Green path planning method for highway transportation of hazardous materials based on multi-source data

Abstract

The present invention discloses a method for planning a green path for the transportation of hazardous goods on highways based on multi-source data, and relates to the field of transportation planning and management. The method comprises: obtaining and processing multi-source data related to the planning process of the green path for the transportation of hazardous goods on highways; calculating the alarm rate, the consequences of accidents in the transportation of hazardous goods, and the risk mitigation factor of each highway section, and constructing a risk assessment model; calculating the emission of greenhouse gases and atmospheric pollutants during the operation of vehicles, and constructing an external cost assessment model in combination with the unit gas monetization price; constructing a green path planning model for the transportation of hazardous goods on highways with the dual objectives of minimum transportation risk and minimum external cost; and designing a hybrid algorithm combining an ant colony algorithm and a simulated annealing algorithm to solve the model and output the result. The present invention can obtain a comprehensive assessment of the risk of the transportation of hazardous goods on highways, and obtain a set of path planning schemes that balance transportation risk and external cost.

Description

Technical Field

[0001] The present invention relates to the field of transportation planning and management, and in particular to a green path planning method for dangerous goods highway transportation based on multi-source data. Background Art

[0002] As the industrialization process continues to advance, the demand and transportation volume of dangerous goods are rising year by year. In 2023, the total volume of dangerous goods transportation in my country will remain at around 1.8 billion tons, of which road transportation accounts for more than 60%. Although dangerous goods transportation accidents are considered typical low-probability high-consequence (LPHC) events, due to their huge transportation volume, if they are not handled properly, accidents such as explosions and leaks will greatly affect human safety and the ecological environment along the route. In recent years, dangerous goods transportation accidents have occurred frequently, causing companies and transportation regulatory authorities to pay special attention to the risk control of dangerous goods during transportation.

[0003] With the advent of the big data era, multi-source heterogeneous data has penetrated into every industry and business functional area today. Dangerous goods transportation process management covers a large amount of multi-source heterogeneous data, such as road data, vehicle data, resident data, traffic volume data, environmental data, etc. By integrating and analyzing these multi-source data, researchers can comprehensively and effectively assess the risks in the dangerous goods transportation process, provide scientific reference for enterprises and regulatory authorities to make transportation plan decisions, and thus prevent and reduce the occurrence of dangerous goods transportation accidents.

[0004] With the popularization of sustainable development theory, countries around the world have begun to pay attention to environmental issues, and the transportation sector has gradually become a key intersection of air quality improvement and greenhouse gas emission control. The external impacts in the transportation sector mainly include traffic congestion, traffic accidents, air pollution, noise pollution and negative changes to the global climate. The costs associated with these negative impacts are called external costs. Through quantitative analysis and model construction, scholars have deeply explored the formation mechanism and impact of these external costs, providing a scientific basis for the formulation of green and sustainable transportation policies. With the popularization of the "dual carbon" goal and the concept of green transportation, more and more studies focus on reducing external costs to promote the sustainable development of the transportation industry.

[0005] The study of vehicle routing problem began in 1959 and has experienced great development since then. However, the study of vehicle routing problem for dangerous goods transportation has gradually become a hot topic since the 21st century. The routing planning problem of dangerous goods transportation usually reflects the trade-off between the interests of carriers and government departments. Therefore, most studies focus on multi-dimensional goals such as time efficiency, transportation safety, and environmental impact, striving to achieve the optimization of dangerous goods transportation routing planning. Summary of the invention

[0006] The purpose of the present invention is to provide a green path planning method for the highway transportation of hazardous goods based on multi-source data. By integrating multi-source data, an accurate assessment of the risks of highway transportation of hazardous goods and the external costs caused by greenhouse gases and atmospheric pollutants emitted during vehicle operation can be achieved. The transportation risk and external cost are used as two objective functions to construct a green path planning model for highway transportation of hazardous goods. The model is solved by using a hybrid algorithm combining an ant colony algorithm and a simulated annealing algorithm, so as to obtain a set of safe path planning schemes with little external impact, providing a scientific and effective reference for transportation companies and regulatory agencies when formulating actual hazardous goods transportation plans.

[0007] The above-mentioned object of the present invention is achieved through the following technical solutions:

[0008] A green path planning method for dangerous goods highway transportation based on multi-source data includes:

[0009] Obtain multi-source data related to the green route planning process for hazardous goods highway transportation through data disclosure websites and online public documents. These multi-source data include highway network data, hazardous goods transportation vehicle alarm data, residential data, highway toll data, land use data, emergency rescue point data, greenhouse gas emissions data, air pollutant emissions data, and unit gas monetization data;

[0010] Use Excel and ArcGIS software to process the acquired multi-source data to form a unified highway transportation dataset of hazardous materials;

[0011] The alarm rate, consequences of dangerous goods transportation accidents and risk mitigation factors of each highway section in the study area were calculated, and then a risk assessment model for dangerous goods highway transportation was constructed by combining the alarm rate, consequences of dangerous goods transportation accidents and risk mitigation factors;

[0012] Calculate the greenhouse gas and air pollutant emissions during vehicle operation, and build an external cost assessment model for hazardous materials highway transportation by combining gas emissions and unit gas monetization prices;

[0013] Combining the risk assessment model and the external cost assessment model, with the dual goals of minimizing transportation risk and minimizing external cost, a green path planning model for highway transportation of hazardous materials is constructed;

[0014] A hybrid algorithm combining ant colony algorithm and simulated annealing algorithm is designed to solve the green path planning model for the transportation of hazardous materials on highways.

[0015] When the hybrid algorithm reaches the termination condition, the green path planning result for the transportation of hazardous materials on highways is output.

[0016] In the implementation method, multi-source data on the transportation of hazardous materials on highways is obtained through data disclosure websites and online public documents, including:

[0017] Obtain highway network data from the Internet public data platform;

[0018] Obtaining alarm data of dangerous goods transport vehicles from government-managed security monitoring systems or platforms;

[0019] Obtain settlement data and emergency rescue point data by calling the application programming interface (API) provided by the map tool;

[0020] Obtain highway toll data from highway management departments;

[0021] Obtain land use data from research results of scientific research institutions;

[0022] Obtain greenhouse gas emissions data, air pollutant emissions data and monetization data per unit of greenhouse gas and air pollutants from policy documents released by the government or research institutions.

[0023] In the implementation method, Excel and ArcGIS software are used to process the acquired multi-source data to form a unified dangerous goods highway transportation data set, including:

[0024] Import highway network data, latitude and longitude coordinates of dangerous goods transport vehicle alarms, longitude and longitude coordinates of residential areas, longitude and longitude coordinates of highway toll stations, land use data, and longitude and longitude coordinates of emergency rescue points into ArcGIS;

[0025] Greenhouse gas emissions data, air pollutant emissions data and unit greenhouse gas and air pollutant monetization data were extracted from public documents on the Internet and then imported into Excel.

[0026] In the implementation method, the alarm rate, consequences of dangerous goods transportation accidents and risk mitigation factors of each highway section in the study area are calculated, and then the risk assessment model of dangerous goods highway transportation is constructed by combining the alarm rate, consequences of dangerous goods transportation accidents and risk mitigation factors, including:

[0027] A runway-type buffer zone with a width of 1 km on both sides was created for each highway section in the study area, and then the alarm rate of hazardous goods transport vehicles on a specific highway section was calculated according to formula (1);

[0028]

[0029] Where: Indicates highway section The alarm rate; Indicates highway section The number of alarm points in the buffer zone; Indicates the total number of alarm points of hazardous goods transport vehicles in the study area.

[0030] A runway-type buffer zone is created for each highway section in the study area to simulate the potential impact area of ​​a hazardous materials transportation accident. The width of the buffer zone on both sides is determined according to the type of hazardous materials. Then, the roadside exposed population of each highway section when a hazardous materials transportation accident occurs is calculated according to formula (2).

[0031]

[0032] Where: Indicates highway section The number of people exposed on the roadside when a dangerous goods transportation accident occurs; Indicates highway section The number of settlements within the buffer zone; represents the total population in the study area; Indicates the total number of settlements in the study area obtained; represents the outdoor exposure rate of residents; It represents the safety risk mitigation effect of indoor environments such as buildings on residents.

[0033] In ArcGIS, a highway network dataset was constructed using length as the cost attribute. The NetworkAnalyst module was then used to perform path optimization analysis on all processed toll station OD pairs to solve the shortest path and derive the highway sections along the way. The traffic flow between each toll station OD pair was then allocated to each section on its corresponding shortest path. After cumulative aggregation, the traffic flow data for each highway section within a certain period of time was obtained. Finally, the exposed population on the road when a dangerous goods transportation accident occurred was calculated according to formula (3).

[0034]

[0035] Where: Indicates highway section The number of people exposed on the road when a dangerous goods transportation accident occurs; Indicates highway section Traffic flow over a certain period of time; represents the average occupancy factor of the vehicle; Indicates highway section Length; Indicates highway section The average speed of all vehicles in the area.

[0036] Call the clipping tool in ArcGIS, take the land use vector data as input elements, and use the runway-shaped buffer created when calculating the roadside exposed population as clipping elements in turn. Clip the corresponding number of new buffers, then export the attribute table of all new buffers, extract the affected land types and corresponding areas near the road section, and finally calculate the affected land area near each highway section when the dangerous goods transportation accident occurs according to formula (4);

[0037]

[0038] Where: Indicates highway section The land area affected by the dangerous goods transportation incident; Indicates Impact weight coefficient of land type; Indicates the number of dangerous goods transport accidents within the affected area. The area of ​​land.

[0039] Create several multi-ring buffers for each highway section in the study area. The multi-ring buffer is composed of multiple runway-type buffers. The number of multi-ring buffers and the width of both sides of the runway-type buffer are set according to specific requirements. The risk mitigation factor of each highway section when a dangerous goods transportation accident occurs is calculated according to formula (5).

