Emergency evacuation method for sea-crossing vehicles based on time-varying road network

Abstract

This invention provides an emergency evacuation method for cross-sea vehicles based on a time-varying road network. The method is activated when a sudden event affecting cross-sea traffic occurs. The method includes the following steps: constructing a time-varying road network that varies with departure time and traffic flow speed; establishing a probabilistic distributed K-shortest path model with the goal of minimizing total evacuation time, dispersing the initial traffic flow in multiple directions along K shortest paths according to assigned probabilities; constructing a chance-constrained road network congestion model using road segment saturation with the goal of minimizing road segment congestion probability; establishing a reverse lane configuration model and using an improved genetic algorithm to jointly solve the above models to obtain the optimal evacuation path and traffic allocation scheme; periodically updating road network parameters, and re-executing the above steps when real-time traffic flow deviates from the planned value by more than a threshold or unexpected congestion occurs, to generate an updated emergency evacuation plan.

Description

Technical Field

[0001] This invention belongs to the field of evacuation route optimization technology, specifically relating to an emergency evacuation method for vehicles crossing the sea based on a time-varying road network. Background Technology

[0002] Ports are crucial hubs in the transportation system, connecting water and land transport and handling massive volumes of cargo. In recent years, the continuous expansion of road networks and the surge in travel demand have significantly increased traffic pressure around ports. In the event of sustained and sudden severe weather, vehicles waiting to cross the sea quickly accumulate outside the port, drastically exacerbating congestion. Emergencies can also cause physical or functional damage to the regional road network, drastically reducing traffic capacity and creating a huge instantaneous demand for evacuation vehicles, leading to severe congestion and reduced evacuation efficiency. Therefore, timely and effective relocation of vehicles from the affected area or potential danger zone (i.e., the study area) to a safe zone is a key measure to prevent emergencies from escalating into major safety problems.

[0003] Emergency evacuation planning is essentially a systemic problem involving complex networks. Transportation networks are inherently dynamic and uncertain: traffic speeds vary with time, road segment flow, and density; road segment capacity also varies with traffic flow and time. Under emergency conditions, these characteristics are amplified. When a sudden event is large-scale, the evacuation traffic flow inevitably places extremely high demands on the regional road network's capacity. Therefore, it is essential to coordinate the coexistence of non-evacuation and evacuation traffic flows under time-varying road network conditions. This not only improves evacuation efficiency but also provides solid theoretical support for emergency command centers. Current research is mainly based on the static road network assumption, treating traffic parameters as constant values ​​or fixed values ​​at specific times. Traditional single evacuation route planning leads to highly concentrated traffic density, violating the principle of efficient evacuation and causing strong disturbances to non-evacuation traffic along the route, thus exacerbating congestion and safety hazards. At the level of emergency traffic organization, the mainstream approach focuses on optimizing intersection signal timing and setting up reverse lanes. While these methods can significantly improve evacuation efficiency, they rarely consider the impact on non-evacuation traffic flow, especially the interference of reverse lanes on non-evacuation traffic flow.

[0004] The core problem with existing technologies is that static models cannot depict the drastic fluctuations in road network capacity and demand during emergency scenarios, while path planning that only pursues evacuation efficiency ignores the cascading impact on background traffic. Although organizational measures such as reverse lanes and signal optimization can quickly release evacuation channel capacity, they lack a systematic assessment of losses in non-evacuation traffic, leading to the frequent emergence of new congestion points. More importantly, there is currently a lack of a method for setting up reverse lanes that simultaneously incorporates evacuation traffic efficiency and non-evacuation traffic demand into the optimization objectives. This makes it impossible to ensure high-speed passage for evacuation vehicles while also taking into account the basic operational needs of background traffic, and it is also difficult to dynamically adjust the scope and duration of reverse flow control based on real-time traffic conditions. As a result, the improvement in evacuation efficiency is often limited, while delays on the non-evacuation side increase significantly, and the overall network performance actually declines. Summary of the Invention

[0005] This invention proposes an emergency evacuation method for vehicles crossing the sea based on a time-varying road network, which solves the problem that traditional evacuation methods ignore the time-varying characteristics of the road network, resulting in unreasonable traffic flow distribution.

[0006] To address the aforementioned technical problems, this invention provides an emergency evacuation method for cross-sea vehicles based on a time-varying road network, which is activated when a sudden event affecting cross-sea traffic occurs. The method is characterized by the following steps:

[0007] Step S1: Perform directed rich attribute node modeling on the evacuation road network area, collect all path nodes in the evacuation road network area and emergency evacuation road sections composed of pairs of different path nodes, collect the road section distance, vehicle speed, free flow travel time and actual traffic capacity of each emergency evacuation road section, and construct a time-varying road network in which the road section travel time changes with departure time and traffic flow speed.

[0008] Step S2: With the goal of minimizing the total evacuation time, construct a probabilistic distributed evacuation system based on a time-varying road network. K The shortest path model divides the initial evacuation traffic flow into multiple traffic segments, with each segment allocated according to its flow probability. K Evacuation in multiple directions along the shortest path;

[0009] Step S3: Taking the minimum congestion probability of road segments in the evacuation road network as the optimization objective, the equilibrium constraint relationship between evacuation traffic flow and non-evacuation traffic flow is quantified by using road segment saturation, and an opportunity-constrained evacuation road network congestion model is constructed.

