CN115527369A

CN115527369A - Large passenger flow early warning and evacuation method under large-area delay condition of airport hub

Info

Publication number: CN115527369A
Application number: CN202211200316.6A
Authority: CN
Inventors: 黄爱玲; 马瑞晨; 王瀚林; 张玄弋; 关伟; 杨文慧; 秦倩; 王子吉安
Original assignee: Beijing Capital International Airport Co ltd; Beijing Jiaotong University
Current assignee: Beijing Capital International Airport Co ltd; Beijing Jiaotong University
Priority date: 2022-09-29
Filing date: 2022-09-29
Publication date: 2022-12-27
Anticipated expiration: 2042-09-29
Also published as: CN115527369B

Abstract

The invention provides a large passenger flow early warning and evacuation method under the condition of large-area delay of an airport terminal. The method comprises the following steps: establishing a large-area delay early warning mechanism and a large-passenger flow early warning mechanism based on flight data; establishing a statistical model of the combination of the passenger check-in and security check physical processes, and predicting passenger aggregation amount through the statistical model based on real-time airport flight data and an early warning mechanism; acquiring travel rules of passengers related to an airport based on taxi track data, determining evacuation directions, evacuation areas and evacuation point positions of passenger flows, and constructing a vehicle scheduling model comprising an evacuation vehicle bearing scheme, path optimization and schedule optimization; and solving the vehicle dispatching model through an improved genetic algorithm, and solving the individual with the minimum fitness function value as a model solving result. The invention aims at evacuation of large passenger flows under the large-area delay background of large-scale airport hubs, improves the emergency response efficiency of airport ground traffic centers and promotes the intelligent construction of flexible traffic hubs.

Description

Large passenger flow early warning and evacuation method under large-area delay condition of airport hub

Technical Field

The invention relates to the technical field of passenger flow early warning and evacuation, in particular to a large passenger flow early warning and evacuation method under the condition of large-area delay of an airport terminal.

Background

With the rapid increase of the urbanization and motorization level of the world, the traffic volume is increasing day by day. The research of a large-passenger-flow early warning and evacuation method system under the background of large-area delay of an airport terminal is an important field for creating a flexible traffic system and improving the emergency handling and recovery capability of the traffic terminal, and is not possible to be generalized with the traditional emergency handling method of fire, public safety events and the like. But at present, the research of the integrated design including emergency response mechanism, evacuation position selection and emergency vehicle dispatching under the background of large-area delay is hardly seen. The existing emergent large-passenger-flow evacuation research is limited to large-passenger-flow risk prediction, operation diagram response adjustment and the like under the urban rail transit scene, verification of real data applied to practical cases is lacked, and the method is deficient in emergent evacuation research under large-area delay of airport terminals. Therefore, it is necessary to design a set of efficient early warning, responding and recovering method system in consideration of the airport operation mechanism.

The evacuation demand prediction work has a lot of influence factors, and most researches directly or indirectly make rough estimation on evacuation demands based on publicly available data, actual survey data and the like. At present, researches are also carried out on the basis of the influence factors and by combining a traditional four-stage method, and the evacuation demand is predicted by adopting a centralized original unit trip generation method, a cross classification method, a gravity model method and the like. Or from the individual selection perspective, a binary selection prediction method of non-ensemble Logistic regression is adopted to predict whether the individual is evacuated or left. In other researches, the evacuation scale is determined by a simulation technology, or a statistical model is utilized to depict the main characteristics (such as sex, age, race, income, reaction time, moving speed and typical behavior) or the environmental characteristics (such as environmental prompt, alarm receiving and distance from disaster) of an evacuation object, so that the group regularity is researched under the environmental disturbance, and the evacuation demand is predicted. The prediction research aiming at the airport passenger aggregation is influenced by multiple factors such as passenger individuality, environmental characteristics and the like, but the arrival distribution, departure distribution and aggregation distribution of the airport passengers have certain regularity, and the individual and environmental characteristics are incorporated into the statistical model parameters, so that the passenger aggregation characteristics can be described in a targeted manner, and the evacuation requirement is determined according to the airport operation and management regulations.

Early studies began to predict aggregated passengers using probability density function methods, considering the passenger arrival distribution as a normal distribution under small sample conditions. According to the scheme, the arrival time of the passengers is obviously influenced by airport reachability by comparing the arrival distribution of the passengers at the airport, and the arrival time distribution of the passengers in the low-reachability system is deviated to the right wing, so that the deviation characteristic of the accumulation time of the passengers at the terminal can be more expressed by using a Pearson III-type distribution function than normal distribution fitting. Because the check-in area and the waiting area in the terminal building have different building functions and service functions, and the passenger gathering characteristics of different areas are different, the Pearson III-type distribution of the passenger stay time in the check-in area and the waiting area is respectively calibrated by using the observation data in three seasons. By introducing a WIFI data source and a Bayesian method, the accuracy of predicting the aggregated passenger volume by the statistical model is further improved. However, the above researches all treat the passenger stay time distribution as a whole, the model is not split into an arrival check-in distribution and a departure distribution, and even if the model can be split, prior distribution information needs to be introduced, so that the coordination of flight delay factors and airport management factors with the passenger arrival and departure processes cannot be flexibly reflected, the variable of delay time cannot be increased in the arrival distribution and the departure distribution, and the complex aggregation scene under large-area flight delay cannot be adapted.

In the traditional emergency scene of city level or regional level, evacuation objects are often distributed dispersedly, people can spontaneously choose to go to hotels, friends or relatives for evacuation and refuge without unified evacuation command, and the individual refuge places are often selected by adopting a Logit model or a logic model with time periods and destinations nested. Under the condition of centralized evacuation with an emergency plan, the traditional methods for selecting the centralized evacuation destination comprise a P median model, a P mean model, an aggregation coverage problem, a maximum coverage problem and the like. Besides the static addressing model, a research and application simulation method is used for capturing the dynamic characteristics of the traffic flow of the road network, and a collaborative optimization method for selecting evacuation points and formulating an evacuation plan based on the dynamic change and the capacity limit of the traffic of the road network is provided. The method belongs to a model-driven evacuation node selection method, the objective function of the method is usually loss minimization or income maximization, and the selected influence factors comprise evacuation distance, evacuation population number and the like. Pure site selection has produced abundant theoretical results, but the factors in the aspects of efficiency, cost and the like of secondary evacuation of returning passenger flows need to be considered in the evacuation point selection under the large passenger flow evacuation background, and the large passenger flow hub is selected as an evacuation destination with certain blindness, so that on one hand, efficient evacuation is not always realized, on the other hand, the hub load is increased, and the normal traffic order is disturbed, so that the site selection research of the large passenger flow evacuation destination is relatively lacked.

