a higher-quality path planning scheme and effectively cope with the complexities brought about by the hazard.

[0058] As the hazard spreads, the evacuation path length increases in the hazard diffusion stage compared to the first two stages. Because exits A and D are blocked, people starting from point a choose exit C, and people starting from point d choose exit B, significantly increasing the path lengths of the first and fourth routes. The table shows that, considering safety and public panic, the improved ACO-GA algorithm produces shorter paths. Shorter paths reduce total evacuation time and increase the chances of survival. The optimal path length of the four routes using the improved ACO-GA algorithm is reduced by an average of approximately 17.73% compared to the basic ant colony algorithm and by an average of 19.51% compared to the sparrow search algorithm.

[0059] The simulation experiments above show that the emergency evacuation route planning scheme obtained by the method of the present invention comprehensively considers three indicators: evacuation time, evacuation route risk, and personnel panic. It adopts the improved ant colony algorithm and genetic algorithm to obtain the optimal emergency evacuation route scheme, which effectively improves the personnel evacuation efficiency and makes the optimal emergency evacuation route scheme obtained by the present invention more valuable for practical application.

Claims

1. An emergency evacuation planning method that combines improved ant colony algorithm and genetic fusion, characterized in that, include: A multi-objective optimization model is constructed with the objectives of minimizing the total evacuation time T, the total road segment risk R, and the total panic index P. The specific objective function of the model is as follows: ； in, f 1. f 2. f 3 represents the shortest total evacuation time, the minimum total road section risk, and the minimum public panic index, respectively. ω 1. ω 2. ω 3 represents the weighting coefficients for the three objectives; The multi-objective optimization model is solved by combining the improved ant colony algorithm and the genetic algorithm, thus obtaining the optimal evacuation path scheme.

2. The emergency evacuation planning method according to claim 1, characterized in that, The solution to the multi-objective optimization model based on the fusion of the improved ant colony algorithm and the genetic algorithm includes: Initialize algorithm parameters, including the population size for the genetic algorithm. N Maximum number of iterations t max pheromone volatility coefficient ρ heuristic factor α 1. α 2. α 3. Expected Heuristic Factors β pheromone constant Q Multi-objective weighting coefficients ω 1. ω 2. ω 3; Initialize the genetic algorithm population, and use the genetic algorithm to perform pre-search iterative solution to the multi-objective optimization model to obtain the initial optimization solution set; The initial high-quality solution set obtained by the genetic algorithm is used to initialize the multidimensional pheromone distribution of the ant colony algorithm. The multi-objective optimization model is then iteratively solved again based on the improved ant colony algorithm. During the iterative solution process of the ant colony algorithm, periodic genetic operations are performed until the final iteration termination condition is reached, at which point the iteration stops and the optimal evacuation path scheme is output.

3. The emergency evacuation planning method according to claim 2, characterized in that, When iteratively solving a multi-objective optimization model using the improved ant colony algorithm, the state transition probability formula is calculated as follows: ； In the above formula, This represents the pheromone concentration over time along the path (i, j). This represents the pheromone concentration related to risk along the path (i, j). This represents the pheromone concentration related to panic along the path (i, j); α 1. α 2. α 3 represent the corresponding heuristic factors. β As a desired heuristic factor; This is an adaptive heuristic function.

4. The emergency evacuation planning method according to claim 3, characterized in that, The adaptive heuristic function The calculation is as follows: ； in, ω d , ω r , ω p These are distance, risk, and panic weights, respectively. ω d + ω r + ω p =1; θ This represents the panic impact coefficient.

5. The emergency evacuation planning method according to any one of claims 2-4, characterized in that, The initial high-quality solution set obtained based on the genetic algorithm is used to initialize the multidimensional pheromone distribution of the ant colony algorithm. The multi-objective optimization model is then iteratively solved again based on the improved ant colony algorithm. Furthermore, periodic genetic operations are performed during the iterative solution process of the ant colony algorithm, specifically including: The initial high-quality solution set obtained based on the genetic algorithm is used to initialize the multidimensional pheromone distribution of the ant colony algorithm, and the initialization parameter is set to k=1. Ant k constructs paths based on improved multi-objective state transition probabilities; Evaluate the three objective values ​​of all paths constructed by the ants and calculate the weighted value of each objective; Based on the single-objective fitness value, an elite selection strategy is executed on the ant path to select high-quality individuals to directly enter the next generation, and the multidimensional pheromone distribution is updated. According to the set periodic trigger interval N GA Perform periodic genetic operations, that is, use the solution constructed by the current ants as the initial population of the genetic algorithm, perform the selection, crossover and mutation operations of the genetic algorithm once to generate a new offspring population, and replace the number of individuals with the lowest fitness in the current ant colony with the individuals with the highest fitness in the offspring population, and continue the iterative operation; Determine if the iteration termination condition has been met. If it is met, stop the iteration and output the optimal solution; otherwise, continue the fusion iteration of genetic algorithm and improved ant colony algorithm to solve the problem.

6. The emergency evacuation planning method according to claim 5, characterized in that, The formulas for calculating the three objectives—total evacuation time, total road segment risk, and total public panic index—are as follows: ； ； ； Where W represents the total number of nodes in the planned evacuation network environment. E represents the total number of channels (or road segments). T represents the evacuation time. d ij For the node i To the node j The distance; V ij ( t ) represents the node at time t. i To the node j The speed is μ, where μ is the inhibition coefficient of panic on speed, and 0 < μ < 1. S i ( t ), I i ( t ), R i ( t ) represent the proportions of susceptible individuals, panicked individuals, and recovered individuals at node i at time t; Let be the decision variable, representing pedestrians. n Is it on road segment (i, j)? For the node i To the node j The inherent risks; λ This represents the amplification factor of panic on risk. x ij ( t ) represents the time from node t. i Move to node j Number of personnel; β i ( t () represents the panic infection rate at node i at time t; N i ( t () represents the total number of people at node i at time t.

7. The emergency evacuation planning method according to claim 5, characterized in that, During the iterative solution process of the ant colony algorithm, the multidimensional pheromone distribution is updated using the following formula: ； This represents the pheromone concentration over time along the path (i, j). This represents the pheromone concentration related to risk along the path (i, j). This represents the pheromone concentration related to panic along the path (i, j).

8. The emergency evacuation planning method according to any one of claims 2-4, characterized in that, The initialization of the genetic algorithm population, based on the genetic algorithm, involves pre-searching and iteratively solving the multi-objective optimization model, including: A genetic algorithm is used for pre-search, nodes i Indicates the starting point and node of personnel evacuation. j Indicate the next target node; first initialize the genetic algorithm population, with a population size of... N Set the number of iterations for the genetic algorithm and start the iteration; For each individual in the initial population, evaluate its three objective values ​​and calculate its weighted single-objective fitness; Genetic operations are performed on the population to generate a new generation of population. These genetic operations include selection, crossover, and mutation, until the pre-search iteration number is reached, which yields the initial optimized solution set of the genetic algorithm.

9. The emergency evacuation planning method according to claim 8, characterized in that, The genetic operation performed on the population is as follows: based on fitness, parent individuals are selected using roulette wheel selection; crossover is performed on the selected parent individuals to generate offspring individuals; node replacement mutation is performed on the offspring individuals with probability; this process is repeated until the pre-search iteration number is reached, thus obtaining the initial optimized solution set of the genetic algorithm.

Images