Fire escape evacuation simulation method based on improved ant colony algorithm

Abstract

The present invention discloses a fire escape evacuation simulation method based on an improved ant colony algorithm. By setting a grid map and environmental parameters such as fire source points, crowd density, smoke diffusion information and other environmental factors, the objective function, constraint conditions, improved heuristic functions, and comprehensive impact index functions, temperature change functions, CO concentration change functions, and visibility change functions are further set. By improving the pheromone content update method, a pheromone update model that conforms to the crowd evacuation process is established, the traditional pheromone update method is improved, the calculation speed is accelerated, and it avoids falling into a local optimal solution. r The setting of needs to comprehensively consider the convergence speed and the global search ability of the algorithm. r The value is adjusted adaptively. It has practical value for the evacuation of personnel in a fire environment.

Description

Technical Field

[0001] The invention belongs to the technical field of cross-application of computer simulation and architecture, and relates to a fire escape evacuation simulation method based on an improved ant colony algorithm in a fire environment. Background Art

[0002] In recent years, fire accidents have occurred frequently in various parts of China, and the number of fires, property losses and casualties caused by them has shown an upward trend. Among them, building fires are the type of fire that causes the most direct economic losses and casualties. Especially in unfamiliar environments, people are not familiar with evacuation routes and fire-fighting equipment, which directly increases the difficulty of evacuation.

[0003] In response to the problems of existing technologies, some researchers have proposed using path planning algorithms such as the A* algorithm and the Dijkstra algorithm. However, traditional path planning algorithms are too simple and are increasingly unsuitable for complex evacuation environments, and are gradually being eliminated by the times. The emergence of intelligent algorithms such as genetic algorithms and ant colony algorithms has given researchers a bright spot, but this method is a choice determined by probability, is blind, and cannot effectively evacuate personnel.

[0004] Since most traditional path planning algorithms consider the optimal walking path of individuals in the current fire situation, they do not combine the entire dynamic evacuation process caused by the fire, and ignore the dynamic changes of the external environment, so that the simulation results are inconsistent with the actual situation. Although the intelligent algorithm can simulate the evacuation behavior, its operation speed is slow and the calculation complexity is high, which can easily cause individuals to fall into local optimality and congest the evacuation path, which will eventually affect the overall evacuation effect. Summary of the invention

[0005] In order to solve the problems existing in the prior art, the object of the present invention is to provide a fire escape evacuation simulation method based on an improved ant colony algorithm.

[0006] In order to achieve the above tasks, the present invention adopts the following technical solutions:

[0007] A fire escape evacuation simulation method based on an improved ant colony algorithm is characterized by comprising the following steps:

[0008] Step 1: Set the grid map and environment parameters;

[0009] Step 2: Set environmental factors such as fire source point, crowd density, and smoke diffusion information;

[0010] Step 3: Set the objective function as:

[0011] ;

[0012] in, mis the total number of people in the building who need to be evacuated; For the k The starting point of the evacuation personnel; ( i, j ) is the building converted into a node in the grid map i To Node j Path; Path k For the k evacuation routes for evacuees; For the k Evacuees passed through the path ( i,j ) For the k Evacuees passed through the path ( i,j ) at the time of

[0013] Step 4: Set the constraints as follows:

[0014]

[0015] In the formula, f (i,j) ( t )for t time( i,j ) population density; f min 、f max are the minimum and maximum population density limits respectively.

[0016] Step 5: Randomly select a person and place him on the initial node to determine whether the maximum number of iterations N is greater than the maximum number of iterations N max .

[0017] Step 6: If N < maximum iteration coefficient N max , then calculate the improved heuristic function.

