A logistics distribution path optimization method and system

Abstract

This invention discloses a method and system for optimizing logistics delivery routes. First, it constructs a flight speed correction model and a payload increase model reflecting the impact of severe weather such as rainfall on the added weight of UAVs. Further, it establishes a flight power correction model, making energy consumption constraints more closely reflect the actual operating environment. Simultaneously, combining the flight speed and flight power correction models, and with minimizing logistics delivery time as the optimization objective, it constructs a route optimization function and a corresponding delivery energy consumption constraint function. Then, based on the aforementioned objective and constraint functions, it generates a logistics delivery optimization model. Finally, by solving the constructed logistics delivery optimization model, it obtains the optimal delivery plan that still ensures timeliness under complex weather conditions. This significantly improves the rationality and adaptability of UAV route planning under complex weather conditions, providing reliable technical support for intelligent and refined scheduling of logistics delivery.

Description

Technical Field

[0001] This invention belongs to the field of logistics distribution route optimization technology, specifically relating to a logistics distribution route optimization method and system. Background Technology

[0002] In recent years, with technological innovation and urban development, my country's logistics market has continued to expand and demand has surged. As a new favorite for low-altitude airspace logistics transportation, drones have a broad market background due to my country's huge logistics transportation market. Compared with traditional logistics transportation methods, drone logistics transportation and delivery have significant advantages, especially in solving the "last mile" delivery problem in rural and mountainous areas. They can avoid the limitations of traditional express logistics transportation routes, have higher transportation efficiency, and can reduce the consumption of human resources and time costs to a certain extent.

[0003] Currently, drone delivery typically optimizes delivery time based on flight speed (i.e., requiring delivery as fast as possible) and considers drone energy consumption as a key operational constraint. Existing drone path planning technologies usually rely on simple, ideal flight environments, using fixed flight speeds and energy consumption to construct objective and constraint functions. However, drones need to operate in more complex weather environments in the real world (such as in severe weather like strong winds and heavy rain, which can hinder drone flight or even cause safety accidents). Existing technologies neglect the impact of weather conditions on drone delivery operations and lack correlation analysis between weather conditions and drone performance and energy consumption, leading to unreasonable path planning (e.g., planning unsuitable routes for drones in complex weather environments such as rural and mountainous areas), making them unsuitable for applications in complex weather environments.

[0004] Therefore, based on the aforementioned shortcomings, there is an urgent need to provide an optimization method that incorporates weather factors into the logistics delivery path of drones, in order to solve the problem of unreasonable path planning caused by the lack of correlation analysis between weather conditions and drone operation performance and energy consumption. Summary of the Invention

[0005] The purpose of this invention is to provide a logistics delivery route optimization method and system to solve the problem of unreasonable route planning caused by the lack of correlation analysis between weather conditions and the performance and energy consumption of drones in the existing technology.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a method for optimizing logistics delivery routes is provided, including: Obtain logistics and delivery information, including the number of delivery stations and the number of drones; A flight speed correction model and a flight load increase model based on weather factors were constructed. Based on the flight load increase model, a flight power correction model was established. The flight load increase model was used to calculate the additional weight caused by rainfall. Based on the number of delivery stations and drones in the logistics delivery information, and the flight speed correction model, a path optimization function is constructed with minimizing logistics delivery time as the optimization objective. Based on the flight power correction model and the flight speed correction model, a delivery energy consumption constraint function for the path optimization function is established. By utilizing the path optimization function and the delivery energy consumption constraint function, a logistics delivery optimization model is constructed, and the logistics delivery optimization model is solved to obtain the optimal logistics delivery scheme that minimizes the logistics delivery time. The optimal logistics delivery scheme includes the optimal delivery station sequence for each drone.

[0007] Based on the aforementioned disclosures, this invention first constructs a flight speed correction model and a payload increase model reflecting the impact of severe weather such as rainfall on the added weight of the UAV, and further establishes a flight power correction model, thereby making the energy consumption constraints closer to the actual operating environment. Simultaneously, combining the flight speed correction model and the flight power correction model, and with minimizing logistics delivery time as the optimization objective, a path optimization function and a corresponding delivery energy consumption constraint function are constructed. Then, based on the aforementioned objective function and constraint function, a logistics delivery optimization model is generated. Finally, by solving the constructed logistics delivery optimization model, the optimal delivery solution that can still guarantee timeliness under complex weather conditions is obtained. Therefore, this invention solves the problem in the prior art of ignoring the correlation between weather conditions and UAV operational performance and energy consumption, significantly improving the rationality and adaptability of UAV path planning under complex weather conditions, effectively reducing flight risks and energy waste caused by weather factors, and providing reliable technical support for intelligent and refined scheduling of logistics delivery.

[0008] In one possible design, the logistics delivery information also includes: the drone's flight speed, wherein a flight speed correction model based on weather factors is constructed, including: For any two delivery stations, obtain the weather information of the airspace between the two delivery stations, wherein the weather information includes airspace wind speed and airspace wind direction angle; Based on the airspace wind direction angle in the weather information, the influence angle of the airspace wind speed on the flight heading of the UAV when flying between any two delivery stations is calculated. Based on the influence angle, the flight speed, and the airspace wind speed, a weather-based flight speed correction model is constructed between any two delivery stations.

[0009] In one possible design, based on the airspace wind direction angle in the weather information, the influence angle of the airspace wind speed on the flight heading of the UAV when flying between any two delivery stations is calculated, including: Obtain the heading angle of the drone when it flies between any two delivery stations; Calculate the angle difference between the heading angle and the airspace wind direction angle, and map the angle difference to the range [-π, π) to obtain the influence angle; Accordingly, based on the influence angle, the flight speed, and the airspace wind speed, a weather-based flight speed correction model is constructed between any two delivery stations, which includes: The flight speed correction model is constructed according to the following formula; ; In the formula, This indicates the corrected flight speed. These represent the flight speed of the drone and the wind speed in the airspace, respectively. This indicates the angle of influence.

