Air transportation logistics transportation management system based on intelligent algorithm optimization

Abstract

The invention discloses an air transportation logistics transportation management system based on intelligent algorithm optimization, which relates to the technical field of air transportation logistics transportation management, and comprises an air transportation logistics transportation data acquisition module, an air transportation logistics transportation data processing module and an air transportation logistics transportation management module, the air transportation logistics transportation data acquisition module acquires air transportation cargo data, initial flight data and flight weather data, and the air transportation logistics transportation data processing module constructs a flight path planning model based on the flight weather data by applying a dynamic planning algorithm, optimizes the flight path planning model in combination with a genetic algorithm and a simulated annealing algorithm, and obtains the flight path planning model. According to the invention, a data acquisition technology and an intelligent algorithm technology are combined with a modern information technology, thereby achieving the real-time and comprehensive monitoring of the air cargo and the flight weather in the air logistics transportation process, and solving a problem that the air cargo information, the flight resource information and the weather information of the flight path are difficult to combine in the prior art. And the problem of managing an air transportation logistics transportation process is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of air freight logistics transportation management, and particularly relates to an air freight logistics transportation management system optimized based on intelligent algorithms. Background Art

[0002] At present, with the accelerating process of global economic integration, air freight logistics, as an efficient transportation method, has become increasingly crucial in the international trade and supply chain systems. Traditional air freight logistics transportation management systems mostly rely on manual experience and simple information technology means, making it difficult to cope with the growing business scale and complex and changeable transportation environment. With the booming development of modern information technologies such as big data and artificial intelligence, the application of intelligent algorithms in the logistics field has gradually deepened. However, existing air freight logistics transportation management systems optimized based on intelligent algorithms still have many deficiencies. Some systems, although they introduce intelligent algorithms, have inaccurate and incomplete data collection, which affects the accuracy of the algorithm model, and the coordination between different algorithms is poor, unable to maximize the overall optimization effect. Therefore, the development of an air freight logistics transportation management system optimized based on intelligent algorithms has emerged as the times require and has become an urgent need to improve air freight logistics efficiency, reduce costs, and enhance transportation safety;

[0003] Although the existing technology has made great progress in the direction of air freight logistics transportation management optimized based on intelligent algorithms, there are still some problems to be optimized. The existing technology is difficult to combine air freight cargo information, flight resource information, and weather information of flight paths to manage air freight logistics transportation, which may lead to logistics stagnation of air freight cargo. Summary of the Invention

[0004] To achieve the above objectives, the present invention is realized through the following technical solutions: An air freight logistics transportation management system optimized based on intelligent algorithms includes an air freight logistics transportation data collection module, an air freight logistics transportation data processing module, and an air freight logistics transportation management module. Among them, each module is communicatively connected, and its characteristics are as follows:

[0005] The air freight logistics transportation data collection module collects air freight cargo data, initial flight data, and flight weather data, providing data support for the realization of the functions of subsequent modules;

[0006] The air freight logistics transportation data processing module consists of an environmental level division module, a flight loadable cargo resource monitoring module, and a flight path planning module;

[0007] The environmental level division module preprocesses the collected data and obtains the maximum cargo weight and maximum cargo volume of the flight, and respectively divides the levels of rainfall intensity, dust concentration, and air pressure;

[0008] The flight available cargo resource monitoring module is divided into a flight available cargo weight monitoring unit, a flight available cargo volume monitoring unit, and a resource combination monitoring unit. Among them, the flight available cargo weight monitoring unit is used to construct a flight available cargo weight monitoring model; the flight available cargo volume monitoring unit is used to construct a flight available cargo volume monitoring model; the resource combination monitoring unit is used to construct a flight available cargo resource monitoring model through the output results of the flight available cargo weight monitoring model and the flight available cargo volume monitoring model;

[0009] The flight path planning module constructs a flight path planning model based on flight weather data, combines genetic algorithm and simulated annealing algorithm to optimize the flight path planning model, and reduces the error of the flight path planning model for planning flight paths according to flight weather levels;

[0010] The air freight logistics transportation management module combines the flight available cargo resource monitoring model and the flight path planning model to manage air freight logistics transportation, solving the problem that it is difficult to combine air freight information, flight resource information, and flight path weather information in the prior art to manage air freight logistics transportation, which may lead to logistics stagnation of air freight.

[0011] A further improvement of the technical solution of the present invention lies in that: the process of the air freight logistics transportation data acquisition module for acquiring air freight data, initial flight data, and flight weather data includes:

[0012] Deploy different types of acquisition devices to acquire air freight data, initial flight data, and flight weather data. Among them, the acquisition devices include electronic platform scales, laser rangefinders, fuel density meters, volumetric flow meters, weighing scales, optical rain sensors, laser dust sensors, piezoresistive pressure sensors, etc.;

[0013] Use an electronic platform scale to acquire the weight of air freight and the weight of the aircraft based on the principle of pressure-gravity conversion; use a laser rangefinder to acquire the length, width, and height of air freight respectively, and calculate the volume of air freight according to the volume calculation formula;

[0014] Use a fuel density meter and a volumetric flow meter to acquire fuel density and fuel volume respectively, and calculate the fuel weight according to the volume calculation formula; use a weighing scale to acquire the weight of the crew, and obtain the initial load of the flight by calculating the sum of the aircraft weight, fuel weight, and crew weight; combine the laser rangefinder and Leica Cyclone software to acquire the initial volume of the flight;

[0015] Collect the real-time rainfall of the flight environment through an optical rain sensor based on the optical principle; collect the real-time dust concentration of the flight environment through a laser dust sensor based on the laser scattering principle; use a piezoresistive pressure sensor to collect the real-time air pressure of the flight environment based on the piezoresistive effect.

[0016] A further improvement of the technical solution of the present invention is that the process of preprocessing the collected data by the environment level division module includes:

[0017] Perform data cleaning and data calibration on the collected weight of air cargo, volume of air cargo, initial load of the flight, initial volume of the flight, real-time rainfall, real-time dust concentration, and real-time air pressure of the flight environment.