[0040]

[0041] Where: Indicates highway section Risk mitigation factors; Sub-indicators representing risk mitigation factors; Indicates sub-indicators The weight value of Indicates sub-indicators The assigned value.

[0042] The alarm rate of each highway section in the study area, the number of people exposed on the roadside and on the road when a dangerous goods transport accident occurs, and the risk mitigation factor of the section are combined to calculate the potential population risk of the highway section when a dangerous goods transport accident occurs according to formula (6);

[0043]

[0044] Where: Indicates that dangerous goods transportation accidents occur on highway sections potential population risks if such incidents occur; Indicates highway section The alarm rate; Indicates highway section The number of settlements within the buffer zone; represents the total population in the study area; Indicates the total number of settlements in the study area obtained; represents the outdoor exposure rate of residents; It indicates the safety risk mitigation effect of indoor environments such as buildings on residents; Indicates highway section Traffic flow over a certain period of time; represents the average occupancy factor of the vehicle; Indicates highway section Length; Indicates highway section The average speed of all vehicles in the area; Indicates highway section risk mitigation factors.

[0045] The potential environmental risk of each highway section in the study area when a dangerous goods transportation accident occurs is calculated according to formula (7) by combining the alarm rate of each highway section in the study area, the land area affected when a dangerous goods transportation accident occurs, and the risk mitigation factor of the section.

[0046]

[0047] Where: Indicates that dangerous goods transportation accidents occur on highway sections Potential environmental risks caused by the above; Indicates highway section The alarm rate; Indicates Impact weight coefficient of land type; Indicates the number of dangerous goods transport accidents within the affected area. The area of ​​land of this type; Indicates highway section risk mitigation factors.

[0048] The total risk value of each highway section when a dangerous goods transportation accident occurs is calculated according to formula (8).

[0049]

[0050] Where: Indicates that dangerous goods transportation accidents occurred on the road section The total risk value occurring on and represent the weight coefficients of population risk and environmental risk respectively; Indicates that dangerous goods transportation accidents occur on highway sections potential population risks if such incidents occur; Indicates that dangerous goods transportation accidents occur on highway sections Potential environmental risks caused by the above; and They respectively represent the maximum value of potential population risk among all highway sections and the maximum value of potential environmental risk among all highway sections.

[0051] In the implementation method, greenhouse gas emissions and air pollutant emissions during vehicle operation are calculated, and the external cost assessment model for hazardous materials highway transportation is constructed by combining gas emissions and unit gas monetization price, including:

[0052] According to formula (9), calculate the carbon dioxide ( ) emissions;

[0053]

[0054] Where: express emissions; Represents the average lower calorific value of fossil fuels; Indicates the mileage of the vehicle; Indicates the vehicle's fuel consumption per 100 kilometers; represents the density of fossil fuels; Indicates the carbon content per unit calorific value of fossil fuels; represents the carbon oxidation rate of fossil fuels; It represents the ratio of the molecular weight of carbon dioxide to carbon; express of global warming potential.

[0055] According to formula (10), the methane ( ) emissions;

[0056]

[0057] Where: express Emissions, in equivalent measure; express Emission factors; express of global warming potential.

[0058] According to formula (11), nitrous oxide ( ) emissions;

[0059]

[0060] Where: express Emissions, in equivalent measure; express Emission factors; express of global warming potential.

[0061] Calculate the amount of atmospheric pollutants emitted in the form of exhaust gas during vehicle operation according to formula (12);

[0062]

[0063] Where: Indicates The tail gas emissions of various air pollutants, including , represents the set of five air pollutants, namely ; Indicates the mileage of the vehicle; Indicates The comprehensive baseline emission coefficient of various air pollutants is related to the vehicle model and fuel type; represents the temperature correction factor; represents the humidity correction factor; Indicates the correction factor of the average speed of the vehicle; Indicates the load factor correction factor of the vehicle.

[0064] The amount of atmospheric pollutants emitted in the form of evaporation during vehicle operation is calculated using formula (13).

[0065]

[0066] Where: express Evaporative emissions; Indicates the evaporative emission coefficient during vehicle driving; Indicates the mileage of the vehicle; Indicates the average speed of the vehicle.

[0067] Combined with the monetary price of unit greenhouse gas, the external cost caused by greenhouse gas emissions during vehicle operation is calculated according to formula (14);

[0068]

[0069] Where: represents the external costs of greenhouse gas emissions; express The unit monetization price of Indicates greenhouse gas emissions; represents the set of three greenhouse gases, namely .

[0070] Combined with the monetary price of air pollutants per unit, the external costs caused by the emission of air pollutants during vehicle operation are calculated according to formula (15);

[0071]

[0072] Where: Represents the external costs caused by the emission of air pollutants; Indicates Unit monetary price of each air pollutant; Indicates The exhaust emissions of various air pollutants, express Evaporative emissions; Represents a collection of five air pollutants.

[0073] Combining the external costs caused by the emission of greenhouse gases and air pollutants, the total external costs caused by gas emissions during vehicle operation are calculated according to formula (16).

[0074]

[0075] Where: It represents the total external cost caused by gas emissions during vehicle operation; represents the external costs of greenhouse gas emissions; Represents the external costs caused by the emission of air pollutants.

[0076] In the implementation method, the risk assessment model and the external cost assessment model are combined to build a green path planning model for the transportation of hazardous materials on highways with the dual goals of minimizing transportation risk and minimizing external costs, including:

[0077] Objective Function Minimize the risks in the transportation of dangerous goods, as shown in formula (17);

[0078]

[0079] Objective Function Minimize the total external cost in the transportation of dangerous goods, as shown in formula (18);

[0080]

[0081]

[0082] Constraints (19) and (20) ensure that vehicles transporting dangerous goods must depart from the specified starting point and arrive at the specified destination;

[0083]

[0084]

[0085] Constraints (21) and (22) ensure that each node is visited at most once;

[0086]

[0087]

[0088] Decision variables , , which means that if a dangerous goods transport vehicle passes through a highway section ,but ;otherwise, .

[0089] In the implementation method, a hybrid algorithm combining an ant colony algorithm and a simulated annealing algorithm is designed to solve the above-mentioned green path planning model for hazardous goods highway transportation, including:

[0090] S1: Initialization:

[0091] S1.1: Initialize algorithm parameters:

[0092] Ant Colony Algorithm Parameters: Number of Ants , maximum number of iterations , pheromone release intensity , pheromone importance factor , heuristic function importance factor 、Maximum number of ant searches and pheromone volatility coefficient ;

[0093] Simulated annealing algorithm parameters: initial temperature , termination temperature , cooling rate , the number of iterations at each temperature .

[0094] S1.2: Initialize the pheromone matrix of the ant colony algorithm:

[0095] Construct a pheromone matrix with the dimension of the number of nodes and initialize the pheromone concentration between all node pairs (𝑖,𝑗) is 0, and then determine whether there is a connection relationship between any two nodes. If node 𝑖 and node 𝑗 are connected, the corresponding pheromone concentration is Set to .

[0096] S1.3: Initialize the transport start point and transport destination .

[0097] S2: Redistribution of pheromones:

[0098] S2.1: Dijkstra algorithm solves the shortest path:

[0099] Based on the connection relationship of the nodes, count the path length between any two connected nodes , and use it as the weight, then apply the Dijkstra algorithm to solve the shortest path for the vehicle to complete the transportation task. For each node pair (i, j) on this shortest path, increase the pheromone concentration on it by Q, which is the updated pheromone concentration. .

[0100] S2.1: Dijkstra algorithm finds the path with the least risk:

[0101] The risk value between two connected nodes obtained by the risk assessment process As the new weight, the Dijkstra algorithm is used again to solve the path with the lowest risk when the vehicle completes the transportation task. For each node pair (i, j) on this path with the lowest risk, the pheromone concentration on it is also increased, and the increase is still Q, that is, the pheromone concentration after the update again .

[0102] S3: Generate initial solution vector:

[0103] Ants start from a designated starting point Start searching and calculate the probability of selecting the next node according to the state transition formula (23) until the ant reaches the specified end point. However, when ants are selecting nodes, a "deadlock" phenomenon sometimes occurs, that is, the ants cannot continue in the search process because there are no adjacent nodes to choose from, causing the algorithm to stagnate and fail. For this reason, the present invention proposes a hybrid strategy with a search step limit to solve the "deadlock" phenomenon in the ant colony algorithm. The hybrid strategy is as follows: before the ants iteratively find the path, the maximum search step of each ant is set to the total number of nodes. Each time the ant finds the path, its search step is increased by 1, and it is checked whether it is stuck in a stagnant state: if it is not stuck, the process ends and the iterative path finding process of the next ant begins; if it is stuck, a marking strategy is adopted in the early stage of the algorithm, that is, the ant is marked as invalid, and all nodes are added to the path search results of the ant in sequence, which means that the length of the solution vector of the ant is the total number of nodes; in the later stage of the algorithm, a recall strategy is adopted to clear all the nodes previously searched by the ant, recall it to the starting node, and allow it to restart the search process to find a feasible path until all ants have completed the path finding and generated a set of initial solution vectors.

[0104]

[0105] In the formula, Indicates Only ants from the node at the current iteration Go to Node The probability of Indicates the current iteration number; Representation Node and nodes The concentration of pheromones on the inter-connecting paths; Representation Node and nodes The length of the connecting path; is the heuristic function, , indicating that the ant slave node Transfer to Node the degree of expectation; Indicates The set of nodes to be selected by the ants; is the pheromone importance factor; is the importance factor of the heuristic function.