[0010] Step S4: Construct a reverse lane configuration model, and use an improved genetic algorithm to process the probabilistic distributed model. K The shortest path model, the opportunity-constrained evacuation network congestion model, and the reverse lane configuration model are jointly solved to obtain the optimal evacuation path and traffic flow allocation scheme.

[0011] Step S5: Every time interval The system collects real-time traffic flow and vehicle speed data for each road segment, updates the parameters of the time-varying road network, and when it detects that the real-time traffic flow of a road segment deviates from the planned traffic flow by more than a threshold or an unexpected congestion occurs on a road segment, it re-executes steps S2-S4 to generate an updated evacuation plan.

[0012] Preferably, the directed rich attribute node modeling of the evacuation road network area described in step S1 includes at least the following steps: splitting a single road network node into multiple entrance and exit nodes, constructing a directed rich attribute network, and using the distance between entrance and exit nodes to represent the actual driving distance at the node.

[0013] Preferably, the actual traffic capacity mentioned in step S1 is obtained by the following formula:

[0014] ;

[0015] ;

[0016] ;

[0017] In the formula, Indicates the capacity of a single lane; Indicates the basic traffic capacity of a single lane; This indicates the lane width correction factor; Indicates lane width; Indicates the lane reduction factor; This indicates the traffic capacity of a road group.

[0018] Preferably, the construction of a time-varying road network in step S1, where the travel time of road segments varies with departure time and traffic flow speed, includes at least: determining the relationship between the road segment weights and departure time of the time-varying road network using a time dependency function, wherein the time dependency function follows a first-in-first-out (FIFO) principle, and the expression of the time dependency function is:

[0019] ;

[0020] In the formula, , These represent the departure times of car A and car B, respectively. , These represent vehicles A and B on the road segment. Travel time on the road.

[0021] Preferably, the vehicle speed in step S1 is a time-dependent speed, and the expression relating the vehicle speed for traffic diversion to the normal traffic flow speed is:

[0022] ;

[0023] In the formula, express At that time, traffic flow In the arc The speed of travel on, of which Indicates the first Each period, Indicates the first Traffic flow of the segment, Indicates the evacuation direction. In order to evacuate the source, As the evacuation endpoint, Indicates the first The shortest path; , The attenuation coefficient; Indicates the first Normal traffic flow speed within a given time period.

[0024] Preferably, in step S2, the traffic flow segments are allocated according to the traffic flow probability. K When conducting multi-directional evacuation along the shortest path, the actual traffic flow is obtained using the following formula:

[0025] ;

[0026] ;

[0027] ;

[0028] In the formula, Represents arc The saturation level of the road section; Represents arc Actual traffic flow; Represents arc The road group's traffic capacity; Represents arc Non-evacuation traffic flow; Indicates the evacuation of traffic flow In the arc Traffic on the internet; For 0-1 variables, when the arc At the evacuation point To the evacuation endpoint of K The value is 1 if it is on the shortest path, and 0 otherwise. To alleviate traffic congestion The number of vehicles; To alleviate traffic congestion In direction Traffic allocation probability; To alleviate traffic congestion In direction From the evacuation point To the evacuation endpoint of K Shortest path flow allocation probability.

[0029] Preferably, the total evacuation time in step S2 is obtained by summing the travel times of each emergency evacuation route, and the expression for the travel time of each route is:

[0030] ;

[0031] In the formula, Arc in time-varying road network exist The permitted travel time for a road segment in the timetable; Indicates the time interval between vehicles. ; Indicates time, , Indicates the first The start and end times of each time period; For free flow Travel time at any given moment; For arc The saturation level of the road section;

[0032] Among them, free-flow travel time The calculation is divided into the following cases based on the different distances the vehicle travels and the time periods it spans:

[0033] (1) When hour:

[0034] ;

[0035] (2) When hour;

[0036] ;

[0037] (3) When hour:

[0038] ;

[0039] In the formula, express At that time, traffic flow In the arc The speed of travel on the road; Indicates the vehicle at time intervals driving distance, ; Indicates the speed at which the vehicle starts moving from the current speed. arrive driving distance, ; Indicates the number of time periods spanned.

[0040] Preferably, the probabilistic distributed method based on time-varying road networks described in step S2... K The constraints of the shortest path model include at least the following:

[0041] Traffic balance constraints at intermediate nodes of the road network: ;

[0042] Traffic balance constraints at the origin and destination points of the road network:

[0043] ;

[0044] Traffic flow constraints for each segment: ;

[0045] Traffic distribution balance constraints: ;

[0046] K Shortest path flow distribution balance constraint: ;

[0047] In the above formula, , , , Representing arcs respectively ,arc , and Actual traffic flow; Represents the set of evacuation source points; Represents the set of intermediate nodes; Indicates the final set of evacuation destinations; To alleviate traffic congestion The number of vehicles; Indicates the total evacuation traffic volume; To alleviate traffic congestion In direction Traffic allocation probability; To alleviate traffic congestion In direction From the evacuation point To the evacuation endpoint of K Shortest path flow allocation probability.

