A dispatching method for airport passenger security channel resource optimization

Abstract

The application discloses a kind of dispatching methods for airport passenger security channel resource optimization, constructs security channel model;Define security channel efficiency index;Calculate the security channel traffic rate under each scheduling scheme;Build security resource multi-objective optimization model, use multi-index balanced optimization, through the resource allocation optimization algorithm of dynamic optimization orientation, reasonably allocate resources;Construct objective function;Set constraint condition: use deep reinforcement learning algorithm to solve objective function, output each security channel optimal opening scheme and security personnel scheduling scheme in each period.Compared with prior art, the application can comprehensively reflect the operation of each link of security system, improve the accuracy and practicality of resource dynamic scheduling, can dynamically respond to the fluctuation of passenger flow in each period to realize resource flexible scheduling, support multi-objective optimization, can output optimal decision by learning the dynamic characteristics of environment, has very high practical value and wide application prospect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of airport terminal operation management, and particularly relates to a dispatching method for airport passenger security channel resource optimization. BACKGROUND

[0002] According to the data of the International Air Transport Association (IATA), the global air passenger volume will exceed 5 billion for the first time in 2025. At the same time, with the acceleration of global economic integration and the vigorous development of the civil aviation industry, the domestic aviation network is continuously improving, and more and more people are willing to choose convenient and efficient air travel, which makes the number of flights and passenger throughput at the airport continue to rise. Beijing Daxing International Airport, as a super hub with an annual passenger throughput of more than 45 million, needs to provide higher operation support capacity for the growing passenger flow. In airport operation management, the travel experience of passengers is directly related to the service quality and passenger satisfaction of the airport, and is an important indicator to measure the operation support level of the airport. Complicated and inefficient security procedures and large-scale passenger flow can cause a large number of passengers to concentrate in the security check in a short time, leading to congestion in the security channel, long waiting time for passenger security check, and seriously affecting the travel experience. The Transportation Security Administration (TSA) of the United States has found that waiting time exceeding 20 minutes will reduce passenger satisfaction by 35%. Therefore, scientifically evaluating the security check efficiency of the security channel to optimize the allocation of security resources has become the key to improving the travel experience of passengers and the operation efficiency of the airport. The current security system of the airport terminal is facing the following common problems:

[0003] (1) Passenger flow fluctuation

[0004] The passenger flow is affected by factors such as time and flight arrangement and shows fluctuation, which may cause congestion in the security channel during peak hours, and resource idling during low peak hours.

[0005] (2) Security resource allocation needs to be optimized

[0006] Currently, the deployment of security resources, such as the allocation of security personnel, the opening and closing of channels, etc., mainly relies on experience values for judgment, and lacks systematic and data-driven optimization strategies. This approach may lead to resource shortages and long queues during peak periods, and low resource utilization during low peak periods, resulting in waste of operating costs. Therefore, it is necessary to introduce a scientific optimization algorithm to improve the efficiency of security resource allocation through data analysis and dynamic adjustment, to ensure smooth passenger traffic, reduce operating costs, and have important practical significance.

[0007] (III) Balance between efficiency and operating cost

[0008] In airport security management, improving security efficiency and controlling operating cost are often a pair of contradictions. Excessive increase of security channels and personnel configuration can shorten passenger waiting time and improve traffic efficiency, but will bring higher operating cost. On the contrary, if resources are too saved, it may lead to too long passenger queuing time, affect travel experience, and even affect the overall operation efficiency of the airport. Therefore, a scientific optimization model and dynamic deployment strategy are needed to achieve a reasonable allocation of security resources while ensuring fast passenger traffic, so as to achieve an optimal balance point between security efficiency and operating cost.

[0009] Therefore, in the management of airport terminal security channel, in order to solve the above problems, a systematic resource optimization allocation algorithm is needed to minimize the number of security channel openings and security personnel configuration under the condition of meeting passenger flow, improve security resource utilization and reduce operating cost. The airport passenger security channel resource optimization scheduling method provided by the present application can effectively solve the above problems. SUMMARY

[0010] The purpose of the present application is to provide an airport passenger security channel resource optimization scheduling method, which can achieve a reasonable allocation of security resources while ensuring fast passenger traffic through a scientific optimization model and dynamic deployment strategy, so as to achieve an optimal balance point between security efficiency and operating cost, minimize the number of security channel openings and security personnel configuration in the management of airport terminal security channel, improve security resource utilization and reduce operating cost.