[0018] Add timestamps to the weight of air cargo, volume of air cargo, initial load of the flight, and initial volume of the flight, uniformly convert the timestamps to UTC time, and through time correction, synchronize the collection times of the weight of air cargo and the initial load of the flight, and the collection times of the volume of air cargo and the initial volume of the flight respectively.

[0019] A further improvement of the technical solution of the present invention is that the process of the environment level division module obtaining the maximum cargo weight of the flight and the maximum cargo volume of the flight and respectively dividing the levels of rainfall intensity, dust concentration, and air pressure includes:

[0020] Set the maximum takeoff load of the flight, calculate the difference between the maximum takeoff load of the flight and the initial load of the flight, and obtain the maximum cargo weight of the flight; set the maximum takeoff cargo volume of the flight, calculate the difference between the maximum takeoff cargo volume of the flight and the initial volume of the flight, and obtain the maximum cargo volume of the flight;

[0021] Statistically calculate the total real-time rainfall in the flight environment per hour to obtain the total hourly rainfall in the flight environment. According to the total hourly rainfall in the flight environment, divide the rainfall intensity level, and divide the total hourly rainfall in the flight environment into low rainfall intensity, medium rainfall intensity, and high rainfall intensity. Among them, low rainfall intensity means that the total hourly rainfall in the flight environment is less than 2.5 mm; medium rainfall intensity means that the total hourly rainfall in the flight environment is between 2.5 mm and 8.0 mm; high rainfall intensity means that the total hourly rainfall in the flight environment is higher than 8.0 mm;

[0022] According to the real-time dust concentration in the flight environment, divide the dust concentration level, and divide the real-time dust concentration in the flight environment into light dust, medium dust, and heavy dust. Among them, light dust means that the real-time dust concentration in the flight environment is less than 1000 μg / m³; medium dust means that the real-time dust concentration in the flight environment is between 1000 and 2000 μg / m³; heavy dust means that the real-time dust concentration in the flight environment is higher than 2000 μg / m³;

[0023] According to the real-time air pressure in the flight environment, divide the real-time air pressure levels, and classify the real-time air pressure in the flight environment into low air pressure, medium air pressure, and high air pressure. Among them, low air pressure means the real-time air pressure in the flight environment is lower than 979 hPa; medium air pressure means the real-time air pressure in the flight environment is between 979 hPa and 1031 hPa; high air pressure means the real-time air pressure in the flight environment is higher than 1031 hPa.

[0024] A further improvement of the technical solution of the present invention lies in: for the flight loadable cargo weight monitoring unit, the construction process of the flight loadable cargo weight monitoring model includes:

[0025] Take the air cargo weight and the maximum cargo weight of the flight as the first data set, divide it into a training set and a test set according to the ratio of 8:2. Combine the training set data with the multiple linear regression algorithm, take the air cargo weight and the maximum cargo weight of the flight as inputs, and take the remaining cargo weight of the flight as the output. Adjust the intercept term, the regression coefficient of the air cargo weight, and the regression coefficient of the maximum cargo weight of the flight respectively to train the flight loadable cargo weight monitoring model;

[0026] Input the test set data into the flight loadable cargo weight monitoring model, evaluate the performance of the flight loadable cargo weight monitoring model, adjust the parameters of the flight loadable cargo weight monitoring model, optimize the flight loadable cargo weight monitoring model, deploy the optimized flight loadable cargo weight monitoring model into the system, and obtain the final flight loadable cargo weight monitoring model.

[0027] A further improvement of the technical solution of the present invention lies in: for the flight loadable cargo volume monitoring unit, the construction process of the flight loadable cargo volume monitoring model includes:

[0028] Take the air cargo volume and the maximum cargo volume of the flight as the second data set, divide it into a training set and a test set according to the ratio of 8:2. Combine the training set data with the multiple linear regression algorithm, take the air cargo volume and the maximum cargo volume of the flight as inputs, and take the remaining cargo volume of the flight as the output. By adjusting the intercept term, the regression coefficient of the air cargo volume, and the regression coefficient of the maximum cargo volume of the flight, train the flight loadable cargo volume monitoring model;

[0029] Input the test set data into the flight loadable cargo volume monitoring model, evaluate the performance of the flight loadable cargo volume monitoring model, adjust the parameters of the flight loadable cargo volume monitoring model, optimize the flight loadable cargo volume monitoring model, deploy the optimized flight loadable cargo volume monitoring model into the system, and obtain the final flight loadable cargo volume monitoring model.

[0030] A further improvement of the technical solution of the present invention lies in: for the resource combination monitoring unit, the construction process of the flight loadable cargo resource monitoring model includes:

[0031] Input the weight of air cargo and the maximum cargo weight of the flight into the flight available cargo weight monitoring model, and the flight available cargo weight monitoring model outputs the remaining cargo weight of the flight; input the volume of air cargo and the maximum cargo volume of the flight into the flight available cargo volume monitoring model, and the flight available cargo volume monitoring model outputs the remaining cargo volume of the flight.

[0032] Take the weight of air cargo, the volume of air cargo, the remaining cargo weight of the flight, and the remaining cargo volume of the flight as the third data set, divide it into a training set and a test set according to a ratio of 7:3, use the training set data and the neural network algorithm, take the weight of air cargo and the volume of air cargo as inputs, take the remaining cargo weight of the flight and the remaining cargo volume of the flight as outputs, calculate the predicted output data through forward propagation, and update the weights and biases of the model using backpropagation to learn the non-linear relationship between the weight of air cargo and the remaining cargo weight of the flight and the non-linear relationship between the volume of air cargo and the remaining cargo volume of the flight, and train the flight available cargo resource monitoring model.

[0033] Input the test set data into the flight available cargo resource monitoring model, use the MSE function to evaluate the errors between the output results of the flight available cargo resource monitoring model and the actual remaining cargo weight and remaining cargo volume of the flight respectively, adjust the parameters of the flight available cargo resource monitoring model according to the evaluation results, optimize the performance of the flight available cargo resource monitoring model, deploy the optimized flight available cargo resource monitoring model into the system, and obtain the final flight available cargo resource monitoring model.

