Civil aviation cargo ship intelligent stowage system based on multi-task model

Through the intelligent loading system of civil aviation cargo aircraft based on multi-task model, the use of constrained multi-objective optimization algorithm to optimize cargo loading, the problem that manual loading is difficult to ensure optimal profits is solved, and profit maximization and cost reduction in the field of aviation logistics are achieved.

CN119090042APending Publication Date: 2024-12-06GUILIN UNIVERSITY OF TECHNOLOGY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410976622.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-21
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The manual loading method in the existing aviation logistics field is difficult to ensure that the profit is approximately optimal or optimal, and cannot meet the current needs of the aviation logistics industry, especially when the scale is expanded and the variety is increased.

Method used

Adopt a civil aviation cargo aircraft intelligent loading system based on multi-task model, and use a constrained multi-objective optimization algorithm (such as EMCMO) combined with an evolutionary multi-task model to optimize the cargo loading scheme, considering the total cost and maximum cargo quantity, while meeting the weight, Volume and special cargo loading constraints are coordinated by the EMCMO algorithm to achieve an optimal or approximately optimal loading scheme.

Benefits of technology

Obtain better cargo loading solutions in the shortest possible time to maximize profits, and increase profits by at least 10% compared to manual loading methods, reducing costs and improving loading efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119090042A_ABST
    Figure CN119090042A_ABST
Patent Text Reader

Abstract

The invention discloses a civil aviation cargo ship intelligent stowage system based on a multi-task model. At present, the mainstream mode of domestic aviation logistics stowage is manual stowage, employees carry out cargo stowage according to own experience, certain flexibility is achieved, however, efficiency, safety and cost cannot be compared with intelligent stowage, and profit difference cannot be maximized. According to the invention, a civil aviation cargo and cargo stowage problem is mainly converted into a constraint multi-objective optimization problem based on a multi-task model, and an intelligent stowage system of the civil aviation cargo and cargo is designed. According to the system, the total cost and the number of goods serve as optimization objectives, meanwhile, factors such as the weight and volume limit of the goods in each flight, the remaining transport capacity and the special goods loading requirement are considered, and a main task model (including constraints) and an auxiliary task model (without constraints) are constructed. Mathematical model modeling is carried out by using a matlab technology, optimal stowage is given through a distribution scheme pushed by intelligent optimization of the system, and finally overall benefit maximization is realized.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of intelligent optimization. According to the actual needs of air logistics, an intelligent loading system for civil aviation cargo aircraft based on a multi-task model is designed, that is, the optimal cargo loading plan is obtained under the premise of meeting the shortest possible time limit, thereby ensuring the maximum profit of the air logistics company. Background Art

[0002] At present, in the field of aviation logistics, manual loading is basically adopted. The manual method mainly optimizes the loading of logistics based on personal experience. With the expansion of scale, the increase of types and the intensification of industry competition, this manual loading method is difficult to guarantee the approximate or optimal profit and it is difficult to meet the needs of the current aviation logistics industry. Based on a detailed investigation of aviation logistics companies, we have developed an intelligent loading system for civil aviation cargo aircraft based on a multi-task model. Compared with the manual allocation method, it can obtain a better allocation plan in the shortest possible time and ultimately maximize profits. Preliminary calculations using the data of an aviation logistics company show that the profit can be increased by at least 10% compared with manual loading. With the globalization of the economy and the rapid growth of aviation logistics business, there is a huge profit space. Intelligent loading of aviation logistics involves relevant information about the goods provided by customers, the capacity resources purchased by aviation logistics companies, and the transportation information of airline flights. The system mainly considers the minimization of total cost and the maximization of the number of goods, where the total cost includes air freight cost, port of departure transshipment cost and port of destination transshipment cost, while considering the weight and volume of the goods in each flight. The constraints and the remaining capacity and the loading requirements of special goods cannot exceed the limit. According to different actual needs, the allocation plan obtained through system intelligent optimization gives the optimal loading and ultimately maximizes the overall benefits. Summary of the invention

[0003] The implementation of the present invention is mainly divided into two parts: actual demand modeling and algorithm optimization.

[0004] 1. Actual demand modeling

[0005] The intelligent loading problem of civil aviation cargo planes is similar to the 0-1 knapsack problem. The current product considers loading as much cargo as possible at the lowest cost, including air freight cost, transit cost at the departure port, and transit cost at the destination port. At the same time, it considers the weight and volume of cargo in each flight that cannot exceed the limit, the remaining capacity, and the loading requirements of special cargo. Of course, if there are other additional or unnecessary requirements, they can also be flexibly added or deleted through the interface.