For large-scale evacuation at regional and urban levels, research is often conducted on how many residents leave dangerous areas at the highest speed by using cars and other vehicles, and evacuation efficiency is measured by road network emptying time. For special dangerous scenes (such as island evacuation and dangerous substance diffusion area evacuation) or scenes that residents lack evacuation tools, it is more efficient to evacuate by adopting large-capacity transportation tools represented by buses. Many researches establish a VRP (Virtual Reality Platform) or TSP (total suspended particle) model to minimize the evacuation time, consider time window constraints, vehicle capacity constraints, line length constraints, and the like, and increase evacuation efficiency metrics such as the evacuation time, queuing length, and the like. Special considerations such as refueling, evacuation of special crowds (disabled people, old people, children and the like), evacuation with pets and the like are increased in research, and the problem of path planning of coordinated scheduling of multiple traffic modes is solved. In order to improve the scene adaptability of the model, some schemes incorporate the evacuation requirement into a continuously evolving system state, so that a dynamic VRP model is constructed. The design that the bus runs along a fixed route is overcome, a return scheduling method for the coordinated optimization of a schedule and an evacuation path is realized, the total evacuation time including the on-board time, the waiting time and the like is shortest, and a double-layer combination algorithm integrating an insertion algorithm and an edge exchange algorithm is designed to solve a model. The evacuation path planning problem is proved to be an NP-hard problem, and the method for solving the VRP problem comprises an accurate algorithm, a heuristic algorithm and the like. When the problem scale is large, heuristic algorithms such as a C-W (computer-to-W) conservation method, a nearest neighbor method, a nearest interpolation method, a genetic algorithm, an ant colony algorithm, a simulated annealing algorithm, a bee colony algorithm and the like are mostly adopted, and the combination of algorithms such as large-field search, variable-field search and the like matched with the model is also available.

The disadvantages of the evacuation strategy and the evacuation planning research scheme in each emergency scene in the prior art include: the method aims at the situation that the evacuation research of sudden large passenger flows caused by large-area delay of airport hubs is less, the research of each evacuation link is relatively independent, and a top-level design and coordination mechanism is lacked. The existing statistical model does not introduce airport operation factors, and then the passenger gathering time-varying condition is difficult to accurately fit.

The evacuation destination selection and the emergency scene are highly dependent, the existing site selection model is more focused on the theory, and the actual evacuation point meeting the evacuation requirement of the large passenger flow is difficult to accurately select; the conventional path planning model does not consider the randomness of the demand of passengers, does not consider the combination of the cost of the passengers and the operation cost of a manager, and makes different path planning schemes under the conditions of different emergency levels.

Disclosure of Invention

The embodiment of the invention provides a large passenger flow early warning and evacuation method under the condition of large-area delay of a airport terminal, so that large passenger flow evacuation can be effectively carried out under the background of large-area delay of the airport terminal.

In order to achieve the purpose, the invention adopts the following technical scheme.

A large passenger flow early warning and evacuation method under the condition of large-area delay of an airport terminal comprises the following steps:

establishing a large-area delay early warning mechanism and a large passenger flow early warning mechanism based on flight data;

establishing a statistical model of the combination of the passenger check-in and security check physical processes, and predicting passenger aggregation amount through the statistical model based on real-time airport flight data and the large-area delay early warning mechanism and the large passenger flow early warning mechanism;

acquiring travel rules of passengers related to an airport based on taxi track data, and determining evacuation directions, evacuation areas and evacuation point positions of passenger flows according to the travel rules of the passengers related to the airport;

constructing a vehicle scheduling model containing an evacuation vehicle bearing scheme, path optimization and schedule optimization according to the passenger gathering amount, the evacuation direction, the evacuation area and the evacuation point location;

and solving the vehicle scheduling model through an improved genetic algorithm, and solving an individual with the minimum fitness function value as a model solving result.

Preferably, the establishing of the large-area delay early warning mechanism and the large-passenger-flow early warning mechanism based on the flight data comprises the following steps:

setting the normal flight passenger aggregation model as A _normal The model of passenger gathering of delayed flights is A _delayed Flight data set is

t ⁽ⁿ⁾ To actual takeoff time, d ⁽ⁿ⁾ To delay duration, m ⁽ⁿ⁾ For carrying passenger, y ⁽ⁿ⁾ For flight attribute labels obtained after cross classification, different attribute labels correspond to different terminal buildings and mathematical models, the accumulated number F 'and accumulated detained passengers S of the flights delayed for over 4 hours are counted, when the accumulated number F' reaches a corresponding threshold value or when the accumulated detained passengers S reach the corresponding threshold value, large-area delay early warning is started, and the method uses A _normal Model estimation of passenger aggregation quantity to use A _delayed The model estimates the passenger gathering amount, and the algorithm outputs whether a large-area delay early warning signal and a passenger gathering distribution map appear in the solar navigation station building.

Preferably, the establishing of the statistical model of the combination of the passenger check-in and security check physical processes predicts the passenger gathering amount through the statistical model based on the real-time airport flight data, the large-area delay early warning mechanism and the large passenger flow early warning mechanism, and comprises the following steps:

establishing a statistical model based on the combination of passenger check-in and security check physical processes, and assuming that the time t of the departure passenger arriving at the terminal building is (mu, sigma) ² ) The lognormal distribution of (a), the probability density function and the distribution function of the statistical model are as follows:

in the formula: t is t ₀ -the time of arrival of the first passenger at the terminal on a flight;

-a probability density function;

Φ (t) -the rate at which a single flight arrives at the check-in area at time t.

The method for calculating the number of the passengers arriving at the check-in area from the departure passenger carried by a single flight at the moment t comprises the following steps:

D _q (t)＝N _q *Φ(t)

in the formula: n is a radical of _q -total number of passengers of the flight;

assuming that the time distribution of the passengers arriving at the security check area and receiving security check is the same as the time distribution of the passengers arriving at the check-in area, the passenger check-in fitting model and the passenger check-in time distribution model are correlated in time, and the integral area of the passenger check-in time distribution model is corrected, then the probability and the number of the passengers who leave the port without delay of the single quasi-point flight q at the time t are calculated by the following method:

in the formula: delta t-time interval for the check-in passenger to enter the security check-in area after taking the boarding check, the time for the first passenger to arrive at the security check-in area by taking a flight is t ₀ + Δ t, at time t the passenger arriving at the security inspection area ₀ -at arrives at the attendant lobby;

-the probability of the departure passenger carried by the punctual flight q at time t passing the security check;

the number of the departed passengers carried by the quasi-point flight q at the moment t for passing the security check;

the method for calculating the number of accumulated safety check-passing passengers of a single flight q at the moment t under the delay condition comprises the following steps:

in the formula: t' ₀ -start of Security check time for delayed flight q, where t' ₀ ＝t ₀ +Δt _d ，Δt _d Delaying the flight for a length of time.

V is the delayed flight security check rate;

delaying the number of passengers who leave the port and go through the security check at the moment t.

The method for calculating the number of the passengers who leave the port and are carried by the single delayed flight and the normal flight at the moment t and are subjected to security inspection by introducing the variable 0-1 is expressed as follows:

in the formula: beta is a _q A variable of 0-1, fetching 1 when flight q is delayed, otherwise fetching 0;

L _q (t) the number of passengers carried by a single flight at time t who have passed the security check;

the difference between the number of the passengers arriving at the flight q and the number of the passengers leaving from the flight q in the check-in area is the number of the passengers gathering at the flight q, and the method for calculating the number of the passengers gathering at the flight q comprises the following steps:

in the formula: h _q (t) -flight q passenger aggregate number;

the calculation method for obtaining the accumulated aggregated passengers in the check-in area of one day by arranging according to the time sequence and accumulating on the time dimension comprises the following steps:

in the formula: h (t) -cumulative aggregated passenger distribution for multiple flights of an airport area;

N ₁ -number of normal flights per day;

N ₂ -delay the number of flights a day.