[0034] A further improvement of the technical solution of the present invention lies in: for the flight path planning module, the construction process of the flight path planning model includes:

[0035] Apply the dynamic programming algorithm, define the dynamic programming state as a triple, which consists of a state point, a time point, and a flight weather level. Divide the flight route into N small route segments, extract route nodes from each small route segment, use the route node as the state point in dynamic programming, record the time point corresponding to the state point and the flight weather level. The flight weather level includes rainfall intensity level, sand and dust level, and air pressure level. Construct a cost function according to the flight weather level and the flight time.

[0036] Initialize the triple, set the cost of the initial state point to 0, and set the costs of other state points to infinity. Starting from the initial state point, obtain all possible transfer points of each state point in chronological order and voyage order, obtain all possible transfer paths of each state point. For all possible transfer points of each state point, calculate the transfer cost using the cost function, compare the transfer costs of all possible transfer paths of each state point, select the minimum transfer cost, optimize the cost function using the minimum transfer cost, and obtain the optimal transfer path of each state point by combining the possible transfer paths corresponding to the minimum transfer costs of each state point. According to the optimal transfer path, plan the flight path of the flight and obtain the flight planning decision.

[0037] Apply the dynamic programming algorithm to combine the triple, the optimized cost function, and the flight planning decision to obtain the flight path planning model, and deploy the flight path planning model to the system to obtain the final flight path planning model.

[0038] A further improvement of the technical solution of the present invention lies in: the process of optimizing the flight path planning model by the flight path planning module in combination with the genetic algorithm and the simulated annealing algorithm includes:

[0039] A1. Extract each state point and its corresponding flight planning decision in the triple of the flight path planning model. Use the genetic algorithm to perform real number encoding on each state point in the triple, and the encoding of each state point corresponds to the flight planning decision to obtain a real number vector. Use this real number vector to represent an individual, randomly generate P individuals, combine the P individuals to obtain the initial population, use the reciprocal of the mean square error as the fitness function, and set the number of iterations and the fitness threshold.

[0040] According to the fitness values of each individual in the population, through the roulette wheel selection method, the greater the fitness value of each individual, the greater the probability that the individual is selected, and select the individuals in the population to participate in the next generation of reproduction; use the arithmetic crossover method to perform gene exchange on the selected individuals; randomly select the flight planning decision corresponding to the individual after gene exchange, randomly change the flight planning decision corresponding to the individual, and repeat the iterative operations of selection, crossover, and mutation until the number of repeated iterative operations reaches the set number of iterations and the fitness values of the individuals in the population reach the fitness threshold, then terminate the repeated iterative operation, select the individual with the highest fitness value in the population, and use the individual with the highest fitness value to update each state point and its corresponding flight planning decision in the flight path planning model to optimize the flight path planning model.

[0041] A2. Use the flight planning decision output by the optimized flight path planning model as the initial solution of the simulated annealing algorithm. Set the number of iterations for simulated annealing optimization, define the neighborhood structure, make a small adjustment to the initial solution to generate a neighborhood solution. Use the cost function in the flight path planning model as the objective function, calculate the objective function values of the initial solution and the neighborhood solution, and compare the two. When the objective function value of the neighborhood solution is better than that of the initial solution, use the neighborhood solution as the current solution; when the objective function value of the initial solution is better than that of the neighborhood solution, according to the probability formula of the simulated annealing algorithm, determine the situation of using the neighborhood solution as the current solution, and then make a small adjustment to the current solution to generate a new neighborhood solution. Use the current solution and the new neighborhood solution to repeat the above iterative process of objective function comparison. When the actual number of iterations reaches the number of iterations for simulated annealing optimization, stop the iterative process of the simulated annealing algorithm and optimize the flight path planning model.

[0042] A further improvement of the technical solution of the present invention is that: in the air freight logistics transportation management module, the process of managing air freight logistics transportation by combining the flight available cargo resource monitoring model and the flight path planning model includes:

[0043] Input the air freight weight and air freight volume into the flight available cargo resource monitoring model, and the flight available cargo resource monitoring model outputs the remaining load weight and remaining load volume of the flight correspondingly;

[0044] According to the remaining load weight of the flight, when the air freight weight is lower than or equal to the remaining load weight of the flight, include all of this air freight in the corresponding flight for transportation; when the air freight weight is higher than the remaining load weight of the flight, screen out the air freight with a weight lower than the remaining load weight of the flight, give priority to arranging the screened air freight to be transported by this flight, and arrange the un-screened air freight to the next flight for transportation;

[0045] According to the remaining load volume of the flight, when the air freight volume is lower than or equal to the remaining load volume of the flight, include all of this air freight in the corresponding flight for transportation; when the air freight volume is higher than the remaining load volume of the flight, screen out the air freight with a volume lower than the remaining load volume of the flight, give priority to arranging the screened air freight to be transported by this flight, and arrange the un-screened air freight to the next flight for transportation;

[0046] Dispatch the flight path planning model. According to the flight weather level and flight time, obtain the transfer path corresponding to the minimum transfer cost for each flight segment, obtain the optimal transfer path for each flight segment, combine the optimal transfer paths for each flight segment, plan the flight path of the flight, obtain the flight planning decision, and conduct air freight logistics transportation according to this flight planning decision.