[0018] Step 6-1: Improve the heuristic function by incorporating factors such as the distance and angle between the node and the target node, the temperature generated during the fire, the concentration of toxic and harmful gases, and visibility into the heuristic function, and set the improved heuristic function formula to:

[0019]

[0020] in, Representation Node i An adjacent node of j To the target node g Corrected distance between θ ig Indicates the current node iThe angle between the direction of the target node and the target node; l ij Representation Node i With Node j The distance between E ij Represents the comprehensive impact index of three influencing factors after a fire occurs: smoke temperature, toxic and harmful gases, and visibility; ξ ij For Node i To Node j The difficulty of passing is as follows:

[0021]

[0022] In the formula, v 0 represents the speed of a person walking on a straight path; v t Indicates the speed of a person walking on a specific path;

[0023] Step 6-2: Set the correction distance formula as follows:

[0024]

[0025] In the formula, d max Representation Node i The maximum distance between each adjacent node and the target node; d min Representation Node i The minimum distance between each adjacent node and the target node; d ( i,g ) is a node i To the target node g The distance value between them; the denominator is added with 0.01 to prevent the distance correction function from being meaningless.

[0026] Step 6-3: According to the three influencing factors of smoke temperature, toxic and harmful gases and visibility after the fire, set the comprehensive impact index function.

[0027]

[0028] P ij ( T ) represents the effect function of smoke temperature on human body; P ij ( ρ co ) represents the effect function of CO concentration on human body; P ij ( kc) indicates the impact of visibility on the human body.

[0029] Step 6-4: Set the temperature change function according to the trend of the impact of temperature on the human body changing over time after a fire occurs.

[0030]

[0031] Step 6-5: Set the CO concentration variation function according to the trend of the impact of CO concentration on human body over time after the fire.

[0032]

[0033] Step 6-6: Set the visibility change function according to the trend of the impact of visibility on human body changing with time after the fire occurs.

[0034]

[0035] Step 7: Improve the pheromone content update method, establish a pheromone update model that conforms to the crowd evacuation process, and improve the traditional pheromone update method;

[0036] Step 7-1: Establish a pheromone update model.

[0037]

[0038]

[0039]

[0040] in , ρ stands for pheromone volatile factor; is the path after this cycle ( i,j )Pheromone content on For the k Personnel stay in ( i,j ) L k For the k The length of the path taken by each person in this cycle; Q The total amount of pheromones released after a person completes a complete path search; λ1 、 λ2 is the adjustment coefficient, which is adjusted according to the actual situation; Path s is the path that can safely reach the target node from the current node in this iteration; Path e It is the path from the current location to the target location in this iteration that contains dangerous factors such as high temperature, high CO concentration, and low visibility.

[0041] Step 7-2: If the optimal solution of the method has not improved after multiple iterations, reduce ρ value to improve global search capabilities, and set ρ Minimum value of to avoid slowing down the convergence speed;

[0042]

[0043] Step 8: Use the roulette method to calculate the transfer probability of personnel as follows:

[0044]

[0045] In the formula, i Indicates the current node; j Indicates the next node to be transferred; k represents the serial number of the ant; p ij k ( t )for t Moment k Ant slave node i To Node j Probability of transfer; τ ij ( t )represent t Time Node i to the pheromone content between nodes; η ij ( t ) represents a node i To Node j The heuristic function between ; α stands for pheromone-inducing factor; β stands for expectation heuristic factor.

[0046] Step 9: If N = the maximum number of iterations, the evacuation distance and evacuation time are output according to the program, and an image of the personnel evacuation path on the grid map is drawn.

[0047] Step 10: Update the taboo table.

[0048] Step 11: Determine whether the person has reached the target node.

[0049] Step 12: If the person reaches the target node, update the pheromone according to step 7.

[0050] Step 13: Increase the number of iterations by 1, that is, N=N+1.

[0051] Step 14: Return to Step 5.

[0052] Step 15: If the person does not reach the target node, return to step 10.

[0053] Step 16: Iterate in sequence until the iteration is completed and the optimal evacuation distance and time are obtained.

[0054] The fire escape evacuation simulation method based on the improved ant colony algorithm of the present invention brings about the following technical innovations:

[0055] 1. Fully consider the dynamic process of fire, improve the ant colony algorithm, take into account factors such as temperature, smoke concentration, visibility, population density, etc. during the fire process, and establish an optimal evacuation path algorithm model that considers dynamic factors.