[0106] S4: Calculate the objective function value:

[0107] The solution vectors of the marked ants are identified based on the length of the solution vectors. The two objective function values ​​of these solution vectors are assigned to Inf (infinity) to distinguish them from the solution vectors of the unmarked ants. For other unmarked ants, the objective function values ​​of their solution vectors are calculated based on the objective function and Perform calculations.

[0108] S5: Initial de-perturbation:

[0109] S5.1: Identify the solution vector of the marked ant and do not perturb it.

[0110] S5.2: At the current temperature, perform multiple perturbations on the solution vectors of other unmarked ants. The specific perturbation rules are as follows: randomly select any node in the solution vector except the starting point and the end point as the perturbation node, and then determine whether there are other nodes between the immediately preceding and succeeding nodes of the perturbation node: if not, reselect the perturbation node for perturbation; if such a node exists, further determine whether there are any nodes among these nodes that have never appeared in the current solution vector; if such a node is found, randomly select one of them to exchange with the perturbation node to generate a new solution vector; otherwise, end the perturbation.

[0111] S6: Trade-offs of the new solution:

[0112] S6.1: Normalize and sum the objective function value of the current solution according to formula (24).

[0113] S6.2: Calculate the two objective function values ​​of the new solution generated after the disturbance, normalize the objective function values ​​of the new solution according to formula (25) and add them up.

[0114] S6.3: Determine whether the new solution is accepted according to the Metropolis criterion of formula (26): if the sum of the normalized objective function values ​​of the new solution ( ) is less than the sum of the normalized objective function values ​​of the current solution ( ), then directly use the new solution to replace the current solution; if Greater than or equal to , then calculate the acceptance probability of the new solution, which is , and then generate a random number rand uniformly distributed in the interval (0,1). If the calculated acceptance probability is greater than the random number rand, the new solution is accepted to replace the current solution; otherwise, the new solution is abandoned and replaced with the current solution.

[0115]

[0116]

[0117]

[0118] In the formula, and Represents the two objective function values ​​of the current solution; and Represent the two objective function values ​​of the new solution; and Represents the sum of the two objective function values ​​after normalization of the current solution and the new solution respectively; Indicates the current temperature; Represents the probability that the new solution replaces the current solution.

[0119] S7: Determine whether the disturbance has been completed for all unmarked ants: if yes, proceed to the next cooling operation; if no, disturb the next unmarked ant.

[0120] S8: Cooling down:

[0121] The temperature is gradually lowered according to the set cooling strategy, and at the same time, the current temperature is compared to see if it has reached the set end temperature. : If the set minimum temperature is reached, end the simulated annealing algorithm; otherwise, return and continue to execute S3;

[0122] S9: Non-dominated solution screening:

[0123] A set of initial solution vectors generated by the ant colony algorithm are perturbed and selected by the simulated annealing algorithm to obtain a set of new solution vectors. This set of new solution vectors is screened for non-dominated solutions and the Pareto optimal solution set is retained.

[0124] S10: Pheromone Update:

[0125] According to the path selection in the Pareto optimal solution set, the pheromone matrix of the ant colony algorithm is updated according to formulas (27) and (28);

[0126]

[0127]

[0128] In the formula, Indicates Only ants are on the node and nodes The concentration of pheromones released on the connecting path between them; is the number of ants in the Pareto optimal solution set; Representation Node and nodes The sum of the pheromone concentrations released by all ants on the connecting path; Indicates the current iteration number of the ant colony algorithm; Indicates The total length of the route taken by the ants.

[0129] S11: Ant Colony Iteration:

[0130] S3-S10 constitutes a complete iteration process of the ant colony algorithm. After each iteration, the number of iterations of the ant colony algorithm is increased by 1, that is, the current number of iterations. , check whether the current number of iterations has reached the preset maximum number of iterations of the ant colony algorithm: if the maximum number of iterations is reached, exit the program and output the result; otherwise, continue to the next iteration.

[0131] In the implementation, when the hybrid algorithm reaches the termination condition, the solution of the algorithm is output, including:

[0132] Repeat S3-S10 until the maximum number of iterations of the ant colony algorithm is reached, exit the program and output the green path planning results for hazardous goods highway transportation, which can provide a reference for scientific and effective path planning solutions for transportation companies and regulatory agencies to balance transportation risks and external costs.

[0133] Compared with the prior art, the beneficial results of the present invention are:

[0134] The present invention proposes for the first time a method for measuring the possibility of dangerous goods transportation accidents based on vehicle alarm data, which provides theoretical support and alternative evaluation methods for quantifying the accident rate of dangerous goods transportation, and solves the problem that the data source of the accident rate of dangerous goods transportation is single and has poor timeliness.

[0135] The present invention innovatively integrates settlement data and toll station data to more accurately calculate the population exposure when a hazardous goods transportation accident occurs, and innovatively uses land use data to calculate the type and area of ​​the environment affected when a hazardous goods transportation accident occurs, enriching the methodological system for hazardous goods transportation risk assessment.

[0136] In the problem of dangerous goods transportation route planning, the present invention takes into account the environmental external costs caused by greenhouse gases and atmospheric pollutants emitted during vehicle transportation, and quantifies these costs based on standard data provided by national technical guidelines, greatly deepening the research on the environmental impact of vehicle emissions in the field of dangerous goods route planning.

[0137] The present invention studies the multi-objective vehicle path planning problem within the framework of dangerous goods transportation, and designs a hybrid algorithm that combines the ant colony algorithm and the simulated annealing algorithm to solve this type of problem. The distributed computing method of the ant colony algorithm can significantly improve the computing power and operating efficiency of the algorithm. At the same time, the simulated annealing algorithm gives the search process a time-varying and eventually zero sudden jump probability, which jumps out of the local optimum and eventually tends to the global optimum, providing a solid algorithmic foundation for scholars' research in the field of dangerous goods transportation path optimization.

[0138] The present invention proposes a hybrid strategy with a search step limit to solve the "deadlock" phenomenon in the ant colony algorithm: by setting a maximum search step limit to avoid the situation where ants are stuck, and at the same time using a marking strategy in the early stage of the algorithm to make the algorithm converge quickly and improve the algorithm's operating efficiency; and using a recall strategy in the later stage of the algorithm to ensure the diversity of solutions and avoid the algorithm from falling into a local optimum. BRIEF DESCRIPTION OF THE DRAWINGS

[0139] The present invention will be further described below in conjunction with the accompanying drawings:

[0140] Figure 1 The highway network topology diagram of Jilin Province described in the present invention;

[0141] Figure 2 It is a flow chart of the method for planning a green path for dangerous goods highway transportation based on multi-source data according to the present invention;

[0142] Figure 3 It is a schematic diagram of a runway-type buffer zone with a width of 1 km on both sides created according to the present invention;

[0143] Figure 4 It is a schematic diagram of the distribution of settlements in Jilin Province according to the present invention;

[0144] Figure 5 is a schematic diagram of the shortest path solution result of the present invention;

[0145] Figure 6 It is a schematic diagram of a runway-type buffer zone with a width of 0.8 km on both sides created according to the present invention;

[0146] Figure 7 It is a schematic diagram of the distribution of emergency rescue points in Jilin Province according to the present invention;

[0147] Figure 8 It is a risk intensity schematic diagram of 38 highway sections described in the present invention;

[0148] Fig. 9 is a flow chart of the hybrid algorithm described in the present invention;

[0149] Fig.10 It is a schematic diagram of the "deadlock" phenomenon described in the present invention;

[0150] Fig.11 It is a schematic diagram of the implementation of the marking strategy of the present invention;

[0151] Fig.12 is a schematic diagram of the disturbance implementation of the present invention;

[0152] Fig.13 It is a schematic diagram of a Pareto solution for green path planning of hazardous goods highway transportation according to the present invention. DETAILED DESCRIPTION

[0153] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below.

[0154] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present invention.

[0155] In this embodiment, a dangerous goods transportation business is involved. The core task of this business is to transport anhydrous ethanol from Taobei District, Baicheng City, Jilin Province to Tumen City, Yanbian Korean Autonomous Prefecture, Jilin Province in August 2023, while ensuring transportation safety and reducing external impacts during the transportation process.

[0156] The transportation task relies on the expressway network throughout the whole process. Some nodes and sections in the expressway network are selected for research. The constructed expressway network topology map of Jilin Province can be found in Figure 1 There are 26 nodes and 38 road sections in total, with the transport starting point being node 1 and the transport end point being node 26.

[0157] The transportation task is undertaken by a special tank truck consisting of a dangerous goods transport semi-trailer tractor and a tank trailer. The model of the dangerous goods transport semi-trailer tractor is CA4250P26K15T1E5A81.

[0158] See also Figure 2 The present invention proposes a green path planning method for dangerous goods highway transportation based on multi-source data, which includes the following steps:

[0159] Step 1: Obtain multi-source data related to the green route planning process for hazardous goods highway transportation through data disclosure websites and online public documents. These multi-source data include highway network data, hazardous goods transportation vehicle alarm data, residential data, highway toll data, land use data, emergency rescue point data, greenhouse gas emissions data, atmospheric pollutant emissions data and unit gas monetization data.

[0160] Step 2: Use Excel and ArcGIS software to process the acquired multi-source data to form a unified dangerous goods highway transportation data set.

[0161] Step 3: Based on the hazardous goods highway transportation dataset, the alarm rate of each highway section in the study area is calculated, the number of potential exposed population when a hazardous goods transportation accident occurs on each highway section in the study area is calculated, including the roadside exposed population and the road exposed population, the potential environmental impact area when a hazardous goods transportation accident occurs on each highway section in the study area, and the risk mitigation factor of each highway section in the study area is calculated. Then, a risk assessment model for hazardous goods highway transportation is constructed by combining the alarm rate, the consequences of hazardous goods transportation accidents and the risk mitigation factor.