[0047] The beneficial effects of the present invention are as follows: In the air freight logistics transportation management system optimized based on intelligent algorithms of the present invention, compared with the traditional air freight logistics transportation management system optimized based on intelligent algorithms, the data acquisition technology and intelligent algorithm technology in the system of the present invention are closely combined with modern information technology, accurately capturing air freight data, initial flight data, and flight weather data, achieving real-time and comprehensive monitoring of air freight and flight weather during the air freight logistics transportation process. By using the multiple linear regression algorithm, a monitoring model for the weight of goods that can be carried by a flight and a monitoring model for the volume of goods that can be carried by a flight are respectively constructed. Through the neural network algorithm, the monitoring model for the weight of goods that can be carried by a flight and the monitoring model for the volume of goods that can be carried by a flight are combined to construct a monitoring model for the resources of goods that can be carried by a flight. Combining the dynamic programming algorithm, genetic algorithm, and simulated annealing algorithm, a flight path planning model is constructed and optimized, and the optimal flight path is planned by comprehensively considering flight weather data, solving the problem that the prior art is difficult to manage air freight logistics transportation by combining air freight information, flight resource information, and weather information of flight paths, which may lead to logistics stagnation of air freight. It ensures that the present invention can refine the dynamic monitoring standards for the air freight logistics transportation management system optimized based on intelligent algorithms within a more accurate range, making the monitored data more accurate indicators under the same conditions. The research and application of this system significantly enhance the degree of intelligence in the air freight logistics transportation management process optimized based on intelligent algorithms, improve the transportation efficiency of air freight logistics, reduce the transportation cost of air freight logistics, and improve the transportation safety of air freight logistics. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0049] Figure 1 It is a block diagram of an air freight logistics transportation management system optimized based on intelligent algorithms of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0051] Such as Figure 1As shown in the figure, the present invention provides an air freight logistics transportation management system optimized based on intelligent algorithms, including an air freight logistics transportation data acquisition module, an air freight logistics transportation data processing module, and an air freight logistics transportation management module. Among them, each module is communicatively connected, and it is characterized in that:

[0052] The air freight logistics transportation data acquisition module collects air freight cargo data, initial flight data, and flight weather data, providing data support for the function realization of subsequent modules;

[0053] The air freight logistics transportation data processing module consists of an environmental level division module, a flight loadable cargo resource monitoring module, and a flight path planning module;

[0054] The environmental level division module preprocesses the collected data, obtains the maximum cargo weight and maximum cargo volume of the flight, and respectively divides the levels of rainfall intensity, dust concentration, and air pressure;

[0055] The flight loadable cargo resource monitoring module is divided into a flight loadable cargo weight monitoring unit, a flight loadable cargo volume monitoring unit, and a resource combination monitoring unit. Among them, the flight loadable cargo weight monitoring unit is used to construct a flight loadable cargo weight monitoring model; the flight loadable cargo volume monitoring unit is used to construct a flight loadable cargo volume monitoring model; the resource combination monitoring unit is used to construct a flight loadable cargo resource monitoring model through the output results of the flight loadable cargo weight monitoring model and the flight loadable cargo volume monitoring model;

[0056] The flight path planning module constructs a flight path planning model based on the flight weather data, combines the genetic algorithm and the simulated annealing algorithm to optimize the flight path planning model, and reduces the error of the flight path planning model for planning the flight path according to the flight weather level;

[0057] The air freight logistics transportation management module combines the flight loadable cargo resource monitoring model and the flight path planning model to manage the air freight logistics transportation, solving the problem that it is difficult to combine air freight cargo information, flight resource information, and flight path weather information in the prior art to manage the air freight logistics transportation, which may lead to the stagnation of air freight logistics.

[0058] Preferably, the process of the air freight logistics transportation data acquisition module collecting air freight cargo data, initial flight data, and flight weather data includes:

[0059] Deploy different types of acquisition devices to collect air freight cargo data, initial flight data, and flight weather data. Among them, the acquisition devices include electronic platform scales, laser rangefinders, fuel density meters, volumetric flow meters, weighing scales, optical rain sensors, laser dust sensors, piezoresistive pressure sensors, etc.;

[0060] Using an electronic platform scale, based on the principle of pressure and gravity conversion, collect the weight of air cargo and the weight of the aircraft; through a laser rangefinder, collect the length, width, and height of the air cargo respectively, and calculate the volume of the air cargo according to the volume calculation formula.

[0061] Use a fuel density meter and a volumetric flow meter to collect the fuel density and fuel volume respectively, and calculate the fuel weight according to the volume calculation formula; use a weighing scale to collect the weight of the crew, and obtain the initial load of the flight by calculating the sum of the aircraft weight, fuel weight, and crew weight; combine the laser rangefinder with Leica Cyclone software to collect the initial volume of the flight.

[0062] Through an optical rain sensor, based on the optical principle, collect the real-time rainfall in the flight environment; a laser dust sensor, based on the principle of laser scattering, collect the real-time dust concentration in the flight environment; use a piezoresistive pressure sensor, based on the piezoresistive effect, collect the real-time air pressure in the flight environment.

[0063] Preferably, the process of preprocessing the collected data by the environmental level division module includes:

[0064] Perform data cleaning and data calibration on the collected weight of air cargo, volume of air cargo, initial load of the flight, initial volume of the flight, real-time rainfall in the flight environment, real-time dust concentration, and real-time air pressure.

[0065] Add time stamps to the weight of air cargo, volume of air cargo, initial load of the flight, and initial volume of the flight, uniformly convert the time stamps to UTC time, and synchronize the collection times of the weight of air cargo and the initial load of the flight and the collection times of the volume of air cargo and the initial volume of the flight respectively through time correction.

[0066] Preferably, the process of obtaining the maximum cargo weight of the flight and the maximum cargo volume of the flight and respectively dividing the levels of rainfall intensity, dust concentration, and air pressure by the environmental level division module includes:

[0067] Set the maximum takeoff load of the flight, calculate the difference between the maximum takeoff load of the flight and the initial load of the flight, and obtain the maximum cargo weight of the flight; set the maximum takeoff cargo volume of the flight, calculate the difference between the maximum takeoff cargo volume of the flight and the initial volume of the flight, and obtain the maximum cargo volume of the flight.

[0068] Statistically sum up the real-time rainfall amount in the flight environment per hour to obtain the total hourly rainfall amount in the flight environment. According to the total hourly rainfall amount in the flight environment, divide the rainfall intensity levels. The total hourly rainfall amount in the flight environment is divided into low rainfall intensity, medium rainfall intensity, and high rainfall intensity. Among them, low rainfall intensity means that the total hourly rainfall amount in the flight environment is less than 2.5 mm; medium rainfall intensity means that the total hourly rainfall amount in the flight environment is between 2.5 mm and 8.0 mm; high rainfall intensity means that the total hourly rainfall amount in the flight environment is higher than 8.0 mm.