[0048] Preferably, the constraints of the opportunity-constrained evacuation network congestion model in step S3 include at least the following:

[0049] ;

[0050] ;

[0051] ;

[0052] ;

[0053] In the formula, Represents the probability of an event. ; This represents the upper limit of road segment saturation. Represents arc The saturation level of the road section; Represents arc The probability of congestion on the device.

[0054] Preferably, the expression for the reverse lane configuration model in step S4 is:

[0055] ;

[0056] ;

[0057] ;

[0058] ;

[0059] ;

[0060] In the formula, This represents the total difference in evacuation time before and after optimization; Indicates evacuation time; The optimized total evacuation time; This indicates the traffic capacity of the road segment after the opposite lane has been set up; Represents arc The road group's traffic capacity; This indicates the change in road capacity after the opposite lane is installed; It is a 0-1 variable, when the road segment The value is 1 when the lane is selected as the opposite lane, and 0 otherwise. Indicates road segment The number of reverse lanes is set in the middle; This represents the path in the initial input scheme; This indicates the total amount of transportation infrastructure and labor resources; This indicates the resources consumed when each road segment is selected as the reverse lane; This indicates the road saturation level after setting up the reverse lane.

[0061] The beneficial effects of the present invention include at least the following:

[0062] 1. By constructing a time-varying road network, the changes in road segment travel time with departure time and traffic flow speed are fully considered, which can more accurately reflect the dynamic characteristics in actual traffic scenarios and has stronger adaptability and accuracy compared with traditional static models.

[0063] 2. In the optimization process, the goal is to minimize the total evacuation time while also considering the minimum probability of road congestion. This balances evacuation efficiency and road network balance, avoiding the local congestion problems caused by simply pursuing evacuation efficiency in traditional methods. It achieves a balanced optimization of evacuation traffic and non-evacuation traffic.

[0064] 3. Adopt probabilistic distributed model K The shortest path model divides evacuation traffic flow into multiple segments and distributes them to multiple paths, increasing the diversity of evacuation paths, reducing dependence on a single path, and improving the robustness of the evacuation system. At the same time, it optimizes the traffic flow distribution of each path by optimizing the traffic allocation probability, further improving evacuation efficiency. Attached Figure Description

[0065] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention. Detailed Implementation

[0066] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0067] like Figure 1 As shown, this embodiment of the invention provides an emergency evacuation method for cross-sea vehicles based on a time-varying road network. When a sudden event affecting cross-sea traffic occurs, the method is activated and includes the following steps:

[0068] Step S1: Perform directed rich attribute node modeling on the evacuation road network area, collect all path nodes within the evacuation road network area and emergency evacuation road segments composed of pairs of different path nodes, collect the road segment distance, vehicle speed, free-flow travel time, and actual traffic capacity of each emergency evacuation road segment, and construct a time-varying road network in which the road segment travel time varies with departure time and traffic flow speed. Specifically, this includes the following sub-steps:

[0069] Step S11: Perform directed rich attribute node modeling for the evacuation road network area.

[0070] Specifically, all path nodes within the evacuation road network area, as well as emergency evacuation road sections composed of pairs of different path nodes, are collected. Roads with the maximum traffic capacity or those with distribution functions are selected as regional dividing lines. In the path nodes, a single road network node is split into multiple entrance and exit nodes, constructing a directed rich attribute network, with the distance between entrance and exit nodes representing the actual travel distance at the node.

[0071] Using directed graphs This represents the regional evacuation road network. Among them, This represents the set of nodes in the evacuation road network, including the set of evacuation source points. Path splitting node set Evacuation convergence point assembly ; This represents the set of arcs in the evacuation road network, namely the connecting road segments between evacuation source points, diversion nodes, and evacuation sink points; Indicate each arc segment The set of road capacity; Represents arc The length of the set; Indicates the total number of vehicles evacuated; This indicates the total number of non-evacuation vehicles; Indicates the evacuation time.

[0072] Multiple origin-destination pairs are set up, and a multi-directional evacuation approach with distributed nodes is adopted to accelerate evacuation efficiency, maximize the utilization of the road network's carrying capacity, and avoid the problem of secondary vehicle aggregation caused by single-path planning. The initial evacuation traffic flow is divided into multiple traffic segments. Each traffic flow has multiple directions Evacuate the area and find the shortest path K to each parking point in each direction. To represent. For example, This indicates that the flow rate of the first segment is... Traffic flow in direction 1 from the evacuation point To the evacuation endpoint The first shortest time path.

[0073] Step S12: Collect basic information on the directed rich attribute road network.

[0074] Data on road segment distances, vehicle speeds, free-flow travel time, and actual traffic capacity of each emergency evacuation road section within the evacuation road network area will be collected. Information on the number of lanes, road grade, single lane width, design speed, and one-way road capacity will be collected through on-site surveys and electronic maps.