Preferably, the obtaining of travel rules of the relevant passengers at the airport based on the taxi track data, and determining the evacuation direction, the evacuation area and the evacuation point location of passenger flow according to the travel rules of the relevant passengers at the airport include:

clearing track data of a taxi navigation system, and extracting pairs from a travel starting point to an end point OD starting point to the end point;

rasterizing a researched city geographic file, loading data into a map to draw a centralized expected line graph and a non-centralized expected line graph, and determining an evacuation direction according to an expected line direction taking an airport as a travel endpoint, wherein the area to which the evacuation direction points is an evacuation belly;

constructing a non-centralized expected line graph into a non-directional network graph, carrying out community division and degree distribution geographical marking on the network through a community discovery algorithm, identifying center nodes and edge nodes according to the degree distribution geographical marking of the travel network, keeping a distance between a destination facing urban central area evacuation and a large-degree area, and enabling the destination facing urban suburban area evacuation to be close to the center nodes;

extracting travel endpoints related to an airport, carrying out point clustering according to the distance between the endpoints, and taking a clustering center as an initial evacuation point;

and adjusting the initial evacuation point according to a set evacuation node selection principle, and determining a final evacuation point position.

Preferably, the constructing a vehicle scheduling model including an evacuation vehicle carrying scheme, a route optimization and a schedule optimization according to the passenger aggregation amount, the evacuation direction, the evacuation area and the evacuation point location includes:

constructing a vehicle dispatching model containing an evacuation vehicle carrying scheme, path optimization and schedule optimization, wherein an objective function of the vehicle dispatching model comprises the following steps:

1) Aiming at minimizing the time cost of the passengers by bypassing the line, the accumulated waste time of the passengers by bypassing is multiplied by a passenger time cost coefficient to be the cost of the bypassing, and the corresponding constraint is expressed as:

in the formula: c ₁ -line detour passenger time cost;

the section passenger flow of the train number k between the ith station and the j station,

a 0-1 variable representing whether the train number k travels from station i to station j;

l _ij -the actual distance between the ith station and the j station;

the number of passengers to destination j is carried by the train number k,

v-average speed of travel of airport bus;

α — passenger time cost coefficient (meta \ hour \. People);

2) The passenger waiting time cost is minimized as a target, the passenger waiting cost is the score of the waiting time and the number of waiting passengers, and finally, the score is multiplied by a time cost coefficient, and the corresponding constraint is expressed as:

in the formula: c ₂ -passenger waiting time becomesThen, the process is carried out;

t ^k -departure time of train number k;

time of departure t of train number k ^k Passenger demand to destination j;

m is the number of evacuation nodes;

3) With the minimum fuel consumption of the vehicle as a target, the emergency vehicle driving distance is multiplied by the fuel consumption coefficient to serve as the fuel consumption cost, and the corresponding constraint expression is as follows:

in the formula: c ₃ -emergency vehicle fuel consumption costs;

L _n -emergency vehicle driving distance;

beta-fuel consumption coefficient (Yuan \ kilometer);

4) Aiming at the minimum of the personnel cost of the groups, the number of the groups corresponding to the emergency vehicles is multiplied by the cost of each task in each group, and the calculation is as follows:

C ₄ ＝γ*n

in the formula: c ₄ -team personnel costs;

γ -team personnel coefficient (Yuan \ ban);

5) Passenger costs include C ₁ And C ₂ The cost of the enterprise includes C ₃ And C ₄ The weighted combination of the passenger cost and the enterprise cost is the total cost, and the calculation is as follows:

P＝C ₁ +C ₂

E＝C ₃ +C ₄

in the formula: p-passenger cost;

e-cost of the enterprise;

z is the weighted total cost.

The setting of the constraint conditions of the vehicle dispatching model comprises the following steps:

1) The distribution relation constraint of the evacuation demand amount comprises the relation between the vehicle capacity and the demand amount of each station, and the station j is accumulated to obtain the passenger capacity Q of each train number k ^k The passenger capacity Q of all the train numbers ^k Accumulating to obtain total evacuation demand Q, and accumulating each train number k to obtain demand Q for evacuation site j _j And accumulating the passenger capacity going to all evacuation places to obtain a total evacuation demand Q, wherein the correlation expression is as follows:

in the formula: k is the set of scheduled train numbers;

Q _i -the evacuation point i is to evacuate the aggregate passenger volume;

Q ^k -the number k carries the passenger volume;

Q-Total evacuation demand;

the variable 0-1 represents whether the transportation task of the point i is assumed by the train number k.

2) Carrying out capacity constraint on vehicles in an emergency dispatching manner, wherein the passenger capacity of each vehicle does not exceed the maximum capacity of the vehicle;

Q ^k ≤θ

in the formula: theta-passenger capacity of airport bus;

3) The quantity of emergency dispatching vehicles is restricted, and the quantity of the vehicles which are put into emergency dispatching is smaller than the quantity of the vehicles reserved in the airport bus;

n≤N

in the formula: n is the number of vehicles invested in emergency dispatching;

n-number of vehicles reserved by airport bus;

4) The relation constraint of the number of passengers carried by the emergency dispatching vehicle and the demand comprises that the number of the passengers carried by each vehicle k to the evacuation point i does not exceed the demand to the evacuation point when the vehicle is sent, the demand is the sum of the passenger volume arriving according to the average arrival rate and the cumulative evacuated passenger volume, the arrival rate of the passenger flow to the evacuation point i can be calculated by dividing the passenger flow of the evacuation point i by the evacuation duration, and the corresponding constraint expression is as follows:

λ _i ＝Q _i /min(60,ΔT _last )

in the formula: lambda _i -arrival rate of passenger flow to evacuation point i;

ΔT _last -length of time in emergency (minutes);

5) The section passenger flow constraint guidance type is characterized in that the section passenger flow among evacuation stations is related to the sequence of vehicle paths passing the evacuation stations, and the section passenger flow from the last section of the route

Deducing the cross-section passenger flow among all evacuation points

And cross-sectional passenger flow between airport and evacuation point

The derivation formula is as follows:

in the formula:

-the cross-sectional passenger flow between evacuation point m and evacuation point r;

P ^k -a set of ride plans for train number k;

6) The method comprises the steps that emergency evacuation starting time is restricted, when large-area delay early warning is started, an emergency department of an airport ground traffic center starts to prepare evacuation vehicles and allocate personnel, the preparation process is completed before emergency evacuation starts, and the vehicles are put into use when the emergency evacuation starts; if the allocation process cannot be completed before the emergency evacuation is started, the evacuation task cannot be started on time and needs to be started after the vehicle is equipped, and the corresponding constraint expression is as follows:

in the formula: t is t ¹ The departure time of the first vehicle is the actual starting time of emergency evacuation;

ΔT _pre -length of time (minutes) spent in stocking the vehicle;

T _eva -emergency evacuation pre-warning start time;

T _alarm large area of delayed early warning start time;

7) The emergency vehicle departure time interval constraint, the departure time interval and the departure time difference between the current train number and the previous train number, the departure is simple and meets the set minimum and maximum departure intervals, and the corresponding constraint expression is as follows:

Δt ^k ＝t ^k -t ^k-1

Δt _min ≤Δt ^k ≤Δt _max

in the formula: Δ t ^k -departure time interval for train number k;

Δt _min -minimum departure interval;

Δt _max -maximum departure interval;

8) The path length of the emergency vehicle is constrained, and the path length should meet the set minimum and maximum distance requirements, and is calculated as follows:

L _min ≤L _n ≤L _max

in the formula: l is _n -vehicle driving path length;

L _min -a minimum path length;

L _max -a maximum path length;

9) Ensuring that each evacuation point is served, the corresponding constraint is expressed as:

10 To ensure that all emergency vehicles return to the airport to accept a subsequent shift after completing a shift of transportation tasks, the corresponding constraint is expressed as:

11 Ensuring that each customer point is on the travel path of the corresponding vehicle, the corresponding constraint is expressed as:

a 0-1 variable indicating whether the train number k traveled from i to j;

a variable 0-1, which indicates whether the transport task at point i is assumed by the train number k;

m is the number of evacuation nodes.

Preferably, the solving the vehicle scheduling model through an improved genetic algorithm, and solving the individual with the minimum fitness function value as a model solving result, includes:

the improved genetic algorithm comprises an evacuation response mechanism, an evacuation demand generation module, a vehicle passenger carrying and routing module and a genetic algorithm engine module, and the improved genetic algorithm outputs an individual with the minimum fitness function value as a solving result of a planning model, and comprises the following steps:

1) The evacuation response mechanism module sets large-scale delay early warning time T _alarm Early warning time T for evacuation of large passenger flow _eva Only when the evacuation vehicle finishes servicing, the evacuation is started;

2) Evacuation early warning time T from large passenger flow by evacuation demand generation module _eva At the beginning, the people to be evacuated are according to the average arrival rate lambda _j Join evacuation queuing and form the evacuation demand Q of the sub-destination _j j (t), each evacuation vehicle carrying the number of passengers d destined for a different destination _j Leave the terminal building, the evacuation demand Q _j (t) dynamically varies with system inputs and outputs;

3) The vehicle passenger carrying and routing module sets the number of passengers randomly boarding a vehicle k to a certain destination as d _j ，d _j Constrained by vehicle capacity and not exceeding t ^k Demand for evacuation while traveling to destination j

If the capacity of the vehicle is free, carrying passengers going to other destinations according to the same strategy, generating all path alternative sets by the routing part by adopting an enumeration method, and selecting a routing scheme which enables a target function to be minimum;

4) After the evacuation response mechanism is started, the genetic algorithm engine module gives an initial or varied schedule, the schedule scheme drives the vehicle passenger carrying and routing module to calculate a carrying scheme and adjust evacuation requirements in real time, vehicle paths passing through evacuation points are calculated, the vehicle carrying and routing scheme is transmitted back to the genetic algorithm engine, a fitness value is calculated, and optimization of the schedule scheme is carried out in an iterative mode.

Solving the vehicle scheduling model through Python language based on genetic algorithm, comprising the following processing steps:

step1 variable coding and parameter setting, wherein an emergency vehicle departure schedule is used as a gene of each chromosome in a genetic algorithm for coding, and the initial population size pop _ size and the cross probability P are obtained _c Mutation probability P _m Setting parameters with the maximum iteration times max _ gen;

initializing a Step2 population, generating a plurality of individuals meeting the population scale by adopting a random initialization mode, wherein each individual corresponds to a different emergency vehicle departure timetable, and the value of each individual is randomly generated from the corresponding range;

step3, driving other modules to calculate a passenger carrying scheme and a passenger carrying path, judging constraint conditions of the generated passenger carrying scheme, path scheme and departure schedule scheme, and if the conditions are not met, regenerating an initialization population;

step4, calculating a fitness function value and calculating the fitness function value of each individual in the population;

step5, selecting operation, namely selecting by adopting a roulette selection method;

step6 cross operation, namely, adopting a single-point cross method when the cross probability P is met _c In the case of (2), a cross point is randomly selected among the randomly paired parent individuals, and the part of variables located before or after the cross point is exchanged to form a new offspring individual;

step7 mutation operation, adopting the method of basic bit mutation and using mutation probability P _m Randomly appointing a value on a certain locus or a plurality of loci to carry out variation operation to form a new offspring individual;

and Step8, checking convergence, judging whether an algorithm termination condition is met, taking the maximum iteration number as the termination condition, if the algorithm termination condition is met, finishing the algorithm, outputting the individual with the minimum fitness function value as a solving result of the model, and if the algorithm termination condition is not met, jumping to Step3, wherein the solving result of the model comprises the number of the train numbers required by the evacuation tasks, the departure time corresponding to each train number, the number of people carrying the train to each evacuation destination and the evacuation path of the train number.

According to the technical scheme provided by the embodiment of the invention, the method provided by the invention aims at large-area large-passenger-flow evacuation under the background of large-area delay of the large-scale airport terminal, improves the emergency response efficiency of the airport ground traffic center, improves the intelligent construction of the flexible traffic terminal, and improves the travel experience of passengers under the condition of sudden disasters.

Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a processing flow chart of a large passenger flow early warning and evacuation method under a large-area delay condition of an airport terminal according to an embodiment of the present invention;

fig. 2 is a flow chart of a data-driven evacuation destination selection according to an embodiment of the present invention;

fig. 3 is a design flow chart of a large-passenger-flow emergency evacuation vehicle dispatching model solving algorithm provided by the embodiment of the invention;

fig. 4 is a flowchart of the work of a genetic algorithm module in a solution algorithm according to an embodiment of the present invention.

Detailed Description

Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention, and are not to be construed as limiting the present invention.

As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or coupled. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

For the convenience of understanding the embodiments of the present invention, the following description will be further explained by taking several specific embodiments as examples in conjunction with the drawings, and the embodiments are not to be construed as limiting the embodiments of the present invention.

Based on the research vacancy, the invention applies a statistical model, an operation research model, a data mining means and theoretical knowledge of traffic design to construct a method system of a systematic large-passenger-flow evacuation scheme from the perspective of a ground traffic center of an airport terminal. The method integrates multi-source heterogeneous data such as flight data, security check data, taxi GPS data, existing statistical data and the like, and aims to realize deployment of a large passenger flow evacuation method in a real situation.

The processing flow chart of the large passenger flow early warning and evacuation method under the condition of large-area delay of the airport terminal provided by the embodiment of the invention is shown in figure 1 and comprises the following processing steps;

and S10, establishing a large-area delay early warning mechanism and a large-passenger-flow early warning mechanism based on flight data.