[0069] According to the real-time dust concentration in the flight environment, divide the dust concentration levels. The real-time dust concentration in the flight environment is divided into mild dust, moderate dust, and severe dust. Among them, mild dust means that the real-time dust concentration in the flight environment is less than 1000 μg / m³; moderate dust means that the real-time dust concentration in the flight environment is between 1000 and 2000 μg / m³; severe dust means that the real-time dust concentration in the flight environment is higher than 2000 μg / m³.

[0070] According to the real-time air pressure in the flight environment, divide the real-time air pressure levels. The real-time air pressure in the flight environment is divided into low air pressure, medium air pressure, and high air pressure. Among them, low air pressure means that the real-time air pressure in the flight environment is less than 979 hPa; medium air pressure means that the real-time air pressure in the flight environment is between 979 hPa and 1031 hPa; high air pressure means that the real-time air pressure in the flight environment is higher than 1031 hPa.

[0071] Preferably, for the flight available cargo weight monitoring unit, the construction process of the flight available cargo weight monitoring model includes:

[0072] Take the air freight weight and the maximum cargo weight of the flight as the first data set, divide it into a training set and a test set according to the ratio of 8:2. Combine the training set data with the multiple linear regression algorithm. Take the air freight weight and the maximum cargo weight of the flight as the input, and the remaining cargo weight of the flight as the output. Adjust the intercept term, the regression coefficient of the air freight weight, and the regression coefficient of the maximum cargo weight of the flight respectively to train the flight available cargo weight monitoring model.

[0073] Input the test set data into the flight available cargo weight monitoring model, evaluate the performance of the flight available cargo weight monitoring model, adjust the parameters of the flight available cargo weight monitoring model, optimize the flight available cargo weight monitoring model, and deploy the optimized flight available cargo weight monitoring model into the system to obtain the final flight available cargo weight monitoring model.

[0074] Preferably, for the flight available cargo volume monitoring unit, the construction process of the flight available cargo volume monitoring model includes:

[0075] Taking the air cargo volume and the maximum load capacity of the flight as the second data set, dividing it into a training set and a test set according to the ratio of 8:2. Combining the training set data with the multiple linear regression algorithm, using the air cargo volume and the maximum load capacity of the flight as inputs and the remaining load volume of the flight as the output, training the monitoring model for the loadable cargo volume of the flight by adjusting the intercept term, the regression coefficient of the air cargo volume, and the regression coefficient of the maximum load capacity of the flight;

[0076] Inputting the test set data into the monitoring model for the loadable cargo volume of the flight, evaluating the performance of the monitoring model for the loadable cargo volume of the flight, adjusting the parameters of the monitoring model for the loadable cargo volume of the flight, optimizing the monitoring model for the loadable cargo volume of the flight, and deploying the optimized monitoring model for the loadable cargo volume of the flight into the system to obtain the final monitoring model for the loadable cargo volume of the flight.

[0077] Preferably, for the resource combination monitoring unit, the construction process of the monitoring model for the loadable cargo resources of the flight includes:

[0078] Inputting the air cargo weight and the maximum load weight of the flight into the monitoring model for the loadable cargo weight of the flight, and the monitoring model for the loadable cargo weight of the flight outputs the remaining load weight of the flight; inputting the air cargo volume and the maximum load capacity of the flight into the monitoring model for the loadable cargo volume of the flight, and the monitoring model for the loadable cargo volume of the flight outputs the remaining load volume of the flight;

[0079] Taking the air cargo weight, the air cargo volume, the remaining load weight of the flight, and the remaining load volume of the flight as the third data set, dividing it into a training set and a test set according to the ratio of 7:3. Using the training set data and the neural network algorithm, taking the air cargo weight and the air cargo volume as inputs and the remaining load weight of the flight and the remaining load volume of the flight as outputs, calculating the predicted output data through forward propagation, and updating the weights and biases of the model using backpropagation to learn the non - linear relationship between the air cargo weight and the remaining load weight of the flight and the non - linear relationship between the air cargo volume and the remaining load volume of the flight, training the monitoring model for the loadable cargo resources of the flight;

[0080] Inputting the test set data into the monitoring model for the loadable cargo resources of the flight, using the MSE function to evaluate the errors between the output results of the monitoring model for the loadable cargo resources of the flight and the actual remaining load weight and remaining load volume of the flight respectively. According to the evaluation results, adjusting the parameters of the monitoring model for the loadable cargo resources of the flight, optimizing the performance of the monitoring model for the loadable cargo resources of the flight, and deploying the optimized monitoring model for the loadable cargo resources of the flight into the system to obtain the final monitoring model for the loadable cargo resources of the flight.

[0081] Preferably, for the flight path planning module, the construction process of the flight path planning model includes:

[0082] Using the dynamic programming algorithm, define the dynamic programming state as a triple, which consists of a state point, a time point, and a flight weather level. Divide the flight route into N small segments of the route, extract the route nodes from each small segment of the route, use the route nodes as the state points in the dynamic programming, record the corresponding time points of the state points and the flight weather levels. The flight weather levels include rainfall intensity level, sand and dust level, and air pressure level. Construct a cost function based on the flight weather levels and the flight time;

[0083] Initialize the triple, set the cost of the initial state point to 0, and set the costs of other state points to infinity. Starting from the initial state point, in the order of time and the order of the route, sequentially obtain all possible transfer points of each state point, obtain all possible transfer paths of each state point. For all possible transfer points of each state point, calculate the transfer cost using the cost function, compare the sizes of the transfer costs of all possible transfer paths of each state point, select the minimum transfer cost, optimize the cost function using the minimum transfer cost, and obtain the optimal transfer path of each state point by combining the possible transfer paths corresponding to the minimum transfer costs of each state point. According to the optimal transfer path, plan the flight path of the flight and obtain the flight planning decision;

[0084] Using the dynamic programming algorithm, combine the triple, the optimized cost function, and the flight planning decision to obtain a flight path planning model, and deploy the flight path planning model to the system to obtain the final flight path planning model.