[0010] In one possible design, a flight payload weight-increasing model is constructed, including: For any two delivery stations, based on the flight speed correction model, the corrected flight speed of the drone between the two delivery stations is obtained, and the flight time of the drone between the two delivery stations is determined based on the corrected flight speed. Obtain the windward area of ​​the drone, and the rainfall in the airspace between any two delivery stations; Based on the corrected flight speed, flight time, rainfall, and windward area, and using the following formula, a flight load increase model between any two delivery stations is constructed. ; In the formula, This indicates the additional weight of the drone due to rainfall. For the water accumulation efficiency of drones, Indicates rainfall amount, Indicates the precipitation coefficient. This indicates the corrected flight speed. This indicates the windward area. This represents the angle of attack of the drone when it flies between any two delivery stations. Indicates the flight time, This indicates the maximum set additional weight.

[0011] In one possible design, a flight power correction model is established, including: Establish a power influence function based on flight altitude and ambient temperature; The actual payload of the UAV is determined using the aforementioned flight payload weight-increasing model; Based on the power influence function and the actual load, the flight power correction model is constructed.

[0012] In one possible design, a power influence function based on flight altitude and ambient temperature is established, including: The power influence function is established according to the following formula; ; In the formula, This represents the power influence function. Indicates standard air density, Let h be the atmospheric pressure at which the drone is at its altitude. Represents the gas constant of air. The ambient temperature at flight altitude h. This indicates the factor affecting sensitivity. Indicates the temperature effect coefficient. Indicates reference temperature; Accordingly, based on the power influence function and the actual load, the flight power correction model is constructed, which includes: The flight power correction model is constructed according to the following formula; ; In the formula, This represents the corrected flight power of the drone when it is in the nth state. This represents the base power of the drone in the nth state. This represents the payload coefficient of the drone when it is in the nth state. This represents the actual payload of the drone in the nth state, where n=1 indicates the drone is in delivery mode, n=2 indicates the drone is in return mode, and n=3 indicates the drone is in unloading hovering mode. Furthermore, when the drone is in either delivery or unloading hovering mode... This is the sum of the total cargo volume and the additional weight of the drone due to rainfall, when the drone is on its return journey. This refers to the additional weight of the drone due to rainfall.

[0013] In one possible design, based on the flight power correction model and the flight speed correction model, a delivery energy consumption constraint function for the path optimization function is established, including: Establish a power penalty function based on rainfall; Based on the flight speed correction model and power penalty function, a delivery energy consumption constraint function is constructed.

[0014] In one possible design, a power penalty function based on rainfall is constructed, including: The power penalty function is constructed according to the following formula; ; In the formula, Denotes the power penalty function, This represents the set minimum power penalty value. Indicates rainfall amount, Indicates the impact coefficient of rainfall; Accordingly, based on the flight speed correction model and power penalty function, the delivery energy consumption constraint function is constructed, which includes: The delivery energy consumption constraint function is constructed according to the following formula; ; ; ; ; In the formula, This represents the total energy consumption for delivery by the k-th drone. This represents the delivery energy consumption of the k-th drone. This represents the return trip energy consumption of the k-th drone. This represents the energy consumption of the k-th drone during unloading and hovering. Let be the maximum available energy of the k-th drone, where This represents the corrected flight power of the k-th drone between delivery station i and delivery station j when the drone is in delivery mode. This represents the power penalty coefficient between delivery station i and delivery station j when the k-th drone is in delivery mode. This represents the flight time of the k-th drone from delivery station i to delivery station j. Let $k$ be the first decision variable, representing whether the kth drone flies from delivery station i to delivery station j. This represents the corrected flight power of the k-th drone between the last delivery station and the delivery center when it is on its return journey. This represents the power penalty coefficient for the drone between the last delivery station and the distribution center. This represents the flight time of the k-th drone from the last delivery station to the delivery center. This represents the corrected flight power of the k-th drone at delivery station i when it is in a loading and unloading hovering state. This represents the power penalty coefficient for the k-th drone at delivery station i. This represents the service time of the k-th drone at delivery station i. Let be the second decision variable, representing whether the k-th drone docks at delivery station i; The power penalty coefficient is calculated based on the power penalty function, and each flight time is calculated based on the corrected flight speed output by the flight speed correction model.

[0015] In a possible design, solve the logistics distribution optimization model to obtain the optimal logistics distribution scheme that minimizes the logistics distribution time, including: Obtain the individual population at the q-th iteration, where when q is 1, the individual population at the q-th iteration is the initial population, and each initial individual in the initial population corresponds to a logistics delivery scheme. Based on the logistics and distribution optimization model, the fitness of each individual in the population at the qth iteration is calculated. Based on the fitness of each individual, the globally optimal individual at the q-th iteration is determined; Determine if the iteration stopping condition is met; If not, then the individual population at the qth iteration will be divided into elite individuals, inferior individuals, and exploratory individuals; A multi-strategy update approach is adopted to update elite individuals, inferior individuals, and exploration individuals separately, so as to obtain the individual population at the (q+1)th iteration after the individual update; Increment q by 1 and reacquire the individual population at the q-th iteration until the iteration stopping condition is met. Based on the globally optimal individual at the iteration stopping condition, determine the optimal logistics and distribution plan.