[0056] 2. The evacuation process of personnel in large buildings is related to many factors. The optimal evacuation path requires that all personnel arrive at the safe evacuation location in the shortest time while not exceeding the maximum congestion level. The objective function and constraints are established according to this requirement.

[0057] 3. The heuristic function in the traditional ant colony algorithm is usually defined as the inverse of the distance between two nodes. In a grid network environment, the distance between two nodes is generally a fixed value, which makes the heuristic function a constant and loses the meaning and heuristic effect of the function itself. The present invention improves the traditional heuristic function in a comprehensive way, and incorporates factors such as the distance and angle between the node and the target node, the temperature generated during the fire, the concentration of toxic and harmful gases and visibility into the heuristic function, so that the function can play a real heuristic role.

[0058] 4. Establish a pheromone update model that conforms to the crowd evacuation process, improve the traditional pheromone update method, speed up the calculation, and avoid falling into the local optimal solution.

[0059] 5. When ρ When the value is too large, the convergence speed is fast, but the algorithm diversity is reduced and the search is prone to stagnation early. ρ If the setting is too small, the convergence speed of the algorithm will slow down, the algorithm diversity will increase, and the search ability will be strengthened. In short, ρ A value that is too large or too small will have a certain impact on the program. ρ The setting of needs to comprehensively consider the convergence speed and the global search ability of the algorithm. ρ The value is adjusted adaptively. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a flow chart of a fire escape evacuation simulation method based on an improved ant colony algorithm of the present invention;

[0061] Figure 2 is the grid map original image.

[0062] Figure 3It is the graph of temperature change function.

[0063] Figure 4 It is the graph of CO concentration change function.

[0064] Figure 5 It is the graph of visibility change function.

[0065] Figure 6 It is the optimal evacuation path image during the fire period.

[0066] Figure 7 It is the optimal evacuation path image during the fire growth period.

[0067] Figure 8 It is the optimal evacuation path image during the peak period of fire.

[0068] Fig. 9 It is the convergence analysis diagram of the algorithm during the fire ignition period.

[0069] Fig.10 It is the convergence analysis diagram of the algorithm during the fire growth period.

[0070] Fig.11 This is the convergence analysis diagram of the algorithm during the peak period of fire.

[0071] The present invention is further described in detail below with reference to the accompanying drawings and embodiments. DETAILED DESCRIPTION

[0072] The overall design idea of ​​the present invention is to study an improved ant colony algorithm for path planning in a complex environment, provide a new pheromone update rule, and improve the global search capability of the algorithm; and consider the dynamic process of the fire, improve the ant colony algorithm, and take into account factors such as temperature, smoke concentration, visibility, and population density during the fire process, and establish an optimal evacuation path algorithm model that considers dynamic factors.

[0073] Taking into account the important factors related to the evacuation process of personnel in large buildings, the optimal evacuation path requires that all personnel arrive at the safe evacuation location in the shortest time while not exceeding the maximum congestion level - that is, the overall evacuation time is the shortest. At the same time, the evacuation distance is required to be the shortest and the passenger flow value on the path is reasonable.

[0074] On the other hand, the heuristic function in the ant colony algorithm is improved in a comprehensive way. Factors such as the distance and angle between the node and the target node, the temperature generated during the fire, the concentration of toxic and harmful gases, and visibility are integrated into the heuristic function to make the function play a real heuristic role. Then, a pheromone update model that conforms to the crowd evacuation process is established, and the traditional pheromone update method is improved. ρ The setting of comprehensively considers the convergence speed and the global search ability of the algorithm.