[0085] Preferably, the process of optimizing the flight path planning model by the flight path planning module in combination with the genetic algorithm and the simulated annealing algorithm includes:

[0086] A1. Extract each state point and its corresponding flight planning decision in the triple of the flight path planning model. Using the genetic algorithm, perform real-number encoding on each state point in the triple, and the encoding of each state point corresponds to the flight planning decision to obtain a real-number vector. Use the real-number vector to represent an individual, randomly generate P individuals, combine the P individuals to obtain an initial population, use the reciprocal of the mean square error as the fitness function, and set the number of iterations and the fitness threshold;

[0087] According to the fitness values of each individual in the population, through the roulette wheel selection method, the greater the fitness value of each individual, the greater the probability of selecting this individual. Select individuals from the population to participate in the reproduction of the next generation; use the arithmetic crossover method to perform gene exchange on the selected individuals; randomly select the flight planning decision corresponding to the individual after gene exchange, randomly change the flight planning decision corresponding to this individual, and repeat the iterative operations of selection, crossover, and mutation until the number of repeated iterative operations reaches the set number of iterations and the fitness values of the individuals in the population reach the fitness threshold, then terminate the repeated iterative operation, select the individual with the highest fitness value in the population, and use this individual with the highest fitness value to update each state point and its corresponding flight planning decision in the flight path planning model to optimize the flight path planning model.

[0088] A2. Use the flight planning decision output by the optimized flight path planning model as the initial solution of the simulated annealing algorithm, set the number of iterations for simulated annealing optimization, define the neighborhood structure, make a small adjustment to the initial solution to generate a neighborhood solution, use the cost function in the flight path planning model as the objective function, calculate the objective function value of the initial solution and the objective function value of the neighborhood solution, compare the objective function value of the initial solution and the objective function value of the neighborhood solution. When the objective function value of the neighborhood solution is better than the objective function value of the initial solution, use the neighborhood solution as the current solution; when the objective function value of the initial solution is better than the objective function value of the neighborhood solution, according to the probability formula of the simulated annealing algorithm, determine the situation of using the neighborhood solution as the current solution, and then make a small adjustment to the current solution to generate a new neighborhood solution. Use the current solution and the new neighborhood solution to repeat the above iterative process of objective function comparison. When the actual number of iterations reaches the number of iterations for simulated annealing optimization, stop the iterative process of the simulated annealing algorithm to optimize the flight path planning model.

[0089] Preferably, the process of the air freight logistics transportation management module combining the flight available cargo resource monitoring model and the flight path planning model to manage air freight logistics transportation includes:

[0090] Input the air freight weight and air freight volume into the flight available cargo resource monitoring model, and the flight available cargo resource monitoring model outputs the remaining cargo weight of the flight and the remaining cargo volume of the flight correspondingly;

[0091] According to the remaining cargo weight of the flight, when the air freight weight is lower than and equal to the remaining cargo weight of the flight, include all of this air freight in the corresponding flight for transportation; when the air freight weight is higher than the remaining cargo weight of the flight, screen out the air freight with a weight lower than the remaining cargo weight of the flight, give priority to arranging the screened air freight to be transported by this flight, and arrange the unscreened air freight to the next flight for transportation;

[0092] According to the remaining cargo volume of the flight, when the volume of the air cargo is lower than or equal to the remaining cargo volume of the flight, all of the air cargo will be included in the corresponding flight for transportation; when the volume of the air cargo is higher than the remaining cargo volume of the flight, the air cargo with a volume lower than the remaining cargo volume of the flight will be screened out, and the screened-out air cargo will be preferentially arranged to be transported by this flight, and the unscreened air cargo will be arranged for the next flight for transportation;

[0093] Dispatch the flight route planning model. According to the flight weather level and flight time of the flight, obtain the transfer path corresponding to the minimum transfer cost for each leg of the voyage, obtain the optimal transfer path for each leg of the voyage, combine the optimal transfer paths for each leg of the voyage, plan the flight path of the flight, obtain the flight planning decision, and conduct air cargo logistics transportation according to this flight planning decision.

[0094] First, through an electronic weighbridge, a laser rangefinder, a fuel density meter, a volumetric flowmeter, a weighing scale, an optical rain sensor, a laser dust sensor, and a piezoresistive pressure sensor, the weight of the air cargo, the volume of the air cargo, the initial load of the flight, the initial volume of the flight, as well as the real-time rainfall, real-time dust concentration, and real-time air pressure of the flight environment were collected; secondly, data cleaning, data standardization, and synchronization processing were performed on the collected data. The maximum takeoff load of the flight and the maximum takeoff cargo volume of the flight were set, and the difference between the maximum takeoff load of the flight and the initial load of the flight and the difference between the maximum takeoff cargo volume of the flight and the initial volume of the flight were calculated respectively to obtain the maximum cargo weight of the flight and the maximum cargo volume of the flight, and the rainfall intensity, dust, and air pressure levels were divided; immediately afterwards, using the multiple linear regression algorithm, a monitoring model for the weight of the cargo that the flight can carry and a monitoring model for the volume of the cargo that the flight can carry were respectively constructed; then, through the neural network algorithm, the monitoring model for the weight of the cargo that the flight can carry and the monitoring model for the volume of the cargo that the flight can carry were combined to construct a monitoring model for the cargo resources that the flight can carry; then, combining the dynamic programming algorithm, the genetic algorithm, and the simulated annealing algorithm, and comprehensively considering the flight weather data, an optimal flight path was planned for the air cargo logistics transportation flight, and a flight route planning model was constructed and optimized. Finally, the air cargo logistics transportation was managed by combining the monitoring model for the cargo resources that the flight can carry and the flight route planning model.

[0095] As described above, it is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art in the technical field disclosed in this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claimed rights.