[0162] Step 4: Based on the hazardous goods highway transportation dataset, calculate the greenhouse gas and air pollutant emissions during vehicle operation, and calculate the external costs caused by the emission of greenhouse gases and air pollutants during vehicle operation in combination with the monetized prices of unit greenhouse gases and air pollutants. Then, combine the external costs caused by the emission of greenhouse gases and air pollutants to construct an external cost assessment model for hazardous goods highway transportation.

[0163] Step 5: Combining the risk assessment model and the external cost assessment model, with the dual goals of minimizing transportation risk and minimizing external cost, a green path planning model for hazardous materials highway transportation is constructed.

[0164] Step 6: Solve the green path planning model for the above-mentioned hazardous materials highway transportation by designing a hybrid algorithm that combines the ant colony algorithm and the simulated annealing algorithm.

[0165] Step 7: When the hybrid algorithm reaches the termination condition, the green path planning result for the transportation of hazardous materials on highways is output.

[0166] Step 1 specifically includes the following steps:

[0167] 1.1. Access the Geofabrik download server and download the road data within Jilin Province in May 2023, including road network data such as highways and bicycle lanes. The highway network data mainly includes information such as road name, road type, and road number.

[0168] 1.2. Visit the "Jilin Province Road Transport Vehicle Dynamic Information Public Service Platform" and download the alarm data of dangerous goods transport vehicles in Jilin Province in August 2023, a total of 387,130. The alarm data of dangerous goods transport vehicles mainly includes the vehicle's license plate number, alarm type, alarm start and end time, alarm duration and alarm address information, see Table 1.

[0169] Table 1 Alarm data point information

[0170]

[0171] 1.3. Call Baidu Map API to obtain residential data. Two input parameters need to be specified before calling the API: the first parameter is the search scope, which is Jilin Province in this embodiment; the second parameter is the keyword, which is rural areas, towns, streets and other residential areas in this embodiment. The residential data mainly includes the name of the residential area and the longitude and latitude coordinate information.

[0172] 1.4. Call Baidu Map API to obtain emergency rescue point data. Before calling the API, two input parameters must also be specified: the first parameter is the search scope, which is Jilin Province in this embodiment; the second parameter is the keyword, which is hospitals, fire departments, traffic police, public security and other emergency rescue points in this embodiment. The emergency rescue point data mainly includes department names and department latitude and longitude coordinate information.

[0173] 1.5. Cooperate with the "Jilin Provincial Department of Transportation" to obtain the name, number and latitude and longitude (WGS84 coordinate system) information of the highway toll stations in the study area, and obtain the highway toll data in Jilin Province in August 2023, totaling 11,276,015 records. The highway toll data mainly includes exit and entry numbers, exit and entry times, vehicle type, number of axles, charging type and license plate number information, see Table 2.

[0174] Table 2 Vehicle charging data information

[0175]

[0176] 1.6. Visit the Resource and Environmental Science and Data Center of the Chinese Academy of Sciences to download the land use raster data with a spatial resolution of 1 km in 2020 from the China Multi-period Land Use Remote Sensing Monitoring Dataset (CNLUCC). The data adopts a two-level classification system: the first level is mainly divided into 6 categories based on land resources and their utilization attributes, including cultivated land, forest land, grassland, water area, construction land and unused land; the second level is mainly divided into 23 types based on the natural attributes of land resources. The land use data mainly includes land use types and land area information of each type.

[0177] 1.7. Visit the National Development and Reform Commission of the People's Republic of China and download the "Guidelines for Calculating and Reporting Greenhouse Gas Emissions from Land Transportation Enterprises". The guidelines provide default values ​​for fuel characteristic parameters, global warming potential of greenhouse gases and emission factors of greenhouse gases, see Tables 3 to 5.

[0178] Table 3 Default values ​​of diesel fuel characteristic parameters

[0179]

[0180] Table 4 Global warming potential of greenhouse gases

[0181]

[0182] Table 5 Vehicle and Emission Factor

[0183]

[0184] 1.8. Visit the Ministry of Environment and Ecology of the People's Republic of China and download the Technical Guidelines for the Compilation of Road Motor Vehicle Emission Inventories. The guidelines provide the comprehensive benchmark emission coefficient of motor vehicles, the temperature correction factor of the comprehensive benchmark emission coefficient, the humidity correction factor of the comprehensive benchmark emission coefficient, the average speed correction factor of the comprehensive benchmark emission coefficient and the load factor correction factor of the comprehensive benchmark emission coefficient, see Tables 6 to 10. The temperature and humidity correction factors are determined according to the meteorological conditions in August 2023, see Table 11.

[0185] Table 6 Comprehensive benchmark emission coefficients for diesel vehicles

[0186]

[0187] Table 7 Diesel vehicle temperature correction factor

[0188]

[0189] Table 8 Humidity correction factors for diesel vehicles

[0190]

[0191] Table 9 Diesel vehicle average speed correction factor

[0192]

[0193] Table 10 Diesel vehicle load factor correction factors

[0194]

[0195] Table 11 Weather conditions

[0196]

[0197] 1.9. Visit the Energy Foundation and download the Technical Guidelines for Coordinated Control of Air Pollutants and Greenhouse Gases from Urban Traffic Version 1-0. The guidelines provide the monetary prices of the impact of unit air pollutants and carbon dioxide emissions, see Table 12.

[0198] Table 12 Monetized prices of gas emissions in Jilin Province

[0199]

[0200] Step 2 specifically includes the following steps:

[0201] 2.1. Highway network data processing includes the following steps:

[0202] Import the highway network data into ArcGIS and save it in the file geographic information library in the form of vector data.

[0203] 2.2. The alarm data processing of dangerous goods transport vehicles includes the following steps:

[0204] 2.2.1. Import the alarm data of dangerous goods transport vehicles into Excel, remove blank data and alarm data with an alarm duration of 0, and finally retain 383,957 valid data;

[0205] 2.2.2. Since the information stored in the alarm address bar is in the form of a combination of geographic names and administrative divisions, it is not suitable for direct processing and analysis. Therefore, the address information of the cleaned alarm data is converted into longitude and latitude coordinates in the WGS84 global coordinate system;

[0206] 2.2.3. Import the converted latitude and longitude coordinates of the alarm point into ArcGIS and save them in the file geographic information database in the form of vector data.

[0207] 2.3. Settlement data processing includes the following steps:

[0208] 2.3.1. The settlement data obtained by keywords may not meet the needs and need further screening. Therefore, the settlement data is first imported into Excel, and then the data that does not belong to the actual settlement such as "XX Town Guoshun Auto Repair" is eliminated, and finally 46,856 settlements are retained;

[0209] 2.3.2. Import the cleaned longitude and latitude coordinates (WGS84 coordinate system) of the settlement into ArcGIS and save them in the file geographic information database in the form of vector data.

[0210] 2.4. Highway toll data processing includes the following steps:

[0211] 2.4.1. Import the names, numbers and latitude and longitude data of the entrances and exits of expressway toll stations in Jilin Province into ArcGIS and save them in the file geographic information database in the form of vector data;

[0212] 2.4.2. Import the expressway toll data into Excel and remove the toll data that are not in Jilin Province in the exit and entry numbers;

[0213] 2.4.3. According to the vehicle type conversion coefficient specified in the Highway Engineering Technical Standards, convert each toll data into standard vehicle equivalent (PCU) data;

[0214] 2.4.4. The PCUs between the OD pairs of the same toll station are accumulated and summarized, and finally the accumulated PCU data between the OD pairs of each toll station is obtained, which is used as the traffic flow between the OD pairs of each toll station.

[0215] 2.5. Land use data processing includes the following steps:

[0216] 2.5.1. Input the land use data into ArcGIS for reclassification, retaining 6 primary types;

[0217] 2.5.2. Call the conversion tool in ArcGIS to convert the land use raster data into land use surface features and save them in the file geographic information library in the form of vector data.

[0218] 2.6. Emergency rescue point data processing includes the following steps:

[0219] 2.6.1. Import the emergency rescue point data into Excel for further screening, remove the data such as "xx Stomatological Hospital" which does not have the emergency rescue capability for dangerous goods transportation accidents, and finally retain 838 emergency rescue points, including 322 traffic police departments, 123 fire departments, and 393 hospitals;

[0220] 2.6.2. Import the cleaned longitude and latitude coordinates (WGS84 coordinate system) of the emergency rescue point into ArcGIS and save them in the file geographic information database in the form of vector data.

[0221] 2.7. Greenhouse gas emission data processing includes the following steps:

[0222] Extract data related to measuring greenhouse gas emissions during vehicle operation from the "Guidelines for Calculating and Reporting Greenhouse Gas Emissions for Land Transportation Enterprises" and save them in Excel in a table format;

[0223] 2.8. Air pollutant emission data processing includes the following steps:

[0224] Extract data related to measuring air pollutant emissions during vehicle operation from the Technical Guidelines for Compilation of Road Motor Vehicle Emission Inventories and save them in Excel in the form of a table;

[0225] 2.9. Unit gas monetization data processing includes the following steps:

[0226] The monetary prices of unit greenhouse gases and air pollutants were extracted from the Technical Guidelines for the Coordinated Control of Air Pollutants and Greenhouse Gases in Urban Traffic Version 1-0 and saved in Excel in the form of a table.