[0011] In order to achieve the above purpose, the technical solution adopted by the present application is: an airport passenger security channel resource optimization scheduling method, the method steps are as follows,

[0012] Step 1, the passenger security process is modeled in multiple dimensions and multiple stages, and a security channel model of dynamic coupling relationship between each link is constructed, the security channel model is based on the difference influencing factors of passenger type, equipment type and its processing capacity, number of security officers and their operation efficiency, the channel flow rate model of different security stages i and the overall flow rate of security channel under the security officer scheduling scheme a are calculated and designed for channel k;

[0013] Step 2, a scientific basis is provided for optimizing the resource deployment of security system in the channel, improving the efficiency of airport security and improving the passenger traffic experience, the security channel efficiency indicators passenger average inspection time τ ka And security officer work intensity r ka ;

[0014] Step 3, set up the security channel set, the security post set, the statistical time period set, the passenger flow set on the time period, the security guard scheduling scheme set, define the number of security guards in each link of the security process under different scheduling schemes, and calculate the security channel passing rate under each scheduling scheme through the security channel model;

[0015] Step 4, construct a security resource multi-objective optimization model, use multi-index balanced optimization, and through multi-objective optimization algorithm, the optimal balance point between different targets can be found, the security resources are reasonably allocated, the joint optimization of dynamic channel configuration and static personnel scheduling is realized, the dynamic resource scheduling of security channel and the static scheduling of security personnel are solved, to realize the efficient operation of the whole security system, through the dynamic optimization oriented resource allocation optimization algorithm, under the conditions of meeting the dynamic passenger flow, the number of security guard posts, the security guard workload limit, etc., the optimization objectives of minimizing the average passenger inspection time, the security manpower cost, and the channel personnel configuration change are balanced to balance the security efficiency and the utilization of human resources, and the resources are reasonably allocated. The dynamic optimization oriented resource allocation optimization algorithm can adjust the weight coefficients of human cost target δ, personnel configuration stability target ψ and passenger satisfaction index (i.e. average passenger inspection time) to implement different terminal building operation strategies to meet different management needs;

[0016] Step 5, introduce binary integer variables to represent the selection of security guard scheduling scheme, and form a binary linear programming problem, so the objective function is constructed as follows:

[0017]

[0018] Where δ, ψ and are the weight coefficients of security manpower operation cost, adjacent time period channel personnel configuration change and average passenger inspection time respectively, c a and N ka are the operation cost of using scheduling scheme a and the total number of security guards required for channel k under scheduling scheme a respectively, is the decision variable, which represents whether channel k adopts scheduling scheme a in period t, τ ka is the average passenger inspection time of channel k under scheduling scheme a;

[0019] Step 6, set the constraint conditions:

[0020] Step 7, optimize the resources of the security channel model, use the deep reinforcement learning algorithm to solve the objective function, and output the optimal opening scheme of each security channel and the security personnel scheduling scheme in each period.

[0021] The scheduling method of the application supports coping with dynamic fluctuations of passenger flow in different periods, can reasonably allocate existing security resources according to passenger flow conditions, dynamically adjust the opening state of the channel and the configuration of security personnel in different periods, and balance the security efficiency and the utilization of security personnel resources, so as to achieve the purpose of scientific and reasonable optimization.

[0022] As preferred, in step 1, the specific construction method of the security channel model is as follows,

[0023] First, the security process is divided into multiple stages in series, including identity verification, passenger body inspection, and luggage inspection key subsystems, and then the identity verification, passenger inspection, and luggage inspection full-link multiple key links are comprehensively modeled,

[0024] Then, the channel flow rate model of channel k in different security stages i under the security personnel scheduling scheme a is established by taking the passenger type, equipment type and its processing capacity, the number of security personnel and its operation efficiency, and the difference influencing factors, as shown below:

[0025]