[0075] The designation of reverse lanes redistributes road segment capacity through right-of-way allocation, thus altering the capacity of the same road segment in different directions. Considering that vehicle driving conditions are significantly influenced by road conditions, lane width correction factors and lane number correction factors are used to calculate road segment capacity. The capacity of a one-way road group is obtained using the following formula:

[0076] ;

[0077] ;

[0078] ;

[0079] In the formula, This indicates the single-lane capacity, expressed in units of equivalent standard passenger cars per hour. ; This indicates the basic traffic capacity of a single lane, in units of... ; This indicates the lane width correction factor; This indicates the lane width, in meters. ; This represents the lane reduction factor, and its value is determined with reference to the "Code for Design of Urban Road Engineering (2016 Edition)" and "Technical Standard for Highway Engineering JTGB01-2014". Indicates the traffic capacity of a road group, in units of .

[0080] Step S13: Determine the time dependency function of the time-varying road network.

[0081] In time-dependent road networks, the edge weights are not fixed values ​​but exhibit dynamic changes. Based on the analysis of the road network and traffic flow characteristics of the study area, traffic flow shows a clear car-following phenomenon during dispersal. The constructed time-varying road network needs to follow the first-in, first-out (FIFO) principle, meaning that vehicles maintain consistent speeds on the same road segment at the same time, and overtaking is not allowed; vehicles entering an intersection first must enter the adjacent intersection first. Therefore, the Ichoua time-dependent function is used to obtain a continuous departure time-travel time function, which is obtained through the following formula:

[0082] ;

[0083] In the formula, , These represent the departure times of car A and car B, respectively. , These represent vehicles A and B on the road segment. The travel time on the road. For any differentiable point in the Ichoua time-dependent function, the derivative is greater than -1.

[0084] Because of the time-dependent changes in vehicle speed across the road network, travel time on a road segment is significantly affected by departure time, meaning that travel time varies across different time periods. Calculating time-dependent road network travel time typically involves discretizing continuous time into a series of equal, continuous time intervals to calculate the travel time of vehicles across different time periods.

[0085] Vehicle arrives at node The moment There are multiple possibilities, therefore calculation When taking the value, it is necessary to first determine the location of the vehicle on the road segment. Number of time periods spanned during the journey The value of . Let . H Time segment set , Indicates the first For a period of time Vehicles on the road section Number of time periods during driving R Obtained through the following formula:

[0086] ;

[0087] In the formula, Indicates that the vehicle has left the node. The moment; Indicates the vehicle has arrived at the node. The moment; Indicates that the vehicle is in From the node To the node Time required; Indicates the number of time periods spanned.

[0088] Step S2: With the goal of minimizing the total evacuation time, construct a probabilistic distributed evacuation system based on a time-varying road network. K The shortest path model divides the initial evacuation traffic flow into multiple traffic segments, with each segment allocated according to its flow probability. K Evacuation is carried out in multiple directions along the shortest path. This includes the following sub-steps:

[0089] Step S21: Set up origin-destination pairs in the road network and determine the evacuation method.

[0090] The initial evacuation traffic flow is divided into multiple segments by using a multi-node, multi-directional evacuation method, with each segment evacuating in multiple directions.

[0091] Step S22: Determine the input values ​​for evacuation traffic flow and non-evacuation traffic flow.

[0092] In evacuation route planning, the scale of evacuation traffic flow and non-evacuation traffic flow are direct factors affecting evacuation efficiency. Before building the model, the input values ​​of these two factors must be determined, including: traffic flow calculation and time-dependent speed function.

[0093] (1) Time-dependent velocity function

[0094] In time-dependent road networks, travel time depends on departure time and changes in traffic flow speed. During emergencies, traffic flow speed is also affected by factors such as the nature of the event and road conditions, including road grade and distance from the accident source. This variation can be described as follows: during emergencies, traffic flow speed decreases to some extent compared to normal traffic flow. To maintain generality, the relationship between evacuation traffic flow speed and normal traffic flow speed is as follows:

[0095] ;

[0096] In the formula, express At that time, traffic flow In the arc The speed of travel on, of which Indicates the first Traffic flow of the segment, Indicates the evacuation direction. In order to evacuate the source, As the evacuation endpoint, Indicates the first The shortest path; , This is the attenuation coefficient, and its specific value depends on the nature of the event. In cases that do not cause road damage or are small- to medium-sized sudden emergencies, it is typically set to [value missing]. ∈[0.8,0.9], ∈[0,0.05]; Indicates the first Normal traffic flow speed within a given time period.

[0097] (2) Calculation of actual traffic flow

[0098] Actual traffic flow is obtained using the following formula:

[0099] ;

[0100] ;

[0101] ;

[0102] In the formula, Represents arc The saturation level of the road section; Represents arc Actual traffic flow, expressed in pcu / h; Represents arc The road group's traffic capacity; Represents arc Non-evacuation traffic flow, in units of pcu / h; Indicates the evacuation of traffic flow In the arc The flow rate is expressed in pcu / h. For 0-1 variables, when the arc At the evacuation point To the evacuation endpoint of K The value is 1 if it is on the shortest path, and 0 otherwise. To alleviate traffic congestion The number of vehicles can be set as the number of evacuation vehicles per unit time in practical applications, with the unit being pcu / h. To alleviate traffic congestion In direction Traffic allocation probability; To alleviate traffic congestion In direction From the evacuation point To the evacuation endpoint of K The shortest path flow allocation probability is used as a decision variable and solved in the genetic algorithm.