Claims

1. An air logistics transportation management system based on intelligent algorithm optimization, including an air logistics transportation data collection module, an air logistics transportation data processing module and an air logistics transportation management module, wherein: Each module is connected in communication, characterized by: The air cargo logistics data collection module collects air cargo data, initial flight data and flight weather data; The air cargo logistics data processing module is composed of an environment classification module, a flight cargo resource monitoring module and a flight path planning module; The environmental grade classification module pre-processes the collected data, obtains the maximum cargo weight and the maximum cargo volume of the flight, and classifies the rainfall intensity, dust concentration and air pressure into grades; The flight cargo resource monitoring module is divided into a flight cargo weight monitoring unit, a flight cargo volume monitoring unit and a resource combination monitoring unit, wherein the flight cargo weight monitoring unit is used to construct a flight cargo weight monitoring model; the flight cargo volume monitoring unit is used to construct a flight cargo volume monitoring model; the resource combination monitoring unit is used to construct a flight cargo resource monitoring model through the output results of the flight cargo weight monitoring model and the flight cargo volume monitoring model; The flight path planning module constructs a flight path planning model based on flight weather data, and optimizes the flight path planning model by combining genetic algorithm and simulated annealing algorithm; The air cargo logistics management module manages air cargo logistics by combining a flight cargo resource monitoring model and a flight path planning model.

2. The air transport logistics management system based on intelligent algorithm optimization according to claim 1 is characterized by: The process of collecting air cargo data, initial flight data and flight weather data by the air cargo logistics transportation data collection module includes: Deploy different types of collection equipment to collect air cargo data, initial flight data, and flight weather data, wherein the collection equipment includes electronic floor scales, laser rangefinders, fuel density meters, volume flow meters, body weight scales, optical rain sensors, laser dust sensors, and piezoresistive air pressure sensors; The air cargo data includes the air cargo weight and the air cargo volume; the initial flight data includes the flight initial load and the flight initial volume; the flight weather data includes the real-time rainfall, real-time dust concentration and real-time air pressure of the flight environment.

3. The air transport logistics management system based on intelligent algorithm optimization according to claim 2 is characterized by: The process of preprocessing the collected data by the environment level classification module includes: Clean and calibrate the collected air cargo weight, air cargo volume, flight initial load, flight initial volume, flight environment real-time rainfall, real-time dust concentration and real-time air pressure data; Add timestamps for air cargo weight, air cargo volume, flight initial load and flight initial volume, convert the timestamps into UTC time, and synchronize the collection time of air cargo weight and flight initial load as well as the collection time of air cargo volume and flight initial volume through time correction.

4. The air transport logistics management system based on intelligent algorithm optimization according to claim 3 is characterized by: The process of the environmental grade classification module obtaining the maximum cargo weight and the maximum cargo volume of the flight and respectively classifying the levels of rainfall intensity, dust concentration and air pressure includes: Set the maximum takeoff load of the flight, calculate the difference between the maximum takeoff load of the flight and the initial load of the flight, and obtain the maximum cargo weight of the flight; set the maximum takeoff cargo volume of the flight, calculate the difference between the maximum takeoff cargo volume of the flight and the initial volume of the flight, and obtain the maximum cargo volume of the flight; Count the total real-time rainfall in the flight environment every hour, obtain the total rainfall per hour in the flight environment, divide the rainfall intensity levels according to the total rainfall per hour in the flight environment, and divide the total rainfall per hour in the flight environment into low rainfall intensity, medium rainfall intensity and high rainfall intensity; According to the real-time dust concentration in the flight environment, the dust concentration level is divided into light dust, moderate dust and heavy dust; According to the real-time air pressure of the flight environment, the real-time air pressure levels are divided, and the real-time air pressure of the flight environment is divided into low pressure, medium pressure and high pressure.

5. The air transport logistics management system based on intelligent algorithm optimization according to claim 4 is characterized by: The flight cargo weight monitoring unit and the flight cargo weight monitoring model construction process include: The air cargo weight and the maximum cargo weight of the flight are used as the first data set, which are divided into a training set and a test set in a ratio of 8:

2. The training set data is combined with the multivariate linear regression algorithm, and the air cargo weight and the maximum cargo weight of the flight are used as inputs, and the remaining cargo weight of the flight is used as output. The intercept term, the regression coefficient of the air cargo weight, and the regression coefficient of the maximum cargo weight of the flight are adjusted respectively to train the flight cargo weight monitoring model. Input the test set data into the flight cargo weight monitoring model, evaluate the performance of the flight cargo weight monitoring model, adjust the flight cargo weight monitoring model parameters, optimize the flight cargo weight monitoring model, and deploy the optimized flight cargo weight monitoring model into the system.

6. The air transport logistics management system based on intelligent algorithm optimization according to claim 5 is characterized by: The flight cargo volume monitoring unit and the flight cargo volume monitoring model construction process include: The air cargo volume and the maximum cargo volume of the flight are used as the second data set, divided into a training set and a test set in a ratio of 8:

2. The training set data is combined with the multivariate linear regression algorithm, the air cargo volume and the maximum cargo volume of the flight are used as input, and the remaining cargo volume of the flight is used as output. By adjusting the intercept term, the regression coefficient of the air cargo volume and the regression coefficient of the maximum cargo volume of the flight, the flight cargo volume monitoring model is trained; Input the test set data into the flight cargo volume monitoring model, evaluate the performance of the flight cargo volume monitoring model, adjust the parameters of the flight cargo volume monitoring model, optimize the flight cargo volume monitoring model, and deploy the optimized flight cargo volume monitoring model to the system.