[0026] Where i represents different security stages, including identity verification (id), passenger body inspection (pass), and luggage inspection (bag), p zk represents the proportion of each passenger type Z in the total passenger flow in channel k, is the processing capacity of the security equipment in channel k in different security links i (unit: passenger / minute), is the influence factor of passenger type Z on security link i in channel k, is the security personnel efficiency function, that is, with the increase of the number of security personnel under the security personnel scheduling scheme a in the security link i of channel k , the personnel efficiency function of the channel flow rate is improved;

[0027] Finally, the overall flow rate of the security channel can be expressed as

[0028] The security process comprehensively models the identity verification, passenger inspection, and luggage inspection full-link multiple key links, extends the traditional single-link optimization to a multi-stage series model, and introduces the dynamic coupling relationship of each link. This modeling method is more in line with the actual security scene and can fully reflect the running status of the security system, providing a more accurate theoretical model basis for the analysis and optimization of security channel resources. The channel model comprehensively considers factors such as security device throughput and personnel operation efficiency, and is a full-link, multi-dimensional, and dynamically coupled joint model. The luggage inspection links are systematically deeply coupled and optimized, so that the model has stronger adaptability in dealing with complex scenarios.

[0029] As preferred, in step 1, the flow rate of the security channel is modeled according to the number of security officers arranged in each link of the security process to form a channel service capacity index under the scheduling scheme.

[0030] As preferred, in step 1, the luggage inspection link model is refined to analyze the sub-processes of the luggage process, such as the forward transmission, X-ray machine detection (including single-view X-ray machine and double-view X-ray machine), and secondary unpacking inspection, and to clarify the influencing factors of each process, such as equipment physical parameters, equipment performance, personnel operation speed, luggage quantity and type, etc., so as to perfect the luggage model and make it have high adaptability in complex scenarios.

[0031] As preferred, in step 2,

[0032] The average passenger inspection time τ ka is expressed as

[0033] This index reflects the average time of passengers from entering the security channel to completing the security, and the smaller the index value, the higher the security efficiency and the shorter the passenger waiting time.

[0034] The security officer work intensity (the number of passengers inspected by each security officer) r ka When the channel k adopts the scheduling scheme a, the index expression is

[0035] Where λ k is the actual number of passengers passing through the security in the statistical period (unit: person), T op is the channel opening time in the statistical period, and N ka is the number of security officers. This index reflects the number of passengers inspected by each security officer in unit time in the statistical period, and can reflect the work load of security officers and intuitively represent the work intensity of security officers.

[0036] As preferred, in step 6, the following constraint conditions are set,

[0037] C1:

[0038] Where θ k is the emergency reservation coefficient of the passenger flow of the channel k, which can be dynamically adjusted according to historical data and real-time situation to cope with sudden changes in passenger flow. This constraint indicates that the theoretical throughput of all open channels in each time period should be able to handle the passenger flow in that time period to meet the passenger demand, s ka is the theoretical throughput of the channel k under the security officer scheduling scheme a;

[0039] C2: The condition is a constraint on the scheduling scheme selected for the channel, meaning that only one scheduling scheme can be used at the current time period;

[0040] C3: Represents the constraint of binary decision variable, only 0 or 1 value selection;

[0041] C4: Ensure that the total number of security guards arranged for each post j in each time period t cannot exceed the total number of available security guards for the post;

[0042] C5: In order to avoid waste of human resources, ensure that the number of passengers checked by each security guard is not less than the minimum value In order to prevent the security guard from being too heavy, which will affect the quality of security, the number of passengers checked by each security guard is required to be not higher than the maximum value of the workload

[0043] As a preferred, in step 7, the resource optimization design and deep learning method of the security channel model is as follows,

[0044] Firstly, the optimization target problem is modeled as a Markov decision process (MDP) of optimal stationary decision, and the Markov decision process is represented by MDP = <S, A, P, R, Π >. The four elements S, A, P, R of the MDP model are state, action, state transition probability set and reward function, respectively. The reward function R(s t , a t ) represents the reward that the system can obtain when the system is in state s t and takes action a t at time period, and Π is the policy, which represents the basis for the system to take action, that is, the action is selected according to the policy.

[0045] Then, a DQN network is constructed to approximate the Q function, and the input of the network is the state vector s t , and the output is the Q value of each action.