[0103] Step S23: Calculate the vehicle travel time across time periods using a time-dependent function based on road network information.

[0104] The calculation of travel time follows the method recommended by the Federal Highway Administration (BPR): the travel time of a road segment and the traffic flow satisfy the BPR function. Combining this with the formula for calculating travel time across time periods in time-dependent road networks, it can be transformed into a calculation method applicable to time-varying road networks. The travel time of a road segment in a time-varying road network is obtained using the following formula:

[0105] ;

[0106] In the formula, Arc in time-varying road network exist The permitted travel time for a road segment in the timetable; Indicates the time interval between vehicles. ; Indicates time, , Indicates the first The start and end times of each time period; For free flow Travel time at any given moment; For arc The saturation level of the road section;

[0107] Among them, free-flow travel time The calculation is divided into the following cases based on the different distances the vehicle travels and the time periods it spans:

[0108] (1) When hour:

[0109] ;

[0110] (2) When hour;

[0111] ;

[0112] (3) When hour:

[0113] ;

[0114] In the formula, express At that time, traffic flow In the arc The speed of travel on the road; Indicates the vehicle at time intervals driving distance, ; Indicates the speed at which the vehicle starts moving from the current speed. arrive driving distance, ; Indicates the number of time periods spanned.

[0115] Based on the above analysis, a probabilistic distributed model based on time-varying road networks is constructed. K The shortest path model has the following objective function and constraints:

[0116] Objective function: ;

[0117] Constraints:

[0118] (1) Traffic balance constraints at intermediate nodes of the road network: ;

[0119] (2) Traffic balance constraints at the origin and destination points of the road network: ;

[0120] (3) Traffic flow constraints for each segment: ;

[0121] (4) Flow distribution balance constraint: ;

[0122] (5) K-shortest path flow distribution balance constraint: ;

[0123] In the above formula, Total evacuation time, in hours. ; Arc in time-varying road network The passage time indicated is relative to the evacuation start time. Traffic flow on road sections The relevant functional relationship; , , , Representing arcs respectively ,arc , and Actual traffic flow; Represents the set of evacuation source points; Represents the set of intermediate nodes; Indicates the final set of evacuation destinations; To alleviate traffic congestion The number of vehicles; Indicates the total evacuation traffic volume; To alleviate traffic congestion In direction Traffic allocation probability; To alleviate traffic congestion In direction From the evacuation point To the evacuation endpoint of K Shortest path flow allocation probability.

[0124] Step S3: Taking the minimum congestion probability of road segments in the evacuation road network as the optimization objective, the equilibrium constraint relationship between evacuation traffic flow and non-evacuation traffic flow is quantified using road segment saturation, and a chance-constrained evacuation road network congestion model is constructed. This includes the following sub-steps:

[0125] Step S31: Use the road segment saturation measure to quantify the equilibrium relationship between evacuation traffic flow and non-evacuation traffic flow, and establish a chance-constrained evacuation road network congestion model.

[0126] Emergency evacuation involves two traffic entities: evacuation traffic and non-evacuation traffic. The equilibrium relationship between them can be equivalently replaced by the relationship between traffic flow and capacity, quantified by the ratio of traffic flow to capacity, i.e., road segment saturation. The traffic flow of non-evacuation traffic is a random variable, and the probability of the congestion level falling below a certain threshold is used as the optimization objective of the model. The chance-constrained evacuation road network congestion model is as follows:

[0127] Objective function:

[0128] Constraints:

[0129] ;

[0130] ;

[0131] ;

[0132] ;

[0133] In the formula, Represents the probability of an event. ; This represents the upper limit of road segment saturation. Represents arc The saturation level of the road section; Represents arc The congestion probability on the arc, i.e. The road section saturation is at The probability is not lower than the confidence level. .

[0134] The calculation is related to the random variable of non-evacuation traffic flow. The pseudocode for the calculation method is as follows:

[0135]

[0136] Step S32: Use the Bonferroni correction method to control the significance level of the constraints in the model, and simplify the solution by performing an exact equivalent transformation.

[0137] It should be noted that probabilistic distributed systems... K The shortest path model is achieved by setting the value of the model parameter K. K It can take values ​​of 1, 2, 3..., the specific value needs to be determined based on the specific situation, but... K The larger the value of , the more complex the model and the more difficult the solution. It can distribute traffic using the entire evacuation network as the load carrier, improving network resource utilization, overcoming the traffic flow re-concentration problem of single-path planning, and ensuring a balanced distribution of traffic flow across the network. Based on a time-varying network, setting different evacuation start times will yield different evacuation schemes, implying the time dependence and rationality of the evacuation schemes. The opportunity-constrained evacuation network congestion model uses a road segment saturation index to measure the equilibrium relationship between evacuated and non-evacuated traffic flows, thereby controlling the rationality of path selection and traffic allocation.

[0138] Step S4: Construct a reverse lane configuration model and use an improved genetic algorithm for probabilistic distributed... K The shortest path model, the opportunity-constrained evacuation network congestion model, and the reverse lane configuration model are jointly solved to obtain the optimal evacuation path and traffic flow allocation scheme.