[0016] Secondly, a logistics delivery route optimization system is provided, including: An acquisition unit is used to acquire logistics and delivery information, wherein the logistics and delivery information includes the number of delivery stations and the number of drones; The model building unit is used to build a flight speed correction model and a flight load increase model based on weather factors, and to establish a flight power correction model based on the flight load increase model. The flight load increase model is used to calculate the additional weight caused by rainfall. The path optimization unit is used to construct a path optimization function with the goal of minimizing logistics delivery time based on the number of delivery stations and the number of drones in the logistics delivery information and the flight speed correction model, and to establish a delivery energy consumption constraint function for the path optimization function based on the flight power correction model and the flight speed correction model. The path optimization unit is also used to construct a logistics delivery optimization model using the path optimization function and the delivery energy consumption constraint function, and solve the logistics delivery optimization model to obtain the optimal logistics delivery scheme that minimizes the logistics delivery time. The optimal logistics delivery scheme includes the optimal delivery station sequence for each drone.

[0017] Thirdly, a logistics distribution route optimization device is provided. Taking the device as an electronic device as an example, it includes a memory, a processor, and a transceiver that are connected in sequence. The memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the logistics distribution route optimization method as described in the first aspect or any possible design in the first aspect.

[0018] Fourthly, a storage medium is provided, on which instructions are stored, which, when executed on a computer, perform the logistics distribution route optimization method as described in the first aspect or any possible design of the first aspect.

[0019] Fifthly, a computer program product containing instructions is provided, which, when executed on a computer, cause the computer to perform the logistics distribution route optimization method as described in the first aspect or any possible design of the first aspect.

[0020] Beneficial effects: (1) This invention first constructs a flight speed correction model and a payload increase model that reflects the impact of severe weather such as rainfall on the added weight of UAVs, and further establishes a flight power correction model, so that the energy consumption constraint is closer to the actual operating environment. At the same time, by combining the flight speed correction model and the flight power correction model, and taking the minimization of logistics delivery time as the optimization objective, a path optimization function and a corresponding delivery energy consumption constraint function are constructed. Then, based on the aforementioned objective function and constraint function, a logistics delivery optimization model is generated. Finally, by solving the constructed logistics delivery optimization model, the optimal delivery scheme that can still guarantee timeliness under complex weather conditions is obtained. Thus, this invention solves the problem of ignoring the correlation between weather conditions and UAV operation performance and energy consumption in the prior art, significantly improves the rationality and adaptability of UAV path planning under complex weather conditions, effectively reduces flight risks and energy waste caused by weather factors, and provides reliable technical support for intelligent and refined scheduling of logistics delivery. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the steps of the logistics delivery route optimization method provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of the logistics distribution route optimization system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0023] It should be understood that although the terms first, second, etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit, without departing from the scope of the exemplary embodiments of the invention.

[0024] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0025] Example: See Figure 1 As shown, the logistics distribution route optimization method provided in this embodiment can be executed by, but is not limited to, computer devices with certain computing resources, such as servers, edge computers, personal computers (PCs, which are multi-purpose computers of a size, price, and performance suitable for personal use; desktop computers, laptops, mini-laptops, tablets, and ultrabooks are all personal computers), smartphones, or personal digital assistants (PDAs). It is understood that the aforementioned execution entities do not constitute a limitation on the embodiments of this application. Accordingly, the operation steps of this method can be, but are not limited to, the steps S1 to S4 below.

[0026] S1. Obtain logistics and delivery information, wherein the logistics and delivery information includes the number of delivery stations and the number of drones; in specific implementation, for example, but not limited to, drones can be loaded onto logistics transport vehicles and used to carry out last-mile logistics delivery (the last mile here is an example, and the delivery distance of drones can be selected according to the actual operation scenario); of course, drones are also set with corresponding parameters, such as flight altitude, maximum payload, maximum available energy, etc. These parameters will be used as subsequent path planning constraints to assist energy consumption constraints in drone path planning.

[0027] After obtaining the logistics and delivery information, the path planning of the drone can be carried out. In this embodiment, weather factors are introduced into the path planning process, so as to make corrections to the flight speed, power and payload based on the actual weather conditions of the drone, so that the drone's energy consumption, flight speed and payload are more in line with the actual environment. Optionally, the construction process of the aforementioned flight speed correction model, flight payload increase model and flight power correction model is as shown in step S2 below.

[0028] S2. Construct a flight speed correction model and a flight load increase model based on weather factors, and establish a flight power correction model based on the flight load increase model. The flight load increase model is used to calculate the additional weight caused by rainfall.

[0029] In specific implementation, logistics delivery information may also include the flight speed of the drone. Each drone can be set to fly at a specified speed. Therefore, based on this, the flight speed can be corrected in combination with weather conditions. The process may be, but is not limited to, the steps S21 to S23 below.

[0030] S21. For any two delivery stations, obtain the weather information of the airspace between the two delivery stations, wherein the weather information includes airspace wind speed and airspace wind direction angle; in this embodiment, in the last-mile delivery scenario, the delivery stations are located in the same area, and the airspace covering all delivery stations can be used as the airspace between any two delivery stations; of course, in complex terrain environments such as mountainous areas, the weather information of the airspace between each delivery station can be collected separately to ensure the accuracy of the constructed flight speed correction model.

[0031] After obtaining the airspace wind speed and airspace wind direction angle, the impact of airspace wind speed on the flight of the UAV can be calculated based on the airspace wind direction angle, as shown in step S22 below.

[0032] S22. Based on the airspace wind direction angle in the weather information, calculate the influence angle of the airspace wind speed on the flight heading of the UAV when flying between any two delivery stations. In specific implementation, the UAV does not fly in a stationary medium. The tailwind in the airspace will improve the propulsion effect relative to the ground, the headwind will reduce it, and the crosswind will change the flight heading. Therefore, this embodiment uses the airspace wind direction angle to measure the influence of the airspace wind speed on the flight heading of the UAV. For example, but not limited to, first obtaining the heading angle of the UAV when flying between any two delivery stations; then, calculating the angle difference between the heading angle and the airspace wind direction angle; finally, mapping the angle difference to the range [-π, π) will yield the influence angle.