[0046] Finally, the DQN algorithm is used for deep learning. The DQN algorithm can well handle the complex scene of resource allocation in the security process, and make the optimal decision by learning the dynamic characteristics of the environment. The history data is used for learning, and the strategy is optimized through continuous trial and error, so as to find a reasonable resource allocation scheme under different passenger flow and other conditions.

[0047] As a preferred, the state set S is represented by s t =(Λ t , N a , S a , M jC1, C2, C3, C4, C5) represent,

[0048] wherein s t ∈S represents the system state at period t, wherein Nk(t) represents the passenger flow of channel k at period t, N a = {N 1a , N 2a , …, N Ka}, a∈A, represents the total number of security officers required for channel k under the scheduling scheme a, S a = {S 1a , S 2a , …, S Ka}, a∈A, represents the theoretical throughput of channel k under the security officer scheduling scheme a, M j = {M1, M2, …, M J}, j∈J, is the total number of security officers available for post j, C1, C2, C3, C4, C5 respectively represent whether the constraint conditions C1-C5 are met under the current state (1 for meeting, 0 for not meeting) ;

[0049] The action set A, i.e., the actions that can be taken at period t based on the current state, can be represented as The action consists of two parts: the opening state of each channel k at each period t (channel opening is marked as 1, and closing is marked as 0); and the selection of the scheduling scheme a by each channel The size of the action space is 2 |K| ×2 |K|×|A| , where K is the channel set and A is the scheduling scheme set.

[0050] The state transition probability P is solved using a model-free learning algorithm.

[0051] The reward function R, since the optimization goal is to find the minimum value, the reward function R is:

[0052]

[0053] The strategy Π, represented by π∈Π, is a stationary decision, under which action a t will be taken when in state s t ,

[0054] When the strategy is uncertain, the probability corresponding to each action is selected as π(a t |s t ) = P[A = a t |S = s t ].

[0055] As preferred, the DQN network has two sub-networks with the same structure, a target network and a prediction network, respectively, and the network parameters of the two sub-networks are iteratively updated according to the result difference (i.e. loss value) between the two sub-networks in the training process, wherein the target network is used to realize the Q actual value, and the prediction network is used to obtain the Q estimated value, and the parameters of the prediction network are updated into the target network every certain iteration, and according to the Q value in the prediction network, the optimal strategy of each state is to select the action that maximizes the Q value, that is, after the opening state of each channel k at each time period t and the security guard scheduling scheme of the channel are selected, the environment feedbacks the instantaneous reward R(s t ,a t ), the current state s t is transferred to the next state s t+1 .

[0056] Compared with the prior art, the advantages of the present application are that:

[0057] (1) By comprehensively modeling the security check whole process in multiple dimensions and multiple stages and dynamically coupling each link, the operation of each link of the security check system can be comprehensively reflected, and the precision and practicality of resource dynamic scheduling are improved.

[0058] (2) The scheduling method of the present application has self-adaptive adjustment capability, can dynamically respond to the fluctuation of passenger flow in each period to realize resource flexible scheduling, increase manpower deployment to guarantee service efficiency during passenger flow peak period, and intelligently reduce redundant configuration to reduce operation cost during low peak period.

[0059] (3) The present application supports multi-objective optimization, can simultaneously meet the indexes of minimizing manpower cost, keeping stable security guard scheduling configuration change, and sustainable security manpower resource, and generates a resource allocation scheme automatically by using a deep reinforcement learning algorithm, can adaptively output an optimal decision by learning the dynamic characteristics of the environment, and significantly improves production efficiency compared with manual scheduling.

[0060] (4) In addition, the system model of the present application has scalability, and post grouping, scheduling rules, channel attributes and the like are parameterized, and can be expanded and adapted according to actual post structure or airport scale. Therefore, the present application has high practical value and wide application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0061] Figure 1 is a method flowchart of the present application;

[0062] Figure 2 is a system model framework diagram of the present application;

[0063] Figure 3 is a security check channel post process schematic diagram of the present application;

[0064] Figure 4 is the DQN algorithm schematic diagram of the present application;

[0065] Figure 5 is the DQN algorithm flow chart of the present application. DETAILED DESCRIPTION

[0066] The multi-index balanced optimization and dynamic optimization oriented security check resource allocation optimization algorithm proposed by the present application is to reduce security personnel cost, stabilize channel personnel configuration, meet passenger flow, balance security officer work load and other multiple goals into the optimization range, as follows:

[0067] (I) Minimizing the use of security resources

[0068] Reducing security personnel cost, reasonably optimizing the number of security channels, the number of security personnel, etc.