And (3) screening flights delayed for more than 4 hours according to the real-time flight data, and starting large-area delay early warning if the accumulated number of the flights delayed for more than 4 hours or the number of passengers detained corresponding to the accumulated number of the flights reaches a corresponding threshold, wherein the related pseudo codes are as follows:

inputting a normal flight passenger aggregation model A _normal And delayed flight passenger aggregation model A _delayed . Flight data set

The key field comprises the actual takeoff time t ⁽ⁿ⁾ Delay time d ^(b) Passenger capacity m ⁽ⁿ⁾ And flight attribute label y obtained after cross classification ⁽ⁿ⁾ Different attribute tags correspond to different terminal buildings and mathematical models (T2 domestic departing flights, T2 international departing flights, T3 domestic departing flights and T3 international departing flights). And counting the accumulated number F 'of flights delaying for more than 4 hours and accumulated detained passengers S, and starting large-area delay early warning when the accumulated number F' reaches a corresponding threshold value or when the accumulated detained passengers S reach the corresponding threshold value. Once the large-area delay early warning is started, the delayed flight passengers can not pass the security check and begin to gather in the check-in area by using the A _normal Model estimation of passenger aggregation quantity to use A _delayed And (4) model estimation, and finally outputting whether a large-area delay early warning signal and a passenger gathering distribution map appear in the solar navigation station building by an algorithm.

And S20, predicting passenger aggregation amount based on a statistical model of passenger check-in and security check physical process combination and real-time airport flight data, and further dividing emergency response time period and evacuation requirement by combining an early warning mechanism.

After the large-area delay early warning is started, the ground traffic center of the airport terminal prepares emergency vehicle resources and dispatches emergency workers. And (3) calculating the accumulation amount of the passengers in the check-in area by using the following statistical model, and immediately starting the evacuation early warning of the large passenger flow if the accumulation amount exceeds the designed number of the passengers in the terminal. And if the emergency resources are completely prepared when the large passenger flow evacuation is started, immediately starting evacuation, and if the emergency resources are not completely prepared when the large passenger flow evacuation is started, starting the large passenger flow evacuation after the resources are scheduled in place.

And establishing a statistical model based on the combination of the passenger check-in process and the security inspection physical process. Suppose that the time t of arrival of a departure passenger at the terminal is subject to a parameter of (mu, sigma) ² ) The lognormal distribution of (2) is such that the probability density function and the distribution function are:

-a probability density function;

D _q (t)＝N _q *Φ(t)

in the formula: n is a radical of _q -total number of passengers on the flight.

Assuming that the time distribution of the passengers arriving at the security check area and receiving security check is the same as the time distribution of the passengers arriving at the check-in area, the passenger check-in fitting model and the passenger check-in time distribution model are linked in time, and the integral area of the passenger check-in fitting model is corrected, so that the probability and the number of the passengers leaving the port of a single quasi-point flight q without delay at the time t are calculated by the following method:

in the formula: delta t-time interval for the check-in passenger to enter the security check-in area after taking the boarding check, the time for the first passenger to arrive at the security check-in area by taking a flight is t ₀ + delta t, at time t the passenger arriving at the security area ₀ -at arrives at the check-in lobby;

probability of departure passenger carried by quasi-point flight q at time t passing security check

The number of the passengers who are carried by the quasi-point flight q at the moment t and go through the security check.

The method for calculating the number of accumulated passengers who pass through the security check of a single flight q at the moment t under the delay condition comprises the following steps:

V is the delayed flight security check rate.

In conclusion, the number of passengers who leave the airport and are carried by a single flight at the moment t can be calculated as follows:

in the formula: beta is a _q A variable of 0-1, fetch 1 when flight q is delayed, otherwise fetch 0.

L _q (t) the number of passengers that were carried by a single flight at time t to exit for security.

The difference between the accumulated number of arriving passengers of the flight q and the accumulated number of leaving passengers of the flight q in the check-in area is the number of accumulated passengers of the flight q, and then the number of accumulated passengers of the flight q can be calculated as:

in the formula: h _q (t) — flight q number of passengers congregate.

Arranging according to the time sequence, and accumulating on the time dimension to obtain the accumulated aggregated passenger distribution of the check-in area in one day, which can be calculated as:

N ₁ -number of normal flights per day;

N ₂ -delay the number of flights a day.

And S30, mining travel rules of passengers related to the airport based on taxi track data, determining evacuation directions, evacuation areas and evacuation point positions of large passenger flows under large-area delay in sequence by a research method combining overall qualitative and quantitative modes, and correcting the evacuation directions, the evacuation areas and the evacuation point positions according to traffic design principles.

The method comprises the following five steps:

step1, cleaning taxi GPS track data and extracting a travel OD (Origin to Destination) pair;

step2, rasterizing a researched city geographic file, loading data into a map to draw a concentrated and non-concentrated expected line graph, determining an evacuation direction by means of an expected line direction taking an airport as a travel endpoint, and simultaneously determining an area pointed by the evacuation direction as an evacuation abdominal land;

step3, constructing the non-aggregative expected line graph into an undirected network graph, and performing community division and degree distribution geographic marking on the network through a community discovery algorithm. The interior of a community structure obtained according to a network community discovery algorithm is a travel intensive area, the contact between communities is less, and the convenience of secondary travel of residents can be ensured only after passengers of an airport group are sent to other large-scale groups, so that the passengers need to be evacuated to a traffic community adjacent to the airport community by an evacuation task. The degree distribution geographical marks of the trip network can identify central nodes and edge nodes, and the distance between the destination evacuated towards the central area of the city and a high-degree area is kept, so that traffic flow is prevented from being further introduced into a busy junction to interfere with normal traffic order; the destination for evacuation in the urban suburban area is close to the central pivot node, so that the destination is convenient to perform self-organizing transfer evacuation;

step4, extracting travel end points related to the airport in the fourth stage, carrying out point clustering according to the distance between the end points, and taking a clustering center as an initial evacuation point.

The clustering center obtained by the Step5 clustering algorithm is only a theoretical reasonable point, and factors such as land utilization, road traffic environment, secondary evacuation transfer efficiency and the like are further considered, and further selection and design are performed near the clustering center point. The proposed evacuation node selection principles are three:

(1) Passengers can conveniently self-organize after getting off the bus, approach to traffic hubs, bus stations, rail traffic stations and the like, and are convenient to transfer and plug into;

(2) The carrying capacity of traffic facilities around the evacuation node meets the requirement of passenger drop, and meanwhile, overpasses with high driving speed and large traffic volume, main roads, intersections, entrances and exits of office units and other places which are easy to cause traffic disorder are avoided;

(3) Avoid quiet places such as schools, hospitals and communities and reduce noise interference as much as possible.

And S40, providing a vehicle scheduling model comprising an evacuation vehicle carrying scheme, path optimization and schedule optimization.

The method comprises the following steps of constructing a vehicle scheduling model containing an evacuation vehicle carrying scheme, path optimization and schedule optimization, and considering two aspects of minimum passenger travel cost and minimum enterprise scheduling cost, wherein the objective function of the vehicle scheduling model comprises the following steps:

1) Aiming at minimizing the time cost of the passengers by the line, the cost spent by the detour is calculated as follows by multiplying the cumulative waste time of the passengers by a passenger time cost coefficient:

in the formula: c ₁ -line detour passenger time cost;

-a 0-1 variable indicating whether the train number k travels from station i to station j;

l _ij -the actual distance between the ith station and the j station;

the number of passengers to destination j is carried by the train number k,

v-average speed of travel of airport bus;

α — passenger time cost coefficient (yuan \ hr. Person).