[0033] After calculating the angle of influence, a flight speed correction model can be constructed by combining flight speed and airspace wind and rain, as shown in step S23 below.

[0034] S23. Based on the influence angle, the flight speed, and the airspace wind speed, construct a weather-based flight speed correction model between any two delivery stations; in specific implementation, the flight speed correction model can be constructed according to, but is not limited to, the following formula.

[0035] ; In the formula, This indicates the corrected flight speed. These represent the flight speed of the drone and the wind speed in the airspace, respectively. This indicates the angle of influence.

[0036] The formula above shows that the influence angle reflects the angle of effect of the wind relative to the drone's flight direction. When it is 0, it indicates the drone is in a tailwind, and its corrected speed equals flight speed + airspeed. When it is π or -π, it indicates the drone is in a headwind, and the corrected speed is flight speed - airspeed. Other values ​​indicate the drone is in a crosswind, and the corrected speed equals... .

[0037] Of course, by inputting the heading angle of the drone between different delivery stations, as well as the airspace wind speed and wind direction between different delivery stations, the corrected flight speed of the drone when flying between different delivery stations can be obtained.

[0038] Thus, based on the aforementioned steps S21 to S23, a flight speed correction model is constructed, thereby incorporating the influence of wind speed in weather factors on the drone's flight; subsequently, the impact of rainfall on the drone's payload can be quantified, and the process can be, but is not limited to, as shown in steps S24 to S26 below.

[0039] S24. For any two delivery stations, based on the flight speed correction model, the corrected flight speed of the drone between the two delivery stations is obtained, and the flight time of the drone between the two delivery stations is determined according to the corrected flight speed. In this embodiment, the flight time of the drone between the two delivery stations can be obtained by dividing the distance between the two delivery stations by the aforementioned corrected flight speed.

[0040] After obtaining the corrected flight speed, the windward area of ​​the UAV and the rainfall in the corresponding flight airspace can be obtained, so that a flight payload weight-increasing model can be constructed based on this, as shown in steps S35 and S26 below.

[0041] S25. Obtain the windward area of ​​the drone and the rainfall in the airspace between any two delivery stations; In this embodiment, when it rains, rainwater will adhere to the surface of the drone, generating additional weight. The additional weight depends on the corrected flight speed of the aircraft, the windward area, the water accumulation efficiency, etc. Therefore, this embodiment constructs the load-weight model based on the aforementioned parameters, and the process is shown in step S26 below.

[0042] S26. Based on the corrected flight speed, flight time, rainfall, and windward area, and using the following formula, construct the flight load increase model between any two delivery stations.

[0043] In practical applications, the flight load weight-increasing model can be represented as: ; In the formula, This indicates the additional weight of the drone due to rainfall (this weight will be added to the drone's payload, affecting hovering power and total energy consumption). The water accumulation efficiency of a drone (representing the actual ability of a drone's surface to capture precipitation, which depends on factors such as the drone's shape and surface material, and can be determined in advance by aerodynamic experiments). Indicates rainfall amount, This indicates the precipitation coefficient (with a value of 0.84, while...). This indicates the liquid water content. This indicates the corrected flight speed. This refers to the windward area (which determines the effective area of ​​contact between the drone and precipitation, and can be preset according to the type of drone). This refers to the angle of attack (also known as the angle of attack) of the drone when it flies between any two delivery stations. It can be preset according to different types of drones, such as the angle of attack of the forward propeller of a rotary-wing drone, which is generally 45°. This refers to the flight time (i.e., the time exposed to rain). This indicates the maximum set additional weight (in this embodiment, it is set to 0.5 kg).

[0044] In this embodiment, when the rainfall is 0, the additional weight is 0, which means that it does not affect the payload of the drone itself.

[0045] Thus, based on the aforementioned steps S24 to S26, the impact of rainfall on the flight payload of the UAV during flight can be quantified.

[0046] After constructing the flight load weight-increasing model, the flight power correction model can be constructed based on it. The process can be, but is not limited to, the steps S27 to S29 below.

[0047] S27. Establish a power influence function based on flight altitude and ambient temperature; in this embodiment, the power influence function can be established by, but is not limited to, the following formula.

[0048] ; In the formula, This represents the power influence function. Indicates standard air density, Let h be the atmospheric pressure at which the drone is at its altitude. Represents the gas constant of air. The ambient temperature at flight altitude h. This represents the sensitivity factor (in this embodiment, the value is 0.8). This represents the temperature effect coefficient (with a value of 0.08). This indicates the reference temperature (which is the set value).

[0049] Thus, the aforementioned power influence function comprehensively considers the impact of flight altitude and flight temperature on power, that is, it quantifies them into a power influence value. At the same time, the aforementioned formula is a general formula. Depending on the flight segment in which the UAV is located, the atmospheric pressure and temperature at the corresponding flight altitude of the flight segment can be substituted into the aforementioned formula to obtain the power influence value corresponding to the flight segment. For example, if the UAV needs to fly from delivery station i to delivery station j, then the atmospheric pressure and temperature at the flight altitude of the UAV at delivery station i and delivery station j can be substituted (the flight altitude of the UAV between different delivery stations can be preset according to the terrain between different delivery stations). Of course, the aforementioned example is only an example, and this embodiment is not limited to this.

[0050] After establishing the power influence function, the actual load of the UAV can be determined based on the flight load weight gain model, as shown in step S28 below.