[0227] Step 3 specifically includes the following steps:

[0228] 3.1. Calculate the alarm rate of each highway section studied in Jilin Province, including the following steps:

[0229] 3.1.1、Acquire the highway network data and alarm point vector data in the file geographic information database. Due to the data conversion error and vector data offset error, not all alarm points are accurately located on the highway sections. In order to eliminate the influence of double errors, a runway-type buffer zone with a width of 1 km on both sides is created for each of the 38 highway sections studied in Jilin Province. The schematic diagram of the creation result of one of the sections is shown in Figure 3 ,The number of alarm points in the buffer can be approximately considered as the number of alarm points on the road section;

[0230] 3.1.2. Calculate the alarm rate of dangerous goods transport vehicles on a specific highway section according to formula (1). The section alarm rate refers to the ratio of the number of alarm points in the buffer zone of the section to the total number of alarm points in the study area. The higher the section alarm rate, the greater the possibility of accidents on the section.

[0231]

[0232] Where: Indicates highway section The alarm rate; Indicates highway section The number of alarm points in the buffer zone; Indicates the total number of alarm points of dangerous goods transport vehicles in Jilin Province.

[0233] 3.2. Calculate the roadside exposed population of each highway section when a dangerous goods transportation accident occurs, including the following steps:

[0234] 3.2.1、Retrieve the highway network data and residential vector data from the file geographic information database. For the residential distribution map, see Figure 4 , and then create a runway-type buffer zone with a width of 0.8 km on both sides for each of the 38 highway sections (the potential impact radius of anhydrous ethanol is 0.8 km in each direction) to simulate the potential impact area of ​​dangerous goods transportation accidents;

[0235] 3.2.2. Calculate the roadside exposed population of each highway section when a dangerous goods transport accident occurs according to formula (2).

[0236]

[0237] Where: Indicates highway section The number of people exposed on the roadside when a dangerous goods transportation accident occurs; Indicates highway section The number of settlements within the buffer zone; represents the total population of Jilin Province; Indicates the total number of settlements in Jilin Province; represents the outdoor exposure rate of residents; It represents the safety risk mitigation effect of indoor environments such as buildings on residents.

[0238] 3.3. Calculate the exposed population on each highway section when a dangerous goods transportation accident occurs, including the following steps:

[0239] 3.3.1. Retrieve the highway network data and toll station vector data from the file geographic information database, and construct a highway network dataset using length as the cost attribute;

[0240] 3.3.2. Use the Network Analyst module to perform path optimization analysis on all processed toll station OD pairs, and then solve the shortest path and export the highway sections along the way. Take one of the toll data as an example. The entrance toll station number is 535 and the exit toll station number is 641. Through path optimization analysis, the shortest path from the entrance toll station to the exit toll station is determined, and then the three highway sections along the shortest path are exported. See the schematic diagram of this solution result. Figure 5 ;

[0241] 3.3.3. Allocate the traffic flow between each toll station OD pair to each road section on its corresponding shortest path, and obtain the monthly traffic flow data of each highway section in August after cumulative aggregation;

[0242] 3.3.4. Calculate the exposed population on the road when a dangerous goods transport accident occurs according to formula (3).

[0243]

[0244] Where: Indicates highway section The number of people exposed on the road when a dangerous goods transportation accident occurs; Indicates highway section Monthly traffic volume in August 2023; represents the average occupancy factor of the vehicle; Indicates highway section Length; Indicates highway section The average speed of all vehicles in the area.

[0245] 3.4. Calculate the affected land area near each highway section when a dangerous goods transportation accident occurs, including the following steps:

[0246] 3.4.1. Retrieve the land use vector data from the file geographic information database, call the clipping tool in ArcGIS, use the land use vector data as the input element, and use the 38 runway-shaped buffers created when calculating the roadside exposed population as clipping elements in turn to clip out 38 new buffers. The clipping result diagram of one road section is shown in Figure 6 ;

[0247] 3.4.2. Export the attribute table of all new buffer zones and extract the affected land types and their corresponding areas near the road sections;

[0248] 3.4.3. Calculate the affected land area near each highway section when a dangerous goods transportation accident occurs according to formula (4).

[0249]

[0250] Where: Indicates highway section The land area affected by the dangerous goods transportation incident; Indicates Impact weight coefficient of land type; Indicates the number of dangerous goods transport accidents within the affected area. The area of ​​land.

[0251] 3.5. Calculate the risk mitigation factor for each highway section when a dangerous goods transportation accident occurs, including the following steps:

[0252] 3.5.1. Retrieve the highway network data and emergency rescue point vector data from the file geographic information database. For the distribution map of emergency rescue points, see Figure 7 . Create two multi-ring buffers for each of the 38 highway sections (the same multi-ring buffer is used to count the distribution of fire stations and hospitals). The ring buffer is composed of multiple runway-type buffers, and the width of each runway-type buffer is set according to the "Range" column in Table 13;

[0253] Table 13 Calculation indexes of factors affecting dangerous goods highway transportation accidents

[0254]

[0255] 3.5.2. Calculate the risk mitigation factor for each highway section when a dangerous goods transport accident occurs according to formula (5).

[0256]

[0257] Where: Indicates highway section Risk mitigation factors; Sub-indicators representing risk mitigation factors; Indicates sub-indicators The weight value of Indicates sub-indicators The assigned value.

[0258] 3.6. Combine the alarm rate of each highway section in the study area, the number of people exposed on the roadside and on the road when a dangerous goods transport accident occurs, and the risk mitigation factor of the section, and calculate the potential population risk of the highway section when a dangerous goods transport accident occurs according to formula (6).

[0259]

[0260] Where: Indicates that dangerous goods transportation accidents occur on highway sections potential population risks if such incidents occur; Indicates highway section The alarm rate; Indicates highway section According to the seventh national census data, the total permanent population in Jilin Province is set as ; Indicates the total number of settlements in Jilin Province; Set the outdoor exposure rate , indoor environmental risk reduction factor ; Indicates highway section Monthly traffic flow in August 2023; In view of the variety of vehicle types and large speed differences on highways, in order to simplify the calculation process, a vehicle type coefficient is set , design speed ; Indicates highway section risk mitigation factors.

[0261] 3.7. The alarm rate of each highway section in the study area, the land area affected when a dangerous goods transport accident occurs, and the risk mitigation factor of the section are combined to calculate the potential environmental risk of the highway section when a dangerous goods transport accident occurs according to formula (7).

[0262]

[0263] Where: Indicates that dangerous goods transportation accidents occur on highway sections Potential environmental risks caused by the above; Indicates highway section The alarm rate; Indicates The impact weight coefficient of land type is set to 1.5, 1.3, 1.2, 3, 2, and 1 for cultivated land, forest land, grassland, water area, construction land, and unused land, respectively; Indicates the number of dangerous goods transport accidents within the affected area. The area of ​​land of this type; Indicates highway section risk mitigation factors.

[0264] 3.8. Calculate the total risk value of each highway section when a dangerous goods transportation accident occurs according to formula (8), then sort the sections according to the risk value and input them into the ArcGIS platform for visualization. The visualization results can be found in Figure 8 , where the depth of the road section’s color reflects the level of risk of the road section.

[0265]

[0266] Where: Indicates that dangerous goods transportation accidents occurred on the road section The total risk value that occurs on the transport of dangerous goods; Based on the current research data and risk analysis, it is believed that population risk and environmental risk are equally important when assessing the risk of dangerous goods transportation. Therefore, the weight coefficients of population risk and environmental risk are set equal, that is, ; Indicates that dangerous goods transportation accidents occur on highway sections potential population risks if such incidents occur; Indicates that dangerous goods transportation accidents occur on highway sections Potential environmental risks caused by the above; and They respectively represent the maximum value of potential population risk among all highway sections and the maximum value of potential environmental risk among all highway sections.

[0267] Step 4 specifically includes the following steps:

[0268] 4.1. Calculating greenhouse gas emissions during vehicle operation includes the following steps:

[0269] 4.1.1. Calculate the carbon dioxide (CO2) during vehicle operation according to formula (9) ) emissions;

[0270]

[0271] Where: express emissions; Indicates the average low calorific value of diesel; Indicates the mileage of the vehicle; Indicates the vehicle's fuel consumption per 100 kilometers; Indicates the density of diesel; Indicates the carbon content per unit calorific value of diesel; Indicates the carbon oxidation rate of diesel; It represents the ratio of the molecular weight of carbon dioxide to carbon; express of global warming potential.

[0272] 4.1.2. Calculate the methane during vehicle operation according to formula (10) ) emissions;

[0273]

[0274] Where: express Emissions, in equivalent measure; express Emission factors; express of global warming potential.

[0275] 4.1.3. Calculate the nitrous oxide ( ) emissions;

[0276]

[0277] Where: express Emissions, in equivalent measure; Respectively Emission factors; express of global warming potential.

[0278] 4.2. Calculating the emission of air pollutants during vehicle operation includes the following steps:

[0279] 4.2.1. Calculate the amount of air pollutants emitted in the form of exhaust gas during vehicle operation according to formula (12);

[0280]

[0281] Where: Indicates The tail gas emissions of various air pollutants, including , represents the set of five air pollutants, namely ; Indicates the mileage of the vehicle; Indicates Comprehensive benchmark emission coefficients for various air pollutants; represents the temperature correction factor; represents the humidity correction factor; Indicates the correction factor of the average speed of the vehicle; Indicates the load factor correction factor of the vehicle.

[0282] 4.2.2. Calculate the amount of atmospheric pollutants emitted in the form of evaporation during vehicle operation according to formula (13).

[0283]

[0284] Where: express Evaporative emissions; Indicates the evaporative emission coefficient during vehicle driving. The Technical Guidelines for Compilation of Road Motor Vehicle Emission Inventories provide reference values. ; Indicates the mileage of the vehicle; Indicates the average speed of the vehicle.

[0285] 4.3. Calculate the external costs caused by greenhouse gas emissions during vehicle operation in combination with the monetized price of greenhouse gas per unit, including the following steps:

[0286] The external costs caused by greenhouse gas emissions during vehicle operation are calculated according to formula (14).