[0069] (II) Stabilizing channel personnel configuration

[0070] Ensure efficient coordination of each link such as identity verification, passenger inspection, luggage inspection, etc. to avoid problems such as efficiency decline or operation inexperience caused by frequent personnel deployment changes. At the same time, stable personnel configuration can keep the passenger inspection time relatively balanced, which has little impact on passenger travel experience, and helps to optimize the security officer scheduling mechanism, reduce invalid personnel scheduling, improve the utilization rate of human resources, thereby reducing the overall operating cost.

[0071] (III) Responding to passenger flow fluctuations

[0072] By dynamically adjusting the security officer scheduling and the number of open channels, it can ensure sufficient resources during peak periods and reasonable control of manpower and equipment use during low peak periods to respond to fluctuating passenger flow. Through the multi-objective optimization algorithm, the optimal balance point between different targets can be found to reasonably allocate security resources, realize the joint optimization of dynamic channel configuration and static personnel scheduling, solve the problems of dynamic resource scheduling of security channels and static scheduling of security personnel, and achieve efficient operation of the overall security system.

[0073] (IV) Sustainable development of security human resources

[0074] By accurately quantifying the number of passengers inspected by each security channel and each period of security officer per capita (i.e. security officer work load) index, the fluctuation of security officer work intensity is limited within a reasonable threshold range to avoid situations such as excessive work intensity of some security officers affecting security quality or low work load of some personnel causing waste of human resources, achieving balanced distribution of security officer work load, improving their job satisfaction and professional sustainability, ultimately achieving the coordinated optimization of security efficiency, service quality and human resource utilization, and ensuring efficient, stable and safe operation of airport security work.

[0075] The invention will be further described below. A scheduling method for optimizing airport passenger security checkpoint resources is described in detail below. Figures 1 to 5 ,

[0076] (I) Model Assumptions

[0077] To facilitate model building and computer simulation algorithm design, the present invention makes the following assumptions:

[0078] (1) Passenger type hypothesis. Different types of passengers require different amounts of time in the security check process. It is assumed that passengers of the same type behave consistently during the security check process. For example, ordinary passengers have similar levels of cooperation when undergoing security checks and similar baggage inspection times, so as to conduct unified analysis and processing of passenger groups.

[0079] (2) Security Check Process Assumptions. It is assumed that the security check process is fixed, including identity verification, passenger screening, baggage inspection, and baggage opening, with the order of these steps remaining constant. The impact of special circumstances (such as temporary adjustments to the security check process due to unforeseen events) on the security check process is not considered.

[0080] To ensure the model's stability and repeatability, the service time for each security checkpoint follows a specific probability distribution, such as an exponential distribution.

[0081] (3) Assumptions regarding security inspection equipment. All security inspection equipment operates normally, the conveyor belt runs at a constant speed, and the luggage trays are used normally. No malfunctions or performance fluctuations will occur during operation.

[0082] (4) Assumptions about security inspectors. The professional skills of security inspectors are stable and remain stable throughout the work process, without significant fluctuations due to factors such as working hours or fatigue.

[0083] (5) Baggage assumption. The number of bags each passenger carries is fixed, and the security check process for one piece of baggage will not cause additional interference to the security check process of other passengers.

[0084] (II) Setting the Objective Function

[0085] Airport terminal needs to ensure the rapid passage of passengers while achieving the rational allocation of security resources in the process of operation, so as to achieve the optimal balance point of security efficiency and operation cost. Reducing the cost of security personnel has been one of the goals of airport operation development, and stable channel personnel allocation can avoid the problems of efficiency decline or operation inexperience caused by frequent personnel allocation, and help to optimize the scheduling mechanism of security personnel, reduce invalid personnel scheduling, improve the utilization rate of human resources, and thus reduce the overall operation cost. At the same time, limiting the fluctuation of security personnel work intensity within a reasonable threshold range can achieve the balanced distribution of security personnel work load, promote the sustainable development of human resources, and ensure the efficient, stable and safe operation of airport security work. Therefore, the objective function is set as formula (1)