[0006] 2. Algorithm optimization

[0007] Intelligent loading optimization of civil aviation cargo aircraft is a constrained multi-objective optimization problem. The various objectives are mutually constrained, which may make the improvement of one objective performance often at the expense of other objective performance. It is impossible to have a solution that makes all objective performance reach the optimal level. Therefore, for multi-objective optimization problems, the solution is usually a set of non-inferior solutions - Pareto solution set. At the same time, the existence of constraints makes the solution set divided into feasible solution set (solution set that satisfies the constraints) and infeasible solution set (solution set that does not satisfy the constraints), which increases the difficulty in overall optimization and the exploration of the final solution set. Constrained multi-objective evolutionary algorithm is an algorithm used to analyze and solve constrained multi-objective optimization problems. Its core is to coordinate the relationship between various objective functions while satisfying the constraints, and find the optimal solution set that makes each objective function reach a relatively large (or relatively small) function value as much as possible. Among many constrained multi-objective evolutionary algorithms, the evolutionary multitasking model (Evolutionary Multitasking—EMT) is a relatively new model. The basic evolutionary algorithm can only solve one task in one run. EMT transfers potential advantage solutions between different tasks to solve multiple related optimization tasks at the same time. The EMCMO algorithm (Evolutionary Multitasking based Constrained Multiobjective Optimization Algorithm) is the first application of EMT in constrained multi-objective optimization problems. The main advantages of this algorithm are as follows:

[0008] 1. An auxiliary task is constructed. The existence of constraints exacerbates the difficulty of solving constrained multi-objective optimization problems. The more constraints there are, the greater the difficulty; the fewer constraints there are, the less difficult it is. EMCMO reduces the number of constraints in the auxiliary task and simplifies the optimization difficulty of the auxiliary task, which can help solve the original problem.

[0009] 2. It realizes the interaction of multiple information. Knowledge transfer is the core part of EMT, and its quality directly affects the final result. EMCMO selects different information for interaction according to different situations, effectively improving the diversity of the population, thus helping to get out of the local optimal area.

[0010] 3. Multiple mating methods. EMCMO uses different mating methods based on the distribution of populations in the early and late stages of evolution, improving diversity in the early stages of evolution and feasibility in the late stages of evolution.

[0011] Therefore, the EMCMO algorithm is more suitable as a solver for intelligent loading optimization of civil aviation cargo aircraft. Of course, the characteristics of the models built according to different actual needs are different, and a more targeted multi-objective evolutionary algorithm can be selected for solving. In order to solve faster, a simple infeasible solution repair method is proposed in the optimization process, so that the population can enter the feasible domain as soon as possible, and at the same time, it can also jump out of the local optimum.

[0012] At present, domestic air logistics mainly adopts manual loading, which is mainly based on personal experience. It is difficult to ensure that the loading is approximately optimal or optimal. Relatively speaking, intelligent loading optimization can ensure that the optimal or approximately optimal is achieved in a shorter time, which not only saves time, but also reduces costs to a certain extent.

[0013] The advantages of the present invention are:

[0014] At present, the domestic aviation logistics field mainly adopts manual loading, which is mainly based on personal experience. It is difficult to ensure that the loading is approximately optimal or optimal. Relatively speaking, intelligent loading optimization can ensure that the optimal or approximately optimal is achieved in a shorter time, which not only saves time, but also reduces costs to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the loading optimization process.

[0016] Figure 2 It is a schematic diagram of the modeling process.

[0017] Figure 3 This is a schematic diagram of the solution process DETAILED DESCRIPTION

[0018] The present invention designs an intelligent loading system for civil aviation cargo aircraft based on a multi-task model. The core design of the present invention is to construct multiple task models, and multiple models collaborate to solve the same problem. According to the cargo-related information provided by the customer, the transportation resources purchased by the aviation logistics company, and the transportation information of the airline's flights, the minimization of the total cost and the maximization of the quantity of cargo are taken as two optimization goals. At the same time, the weight and volume of the cargo in each flight cannot exceed the flight limit, and the remaining transportation capacity and the loading requirements of special cargo are taken as constraints to construct a constrained multi-objective optimization model, which is set as the main task, and then the constraints are ignored to construct an unconstrained multi-objective optimization model, which is set as an auxiliary task. Finally, the EMCMO algorithm is used as the solver of the intelligent loading system for civil aviation cargo aircraft based on the multi-task model to optimize the two objective functions.

[0019] The detailed steps are as follows:

[0020] Step 1: Input Data

[0021] Convert the data in Excel to .mat file and read it into the matlab buffer.