[0075] Part I:

[0076] See also Figure 1 This embodiment provides a fire escape evacuation simulation method based on an improved ant colony algorithm, and the steps are as follows:

[0077] Step 1: Set up the grid map and environment parameters

[0078] In the grid environment, this embodiment uses a 20*20 grid map. Assuming that the building under study is regarded as this simple grid map, the initial image is as follows: Figure 2 As shown. The white grid represents no obstacles and safe passage, while the black grid represents obstacles and inaccessibility. The obstacles here represent the buildings and display cabinets displaying exhibits in the venue. When designing obstacles, factors such as the circuitous tour routes of the exhibition venue are taken into consideration. Exit 1 and Exit 2 on the top and Exit 3 on the bottom represent the established safe exits, and the cross block in the middle represents the initial escape position.

[0079] Step 2: Set the fire source point, crowd density, smoke diffusion information and other environmental factors as shown in the following table:

[0080]

[0081] Step 3: Set the objective function

[0082] The objective function is:

[0083] ; ;

[0084] in, m is the total number of people in the building who need to be evacuated; For the k The starting point of the evacuation personnel; ( i, j ) is the building converted into a node in the grid map i To Node j Path; Path k For the k evacuation routes for evacuees; For the k Evacuees passed through the path ( i,j ) For the k Evacuees passed through the path ( i,j ) at the time of

[0085] Step 4: Set constraints

[0086] The constraints are:

[0087]

[0088] in, f (i,j) ( t )for t time( i, j ) population density; f min 、f max are the minimum and maximum population density limits respectively.

[0089] Step 5: Randomly select people and place them on the initial node to determine whether the maximum number of iterations N is greater than 100.

[0090] Step 6: If N<100, calculate the improved heuristic function.

[0091] Step 6-1: Improve the heuristic function by incorporating factors such as the distance and angle between the node and the target node, the temperature generated during the fire, the concentration of toxic and harmful gases, and visibility into the heuristic function, and set the improved heuristic function formula to:

[0092]

[0093] in, Representation Node i An adjacent node of j To the target node g The corrected distance between θ ig Indicates the current node i The angle between the direction of the target node and the target node; l ij Representation Node i With Node j The distance between E ij Represents the comprehensive impact index of three influencing factors after a fire occurs: smoke temperature, toxic and harmful gases, and visibility; ξ ij For Node i To Node j The difficulty of passing is as follows:

[0094]

[0095] In the formula, v 0 represents the speed of a person walking on a straight path; v t Indicates the speed of a person walking on a specific path;

[0096] Step 6-2: Set the correction distance formula.

[0097]

[0098] d max Representation Node i The maximum distance between each adjacent node and the target node; d min Representation Node i The minimum distance between each adjacent node and the target node; d ( i,g ) is a node i To the target node g The distance value between them; the denominator is added with 0.01 to prevent the distance correction function from being meaningless.

[0099] Step 6-3: According to the three influencing factors of smoke temperature, toxic and harmful gases and visibility after the fire, set the comprehensive impact index function.

[0100]

[0101] P ij ( T ) represents the effect function of smoke temperature on human body; P ij ( ρ co ) represents the effect function of CO concentration on human body; P ij ( kc ) indicates the impact of visibility on the human body.

[0102] Step 6-4: Set the temperature change function according to the trend of the impact of temperature on the human body changing over time after a fire occurs.

[0103]

[0104] In this embodiment, the temperature change function graph is as follows: Figure 3 shown.

[0105] Step 6-5: Set the CO concentration variation function according to the trend of the impact of CO concentration on human body over time after the fire.

[0106]

[0107] CO concentration change function graph see Figure 4 .

[0108] Step 6-6: Set the visibility change function according to the trend of the impact of visibility on human body changing with time after the fire occurs.

[0109]

[0110] For the graph of visibility change function, see Figure 5 .

[0111] Step 7: Improve the pheromone content update method, establish a pheromone update model that conforms to the crowd evacuation process, and improve the traditional pheromone update method.

[0112] Step 7-1: Establish a pheromone update model.