[0139] Specifically, in real-world emergency evacuation cases, in addition to obtaining emergency evacuation plans through route optimization, temporary traffic organization strategies can be combined to further improve the efficiency of the evacuation plan. Based on the results of the emergency evacuation route optimization model, this invention utilizes a temporary reverse lane traffic organization method to redistribute available traffic resources, constructs a reverse lane configuration model, and further optimizes the emergency evacuation plan by redistributing available traffic resources.

[0140] The advantages of configuring reverse lanes are twofold: firstly, it can improve the capacity of the congested side, thereby reducing travel time; secondly, it can further reduce the probability of system congestion. It is important to note that, considering the balance between the two main traffic drivers, the configuration of reverse lanes must take into account the needs of both non-evacuation and evacuation traffic flows. Furthermore, the number of reverse lanes should not be excessive, and traffic organization strategies need to optimize evacuation plans within the constraint of the maximum number of reverse lanes.

[0141] The configuration model for the oncoming lane is as follows:

[0142] Objective function: ;

[0143] Constraints:

[0144] ;

[0145] ;

[0146] ;

[0147] ;

[0148] In the formula, This represents the total difference in evacuation time before and after optimization; Indicates evacuation time; The optimized total evacuation time; This indicates the traffic capacity of the road segment after the opposite lane has been set up; Represents arc The road group's traffic capacity; This indicates the change in road capacity after the opposite lane is installed; It is a 0-1 variable, when the road segment The value is 1 when the lane is selected as the opposite lane, and 0 otherwise. Indicates road segment The number of reverse lanes is set in the middle; This represents the path in the initial input scheme; This indicates the total amount of transportation infrastructure and labor resources; This indicates the resources consumed when each road segment is selected as the reverse lane; This indicates the road saturation level after setting up the reverse lane.

[0149] In the above model, the objective function represents the optimization objective of maximizing the time savings before and after optimization, with the total evacuation time difference as the optimization goal. The first and second constraints are the formulas for calculating the traffic capacity of road segments after setting the reverse lanes. For road segments belonging to the initial scheme, the traffic capacity increases, and for road segments not belonging to the initial scheme, the traffic capacity decreases. The third constraint represents setting an upper limit on the number of reverse lanes. The fourth constraint represents the range of values ​​for the saturation of road segments after setting the reverse lanes. Within this range, the redistribution of reverse road resources can be considered reasonable.

[0150] Based on the path planning model, an improved genetic algorithm is used to solve the multi-objective optimization model, obtaining the optimal evacuation routes and traffic flow allocation schemes. The algorithm will produce a Pareto solution set, where each solution in the solution set improves one objective function but deteriorates another objective function. Compared with the weighted method, it is more objective and operable, and any solution can be selected as the output scheme according to actual needs.

[0151] The improved genetic algorithm in this embodiment of the invention employs the improved multi-objective genetic algorithm NSGA II, and includes the following steps:

[0152] Step S41: Individual coding, for continuous variables Its encoding length is This refers to the number of evacuation traffic flows in different directions.

[0153] Step S42: Population initialization, generating an initial population and setting the population size;

[0154] Step S43: The Pareto front method is used to determine the merits of individuals through non-dominated ranking, resulting in a Pareto solution set. Each solution in the set improves one objective function while deteriorating another. Compared to the weighted method, this approach is more objective and operable, allowing any solution to be selected as the output scheme based on actual needs. The non-dominated ranking using the Pareto front method includes the following steps:

[0155] Step S431: For each individual Allocate two key quantities: Dominance Number of solutions With The solution set of domination .

[0156] Step S432: Setting ,Will Individual classification ;

[0157] Step S433: For The individuals in the solution, traversing each solution of , and each of the solutions Subtract 1;

[0158] Step S434: Let ,Will Solution classification ;

[0159] Step S435: Repeat Step 3 and Step 4 until all individuals have been grouped. .

[0160] Step S44: Select superior individuals using a binary tournament selection method for subsequent genetic operations, including the following steps:

[0161] Step S441: Randomly select two individuals from the parent population and compare their non-dominance order and crowding distance, prioritizing the individual with the higher non-dominance level; if the two individuals have the same non-dominance level, compare their crowding distance and select the individual with the higher crowding distance; if both individuals have the same non-dominance level and crowding distance, arbitrarily select one of them.

[0162] Step S4422: Repeat step S441 until the set population size is reached.

[0163] Step S45: Based on the existing adaptive strategy, an adaptive crossover and mutation strategy based on genetic parameters was designed to dynamically adjust the crossover probability according to the algorithm's evolutionary stage and the dominance of individuals in the population. and mutation probability :

[0164] ;

[0165] ;

[0166] In the formula, Represents two overlapping individuals and The smaller value of the non-dominant level; This represents the mean of non-dominant ranks in the current population; This represents the highest non-dominant level in the current population; This indicates the current individual's non-dominant level; Indicates the current evolutionary stage of the algorithm; This indicates the number of iterations set for the algorithm.

[0167] Adaptive crossover and mutation based on genetic parameters dynamically adjust parameter probabilities according to population quality and the stage of the algorithm, specifically including the following:

[0168] (1) When the non-dominant level of the crossover and mutation objects is high, that is, the current population quality is poor, crossover and mutation operations are performed with the highest probability to promote the evolution of individuals towards a better solution.