[0051] S28. Using the aforementioned flight load weight-increasing model, determine the actual load of the UAV. In specific implementation, substitute parameters such as rainfall and corrected flight speed between different delivery stations into the aforementioned flight load weight-increasing model to obtain the additional weight generated by the UAV when flying between different delivery stations. Then, add it to its current cargo load to obtain its actual load. That is, when the UAV is in delivery mode and unloading hovering mode, the actual load is the sum of the total cargo volume and the additional weight generated by the UAV due to rainfall. When the UAV is in return mode, the actual load is the additional weight generated by the UAV due to rainfall (on the return trip, the cargo is unloaded, so the load generated is only the additional weight generated by rainfall).

[0052] Thus, after determining the actual payload of the UAV, the flight power correction model can be constructed by combining the aforementioned power influence virtual function, as shown in step S29 below.

[0053] S29. Based on the power influence function and the actual load, the flight power correction model is constructed.

[0054] In practical implementation, for example, but not limited to, constructing a flight power correction model according to the following formula.

[0055] ; In the formula, This represents the corrected flight power of the drone when it is in the nth state. This represents the reference power of the drone in the nth state (the reference power is the same when the drone is in delivery state and return state). This represents the payload coefficient of the drone when it is in the nth state. This represents the actual payload of the drone in the nth state, where n=1 indicates the drone is in delivery mode, n=2 indicates the drone is in return mode, and n=3 indicates the drone is in unloading hovering mode. Furthermore, when the drone is in either delivery or unloading hovering mode... This is the sum of the total cargo volume and the additional weight of the drone due to rainfall, when the drone is on its return journey. The additional weight of the drone due to rainfall; in this embodiment, the reference power and load factor under the aforementioned different states can be preset, and this embodiment does not impose specific limitations.

[0056] Of course, the aforementioned flight power correction model is also a general formula. By substituting the actual load and power impact values ​​for different flight segments, the corrected flight power corresponding to different flight segments can be obtained.

[0057] Therefore, after constructing the flight power correction model, flight speed correction model, and flight load increase model through the aforementioned steps S2 and their sub-steps, the path optimization function and the corresponding energy consumption constraint function can be constructed based on these. The process can be, but is not limited to, the steps S3 below.

[0058] S3. Based on the number of delivery stations and the number of drones in the logistics delivery information, and the flight speed correction model, construct a path optimization function with the goal of minimizing logistics delivery time, and establish a delivery energy consumption constraint function for the path optimization function based on the flight power correction model and the flight speed correction model.

[0059] In practical implementation, the path optimization function can be, but is not limited to, the flight time of the drone during the delivery process, and can be expressed as: ; In the formula, For logistics delivery time, This represents the flight distance of the k-th drone from delivery station i to delivery station j. This represents the corrected flight speed of the k-th drone when making deliveries between delivery station i and delivery station j. Let be the first decision variable, representing whether the k-th drone flies from delivery station i to delivery station j. For the number of drones, Let i represent the number of delivery stations. When i or j equals 0, it indicates a delivery center. Specifically, when a drone flies from delivery station i to delivery station j... If it is 1, then it is 0.

[0060] Furthermore, based on the foregoing, this embodiment provides a further optimization scheme, namely, incorporating the takeoff and landing times of the drone, as well as the hovering and unloading time, to obtain a more accurate logistics delivery time; therefore, the improved path optimization function can be expressed as: ; In the formula, This indicates the required flight altitude for the drone (a set value, such as the altitude from which to land, the altitude required for takeoff, etc.). This indicates the takeoff and landing speed of the drone (a set value). This represents the load influence factor (which is a constant). This represents the quantity of goods at the i-th delivery station. This indicates the unloading speed (unit: seconds / piece). Let be the second decision variable, representing whether the k-th drone docks at delivery station i, where When the value is 1, it indicates that the k-th drone docks at delivery station i (i.e., hovers and unloads at delivery station i); otherwise, it is 0. Therefore, we introduce... Then, the time spent by the drone hovering and unloading at each of the actual assigned delivery stations can be calculated.

[0061] In this embodiment, the first term on the left side of the formula is the flight time of the drone during the delivery process, the second term is the take-off and landing time of the drone during the delivery process, and the third term is the docking and unloading time of the drone at each delivery point. Thus, the sum of the three terms can be used as the logistics delivery time. Therefore, the optimization goal of this embodiment is to minimize the logistics delivery time, thereby ensuring the timeliness of logistics delivery.

[0062] Based on the aforementioned formula, after constructing the path optimization function, a delivery energy consumption constraint function that takes into account weather factors can be constructed based on the flight power correction model. For example, but not limited to, the following steps S31 and S32 can be used to construct the aforementioned energy consumption constraint function.

[0063] S31. Establish a power penalty function based on rainfall; in specific implementation, rainfall will lead to a decrease in propulsion efficiency, therefore, it is necessary to establish a power penalty function based on rainfall; for example, but not limited to, the power penalty function can be constructed according to the following formula.

[0064] ; In the formula, Denotes the power penalty function, This represents the set minimum power penalty value. Indicates rainfall amount, This represents the rainfall impact coefficient. In this embodiment, by substituting the rainfall amount for different flight segments, the power penalty coefficient for different flight segments can be obtained. Different flight segments can be flight segments between two delivery stations, flight segments from the last delivery station to the delivery center, or hovering unloading flight segments, which means that the rainfall encountered during hovering unloading is also taken into account in the power correction process for hovering unloading.

[0065] Thus, based on the aforementioned formula, after constructing the power penalty function, the delivery energy consumption constraint function can be constructed by combining it with the flight speed correction model, as shown in step S32 below.