2) The passenger waiting time cost is minimized as the target, the passenger waiting cost is the result of waiting time and the number of waiting passengers, and finally, the result is multiplied by a time cost coefficient, and the calculation is as follows:

in the formula: c ₂ -passenger waiting time cost;

t ^k -departure time of train number k;

time of departure t of train number k ^k Passenger demand to destination j;

m is the number of evacuation nodes;

3) The minimum fuel consumption of the vehicle is taken as a target, the emergency vehicle driving distance is multiplied by the fuel consumption coefficient to obtain the fuel consumption cost, and the calculation is as follows:

in the formula: c ₃ -emergency vehicle fuel consumption costs;

L _n -emergency vehicle driving distance;

beta-fuel consumption coefficient (Yuan \ km).

4) Aiming at the minimum of the personnel cost of the groups, the number of the groups corresponding to the emergency vehicles is multiplied by the cost of each task, and the calculation is as follows:

C ₄ ＝γ*n

in the formula: c ₄ -team personnel costs;

gamma-team personnel coefficient (Yuan \ team).

5) Passenger costs include C ₁ And C ₂ Enterprise costs include C ₃ And C ₄ The weighted combination of the passenger cost and the enterprise cost is the total cost, and is calculated as follows:

P＝C ₁ +C ₂

E＝C ₃ +C ₄

in the formula: p-passenger cost;

e-cost of the enterprise;

z-the weighted total cost.

The constraint condition setting of the vehicle scheduling model comprises the following steps:

1) The distribution relation constraint of the evacuation demand amount comprises the relation between the vehicle capacity and the demand amount of each station, and the station j is accumulated to obtain the passenger capacity Q of each train number k ^k The passenger capacity Q of all the train numbers ^k And accumulating to obtain the total evacuation demand Q. Accumulating each train number k to obtain the demand Q of the evacuation place j _j Accumulating the passenger capacity going to all evacuation places to obtain a total evacuation demand Q, wherein the correlation relationship is expressed as follows:

in the formula: k is the set of scheduled train numbers;

Q _i -the evacuation point i is to evacuate the aggregate passenger volume;

Q ^k the number k carries the passenger volume;

Q-Total evacuation demand;

the variable 0-1 represents whether the transport task at point i is assumed by the train number k.

2) And (4) carrying out emergency dispatching on the capacity constraint of the vehicles, wherein the passenger capacity of each vehicle does not exceed the maximum capacity of the vehicle.

Q ^k ≤θ

In the formula: theta-passenger carrying capacity of airport bus.

3) The quantity of emergency dispatching vehicles is restricted, and the quantity of the vehicles which are put into emergency dispatching is smaller than the quantity of reserved vehicles of the airport bus.

n≤N

In the formula: n is the number of vehicles invested in emergency dispatching;

n-number of vehicles reserved at airport bus.

4) And the relation constraint of the number of passengers carried by the emergency dispatching vehicle and the demand comprises that the passenger carrying capacity to the evacuation point i in each vehicle number k does not exceed the demand to the evacuation point when the vehicle is sent, and the demand is the passenger carrying capacity arriving according to the average arrival rate minus the accumulated evacuated passenger carrying capacity. The arrival rate of the passenger flow going to the evacuation point i can be calculated by dividing the passenger flow of the evacuation point i by the evacuation duration, when the duration is longer, the evacuation is finished within one hour, and the corresponding constraint expression is as follows:

λ _i ＝Q _i /min(60,ΔT _last )

in the formula: lambda [ alpha ] _i -arrival rate of passenger flow to evacuation point i;

ΔT _last -length of time in emergency (minutes);

5) The section passenger flow constraint guidance mode is characterized in that the section passenger flow among evacuation stations is related to the sequence of vehicle paths passing through the evacuation stations. When going to the first evacuation point, the cross section passenger flow is the sum of passengers carried by the vehicles of the train number, and the cross section passenger flow is reduced in a step manner along with the passengers falling through each evacuation point. After the passengers at the last evacuation point are delivered, the empty bus at the airport returns, and the cross-section passenger flow at the moment is 0. Derived profile traffic from last leg

Deducing the cross-section passenger flow among all evacuation points

And cross-sectional passenger flow between airport and evacuation point

The derivation formula is as follows:

in the formula:

P ^k -set of loading schemes for train number k.

6) The emergency evacuation starting time is restricted, and the actual emergency evacuation starting time depends on whether the vehicle is completely equipped or not. When the large-area delay early warning is started, an emergency department of the airport ground traffic center starts to prepare the evacuation vehicles and allocate personnel, the preparation process is completed before the emergency evacuation starts, and the vehicles are put into use when the emergency evacuation starts. If the allocation process cannot be completed before the emergency evacuation is started, the evacuation task cannot be started on time and needs to be started after the vehicle is equipped.

In the formula: t is t ¹ The departure time of the first vehicle number is the actual starting time of emergency evacuation;

ΔT _pre -length of time (minutes) spent in stocking the vehicle;

T _eva -emergency evacuation pre-warning start time;

T _alarm large area delayed early warning start time.

7) The emergency vehicle departure time interval is restrained, the departure time interval and the current train number and the last train number departure time difference, the departure is simple, the set minimum and maximum departure interval is satisfied, and the calculation formula is as follows:

Δt ^k ＝t ^k -t ^k-1

Δt _min ≤Δt ^k ≤Δt _max

in the formula: Δ t ^k -departure time interval of train number k;

Δt _min -minimum departure interval;

Δt _max -maximum departure interval.

8) The emergency vehicle path length constraint is that the path length meets the set minimum and maximum distance requirements, and the calculation is as follows:

L _min ≤L _n ≤L _max

in the formula: l is _n -vehicle driving path length;

L _min -a minimum path length;

L _max -maximum path length.

9) Other constraints, including ensuring that each evacuation point is served, are calculated as follows:

10 Other constraints to ensure that all emergency vehicles return to the airport for subsequent shifts after completing a shift of transportation tasks, as calculated below:

11 Other constraints that ensure that each customer point is on the corresponding vehicle's travel path, are calculated as follows:

a 0-1 variable indicating whether the train number k traveled from i to j;

a variable 0-1, representing whether the transport task at point i is assumed by the train number k;

m is the number of evacuation nodes.

And S50, solving the vehicle scheduling model through an improved genetic algorithm, and outputting an individual with the minimum fitness function value as a model solving result.

The improved genetic algorithm comprises an evacuation response mechanism, an evacuation demand generation module, a vehicle passenger carrying and routing module and a genetic algorithm engine module, and outputs an individual with the minimum fitness function value as a solving result of a planning model, and the method comprises the following steps:

1) An evacuation response mechanism module which is a precondition for triggering large passenger flow emergency evacuation and delays the early warning time T in large scale _alarm Early warning time T for evacuation of large passenger flow _eva And evacuation is started only after the evacuation vehicle is completely equipped.

2) The evacuation demand generation module is a dynamic unit matched with the whole evacuation scheduling system and evacuates the early warning time T from the large passenger flow _eva At the beginning, the persons to be evacuated are according to the average arrival rate lambda _j Join evacuation queue and form the evacuation demand Q of the sub-destination _j (t) each evacuation vehicle carries the number d of passengers to a different destination _j Leave the terminal building, the evacuation demand Q _j (t) dynamically varies with system input and output.