[0169] (2) When the non-dominated level of the crossover and mutation objects is low, that is, the current population quality is good. At the same time, considering the current evolution stage of the algorithm and the overall non-dominated level of the population, if the current evolution stage gen is large, the crossover and mutation probability is appropriately reduced in order to ensure the convergence of the algorithm; if the overall maximum non-dominated level of the population is large, that is, the algorithm has not yet converged, the crossover and mutation probability is appropriately increased.

[0170] (3) In particular, if individuals in the current population do not dominate each other during the algorithm process, crossover and mutation are performed with the highest probability to improve population diversity.

[0171] Step S46: Perform crossover using a crossover operator based on the normal distribution;

[0172] Step S47: Perform mutation using the Cauchy mutation operator.

[0173] In the Pareto front method, to ensure a uniform distribution of solutions along the Pareto front, it is necessary to calculate the crowding distance for solutions on the same front, i.e., the density of individuals within the same level. The calculation steps are as follows:

[0174] Step 1: Sort the individuals in a single frontier according to the values ​​of a certain objective function from smallest to largest;

[0175] Step 2: Record the maximum value obtained after sorting as The minimum value is denoted as .

[0176] Step 3: Calculate each individual The distance between adjacent individuals, i.e. and The difference between the target values, after normalization, is obtained using the following formula:

[0177] ;

[0178] ;

[0179] In the formula, Represents an individual Crowded distance; Represents the corresponding sub-objective function Crowded distance; Represent the corresponding sub-objective function;

[0180] Step 4: Traverse all individuals in the current frontier and sum the crowding distances of each individual;

[0181] Step 5: Proceed to the next front edge and repeat steps 1 to 4 until the congestion distance of all front edges is obtained;

[0182] Step 6: Reorder each layer according to the crowding distance, and then sort the population. The larger the crowding distance, the more evenly the solutions are distributed on the frontier.

[0183] Furthermore, this embodiment of the invention introduces a crossover operator based on a normal distribution into the traditional genetic algorithm to enhance the algorithm's search capability. The crossover process is as follows:

[0184] Step 1: Assume the two parent chromosomes undergoing the crossover operation are... and Generate a ;

[0185] Step 2: If ,but

[0186] ;

[0187] ;

[0188] like ,but

[0189] ;

[0190] ;

[0191] In the formula, This represents a normally distributed random variable.

[0192] The mutation process using the Cauchy mutation operator is as follows:

[0193] .

[0194] Step S5: Every time interval The system collects real-time traffic flow and vehicle speed data for each road segment, updates the parameters of the time-varying road network, and when it detects that the real-time traffic flow of a road segment deviates from the planned traffic flow by more than a threshold or an unexpected congestion occurs on a road segment, it re-executes steps S2-S4 to generate an updated evacuation plan.

[0195] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described; only preferred embodiments of the present invention are illustrated. The descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. As long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0196] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this invention should be determined by the appended claims.

Claims

1. A method for emergency evacuation of vehicles crossing the sea based on a time-varying road network, wherein the method is activated when a sudden event affecting sea-crossing traffic occurs, characterized in that... Includes the following steps: Step S1: Perform directed rich attribute node modeling on the evacuation road network area, collect all path nodes in the evacuation road network area and emergency evacuation road sections composed of pairs of different path nodes, collect the road section distance, vehicle speed, free flow travel time and actual traffic capacity of each emergency evacuation road section, and construct a time-varying road network in which the road section travel time changes with departure time and traffic flow speed. Step S2: With the goal of minimizing the total evacuation time, construct a probabilistic distributed evacuation system based on a time-varying road network. K The shortest path model divides the initial evacuation traffic flow into multiple traffic segments, with each segment allocated according to its flow probability. K Evacuation in multiple directions along the shortest path; The probabilistic distributed system based on time-varying road networks K The constraints of the shortest path model include at least the following: Traffic balance constraints at intermediate nodes of the road network: ; Traffic balance constraints at the origin and destination points of the road network: ； Traffic flow constraints for each segment: ; Traffic distribution balance constraints: ; K Shortest path flow distribution balance constraint: ; In the above formula, , , , Representing arcs respectively ,arc , and Actual traffic flow; Represents the set of evacuation source points; Represents the set of intermediate nodes; Indicates the final set of evacuation destinations; To alleviate traffic congestion The number of vehicles; Indicates the total evacuation traffic volume; To alleviate traffic congestion In direction Traffic allocation probability; To alleviate traffic congestion In direction From the evacuation point To the evacuation endpoint of K Shortest path flow allocation probability; Step S3: Taking the minimum congestion probability of road segments in the evacuation road network as the optimization objective, the equilibrium constraint relationship between evacuation traffic flow and non-evacuation traffic flow is quantified by using road segment saturation, and an opportunity-constrained evacuation road network congestion model is constructed. Step S4: Construct a reverse lane configuration model, and use an improved genetic algorithm to process the probabilistic distributed model. K The shortest path model, the opportunity-constrained evacuation network congestion model, and the reverse lane configuration model are jointly solved to obtain the optimal evacuation path and traffic flow allocation scheme. Step S5: Every time interval The system collects real-time traffic flow and vehicle speed data for each road segment, updates the parameters of the time-varying road network, and when it detects that the real-time traffic flow of a road segment deviates from the planned traffic flow by more than a threshold or an unexpected congestion occurs on a road segment, it re-executes steps S2-S4 to generate an updated evacuation plan.