[0086]

[0087] Where δ, ψ and are the weight coefficients of security personnel operation cost, adjacent time period channel personnel allocation change and average passenger screening time respectively, c a and N ka are the operation cost required by the scheduling scheme a and the total number of security personnel required by the channel k under the scheduling scheme a respectively, is the decision variable, which represents whether the channel k adopts the scheduling scheme a in period t. τ ka is the average passenger screening time of channel k under the scheduling scheme a. T is the statistical time period, K represents the set of all available security channels in the airport security system, and A is the set of security personnel scheduling schemes.

[0088] (III) Constraint conditions

[0089] (1) In each time period, the theoretical throughput of all open channels should be able to handle the passenger flow of the time period, meet the demand of passenger volume, and need to apply the sudden changes of passenger flow.

[0090]

[0091] (2) The constraint of the scheduling scheme selected by the channel, a channel can only adopt one scheduling scheme in the current time period.

[0092]

[0093] (3) is the constraint of binary decision variable, which represents whether the channel k adopts the scheduling scheme a in period t, so only 0 or 1 value is selected.

[0094]

[0095] (4) Since the total number of security inspectors is fixed, the total number of security inspectors assigned to each position j at each time period t cannot exceed the total number of available security inspectors for that position.

[0096]

[0097] (5) To avoid waste of human resources and prevent security inspectors from being overburdened, which may affect the quality of security inspection, the work intensity of security inspectors (the number of passengers inspected per capita) is required to be restricted within a certain range.

[0098]

[0099] (IV) Design of the security inspection channel resource optimization algorithm based on DQN

[0100] Model the above optimization objective problem as a Markov decision (MDP) process of optimal stationary decision-making. The Markov decision process is represented by MDP = <S, A, P, R, Π>. The four elements S, A, P, and R of the MDP model are state, action, state transition probability set, and reward function respectively. The reward function R(s t , a t ) represents the reward that the system can obtain when the system is in state s t and takes action a t at a certain time period. Π is the policy, which represents the basis for the system to take actions, that is, actions are selected according to the policy, so it is a mapping from state to action.

[0101] (1) State set

[0102] Use s t ∈ S to represent the system state at time period t, where represents the passenger flow of channel k at time period t, N a = {N 1a , N 2a , …, N Ka}, a ∈ A, represents the total number of security inspectors required for channel k under the scheduling plan a, S a = {S 1a , S 2a , …, S Ka}, a ∈ A, represents the theoretical throughput of channel k under the security inspector scheduling plan a, M j = {M1, M2, …, M I}, j ∈ J, is the total number of available security inspectors for position j, and C1, C2, C3, C4, C5 respectively represent whether the constraint conditions C1 - C5 are satisfied in the current state (marked as 1 for satisfaction and 0 for dissatisfaction).

[0103] (2) Action set

[0104] The action taken based on the current state during time period t can be represented as: The action consists of two parts: first, the on / off state of each channel k in each time period t. (Channel opening is marked as 1, and closing is marked as 0); secondly, each channel selects scheduling scheme a. The size of the motion space is 2. |K| ×2 |K|×|A| , where K is the channel set and A is the scheduling scheme set.

[0105] (3) State transition probability

[0106] The state transition profile of this problem is unknown, so it can be solved using a model-free learning algorithm (such as Q-learning).

[0107] (4) Reward function

[0108] Since the optimization objective is to find the minimum value, the reward function is:

[0109]

[0110] (5) Strategy

[0111] Let π∈Π denote a stationary decision, under the policy π, in state s. t Action a will be taken at that time. t When the strategy is uncertain, the probability of choosing each action is π(a). t |s t )=P[A=a t |s=s t ].

[0112] (6) DQN network

[0113] Construct a deep neural network to approximate the Q-function; the input of this network is the state vector s. t The output is the Q value for each action.