[0066] S32. Based on the flight speed correction model and power penalty function, a delivery energy consumption constraint function is constructed. In practical implementation, as shown by the aforementioned flight power correction model, the UAV is divided into three states: delivery state, return state, and hovering state. Therefore, the delivery energy consumption of a UAV also consists of three parts: delivery, return, and hovering, expressed by the following formula: ; In the formula, This represents the total energy consumption for delivery by the k-th drone. This represents the delivery energy consumption of the k-th drone. This represents the return trip energy consumption of the k-th drone. This represents the energy consumption of the k-th drone during unloading and hovering. This represents the maximum available energy of the kth drone (e.g., set to 1.3 kWh).

[0067] Furthermore, , and The specific calculation formula is as follows: ; ; ; In the formula, This represents the corrected flight power of the k-th drone between delivery station i and delivery station j when the drone is in delivery mode. This represents the power penalty coefficient between delivery station i and delivery station j when the k-th drone is in delivery mode. This represents the flight time of the k-th drone from delivery station i to delivery station j. Let be the first decision variable, representing whether the k-th drone flies from delivery station i to delivery station j.

[0068] Similarly, This represents the corrected flight power of the k-th drone between the last delivery station and the distribution center when it is in the return flight state (i.e., after the delivery is completed, it flies directly back to the distribution center, i.e., the logistics vehicle, from the last delivery station). This represents the power penalty coefficient for the drone between the last delivery station and the distribution center. This represents the flight time of the k-th drone from the last delivery station to the delivery center.

[0069] This represents the corrected flight power of the k-th drone at delivery station i when it is in a loading and unloading hovering state. This represents the power penalty coefficient for the k-th drone at delivery station i. This represents the service time of the k-th drone at delivery station i (i.e., the aforementioned hovering and unloading time, calculated by dividing the amount of cargo at the delivery station by the unloading speed). The second decision variable represents whether the k-th drone docks at delivery station i. Specifically, if the k-th drone docks at delivery station i, the hovering energy consumption at that station is calculated; otherwise, if the k-th drone does not dock at delivery station i, the hovering energy consumption is not calculated. If the value is 0, then the hovering energy consumption of the kth drone at delivery station i is 0; based on this, the total hovering energy consumption of the kth drone at each of its actual assigned stations can be calculated.

[0070] It should be noted that the aforementioned power penalty coefficient is calculated based on the power penalty function, and each flight time is calculated based on the corrected flight speed output by the flight speed correction model.

[0071] Thus, this embodiment incorporates meteorological influences (wind affects flight speed, temperature, and air pressure, which affect power, while precipitation is used to introduce a penalty coefficient) and the UAV payload status into the energy consumption constraint function, making the constructed energy constraint closer to the actual operating environment of the UAV. Therefore, this constraint function is used to determine whether the mission is feasible within the battery capacity.

[0072] Of course, this path optimization function also has other constraints, such as the number of drones (the maximum number of drones departing from the distribution center is K), drone flight altitude constraints (flight altitude greater than the minimum set altitude and less than or equal to the maximum set altitude), and battery constraints (such as the drones having enough remaining battery power to complete the delivery, and the drones having enough energy after charging to meet the delivery flight and delivery tasks). Additionally, the constraints also include: ; This constraint means that all drones dispatched for delivery must eventually return to the delivery center.

[0073] ; ; The aforementioned two constraints mean that only one drone is used for delivery at each delivery station.

[0074] ; This constraint represents the total demand for goods at delivery stations along the delivery route, plus the sum of the additional weight incurred between delivery stations due to rainfall. This means that the additional weight of the drone due to rainfall during its flight from delivery station i to delivery station j must be less than the rated maximum payload M.

[0075] , It is a constant value and belongs to {1,2,...,L}; This constraint indicates that a drone arrives at a delivery station. Leave the delivery station after completing the delivery.

[0076] Of course, the aforementioned constraints are all common constraints for UAV path planning, and will not be repeated in this embodiment.

[0077] Therefore, after constructing the path optimization function and its corresponding delivery energy constraint function based on the aforementioned step S3, a logistics delivery optimization model can be generated. Then, by solving the model, the optimal logistics delivery scheme that minimizes the logistics delivery time can be obtained, as shown in step S4 below.

[0078] S4. Using the path optimization function and the delivery energy consumption constraint function, construct a logistics delivery optimization model and solve the logistics delivery optimization model to obtain the optimal logistics delivery scheme that minimizes the logistics delivery time. The optimal logistics delivery scheme includes the optimal delivery station sequence for each drone.

[0079] In practical applications, this embodiment uses a multi-strategy intelligent population algorithm to solve the logistics distribution optimization model. The process can be, but is not limited to, the steps S41 to S47 below.

[0080] S41. Obtain the individual population at the q-th iteration, where when q is 1, the individual population at the q-th iteration is the initial population, and each initial individual in the initial population corresponds to a logistics delivery scheme. In this embodiment, the drones and delivery stations can be initialized and allocated multiple times, and each initialization will result in a logistics delivery scheme, such as drone 1: allocated station S→M1→M4→M7→S, drone 2: S→M2→M5→S, drone 3: S→M3→M5→S, where M1-M7 represent delivery stations and S is the delivery center. Therefore, the aforementioned logistics delivery scheme can be encoded as an initial individual. Thus, based on the aforementioned initialization process, the initial population can be obtained.

[0081] After obtaining the individual population at the qth iteration, multiple logistics and distribution schemes are determined. Then, the fitness of individuals can be calculated by combining the aforementioned logistics and distribution optimization model, as shown in step S42 below.