2. The emergency evacuation method for cross-sea vehicles based on a time-varying road network according to claim 1, characterized in that: The directed rich attribute node modeling of the evacuation road network area described in step S1 includes at least the following steps: splitting a single road network node into multiple entrance and exit nodes, constructing a directed rich attribute network, and using the distance between entrance and exit nodes to represent the actual driving distance at the node.

3. The emergency evacuation method for cross-sea vehicles based on a time-varying road network according to claim 1, characterized in that: The actual traffic capacity mentioned in step S1 is obtained by the following formula: ； ； ； In the formula, Indicates the capacity of a single lane; Indicates the basic traffic capacity of a single lane; This indicates the lane width correction factor; Indicates lane width; Indicates the lane reduction factor; This indicates the traffic capacity of a road group.

4. The emergency evacuation method for cross-sea vehicles based on a time-varying road network according to claim 1, characterized in that: The construction of a time-varying road network in step S1, where road segment travel time varies with departure time and traffic flow speed, includes at least: determining the relationship between road segment weights and departure times using a time dependency function, wherein the time dependency function follows a first-in-first-out (FIFO) principle, and the expression of the time dependency function is: ； In the formula, , These represent the departure times of car A and car B, respectively. , These represent vehicles A and B on the road segment. Travel time on the road.

5. The emergency evacuation method for cross-sea vehicles based on a time-varying road network according to claim 1, characterized in that: The vehicle speed mentioned in step S1 is a time-dependent speed. The expression for the relationship between the vehicle speed for traffic diversion and the normal traffic flow speed is: ； In the formula, express At that time, traffic flow In the arc The speed of travel on, of which Indicates the first Each period, Indicates the first Traffic flow of the segment, Indicates the evacuation direction. In order to evacuate the source, As the evacuation endpoint, Indicates the first The shortest path; , The attenuation coefficient; Indicates the first Normal traffic flow speed within a given time period.

6. The emergency evacuation method for cross-sea vehicles based on a time-varying road network according to claim 1, characterized in that: In step S2, the traffic flow segments are allocated according to the flow distribution probability. K When conducting multi-directional evacuation along the shortest path, the actual traffic flow is obtained using the following formula: ； ； ； In the formula, Represents arc The saturation level of the road section; Represents arc Actual traffic flow; Represents arc The road group's traffic capacity; Represents arc Non-evacuation traffic flow; Indicates the evacuation of traffic flow In the arc Traffic on the internet; For 0-1 variables, when the arc At the evacuation point To the evacuation endpoint of K The value is 1 if it is on the shortest path, and 0 otherwise. To alleviate traffic congestion The number of vehicles; To alleviate traffic congestion In direction Traffic allocation probability; To alleviate traffic congestion In direction From the evacuation point To the evacuation endpoint of K Shortest path flow allocation probability.

7. The emergency evacuation method for cross-sea vehicles based on a time-varying road network according to claim 1, characterized in that: The total evacuation time mentioned in step S2 is obtained by summing the travel times of each emergency evacuation route. The expression for the travel time of each route is: ； In the formula, Arc in time-varying road network exist The permitted travel time for a road segment in the timetable; Indicates the time interval between vehicles. ; Indicates time, , Indicates the first The start and end times of each time period; For free flow Travel time at any given moment; For arc The saturation level of the road section; Among them, free-flow travel time The calculation is divided into the following cases based on the different distances the vehicle travels and the time periods it spans: (1) When hour: ； (2) When hour; ； (3) When hour: ； In the formula, express At that time, traffic flow In the arc The speed of travel on the road; Indicates the speed at which the vehicle starts moving from the current speed. arrive driving distance, ; Indicates the number of time periods spanned.

8. The emergency evacuation method for cross-sea vehicles based on a time-varying road network according to claim 1, characterized in that: The constraints of the opportunity-constrained evacuation network congestion model described in step S3 include at least the following: ； ； ； ； In the formula, Represents the probability of an event. ; This represents the upper limit of road segment saturation. Represents arc The saturation level of the road section; Represents arc The probability of congestion on the device.

9. The emergency evacuation method for cross-sea vehicles based on a time-varying road network according to claim 1, characterized in that: The expression for the reverse lane configuration model in step S4 is: ； ； ； ； ； In the formula, This represents the total difference in evacuation time before and after optimization; Indicates evacuation time; The optimized total evacuation time; This indicates the traffic capacity of the road segment after the opposite lane has been set up; Represents arc The road group's traffic capacity; This indicates the change in road capacity after the opposite lane is installed; It is a 0-1 variable, when the road segment The value is 1 when the lane is selected as the opposite lane, and 0 otherwise. Indicates road segment The number of reverse lanes is set in the middle; This represents the path in the initial input scheme; This indicates the total amount of transportation infrastructure and labor resources; This indicates the resources consumed when each road segment is selected as the reverse lane; This indicates the road saturation level after setting up the reverse lane.

Images