[0114] (7) Network Structure

[0115] This network has two identical sub-networks: a target network and a prediction network. During training, the network parameters of these two sub-networks are iteratively updated based on the difference between their results (i.e., the loss value). The target network is used to realize the Q-value, and the prediction network is used to obtain the Q-estimate. After a certain number of iterations, the parameters of the prediction network are updated in the target network. Based on the Q-value in the prediction network, the optimal strategy for each state is to choose the action that maximizes that Q-value. That is, after selecting the opening state of each channel k in each time period t and the security personnel scheduling scheme for that channel, the environment provides the instantaneous reward R(s). t, a t ), current state s t transition to next state s t+1 .

[0116] (8) Pseudocode of DQN-based security channel resource optimization algorithm

[0117]

[0118]

[0119] The above describes in detail a dispatching method for airport passenger security channel resource optimization provided by the present application. The principles and implementation modes of the present application are described by using specific examples. The above example is only used to help understand the method and core idea of the present application. Meanwhile, for those skilled in the art, the specific implementation mode and application range can be changed according to the idea of the present application, and the present application can be changed and improved without exceeding the concept and range defined in the appended claims. In summary, the content of the present description should not be understood as a limitation of the present application.

Claims

1. A scheduling method for optimizing airport passenger security checkpoint resources, characterized in that: The steps are as follows: Step 1 involves performing a multi-dimensional, multi-stage integrated model of the passenger security check process, and constructing a security check channel model that dynamically couples the various stages. This model is based on factors influencing passenger type, equipment type and its processing capacity, and the number of security personnel and their operational efficiency, thus affecting the channel's performance. Security inspector scheduling plan Below are different security check stages. The flow velocity model of the channel and the overall flow velocity of the security check channel are calculated and designed. Step 2: Define the efficiency indicators of security checkpoints, specifically the average passenger screening time. and the workload of security inspectors ; Step 3: Set up a set of security check lanes, a set of security check posts, a set of statistical time periods, a set of passenger flow during that time period, and a set of security personnel scheduling schemes. Define the number of security personnel configured for each stage of the security check process under different scheduling schemes, and calculate the passage speed of the security check lanes under each scheduling scheme through the security check lane model. Step 4: Construct a multi-objective optimization model for security check resources. Use multi-index balance optimization and a dynamic optimization-oriented resource allocation optimization algorithm to minimize the average passenger check time, security check manpower cost, and channel personnel configuration changes under the conditions of dynamic passenger flow, security check staff number limit, and security check staff workload limit. Balance security check efficiency and human resource utilization, and allocate resources rationally. Step 5, construct the objective function: ； in , and These are the weighting coefficients for security check manpower operating costs, changes in staffing levels at checkpoints in adjacent time periods, and average passenger check-through time. and These are the use of a scheduling scheme. The operating costs and channel required for this security checkpoint In the scheduling plan The total number of security personnel required below For decision variables, representing the time period Central Channel Should a shift scheduling scheme be adopted? , For channel In the scheduling plan Average inspection time for disembarking passengers; Step 6, set constraints: Step 7: Perform resource optimization design on the security checkpoint model, use deep reinforcement learning algorithm to solve the objective function, and output the optimal opening scheme and security personnel scheduling scheme for each security checkpoint in each time period.

2. The scheduling method for optimizing airport passenger security checkpoint resources according to claim 1, characterized in that: In step 1, the specific method for constructing the security checkpoint model is as follows: First, the security check process is divided into multiple sequential stages, including key subsystems for identity verification, passenger body inspection, and baggage inspection. Then, a comprehensive model is created for all key links in the entire chain, from identity verification and passenger screening to baggage screening. Then, based on the differences in passenger type, equipment type and its processing capacity, and the number of security personnel and their operational efficiency, a channel was established. Security inspector scheduling plan Below are different security check stages. The channel flow velocity model is shown below: ； in This indicates different stages of the security check process, including identity verification, passenger body search, and baggage inspection. Indicates in the channel Each type of passenger The proportion of total passenger traffic For channel Different security check stages The processing capacity of the security inspection equipment For channel Chinese passenger type For security check process Influence factors For the efficiency function of security personnel, i.e., in the channel Security check process In the meantime, with the security inspector shift schedule... Number of security personnel The increase in flow rate in the channel is a function of personnel efficiency. Finally, the overall flow velocity of the security checkpoint can be expressed as .

3. The scheduling method for optimizing airport passenger security checkpoint resources according to claim 1, characterized in that: In step 2, the flow rate of the security checkpoint is modeled by combining the number of security personnel configured in each stage of the security check process to form the channel service capacity index under the scheduling plan.