[0082] S42. Based on the logistics distribution optimization model, calculate the fitness of each individual in the population at the q-th iteration. In this embodiment, each individual at the q-th iteration corresponds to a logistics distribution scheme. Therefore, the delivery station sequence for each drone has been determined. Based on this, it can be substituted into the path optimization function in the logistics distribution optimization model to obtain the logistics distribution time for each individual. Then, the reciprocal of the logistics distribution time is used as the fitness, that is, the shorter the delivery time, the greater the fitness, indicating that the individual is better. At the same time, check whether the logistics distribution scheme corresponding to each individual meets the aforementioned delivery energy constraint function and the aforementioned constraint conditions. If it meets the constraints, retain it. If it does not meet the constraints, repair (e.g., crossover) the individuals that do not meet the constraints. Then, re-determine whether the constraints are met. If they are met, retain them. Otherwise, remove the individual and crossover and mutation are performed on the retained individuals to generate new individuals to keep the population size unchanged.

[0083] Thus, after calculating the fitness of each individual in the population at the q-th iteration, the globally optimal individual at the q-th iteration can be determined, as shown in step S43 below.

[0084] S43. Based on the fitness of each individual, determine the globally optimal individual at the q-th iteration. In practice, first, select the individual with the maximum fitness at the q-th iteration; then, determine whether the maximum fitness is greater than the fitness of the globally optimal individual at the (q-1)-th iteration. If it is greater, the individual with the maximum fitness is taken as the globally optimal individual at the q-th iteration; otherwise, the globally optimal individual at the (q-1)-th iteration is taken as the globally optimal individual at the q-th iteration. Of course, when q is 1, the globally optimal individual at the q-th iteration is the individual with the highest fitness in that iteration. After determining the globally optimal individual, it can be determined whether the iteration stopping condition is met, as shown in step S44 below.

[0085] S44. Determine whether the iteration stopping condition is met; in specific implementation, the iteration stopping condition is that q reaches the maximum number of iterations or the fitness of the globally optimal individual at the qth iteration is greater than or equal to the fitness threshold.

[0086] If the aforementioned iteration stopping condition is not met, then population update is required. In specific applications, this embodiment addresses the problems of premature convergence (i.e., easy to get trapped in local optima), insufficient population diversity, and imbalance between exploration and development in traditional population optimization algorithms. It provides an optimization algorithm that divides the population into different types of individuals and uses different strategies to update different types of individuals, thereby solving the aforementioned shortcomings.

[0087] Optionally, the population update process is as shown in steps S45 and S46 below.

[0088] S45. If not, then divide the individual population at the q-th iteration into elite individuals, inferior individuals, and exploratory individuals; in specific implementation, for example, but not limited to, sorting the individual population in descending order of fitness to obtain a sorting sequence; then, the top 30% of individuals in the sorting sequence are regarded as elite individuals, the bottom 20% are regarded as inferior individuals, and the remaining individuals in the sorting sequence are regarded as exploratory individuals.

[0089] Thus, after the individual types in the population are classified, the corresponding update strategies can be used to update the position for different types of individuals, as shown in step S46 below.

[0090] S46. Employ a multi-strategy update approach to update elite individuals, inferior individuals, and exploration individuals separately, so as to obtain the individual population at the (q+1)th iteration after the individual update.

[0091] In specific implementation, for elite individuals, this embodiment adopts an update strategy based on a combination of neighborhood individuals and global individuals, thereby preventing the traditional over-reliance on a single global optimal position for position updates, which weakens the search capability and causes the exploration range to be limited due to premature convergence in the early stages of the search.

[0092] For any elite individual, the individual update process can be, but is not limited to, as shown in S46a to S46j below.

[0093] S46a. For any elite individual, obtain the neighborhood set of that elite individual from the individual population at the q-th iteration; in specific implementation, take that elite individual as the center and move to the left... Each individual, take the right. Each individual is an elite individual, thus forming a neighborhood of individuals, where NS is the set size of the neighborhood set. To round down, This is for rounding up.

[0094] Furthermore, when the number of individuals to the left or right of any elite individual is insufficient, individuals are added from the other end of the population to make the size of the entire neighborhood set of individuals reach NS.

[0095] Thus, after obtaining the neighborhood set of any elite individual, the selection of the optimal neighborhood individual can be carried out, as shown in step S46b below.

[0096] S46b. Select the optimal neighboring individual from the set of neighboring individuals; in this embodiment, the neighboring individual with the highest fitness is selected from the set of neighboring individuals as the optimal neighboring individual; then, the search weight and search coefficient can be calculated, as shown in step S46c below.

[0097] S46c. Calculate the elite search weight and elite search coefficient at the q-th iteration.

[0098] In practical implementation, examples, but not limited to the following formulas, can be used to calculate the elite search weight and elite search coefficient.

[0099] ; ; ; ; In the formula, All of these represent the search weight of elites. This represents the initial values ​​for the three elite search weights. The maximum number of iterations, Indicates the elite search coefficient. Represents constants (with values ​​of 0.7 and 0.3 respectively).

[0100] As can be seen from the aforementioned formula for calculating elite search weight, The step size decreases linearly with the number of iterations to ensure a large exploration step size in the early stages, while gradually decreasing the step size in later stages to facilitate local exploration; similarly, An exponential decay mechanism is employed, which initially encourages individuals to learn from their historical best practices, maintaining individual diversity, and gradually diminishing this decay as iterations progress. Then, a sine function is introduced for adjustment, so that it can fluctuate dynamically in different iteration stages. This fluctuation not only helps to escape local optima, but also provides additional perturbation when stuck, stimulates new search directions, and enhances global optimization ability.

[0101] Furthermore, this embodiment also introduces an elite search coefficient, which is used to combine the globally optimal individual and the elite individual to calculate the search vector, so that the dynamic search step size can be calculated based on the search vector, thereby realizing the position update of any elite individual; wherein, the search vector calculation process is as shown in step S46d below.