3) The vehicle passenger carrying and routing module is a dynamic unit with random factors, and the number of passengers for a certain destination which are randomly picked up by a vehicle k is d _j ，d _j Constrained by vehicle capacity and not exceeding t ^k Timely evacuation demand to destination j

If there is still room for the vehicle capacity, passengers destined for other destinations are carried according to the same strategy. And the routing part generates all path alternative sets by adopting an enumeration method and selects a routing scheme which enables the objective function to be minimum.

4) The genetic algorithm engine is a driving module of the optimization algorithm, after the evacuation response mechanism is started, the genetic algorithm gives an initial or varied schedule, the schedule scheme drives the vehicle passenger carrying and routing module to calculate a carrying scheme, adjust the evacuation requirements in real time and calculate the vehicle paths passing through the evacuation points. And finally, transmitting the vehicle bearing and routing scheme back to the genetic algorithm engine, calculating a fitness value, and iteratively optimizing the schedule scheme.

Solving the model through Python language based on genetic algorithm, comprising the following processing steps:

step1 variable coding and parameter setting, wherein an emergency vehicle departure timetable is used as a gene of each chromosome in a genetic algorithm to be coded, and the initial population size pop _ size and the cross probability P are obtained _c Mutation probability P _m Setting parameters with the maximum iteration times max _ gen;

step4, calculating a fitness function value, and calculating the fitness function value of each individual in the population;

and Step8, convergence checking, namely judging whether an algorithm termination condition is met, taking the maximum iteration number as the termination condition, if so, finishing the algorithm, outputting the individual with the minimum fitness function value as a solving result of the model, and if not, skipping to Step3.

The solving result of the algorithm comprises the number of the train numbers required by the evacuation tasks, the departure time corresponding to each train number, the number of people carrying the trains to the evacuation destinations and the evacuation path of the train number. The results of the solution can be shown in the following table.

In summary, in the existing research, the evacuation research for sudden large passenger flows caused by large-area delay of airport terminals is less, the research on each evacuation link is relatively independent, and a top-level design and coordination mechanism is lacked. (1) The statistical model for estimating the passenger gathering distribution introduces airport operation factors, and can accurately fit the passenger gathering time-varying condition; (2) The evacuation destination selection adopts a method of combining quantification and qualification, so that the actual evacuation point meeting the evacuation requirement of the large passenger flow can be accurately selected; (3) The vehicle dispatching model considers the randomness of passenger demands, the passenger cost and the operation cost of a manager, and can adjust the weight combination under different emergency levels to make different vehicle dispatching schemes.

Those of ordinary skill in the art will understand that: the figures are merely schematic representations of one embodiment, and the blocks or flow diagrams in the figures are not necessarily required to practice the present invention.

From the above description of the embodiments, it is clear to those skilled in the art that the present invention can be implemented by software plus necessary general hardware platform. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which may be stored in a storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments or some parts of the embodiments.

All the embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, apparatus or system embodiments, which are substantially similar to method embodiments, are described in relative ease, and reference may be made to some descriptions of method embodiments for related points. The above-described embodiments of the apparatus and system are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. One of ordinary skill in the art can understand and implement without inventive effort.

While the invention has been described with reference to specific preferred embodiments, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined in the following claims. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A large passenger flow early warning and evacuation method under the condition of large-area delay of an airport terminal is characterized by comprising the following steps:

establishing a statistical model of a combination of physical processes of passenger check-in and security check, and predicting passenger aggregation amount through the statistical model based on real-time airport flight data and the large-area delay early warning mechanism and the large-passenger-flow early warning mechanism;

and solving the vehicle scheduling model through an improved genetic algorithm, and solving the individual with the minimum fitness function value as a model solving result.

2. The method of claim 1, wherein the establishing of the large-area delay early warning mechanism and the large passenger flow early warning mechanism based on the flight data comprises:

setting the normal flight passenger aggregation model as A _normal The model of passenger gathering of delayed flights is A _delayed The flight data set is

t ⁽ⁿ⁾ To actual takeoff time, d ⁽ⁿ⁾ For the delay time, m ⁽ⁿ⁾ For passenger capacity, y ⁽ⁿ⁾ For flight attribute labels obtained after cross classification, different attribute labels correspond to different terminal buildings and mathematical models, the accumulated number F 'and accumulated detained passengers S of the flights delayed for over 4 hours are counted, when the accumulated number F' reaches a corresponding threshold value or when the accumulated detained passengers S reach the corresponding threshold value, large-area delay early warning is started, and the method uses A _normal Model estimation of passenger aggregation quantity to use A _delayed The model estimates the passenger gathering amount, and the algorithm outputs whether a large-area delay early warning signal and a passenger gathering distribution map appear in the solar navigation station building.

3. The method as claimed in claim 2, wherein the establishing of the statistical model of the combination of the passenger check-in and security check physical process, and the prediction of the passenger gathering amount through the statistical model based on the real-time airport flight data and the large-area delay early warning mechanism and the large passenger flow early warning mechanism comprise:

establishing a statistical model based on the combination of passenger check-in and security check physical processes, and assuming that the time t of the departure passenger arriving at the terminal building is (mu, sigma) ² ) The lognormal distribution of (2), the probability density function and the distribution function of the statistical model are as follows:

in the formula: t is t ₀ -the time of arrival of the first passenger on a flight at the terminal;

-a probability density function;

-the rate at which passengers carried by an individual flight arrive at the boarding area at time t;

D _q (t)＝N _q *Φ(t)

in the formula: n is a radical of _q -total number of passengers of the flight;

assuming that the time distribution of the passengers arriving at the security check area and receiving security check is the same as the time distribution of the passengers arriving at the check-in area, the passenger check-in fitting model and the passenger check-over time distribution model are associated in time, and the integral area of the passenger check-over time distribution model is corrected, so that the probability and the number of passengers leaving the port of the single punctual flight q without delay at the time t are calculated by the method:

in the formula: t' ₀ -start of Security check time for delayed flight q, where t' ₀ ＝t ₀ +Δt _d ，Δt _d Delaying the flight for a duration;

v is the delayed flight security check rate;

delaying the number of passengers who leave the port and go through the security check at the moment t;

the method for calculating the number of the passengers who leave the port and are carried by the single delayed flight and the normal flight at the moment t and are subjected to security check by introducing a variable 0-1 is expressed as follows:

in the formula: beta is a _q A variable 0-1, taking 1 when flight q is delayed, otherwise taking 0;

in the formula: h _q (t) -flight q number of passengers congregate;

the calculation method for obtaining the accumulated passengers gathered in the check-in area of one day by arranging according to the time sequence and accumulating on the time dimension comprises the following steps:

in the formula: h (t) -cumulative aggregated passenger distribution for multiple flights in one-day check-in area;

N ₁ -number of normal flights per day;

N ₂ -delay the number of flights a day.