4. The scheduling method for optimizing airport passenger security checkpoint resources according to claim 1, characterized in that: In step 1, the baggage inspection process model is refined, and the baggage process sub-processes of pre-shipment, X-ray machine inspection, and secondary baggage inspection are analyzed. The influencing factors of each process, such as equipment physical parameters, equipment performance, personnel operation speed, baggage quantity and type, are identified to improve the baggage model.

5. The scheduling method for optimizing airport passenger security checkpoint resources according to claim 1, characterized in that: In step 2, The average passenger screening time The expression is , The workload of security inspectors ,aisle Adopt a scheduling scheme When, the expression for this indicator is: , in To count the actual number of passengers passing through security during a given time period, This refers to the channel opening hours within the statistical period. It refers to the number of security personnel.

6. The scheduling method for optimizing airport passenger security checkpoint resources according to claim 1, characterized in that: In step 6, set the following constraints: C1: ， in, For channel Up passenger flow The emergency reserve coefficient can be dynamically adjusted based on historical data and real-time conditions to cope with sudden changes in passenger flow. For channel Security inspector scheduling plan Theoretical throughput below; C2: This condition is a constraint on the scheduling scheme for channel selection; C3: This represents the constraint of the binary decision variable, which can only have a value of 0 or 1; C4: Ensure that in each time period Up and every position The total number of security inspectors assigned to this position shall not exceed the total number of security inspectors available for that position; C5: To avoid wasting human resources, ensure that each security inspector checks at least the minimum number of passengers. To prevent security personnel from being overburdened and affecting the quality of security checks, it is required that the average number of passengers checked per security officer should not exceed the maximum workload. .

7. A scheduling method for optimizing airport passenger security checkpoint resources according to claim 1, characterized in that: In step 7, the method for resource optimization design and deep learning of the security checkpoint model is as follows: First, the optimization problem is modeled as a Markov decision process with optimal stationary decision-making. The Markov decision process uses... The MDP model consists of four elements: S, A, P, and R, representing the state, action, set of state transition probabilities, and reward function, respectively. This indicates the state when the system is in a given time period. And take action The rewards that the system can obtain. It is a strategy, representing the basis for the system's actions; Then, a DQN network is constructed, which is a deep neural network to approximate the Q function. The input of this network is the state vector. The output is the Q value for each action; Finally, the DQN algorithm is used for deep learning, which utilizes historical data for learning and optimizes the strategy through continuous trial and error, thereby finding a reasonable resource allocation scheme under different passenger flow and other conditions.

8. A scheduling method for optimizing airport passenger security checkpoint resources according to claim 7, characterized in that: The state set S is used express, in, Indicates time period The system status, where , indicates a time period aisle Passenger flow , represents a channel In the scheduling plan The total number of security personnel required below , indicating channel Security inspector scheduling plan Theoretical throughput below, It is a job position. Total number of available security personnel , , , , These indicate whether constraints C1 - C5 are satisfied under the current state; The action set A can be within a time period The action taken based on the current state can be represented as The action consists of two parts: one is each channel In each time period On status Secondly, the scheduling plan for each channel. choose The size of the action space is ,in A is a set of channels, and A is a set of scheduling schemes; The state transition probability Solve using a model-free learning algorithm; The reward function R is: ； The strategy ,use This indicates a smooth decision-making process, in this strategy Below, in a state Action will be taken at that time. When the strategy is uncertain, the probability of choosing each action is: .

9. A scheduling method for optimizing airport passenger security checkpoint resources according to claim 7, characterized in that: The DQN network has two structurally identical sub-networks: a target network and a prediction network. During training, the network parameters of the two sub-networks are iteratively updated based on the difference between their results. The target network is used to realize the Q-value, and the prediction network is used to obtain the Q-estimate. After a certain number of iterations, the parameters of the prediction network are updated in the target network. Based on the Q-value in the prediction network, the optimal strategy for each state is to select the action that maximizes that Q-value, i.e., to select the action that maximizes each channel. In each time period After the activation status and security personnel scheduling plan for this channel are determined, an instantaneous reward is generated based on the environmental feedback. Current state Transition to the next state .

Images