[0287]

[0288] Where: represents the external costs of greenhouse gas emissions; express The unit monetization price of Indicates greenhouse gas emissions; represents the set of three greenhouse gases, namely .

[0289] 4.4. Based on the monetary price of unit atmospheric pollutants, calculate the external costs caused by the emission of warm atmospheric pollutants during vehicle operation.

[0290] The external costs caused by the emission of air pollutants during vehicle transportation are calculated according to formula (15).

[0291]

[0292] Where: Represents the external costs caused by the emission of air pollutants; Indicates Unit monetary price of each air pollutant; Indicates Exhaust emissions of various air pollutants; express Evaporative emissions; Represents a collection of five air pollutants.

[0293] 4.5. Combined with the external costs caused by the emission of greenhouse gases and air pollutants, an external cost assessment model for the transportation of hazardous materials on highways is constructed, including the following steps:

[0294] The total external cost caused by gas emissions during vehicle transportation is calculated according to formula (16).

[0295]

[0296] Where: represents the total external cost caused by gas emissions during vehicle transportation; represents the external costs of greenhouse gas emissions; Represents the external costs caused by the emission of air pollutants.

[0297] Step 5 specifically includes the following steps:

[0298] Combining the risk assessment model and the external cost assessment model, with the dual goals of minimizing transportation risk and minimizing external cost, a green path planning model for highway transportation of hazardous materials is constructed, which also includes:

[0299] Objective Function Minimize the risks in the transportation of dangerous goods, as shown in formula (17);

[0300]

[0301] Objective Function Minimize the total external cost in the transportation of dangerous goods, as shown in formula (18);

[0302]

[0303]

[0304] Constraints (19) and (20) ensure that vehicles transporting dangerous goods must depart from the specified starting point and arrive at the specified destination;

[0305]

[0306]

[0307] Constraints (21) and (22) ensure that each node is visited at most once;

[0308]

[0309]

[0310] Decision variables , , which means that if a dangerous goods transport vehicle passes through a highway section ,but ;otherwise, .

[0311] Step 6 specifically includes the following steps:

[0312] A hybrid algorithm combining the ant colony algorithm and the simulated annealing algorithm is designed to solve the above-mentioned green path planning model for hazardous goods highway transportation. The overall flow chart of the hybrid algorithm can be found in Fig. 9 ,include:

[0313] S1: Initialization:

[0314] S1.1: Initialize algorithm parameters:

[0315] Ant Colony Algorithm Parameters: Number of Ants , maximum number of iterations , pheromone release intensity , pheromone importance factor , heuristic function importance factor 、Maximum number of ant searches and pheromone volatility coefficient ;

[0316] Simulated annealing algorithm parameters: initial temperature , termination temperature , cooling rate and the number of iterations at each temperature .

[0317] S1.2: Initialize the pheromone matrix of the ant colony algorithm:

[0318] Construct a pheromone matrix with the dimension of the number of nodes and initialize the pheromone concentration between all node pairs (𝑖, 𝑗) is 0, and then determine whether there is a connection relationship between any two nodes. If node 𝑖 and node 𝑗 are connected, the corresponding pheromone concentration is Set to .

[0319] S1.3: Initialize the transport start point and transport destination .

[0320] S2: Redistribution of pheromones:

[0321] S2.1: Dijkstra algorithm solves the shortest path:

[0322] Based on the connection relationship of the nodes, count the path length between any two connected nodes , and use it as the weight, then apply the Dijkstra algorithm to solve the shortest path for the vehicle to complete the transportation task. For each node pair (i, j) on this shortest path, increase the pheromone concentration on it by Q, which is the updated pheromone concentration. .

[0323] S2.1: Dijkstra algorithm finds the path with the least risk:

[0324] The risk value between two connected nodes obtained by the risk assessment process As the new weight, the Dijkstra algorithm is used again to solve the path with the lowest risk when the vehicle completes the transportation task. For each node pair (i, j) on this path with the lowest risk, the pheromone concentration on it is also increased, and the increase is still Q, that is, the pheromone concentration after the update again .

[0325] S3: Generate initial solution vector:

[0326] Ants start from a designated starting point Start searching and calculate the probability of selecting the next node according to the state transition formula (23) until the ant reaches the specified end point. However, ants sometimes experience a "deadlock" phenomenon when selecting nodes, that is, the ants cannot continue searching because there are no adjacent nodes to choose from, causing the algorithm to stagnate and fail. Fig.10 ,Since nodes 15 and 19 have been selected and put into the taboo table, node 20 no longer has any feasible adjacent nodes. At this time, the algorithm is stuck, and node 20 will continue to appear in the subsequent part of the ant solution path, forming a cycle.

[0327] To this end, the present invention proposes a hybrid strategy with search step limit to solve the "deadlock" phenomenon in the ant colony algorithm. The hybrid strategy is as follows: before the ants iteratively search for paths, the maximum search step number of each ant is set to the total number of nodes, that is, , ensuring that only those ants that are truly stuck are screened out. Each time an ant finds a path, its search step count is increased by 1, and it is checked whether it is stuck. If it is not stuck, the process ends and the next ant's iterative pathfinding process begins; if it is stuck, a marking strategy is adopted in the early stage of the algorithm, that is, the ant is marked as invalid, and all nodes are added to the ant's path search results in order. The schematic diagram of the implementation result of the marking strategy can be seen in Fig.11 ; In the later stage of the algorithm, a recall strategy is adopted to clear all the nodes previously searched by the ant, recall it to the starting node, and allow it to restart the search process to find a feasible path. This process continues until all ants have completed pathfinding and generated a set of initial solution vectors.

[0328]

[0329] In the formula, Indicates Only ants from the node at the current iteration Go to Node The probability of Indicates the current iteration number; Representation Node and nodes The concentration of pheromones on the inter-connecting paths; Representation Node and nodes The length of the connecting path; is the heuristic function, , indicating that the ant slave node Transfer to Node the degree of expectation; Indicates The set of nodes to be selected by the ants; is the pheromone importance factor; is the importance factor of the heuristic function.

[0330] S4: Calculate the objective function value:

[0331] The solution vectors of the marked ants are identified based on the length of the solution vectors. The two objective function values ​​of these solution vectors are assigned to Inf (infinity) to distinguish them from the solution vectors of the unmarked ants. For other unmarked ants, the objective function values ​​of their solution vectors are calculated based on the objective function and Perform calculations.

[0332] S5: Initial de-perturbation:

[0333] S5.1: Identify the solution vector of the marked ant and do not perturb it.

[0334] S5.2: At the current temperature, perform multiple perturbations on the solution vectors of other unmarked ants. The specific perturbation rules are as follows: randomly select any node in the solution vector except the starting point and the end point as the perturbation node, and then determine whether there are other nodes between the immediately preceding and immediately succeeding nodes of the perturbation node: if not, reselect the perturbation node for perturbation; if such a node exists, further determine whether there are any nodes among these nodes that have never appeared in the current solution vector. If such a node is found, randomly select one of them to exchange with the perturbation node to generate a new solution vector. Otherwise, end the perturbation. See the schematic diagram of the perturbation implementation for details. Fig.12 .

[0335] S6: Trade-offs of the new solution:

[0336] S6.1: Normalize and sum the objective function value of the current solution according to formula (24).

[0337] S6.2: Calculate the two objective function values ​​of the new solution generated after the disturbance, normalize the objective function values ​​of the new solution according to formula (25) and add them up.

[0338] S6.3: Determine whether the new solution is accepted according to the Metropolis criterion of formula (26): if the sum of the normalized objective function values ​​of the new solution ( ) is less than the sum of the normalized objective function values ​​of the current solution ( ), then directly use the new solution to replace the current solution; if Greater than or equal to , then calculate the acceptance probability of the new solution, which is , and then generate a random number rand uniformly distributed in the interval (0,1). If the calculated acceptance probability is greater than the random number rand, the new solution is accepted to replace the current solution; otherwise, the new solution is abandoned and replaced with the current solution.

[0339]

[0340]

[0341]

[0342] In the formula, and Represents the two objective function values ​​of the current solution; and Represent the two objective function values ​​of the new solution; and Represents the sum of the two objective function values ​​after normalization of the current solution and the new solution respectively; Indicates the current temperature; Represents the probability that the new solution replaces the current solution.

[0343] S7: Determine whether the disturbance has been completed for all unmarked ants: if yes, proceed to the next cooling operation; if no, disturb the next unmarked ant.

[0344] S8: Cooling down:

[0345] The temperature is gradually reduced according to the set exponential cooling strategy, that is, the current temperature , and compare whether the current temperature has reached the set end temperature : If the set minimum temperature is reached, end the simulated annealing algorithm; otherwise, return and continue to execute S3;

[0346] S9: Non-dominated solution screening:

[0347] A set of initial solution vectors generated by the ant colony algorithm are perturbed and selected by the simulated annealing algorithm to obtain a set of new solution vectors. This set of new solution vectors is screened for non-dominated solutions and the Pareto optimal solution set is retained.

[0348] S10: Pheromone Update:

[0349] According to the path selection in the Pareto optimal solution set, the pheromone matrix of the ant colony algorithm is updated according to formulas (27) and (28);

[0350]

[0351]

[0352] In the formula, Indicates Only ants are on the node and nodes The concentration of pheromones released on the connecting path between them; is the number of ants in the Pareto optimal solution set; Representation Node and nodes The sum of the pheromone concentrations released by all ants on the connecting path; Indicates the current iteration number of the ant colony algorithm; Indicates The total length of the route taken by the ants.

[0353] S11: Ant Colony Iteration:

[0354] S3-S10 constitutes a complete iteration process of the ant colony algorithm. After each iteration, the number of iterations of the ant colony algorithm is increased by 1, that is, the current number of iterations. , check whether the current number of iterations has reached the preset maximum number of iterations of the ant colony algorithm: if the maximum number of iterations is reached, exit the program and output the result; otherwise, continue to the next iteration.