[0022] Step 2: Modeling

[0023] The goods loaded in the crate are considered 1, and those not loaded in the crate are considered 0. Assuming there are 3 goods, named a, b, and c; 2 crates, named A and B, then 10 can represent a certain goods loaded in crate A. There are eight loading schemes: 101010, 101001, 100110, 100101, 011010, 011001, 010110, and 010101 (each goods can only be loaded in one crate). At the same time, there is a requirement that goods can only be loaded in certain crates, so the goods that can be loaded in the crate are considered 1, and those that cannot be loaded are considered 0. For example, if there are 101011 goods (i.e. a and b can only be loaded in A, while c can be loaded in A and B), then after multiplication, eight loading schemes will be obtained based on the above conditions, namely 101010, 101001, 100010, 100001, 001010, 001001, 000010, 000001. The model contains two objective functions and multiple constraints. The objective functions are total cost (total cost includes transshipment cost at the departure port, transshipment cost at the destination port, and air freight cost. When the loading weight of the flight without an agreement does not reach the specified weight, a corresponding fine will be incurred) and the loading quantity. The constraints are the weight and volume constraints of each box and the loading quantity constraints. The goal of the solution is to obtain the solution with the minimum total cost and the maximum loading quantity under the premise of satisfying the constraints.

[0024] Step 3: Solve

[0025] The EMCMO algorithm is used for optimization.

[0026] In the process of generating solutions, a large number of infeasible solutions may be generated. In order to solve the problem faster, an infeasible solution repair method is proposed. If the generated solution is infeasible, first calculate the weight excess ratio Rw and volume excess ratio Rv of the infeasible solution. If Rw is greater than Rv, first load the goods according to the weight sorting of the goods. If the weight of the goods meets the requirements, then load the goods according to the volume sorting of the goods. If Rv is greater than Rw, first sort by volume and then unload by weight. Then, check the loading status of the goods. If there are goods loaded in multiple boxes at the same time, randomly select a box to unload the goods until the goods are loaded in only one box. If a certain cargo is not loaded, load the cargo according to the air freight price of each box of the cargo. If all boxes cannot be loaded, randomly unload the loaded cargo from the boxes that can be loaded with the cargo until the cargo is loaded. When all the cargo has been checked, the infeasible solution repair process ends. If the cargo meets the loading requirements, the infeasible solution repair process can be jumped out in advance.

Claims

1. The intelligent loading system for civil aviation cargo aircraft based on multi-task model is characterized by: Construct multiple task models, and use them to collaborate to solve the same problem. Based on the cargo-related information provided by the customer, the transportation resources purchased by the air logistics company, and the transportation information of the airline's flights, the total cost and the quantity of cargo are used as optimization targets, and the cargo aircraft's load limit, size limit, remaining capacity, and special needs are used as constraints to construct a constrained multi-objective optimization model, which is set as the main task. Then, ignore the constraints to construct an unconstrained multi-objective optimization model, which is set as the auxiliary task. Finally, use the EMCMO algorithm as a solver for the intelligent loading system of civil aviation cargo aircraft based on the multi-task model to optimize the two objective functions.

2. The intelligent loading system for civil aviation cargo aircraft based on multi-task model according to claim 1 is characterized in that: The specific modeling steps are as follows: Step 1: Coding Set the value of 1 if the goods can be loaded into the crate, and 0 if they cannot be loaded into the crate; suppose there are 3 goods a, b, c, and 2 crates A and B; then all the relationships between goods being loaded into the crates can be expressed as eight relationships: 101010, 101001, 100110, 100101, 011010, 011001, 010110, 010101; for example, 101010 means that the three goods can only be loaded into crate A; Step 2: Dot product operation By performing dot multiplication operations through Matlab, eight loading schemes can be obtained based on the above conditions, namely 101010, 101001, 100010, 100001, 001010, 001001, 000010, 000001; Step 3: Determine the objective function and constraint function, and establish the main task model and auxiliary task model According to the actual situation, the objective functions of aviation logistics are established as total cost and cargo loading volume; the constraint functions are set as weight limit and volume limit of each box; the main task contains constraints, while the auxiliary task does not contain constraints; Step 4: Optimization solution Through MATLAB programming, the EMCMO constrained multi-objective optimization algorithm is introduced to obtain the solution with the minimum total cost and the maximum loading volume under the conditions of satisfying the above constraints; when all the goods have been checked, the infeasible solution repair process is terminated; if the goods meet the loading requirements, the infeasible solution repair process can be jumped out in advance.