7. The air transport logistics management system based on intelligent algorithm optimization according to claim 6 is characterized by: The resource combination monitoring unit, the process of constructing the flight cargo resource monitoring model includes: The weight of air cargo and the maximum cargo weight of the flight are input into the flight cargo weight monitoring model, and the flight cargo weight monitoring model outputs the remaining cargo weight of the flight; the volume of air cargo and the maximum cargo volume of the flight are input into the flight cargo volume monitoring model, and the flight cargo volume monitoring model outputs the remaining cargo volume of the flight; The air cargo weight, air cargo volume, flight remaining cargo weight and flight remaining cargo volume are used as the third data set, and divided into a training set and a test set in a ratio of 7:

3. The training set data and the neural network algorithm are used to take the air cargo weight and air cargo volume as input, and the flight remaining cargo weight and flight remaining cargo volume as output. The output data is predicted by forward propagation calculation, and the weight and bias of the model are updated by back propagation. The nonlinear relationship between the air cargo weight and the flight remaining cargo weight and the nonlinear relationship between the air cargo volume and the flight remaining cargo volume are learned, and the flight cargo resource monitoring model is trained. The test set data is input into the flight cargo resource monitoring model. The MSE function is used to evaluate the errors between the output results of the flight cargo resource monitoring model and the actual flight remaining cargo weight and flight remaining cargo volume. According to the evaluation results, the parameters of the flight cargo resource monitoring model are adjusted to optimize the performance of the flight cargo resource monitoring model. The optimized flight cargo resource monitoring model is deployed into the system.

8. The air transport logistics management system based on intelligent algorithm optimization according to claim 7 is characterized by: The flight path planning module and the construction process of the flight path planning model include: Using the dynamic programming algorithm, the dynamic programming state is defined as a triplet consisting of a state point, a time point, and a flight weather level. The flight route is divided into N small segments, and a route node is extracted from each small segment. The route node is used as the state point in the dynamic programming. The time point corresponding to the state point and the flight weather level are recorded. The flight weather level includes the rainfall intensity level, the dust level, and the air pressure level. According to the flight weather level and the flight time, a cost function is constructed; Initialize the triples, set the initial state point cost to 0, set the costs of other state points to infinity, start from the initial state point, and obtain all possible transfer points of each state point in chronological order and flight order, obtain all possible transfer paths of each state point, calculate the transfer cost for all possible transfer points of each state point using the cost function, compare the transfer cost of all possible transfer paths of each state point, select the minimum transfer cost, optimize the cost function using the minimum transfer cost, obtain the optimal transfer path for each state point by combining the possible transfer paths corresponding to the minimum transfer cost of each state point, plan the flight path according to the optimal transfer path, and obtain the flight planning decision; The dynamic programming algorithm is used to combine the triples, the optimized cost function and the flight planning decision to obtain the flight path planning model, which is then deployed into the system to obtain the final flight path planning model.

9. The air transport logistics management system based on intelligent algorithm optimization according to claim 8 is characterized by: The flight path planning module combines the genetic algorithm with the simulated annealing algorithm to optimize the flight path planning model, including: A1. Extract each state point and its corresponding flight planning decision in the triple of the flight path planning model, use the genetic algorithm to encode each state point in the triple, and the encoding of each state point corresponds to the flight planning decision, obtain a real number vector, use the real number vector to represent the individual, randomly generate P individuals, combine the P individuals, obtain the initial population, use the inverse of the mean square error as the fitness function, and set the number of iterations and the fitness threshold; According to the fitness value of each individual in the population, the roulette wheel selection method is used, and the individuals in the population are selected to participate in the next generation of reproduction according to the fact that the larger the fitness value of each individual, the greater the probability of the individual being selected; the selected individuals are subjected to gene exchange using the arithmetic crossover method; the flight planning decision corresponding to the individual after the gene exchange is randomly selected, and the flight planning decision corresponding to the individual is randomly changed, and the iterative operations of selection, crossover and mutation are repeated until the number of repeated iterative operations reaches the set number of iterations and the fitness value of the individual in the population reaches the fitness threshold, the repeated iterative operation is terminated, and the individual with the highest fitness value in the population is selected, and each state point and its corresponding flight planning decision in the flight path planning model are updated using the individual with the highest fitness value, so as to optimize the flight path planning model; A2. The flight planning decision output by the optimized flight path planning model is used as the initial solution of the simulated annealing algorithm. The number of simulated annealing optimization iterations is set, the domain structure is defined, the initial solution is slightly adjusted, and a domain solution is generated. The cost function in the flight path planning model is used as the objective function. The objective function value of the initial solution and the objective function value of the domain solution are calculated and compared. When the objective function value of the domain solution is better than the objective function value of the initial solution, the domain solution is used as the current solution. When the objective function value of the initial solution is better than the objective function value of the domain solution, the domain solution is determined as the current solution according to the probability formula of the simulated annealing algorithm, and the current solution is slightly adjusted to generate a new domain solution. The above-mentioned iterative process of objective function comparison is repeated using the current solution and the new domain solution. When the actual number of iterations reaches the number of simulated annealing optimization iterations, the iterative process of the simulated annealing algorithm is stopped to obtain the optimized flight path planning model.

10. The air transport logistics management system based on intelligent algorithm optimization according to claim 9, characterized in that: The air logistics transportation management module combines the flight cargo resource monitoring model and the flight path planning model to manage air logistics transportation, including: The air cargo weight and air cargo volume are input into the flight cargo resource monitoring model, and the flight cargo resource monitoring model accordingly outputs the flight remaining cargo weight and the flight remaining cargo volume; According to the remaining cargo weight of the flight, when the weight of the air cargo is lower than or equal to the remaining cargo weight of the flight, all the air cargo will be included in the corresponding flight for transportation; when the weight of the air cargo is higher than the remaining cargo weight of the flight, the air cargo with a weight lower than the remaining cargo weight of the flight will be screened out, and the screened air cargo will be given priority for transportation through the flight, and the air cargo that has not been screened will be arranged to the next flight for transportation; According to the remaining cargo volume of the flight, when the air cargo volume is lower than or equal to the remaining cargo volume of the flight, all the air cargo will be included in the corresponding flight for transportation; when the air cargo volume is higher than the remaining cargo volume of the flight, the air cargo with a volume lower than the remaining cargo volume of the flight will be screened out, and the screened air cargo will be given priority for transportation through the flight, and the air cargo that has not been screened will be arranged to the next flight for transportation; The flight path planning model is used to schedule flights. According to the flight weather level and flight time, the transfer path corresponding to the minimum transfer cost of each flight segment is obtained, and the optimal transfer path for each flight segment is obtained. Combined with the optimal transfer path for each flight segment, the flight path is planned, the flight planning decision is obtained, and air logistics transportation is carried out according to the flight planning decision.

Images