[0355] Step 7 specifically includes the following steps:

[0356] In the implementation, when the hybrid algorithm reaches the termination condition, the solution of the algorithm is output, including:

[0357] Repeat S3-S10 until the maximum number of iterations of the ant colony algorithm is reached, exit the program and output a set of Pareto solutions for the green path planning of dangerous goods highway transportation. Fig.13 ,The results can provide a reference for transportation companies and regulators to develop scientific and effective route planning solutions to balance transportation risks and external costs.

Claims

1. A green path planning method for dangerous goods highway transportation based on multi-source data, characterized in that: include: Obtain multi-source data related to the green route planning process for hazardous goods highway transportation through data disclosure websites and online public documents. These multi-source data include highway network data, hazardous goods transportation vehicle alarm data, residential data, highway toll data, land use data, emergency rescue point data, greenhouse gas emissions data, air pollutant emissions data, and unit gas monetization data; Use Excel and ArcGIS software to process the acquired multi-source data to form a unified highway transportation dataset of hazardous materials; The alarm rate, consequences of dangerous goods transportation accidents and risk mitigation factors of each highway section in the study area were calculated, and then a risk assessment model for dangerous goods highway transportation was constructed by combining the alarm rate, consequences of dangerous goods transportation accidents and risk mitigation factors; Calculate the greenhouse gas and air pollutant emissions during vehicle operation, and build an external cost assessment model for hazardous materials highway transportation by combining gas emissions and unit gas monetization prices; Combining the risk assessment model and the external cost assessment model, with the dual goals of minimizing transportation risk and minimizing external cost, a green path planning model for highway transportation of hazardous materials is constructed; A hybrid algorithm combining ant colony algorithm and simulated annealing algorithm is designed to solve the green path planning model for the transportation of hazardous materials on highways. When the hybrid algorithm reaches the termination condition, the green path planning result for the transportation of hazardous materials on highways is output; In the construction of the risk assessment model for dangerous goods highway transportation, the alarm rate is used instead of the accident rate for dangerous goods transportation. The consequences of dangerous goods transportation accidents are measured from two aspects: the number of potentially exposed people and the potential environmental impact area. The mitigation effect of emergency rescue forces on dangerous goods transportation risks is considered, including: A runway-type buffer zone with a width of 1 km on both sides was created for each highway section in the study area, and then the alarm rate of hazardous goods transport vehicles on a specific highway section was calculated according to formula (1); Where: P ij represents the alarm rate of highway section ij; k ij represents the number of alarm points in the buffer zone of highway section ij; K represents the total number of alarm points of hazardous goods transport vehicles obtained in the study area; A runway-type buffer zone is created for each highway section in the study area to simulate the potential impact area of ​​a hazardous goods transportation accident. The width of the buffer zone on both sides is determined according to the type of hazardous goods. Then, the roadside exposed population of each highway section when a hazardous goods transportation accident occurs is calculated according to formula (2). Where: E represents the number of people exposed on the roadside when a dangerous goods transportation accident occurs on highway section ij; ij POP represents the number of settlements in the buffer zone of highway section ij; total represents the total population in the study area; E Total Indicates the total number of settlements in the study area; p out represents the outdoor exposure rate of residents; in It indicates the safety risk mitigation effect of indoor environments such as buildings on residents; In ArcGIS, a highway network dataset was constructed using length as the cost attribute. Then, the NetworkAnalyst module was used to perform path optimization analysis on all the processed toll station OD pairs to solve the shortest path and derive the highway sections along the way. The traffic flow between each toll station OD pair was then allocated to each section on its corresponding shortest path. After cumulative aggregation, the traffic flow data for each highway section within a certain period of time was obtained. Finally, the exposed population on the road when a dangerous goods transportation accident occurred was calculated according to formula (3). Where: represents the number of people exposed on the highway section ij when a dangerous goods transportation accident occurs; Q ij It represents the traffic flow of highway section ij in a certain period of time; represents the average occupancy factor of the vehicle; L ij represents the length of the highway section ij; V ij represents the average speed of all vehicles traveling on the highway section ij; Call the clipping tool in ArcGIS, take the land use data as input elements, and use the runway-shaped buffer created when calculating the roadside exposed population as clipping elements in turn. Clip the corresponding number of new buffers, then export the attribute table of all new buffers, extract the affected land types and their corresponding areas near the road section, and finally calculate the affected land area near each highway section when a dangerous goods transportation accident occurs according to formula (4); Where: S ij represents the land area of ​​highway section ij affected by the dangerous goods transportation accident; θ k represents the impact weight coefficient of the kth land type; S k It represents the area of ​​the kth type of land within the scope of the dangerous goods transportation accident; Create several multi-ring buffers for each highway section in the study area. The multi-ring buffer is composed of multiple runway-type buffers. The number of multi-ring buffers and the width of both sides of the runway-type buffer are set according to specific requirements. The risk mitigation factor of each highway section when a dangerous goods transportation accident occurs is calculated according to formula (5); Where: ε ij represents the risk mitigation factor for highway section ij; ε t Sub-indicator representing the risk mitigation factor; τ t Represents the sub-index ε t The weight value of VA t Represents the sub-index ε t The assigned value of The alarm rate of each highway section in the study area, the number of people exposed on the roadside and on the road when a dangerous goods transportation accident occurs, and the risk mitigation factor of the section are combined to calculate the potential population risk of the highway section when a dangerous goods transportation accident occurs according to formula (6); Where: P represents the potential population risk caused by a dangerous goods transportation accident on the highway section ij; ij represents the alarm rate of highway section ij; E ij POP represents the number of settlements in the buffer zone of highway section ij; total represents the total population in the study area; E Total Indicates the total number of settlements in the study area; p out represents the outdoor exposure rate of residents; in It indicates the safety risk mitigation effect of indoor environment such as buildings on residents; Q ij It represents the traffic flow of highway section ij in a certain period of time; represents the average occupancy factor of the vehicle; L ij represents the length of the highway section ij; V ij represents the average speed of all vehicles traveling on the highway section ij; ε ij represents the risk mitigation factor for highway section ij; The alarm rate of each highway section in the study area, the land area affected when a dangerous goods transportation accident occurs, and the risk mitigation factor of the section are combined to calculate the potential environmental risk of the highway section when a dangerous goods transportation accident occurs according to formula (7); Where: P represents the potential environmental risk caused by a dangerous goods transportation accident on the highway section ij; ij represents the alarm rate of highway section ij; θ k represents the impact weight coefficient of the kth land type; S k represents the area of ​​the kth type of land within the scope of the dangerous goods transportation accident; ε ij represents the risk mitigation factor for highway section ij; According to formula (8), the total risk value of each highway section when a dangerous goods transportation accident occurs is calculated: Where: R ij represents the total risk value of dangerous goods transportation accidents occurring on road section ij; ω1 and ω2 represent the weight coefficients of population risk and environmental risk respectively; It represents the potential population risk caused by a dangerous goods transportation accident on the highway section ij; It represents the potential environmental risk caused by a dangerous goods transportation accident on the highway section ij; and They represent the maximum value of potential population risk among all highway sections and the maximum value of potential environmental risk among all highway sections respectively; In the construction of the external cost assessment model for hazardous materials highway transportation, the greenhouse gas emissions and air pollutant emissions during vehicle operation are calculated, and the total external costs are calculated in combination with the monetized price of unit gas, including: Calculate the carbon dioxide CO2 emissions during vehicle operation according to formula (9); Where: represents the emission of CO2; NCV represents the average lower calorific value of fossil fuel; D represents the mileage of the vehicle; OC represents the fuel consumption per 100 kilometers of the vehicle; ρ represents the density of fossil fuel; CC represents the carbon content per unit calorific value of fossil fuel; OF represents the carbon oxidation rate of fossil fuel; 44 / 12 represents the molecular weight ratio of carbon dioxide to carbon; represents the global warming potential of CO2; Calculate the methane CH4 emissions during vehicle operation according to formula (10); Where: It represents the emission of CH4, measured in CO2 equivalent; represents the emission factor of CH4; represents the global warming potential of CH4; Calculate the emission of nitrous oxide (N2O) during vehicle operation according to formula (11); Where: It represents the emission of N2O, measured in CO2 equivalent; represents the emission factor of N2O; represents the global warming potential of N2O; Calculate the amount of air pollutants emitted in the form of exhaust gas during vehicle operation according to formula (12); Where: represents the tail gas emission of the mth air pollutant, where m∈M, M represents the set of five air pollutants, including carbon monoxide CO, nitrogen oxides NO x , fine particulate matter PM 2.5 , inhalable particulate matter PM 10 and volatile organic compounds VOCs, namely M = {CO, NO x 、PM 2.5 、PM 10 , VOCs}; D represents the vehicle's mileage; BEF m It represents the comprehensive benchmark emission coefficient of the mth type of air pollutant, which is related to the vehicle model and fuel type; represents the temperature correction factor; represents the humidity correction factor; γ represents the average speed correction factor of the vehicle; θ represents the load factor correction factor of the vehicle; Calculate the amount of atmospheric pollutants emitted in the form of evaporation during vehicle operation according to formula (13); Where: Indicates the evaporative emission of VOCs; It represents the evaporative emission coefficient during the driving process of the vehicle; D represents the mileage of the vehicle; V represents the average driving speed of the vehicle; Combined with the monetary price of unit greenhouse gas, the external cost caused by greenhouse gas emissions during vehicle operation is calculated according to formula (14); Where: C GHG represents the external costs of greenhouse gas emissions; Indicates the unit monetary price of CO2; represents the emission of the hth greenhouse gas; H represents the set of three greenhouse gases, i.e., H = {CO2, CH4, N2O}; Combined with the monetary price of unit air pollutants, the external costs caused by the emission of air pollutants during vehicle operation are calculated according to formula (15); Where: C AP Represents the external costs caused by the emission of air pollutants; represents the unit monetary price of the mth air pollutant; represents the tail gas emission of the mth type of air pollutant, represents the evaporation emission of VOCs; M represents the set of five air pollutants, namely, M = {CO, NO x 、PM 2.5 、PM 10 , VOCs}; Combined with the external costs caused by the emission of greenhouse gases and air pollutants, the total external costs caused by gas emissions during vehicle operation are calculated according to formula (16); C=C GHG +C AP (16); Where: C represents the total external cost caused by gas emissions during vehicle operation; C GHG represents the external cost caused by greenhouse gas emissions; C AP Represents the external costs caused by the emission of air pollutants.