[0113]

[0114]

[0115]

[0116] in , ρ stands for pheromone volatile factor; is the path after this cycle ( i,j )Pheromone content on For the k Personnel stay in ( i,j ) L k For the k The length of the path taken by each person in this cycle; Q The total amount of pheromones released after a person completes a complete path search; λ1 、 λ2 is the adjustment coefficient, which is adjusted according to the actual situation; Path s is the path that can safely reach the target node from the current node in this iteration; Path e It is the path from the current location to the target location in this iteration that contains dangerous factors such as high temperature, high CO concentration, and low visibility.

[0117] Step 7-2: If the optimal solution of the method has not improved after multiple iterations, reduce ρ value to improve global search capabilities, and set ρ to avoid slowing down the convergence speed.

[0118]

[0119] Step 8: Use the roulette method to calculate the transfer probability of personnel.

[0120]

[0121] In the formula,i Indicates the current node; j Indicates the next node to be transferred; k represents the serial number of the ant; p ij k ( t )for t Moment k Ant slave node i To Node j Probability of transfer; τ ij ( t )represent t Time Node i to the pheromone content between nodes; η ij ( t ) represents a node i To Node j The heuristic function between ; α stands for pheromone-inducing factor; β stands for expectation heuristic factor.

[0122] Step 9: If N=100, the evacuation distance and evacuation time are output according to the program, and an image of the personnel evacuation path on the grid map is drawn.

[0123] Step 10: Update the taboo table.

[0124] Step 11: Determine whether the person has reached the target node.

[0125] Step 12: If the person reaches the target node, update the pheromone according to step 7.

[0126] Step 13: Increase the number of iterations by 1, that is, N=N+1.

[0127] Step 14: Return to Step 5.

[0128] Step 15: If the person does not reach the target node, return to step 10.

[0129] Step 16: Iterate in sequence until the iteration is completed, and obtain the optimal evacuation distance and time for different exits at different stages of the fire, the change trend of the convergence curve and the evacuation image as shown in the following table.

[0130]

[0131] Part II:

[0132] (1) Establish an appropriate personnel evacuation model, consider the shortest evacuation distance, the shortest evacuation time and the appropriate crowd density, and conduct mathematical modeling, which can be well combined with the improved ant colony algorithm.

[0133] (2) By simulating the influence functions of factors such as smoke temperature, CO concentration and visibility on the human body and incorporating them into the ant colony algorithm, the algorithm is made more in line with the actual situation and its applicability is improved.

[0134] (3) The heuristic function, pheromone update method and pheromone volatilization factor in the ant colony algorithm are improved so that its global search capability and convergence speed are not disturbed by external factors, thereby improving the algorithm's optimization ability.

[0135] Part III:

[0136] In this embodiment, Figure 6-Figure 11 The optimal evacuation path images during the fire ignition period are given respectively ( Figure 6 ), the optimal evacuation path image during the fire growth period ( Figure 7 ), the optimal evacuation path image during the peak period of fire ( Figure 8 ). And the convergence analysis diagram of the fire ignition period algorithm ( Fig. 9 ), Fire growth period algorithm convergence analysis diagram ( Fig.10 ) and the convergence analysis diagram of the algorithm during the peak period of fire ( Fig.11 ). From the results of the embodiment, different stages of fire, the fire start period, the growth period and the heyday, affect the choice of escape routes and exits of people in the building. In this embodiment, the improved ant colony algorithm adds external environmental factors. Under the operation of the algorithm, it has a better effect and has practical value for the evacuation of people in a fire environment.