4. The method according to claim 3, wherein the step of obtaining travel rules of passengers related to the airport based on taxi track data and determining evacuation directions, evacuation areas and evacuation points of passenger flows according to the travel rules of the passengers related to the airport comprises the steps of:

constructing a non-centralized expected line graph into a undirected network graph, carrying out community division and degree distribution geographical marking on the network through a community discovery algorithm, identifying center nodes and edge nodes according to the degree distribution geographical marking of the trip network, keeping a distance between a destination facing the evacuation of a central area of a city and a region of a large degree, and enabling the destination facing the evacuation of a suburban area of the city to be close to the center nodes;

5. The method of claim 4, wherein constructing a vehicle scheduling model including an evacuation vehicle loading scheme, path optimization and schedule optimization according to the passenger aggregation amount, evacuation direction, evacuation area and evacuation point location comprises:

constructing a vehicle scheduling model including an evacuation vehicle carrying scheme, path optimization and schedule optimization, wherein an objective function of the vehicle scheduling model comprises the following steps:

in the formula: c ₁ -line detour passenger time cost;

l _ij -the actual distance between the ith station and the j station;

the number of passengers to destination j is carried by the train number k,

v-average speed of travel of airport bus;

α — passenger time cost coefficient (meta \ hour \. People);

in the formula: c ₂ -passenger waiting time cost;

t ^k the departure time of the train number k;

time t of departure for train number k ^k To the purpose ofPassenger demand at ground j;

m is the number of evacuation nodes;

3) With the minimum fuel consumption of the vehicle as a target, the emergency vehicle driving distance is multiplied by the fuel consumption coefficient to be used as the fuel consumption cost, and the corresponding constraint is expressed as:

in the formula: c ₃ -emergency vehicle fuel consumption costs;

L _n -emergency vehicle driving distance;

beta-fuel consumption coefficient (Yuan \ kilometer);

C ₄ ＝γ*n

in the formula: c ₄ -team personnel costs;

γ -team personnel coefficient (Yuan \ ban);

5) Passenger costs include C ₁ And C ₂ Enterprise costs include C ₃ And C ₄ The weighted combination of the passenger cost and the enterprise cost is the total cost, and the calculation is as follows:

P＝C ₁ +C ₂

E＝C ₃ +C ₄

in the formula: p-passenger cost;

e-cost of the enterprise;

z-weighted total cost;

setting the constraint conditions of the vehicle scheduling model comprises the following steps:

1) And (4) carrying out distribution relation constraint on evacuation demand, including the relation between the vehicle capacity and the demand of each station, and accumulating the stations j to obtain the passenger carrying capacity of each train number kQuantity Q ^k The passenger capacity Q of all the train numbers ^k Accumulating to obtain total evacuation demand Q, and accumulating each train number k to obtain demand Q for evacuation site j _j And accumulating the passenger capacity going to all evacuation places to obtain a total evacuation demand Q, wherein the correlation expression is as follows:

in the formula: k is the set of scheduled train numbers;

Q _i -the evacuation point i is to evacuate the aggregate passenger volume;

Q ^k the number k carries the passenger volume;

Q-Total evacuation demand;

Q ^k ≤θ

in the formula: theta-passenger capacity of airport bus;

3) The quantity of emergency dispatching vehicles is restricted, and the quantity of the vehicles which are put into emergency dispatching is smaller than the quantity of reserved vehicles of the airport bus;

n≤N

in the formula: n is the number of vehicles invested in emergency dispatching;

n-airport bus reserve vehicle number;

4) The relation constraint of the number of passengers carried by the emergency dispatching vehicle and the demand comprises that the number of the passengers carried by each vehicle k to the evacuation point i does not exceed the demand to the evacuation point when the vehicle is sent, the demand is the sum of the passenger quantity arriving according to the average arrival rate and the cumulative evacuation passenger quantity, the arrival rate of the passenger flow to the evacuation point i can be calculated by dividing the passenger flow of the evacuation point i by the evacuation duration, and the corresponding constraint is expressed as follows:

λ _i ＝Q _i /min(60，ΔT _last )

ΔT _last -length of time in emergency (minutes);

Deducing the cross-section passenger flow among all evacuation points

And cross-sectional passenger flow between airport and evacuation point

The derivation formula is as follows:

in the formula:

the section passenger flow between the evacuation point m and the evacuation point r;

P ^k -a set of ride plans for train number k;

6) The method comprises the following steps that emergency evacuation starting time is restricted, when large-area delay early warning is started, an airport ground traffic center emergency department starts to prepare evacuation vehicles and allocate personnel, the preparation process is completed before emergency evacuation starts, and the vehicles are put into use when the emergency evacuation starts; if the allocation process cannot be completed before the emergency evacuation is started, the evacuation task cannot be started on time and needs to be started after the vehicle is equipped, and the corresponding constraint expression is as follows:

ΔT _pre -length of time (minutes) spent in stocking the vehicles for reconditioning;

T _eva -emergency evacuation early warning start time;

T _alarm large area of delayed early warning start time;

Δt ^k ＝t ^k -t ^k-1

Δt _min ≤Δt ^k ≤Δt _max

in the formula: Δ t ^k -departure time interval of train number k;

Δt _min -most preferablyA small departure interval;

Δt _max -maximum departure interval;

L _min ≤L _n ≤L _max

in the formula: l is _n -vehicle driving path length;

L _min -a minimum path length;

L _max -a maximum path length;

a 0-1 variable indicating whether the vehicle number k traveled from i to j;

m is the number of evacuation nodes.

6. The method of claim 5, wherein solving the vehicle dispatch model through a modified genetic algorithm to obtain the individual with the smallest fitness function value as the model solution result comprises:

2) Evacuation early warning time T from large passenger flow by evacuation demand generation module _eva At the beginning, the people to be evacuated are according to the average arrival rate lambda _j Join evacuation queuing and form the evacuation demand Q of the sub-destination _j (t) each evacuation vehicle carries the number d of passengers to a different destination _j Leave the terminal building, the evacuation demand Q _j (t) dynamically varies with system input and output;

4) After the evacuation response mechanism is started, the genetic algorithm engine module gives an initial or varied schedule, the schedule scheme drives the vehicle passenger carrying and routing module to calculate a carrying scheme and adjust evacuation requirements in real time, vehicle paths passing through evacuation points are calculated, the vehicle carrying and routing scheme is transmitted back to the genetic algorithm engine, a fitness value is calculated, and optimization of the schedule scheme is carried out in an iterative mode;

step1 variable coding and parameter setting, wherein an emergency vehicle departure schedule is used as a gene of each chromosome in a genetic algorithm for coding, and the initial population size pop _ size and the cross probability P are obtained _c Probability of mutation P _m Setting parameters with the maximum iteration times max _ gen;

step2, initializing a population, namely generating a plurality of individuals meeting the population scale by adopting a random initialization mode, wherein each individual corresponds to different emergency vehicle departure time tables, and the value of each individual is randomly generated from the corresponding range;

step6 cross operation, namely, adopting a single-point cross method when the cross probability P is met _c In the case of (2), a cross point is randomly selected among the randomly paired parent individuals, and the portion of the variables located before or after the cross point is swapped to form a new offspring individual;

and Step8, convergence checking, namely judging whether an algorithm termination condition is met or not, taking the maximum iteration number as the termination condition, if the algorithm termination condition is met, ending the algorithm, outputting the individual with the minimum fitness function value as a solving result of the model, and if the algorithm termination condition is not met, skipping to Step3, wherein the solving result of the model comprises the number of the train numbers required by the evacuation tasks, the departure time corresponding to each train number, the number of people carried to each evacuation destination and the evacuation path of the train number.