2. The method for planning a green path for dangerous goods highway transportation based on multi-source data according to claim 1 is characterized in that: Process acquired multi-source data and import it into Excel and ArcGIS, including: Import highway network data, latitude and longitude coordinates of dangerous goods transport vehicle alarms, longitude and longitude coordinates of residential areas, longitude and longitude coordinates of highway toll stations, land use data, and longitude and longitude coordinates of emergency rescue points into ArcGIS; Greenhouse gas emission data, air pollutant emission data and unit greenhouse gas and air pollutant monetization data were extracted from public documents on the Internet and then imported into Excel.

3. The method for planning a green path for dangerous goods highway transportation based on multi-source data according to claim 1 is characterized in that: Combining the risk assessment model and the external cost assessment model, a green path planning model for highway transportation of hazardous materials is constructed with the dual objectives of minimizing transportation risk and minimizing external cost, including: The objective function f1 minimizes the risk in the transportation of dangerous goods, as shown in formula (17); The objective function f2 minimizes the total external cost in the process of dangerous goods transportation, as shown in formula (18); Constraints (19) and (20) ensure that vehicles transporting dangerous goods must depart from the specified starting point and arrive at the specified destination; Constraints (21) and (22) ensure that each node is visited at most once; Decision variables If a dangerous goods transport vehicle passes through the highway section ij, then x ij =1; otherwise, x ii =0.

4. The method for planning green routes for dangerous goods highway transportation based on multi-source data according to claim 1 is characterized in that: A hybrid algorithm combining ant colony algorithm and simulated annealing algorithm is designed to solve the constructed green path planning model for hazardous goods highway transportation, including: S1: Initialization: S1.1: Initialize algorithm parameters: Ant colony algorithm parameters: number of ants ant num , maximum number of iterations iter max , pheromone release intensity Q, pheromone importance factor α, heuristic function importance factor β, maximum number of ant searches search total and pheromone volatility coefficient ρ; Simulated annealing algorithm parameters: initial temperature T0, termination temperature T end , cooling rate q and the number of iterations iter at each temperature tem ; S1.2: Initialize the pheromone matrix of the ant colony algorithm: Construct a pheromone matrix with the dimension of the number of nodes and initialize the pheromone concentration τ between all node pairs (i, j) ij is 0, and then determine whether there is a connection relationship between any two nodes. If node i and node j are connected, the corresponding pheromone concentration τ ij Set to Q; S1.3: Initialize the transport starting point N start and transport destination N end ; S2: Redistribution of pheromones: S2.1: Dijkstra algorithm solves the shortest path: Based on the connection relationship between nodes, the path length D between any two connected nodes is counted ij , and use it as the weight, then apply the Dijkstra algorithm to solve the shortest path for the vehicle to complete the transportation task. For each node pair (i, j) on this shortest path, increase the pheromone concentration on it by Q, that is, the updated pheromone concentration τ ij =τ ij +Q; S2.1: Dijkstra algorithm finds the path with the least risk: The risk value R between two connected nodes obtained by the risk assessment process ij As the new weight, the Dijkstra algorithm is used again to solve the path with the lowest risk when the vehicle completes the transportation task. For each node pair (i, j) on this path with the lowest risk, the pheromone concentration on it is also increased, and the increase is still Q, that is, the pheromone concentration τ after the update again ij =τ ij +Q; S3: Generate initial solution vector: The ant starts from the specified starting point N. start Start searching and calculate the probability of selecting the next node according to the state transition formula (23) until the ant reaches the specified end point N. end , the present invention proposes a hybrid strategy with search step limit to solve the "deadlock" phenomenon in the ant colony algorithm. The hybrid strategy is as follows: before the ants iteratively search for paths, the maximum search step of each ant is set to the total number of nodes. Each time the ant searches for paths, its search step is increased by 1, and it is checked whether it is stuck: if it is not stuck, the process ends and the iterative path finding process of the next ant begins; if it is stuck, a marking strategy is adopted in the early stage of the algorithm, that is, the ant is marked as invalid, and all nodes are added to the path search results of the ant in sequence, which means that the length of the solution vector of the ant is the total number of nodes; in the later stage of the algorithm, a recall strategy is adopted to clear all the nodes previously searched by the ant, recall it to the starting node, and allow it to restart the search process to find a feasible path, until all ants have completed path finding and generated a set of initial solution vectors; In the formula, represents the probability of the kth ant going from node i to node j at the current number of iterations; iter represents the current number of iterations; τ ij (iter) represents the pheromone concentration on the connection path between node i and node j; D ij represents the length of the connection path between node i and node j; η ij (iter) is the heuristic function, η ij (iter) = 1 / D ij , represents the expected degree of ant transfer from node i to node j; allow k represents the set of nodes to be selected by the kth ant; α is the pheromone importance factor; β is the heuristic function importance factor; S4: Calculate the objective function value: The solution vectors of the marked ants are identified according to the length of the solution vectors, and the two objective function values ​​of these solution vectors are assigned to Inf to distinguish them from the solution vectors of the unmarked ants. For other unmarked ants, the objective function values ​​of their solution vectors are calculated according to the objective functions f1 and f2; S5: Initial de-perturbation: S5.1: Identify the solution vector of the marked ant and do not perturb it; S5.2: At the current temperature, perform multiple perturbations on the solution vectors of other unmarked ants. The specific perturbation rules are as follows: randomly select any node in the solution vector except the starting point and the end point as the perturbation node, and then determine whether there are other nodes between the immediately preceding and immediately succeeding nodes of the perturbation node: if not, reselect the perturbation node for perturbation; if such a node exists, further determine whether there are any nodes among these nodes that have never appeared in the current solution vector. If such a node is found, randomly select one of them to exchange with the perturbation node to generate a new solution vector; otherwise, end the perturbation; S6: Trade-offs of the new solution: S6.1: normalize and sum the objective function value of the current solution according to formula (24); S6.2: Calculate the two objective function values ​​of the new solution generated after the disturbance, normalize the objective function values ​​of the new solution according to formula (25) and add them up; S6.3: Determine whether the new solution is accepted according to the Metropolis criterion of formula (26): If the sum of the normalized objective function values ​​E2 of the new solution is less than the sum of the normalized objective function values ​​E1 of the current solution, then directly use the new solution to replace the current solution; if E2 is greater than or equal to E1, calculate the acceptance probability of the new solution, which is exp(-(E2-E1) / T), and then generate a random number rand uniformly distributed in the interval (0,1). If the calculated acceptance probability is greater than the random number rand, the new solution is also accepted to replace the current solution; otherwise, the new solution is abandoned to replace the current solution; In the formula, and Represents the two objective function values ​​of the current solution; and Represents the two objective function values ​​of the new solution; E1 and E2 represent the sum of the two objective function values ​​after normalization of the current solution and the new solution respectively; T represents the current temperature; P represents the probability of the new solution replacing the current solution; S7: Determine whether the disturbance has been completed for all unmarked ants: if yes, proceed to the next cooling operation; if no, disturb the next unmarked ant; S8: Cooling down: The temperature is gradually lowered according to the set cooling strategy, and at the same time, the current temperature is compared to see if it has reached the set end temperature T. end : If the set minimum temperature is reached, end the simulated annealing algorithm; otherwise, return and continue to execute S3; S9: Non-dominated solution screening: After the disturbance and new solution selection of the simulated annealing algorithm, a set of initial solution vectors generated by the ant colony algorithm is used to obtain a set of new solution vectors. The non-dominated solutions of this set of new solution vectors are screened to retain the Pareto optimal solution set. S10: Pheromone Update: According to the path selection in the Pareto optimal solution set, the pheromone matrix of the ant colony algorithm is updated according to formulas (27) and (28); In the formula, Δτ ij k represents the pheromone concentration released by the kth ant on the connection path between nodes i and j; n Pareto is the number of ants in the Pareto optimal solution set; Δτ ij represents the sum of the pheromone concentrations released by all ants on the path connecting node i and node j; iter represents the current iteration number of the ant colony algorithm; L k represents the total length of the route taken by the kth ant; S11: Ant Colony Iteration: S3-S10 constitute a complete iterative process of the ant colony algorithm. After each iteration, the number of iterations of the ant colony algorithm is increased by 1, that is, the current number of iterations iter = iter + 1. Check whether the current number of iterations has reached the preset maximum number of iterations of the ant colony algorithm: if the maximum number of iterations is reached, exit the program and output the result; otherwise, continue to the next iteration.

5. The method for planning green routes for dangerous goods highway transportation based on multi-source data according to claim 4 is characterized in that: When the hybrid algorithm reaches the termination condition, the algorithm's solution results are output, including: Repeat S3-S10 until the maximum number of iterations of the ant colony algorithm is reached, exit the program and output the green path planning results for hazardous goods highway transportation, which can provide a reference for scientific and effective path planning solutions for transportation companies and regulators to balance transportation risks and external costs.

Images