Claims

1. A fire escape evacuation simulation method based on improved ant colony algorithm, characterized in that: The specific steps include: Step 1: Set the grid map and environment parameters; Step 2: Set the fire source point, crowd density, and smoke diffusion information; Step 3: Set the objective function as: ； ； in, m is the total number of people in the building who need to be evacuated; For the k The starting point of the evacuation personnel; ( i,j ) is the building converted into a node in the grid map i To Node j Path; Path k For the k evacuation routes for evacuees; For the k Evacuees through the path ( i,j ) For the k Evacuees through the path ( i,j ) at the time of Step 4: Set the constraints as: In the formula, f (i,j) ( t )for t time( i,j ) population density; f min 、f max are the minimum and maximum population density limits, respectively; Step 5: Randomly select a person and place him on the initial node to determine whether the maximum number of iterations N is greater than the maximum number of iterations N max ; Step 6: If N < maximum iteration coefficient N max , then calculate the improved heuristic function; Step 6-1: The improved heuristic function incorporates factors such as the distance and angle between the node and the target node, the temperature generated during the fire, the concentration of toxic and harmful gases, and visibility into the heuristic function, and sets the improved heuristic function formula as follows: in, Representation Node i An adjacent node of j To the target node g Corrected distance between θ ig Indicates the current node i The angle between the direction of the target node and the target node; l ij Representation Node i With Node j The distance between E ij Represents the comprehensive impact index of three influencing factors after a fire occurs: smoke temperature, toxic and harmful gases, and visibility; ξ ij For Node i To Node j The difficulty of passing is as follows: In the formula, v 0 represents the speed of a person walking on a straight path; v t Indicates the speed of a person walking on a specific path; Step 6-2: Set the correction distance formula as follows: In the formula, d max Representation Node i The maximum distance between each adjacent node and the target node; d min Representation Node i The minimum distance between each adjacent node and the target node; d ( i,g ) is a node i To the target node g The distance value between them; the denominator is added with 0.01 to prevent the distance correction function from being meaningless; Step 6-3: According to the three influencing factors of smoke temperature, toxic and harmful gases and visibility after the fire, the comprehensive impact index function is set as follows: In the formula, P ij ( T ) represents the effect function of smoke temperature on human body; P ij ( ρ co ) represents the effect function of CO concentration on human body; P ij ( kc ) indicates the degree of impact of visibility on the human body; Step 6-4: According to the trend of the temperature effect on the human body after the fire, the temperature change function is set as: Step 6-5: According to the trend of the influence of CO concentration on human body after fire, the change function of CO concentration is set as: Step 6-6: According to the trend of the impact of visibility on human body after the fire, the visibility change function is set as follows: Step 7: Improve the pheromone content update method, establish a pheromone update model that meets the crowd evacuation process, and improve the traditional pheromone update method; among which: Step 7-1: Establish the pheromone update model as follows: in , ρ stands for pheromone volatile factor; is the path after this cycle ( i,j )Pheromone content on For the k Personnel stay in ( i, j ) L k For the k The length of the path taken by each person in this cycle; Q The total amount of pheromones released after a person completes a complete path search; λ1 、 λ2 is the adjustment coefficient, which is adjusted according to the actual situation; Path s is the path that can safely reach the target node from the current node in this iteration; Path e The path from the current location to the target location in this iteration contains dangerous factors such as high temperature, high CO concentration, and low visibility; Step 7-2: If the optimal solution of the method has not improved after multiple iterations, reduce ρ value to improve global search capabilities, and set ρ Minimum value of to avoid slowing down convergence: Step 8: Use the roulette method to calculate the transfer probability of personnel as follows: In the formula, i represents the current node; j represents the next node to be transferred; k represents the sequence number of the ant; p ij k ( t )for t Moment k Ant slave node i To Node j Probability of transfer; τ ij ( t )represent t Time Node i to the pheromone content between nodes; η ij ( t ) represents a node i To Node j The heuristic function between ; α stands for pheromone heuristic factor; β represents the expectation heuristic factor; Step 9: If N = the maximum number of iterations, the evacuation distance and evacuation time are output according to the program, and an image of the personnel evacuation path on the grid map is drawn; Step 10: Update the taboo table; Step 11: Determine whether the person has reached the target node; Step 12: If the person reaches the target node, update the pheromone according to step 7; Step 13: Increase the number of iterations by 1, that is, N=N+1; Step 14: Return to step 5; Step 15: If the person has not reached the target node, return to step 10; Step 16: Iterate in sequence until the iteration is completed and the optimal evacuation distance and time are obtained.

Images