[0102] S46d. Calculate the search vector corresponding to any elite individual based on the elite search coefficient, any elite individual, and the globally optimal individual at the qth iteration.

[0103] In specific implementation, the search vector can be calculated using, for example, but not limited to, the following formula.

[0104] ; In the formula, Represents the search vector. This represents the globally optimal individual at the q-th iteration. Represents any of the elite individuals mentioned above. Let represent a random vector (whose dimension is the same as that of any elite individual); thus, according to this formula, the elite search coefficient is used to adjust the weight of global guidance and random perturbation based on the number of iterations, that is, in the early stages of the search, When the value is small, it relies more on random vectors (i.e. random perturbations) to expand the search range. As the iteration progresses, its value gradually increases, and it gradually guides the global search towards the global optimum, thus achieving a smooth transition between exploration and development.

[0105] After calculating the search vector, the guiding individual can be determined using the optimal neighborhood individuals, as shown in step S46e below.

[0106] S46e. Using the optimal neighboring individuals, a guiding individual is determined; in this embodiment, if If the neighborhood of the best individual is selected as the guiding individual, then the global best individual will be selected as the guiding individual.

[0107] After identifying the guiding individual, the dynamic search step size can be calculated by combining the aforementioned search vector and elite search weight, as shown in step S46f below.

[0108] S46f. Calculate the dynamic search step size based on the search vector, the elite search weight, and the guiding individual; in specific implementations, the following formula may be used, but is not limited to, to calculate the dynamic search step size.

[0109] ; In the formula, Indicates the dynamic search step size. This represents the historical best position of any of the elite individuals. This refers to the guiding individual. These represent the first random number and the second random number, respectively.

[0110] Furthermore, ; In the formula, Represents the standard normal distribution. It indicates a uniform distribution.

[0111] Thus, based on the aforementioned formula, after calculating the dynamic search step size, the position can be updated accordingly, as shown in step S46g below.

[0112] S46g. Based on the dynamic search step size, the position of any elite individual is updated to obtain the initially updated elite individual, and after updating all elite individuals, the initial updated population is obtained; in this embodiment, the dynamic search step size is used plus the position vector of the elite individual (i.e., If the initial update of any of the elite individuals is obtained, then the initial updated population can be obtained after updating all the other elite individuals in the aforementioned manner.

[0113] As can be seen from the aforementioned initial update method for elite individuals, this embodiment uses the optimal neighboring individual or the global optimal individual to guide the individual update. That is, in the early stage of the iteration, the optimal neighboring individual is relied upon, and in the later stage of the iteration, the global optimal individual is relied upon for position update. Based on this, the two work together to not only amplify the diversity of the population, but also achieve a smooth transition from exploration to development, thereby improving the overall search efficiency.

[0114] After completing the initial update of the elite individuals, this embodiment also introduces a local fine-grained search of the elite individuals to quickly locate high-quality areas, as shown in steps S46h to S46j below.

[0115] S46h. Select perturbation individuals from the initial updated population; in this embodiment, the top 20% of elite individuals in terms of fitness are selected as target individuals, and then the corresponding initially updated elite individuals are selected from the initial updated population to serve as perturbation individuals.

[0116] Thus, after obtaining each perturbation individual, a local fine search can be performed on each perturbation individual, as shown in step S46i below.

[0117] S46i. Perform a local fine-grained search on the perturbation individual to obtain an updated perturbation individual; in this embodiment, the following formula may be used, but is not limited to, to perform a local fine-grained search on any perturbation individual.

[0118] ; In the formula, This represents any of the perturbation individuals after the update. This refers to any of the disturbed individuals. This indicates the local search intensity (with a value of 0.9). It is the sixth random number that conforms to the standard normal distribution.

[0119] In this embodiment, the perturbation individuals (i.e., the individuals obtained after the initial update of the top 20% of elite individuals with the highest fitness) use global optimal information to perform a fine-grained local search for perturbation, thereby enhancing the local development intensity and enabling the rapid locking of the optimal solution region; thus, the development capability can be improved.

[0120] After updating each disturbed individual based on the aforementioned formula, the individuals not selected in the initial updated population can be combined to form the updated elite individuals, as shown in step S46j below.

[0121] S46j. The updated perturbation individuals and the initially updated elite individuals that were not selected from the initial updated population are used as the updated elite individuals.

[0122] By following the steps S46a to S46j, the position update of elite individuals can be completed; then, the position update of exploration individuals can be performed, as shown in the following steps.

[0123] Step 1: For any given exploration individual, randomly select one elite individual from all elite individuals to serve as the leader individual.

[0124] After the leader is selected, random individuals can be selected, as shown in steps two and three below.

[0125] Step 2: Randomly select two exploration individuals from all exploration individuals to serve as the first random individual and the second random individual, respectively. In practice, after selecting the two random individuals, position difference calculation can be performed, as shown below.

[0126] Step 3: Calculate the position difference vector between the first random individual and the second random individual.

[0127] After calculating the position difference vector, the weights can be updated, as shown in step four below.

[0128] Step 4: Calculate the first update weight and the second update weight for any of the exploration individuals.

[0129] Optionally, for example, but not limited to, the first update weight and the second update weight can be calculated according to the following formula.

[0130] ; ; In the formula, These represent the first update weight and the second update weight, respectively.

[0131] Thus, based on the aforementioned formula, after calculating the two update weights, the position update of any exploration individual can be performed by combining the aforementioned leader individual and position difference vector, as shown in step 5 below.

[0132] Step 5: Based on the leader individual, the position difference vector, the first update weight, and the second update weight, perform an individual update on any one of the exploration individuals to obtain the updated one of the exploration individuals.

[0133] In this embodiment, for example, but not limited to, the following formula can be used to update any of the exploration individuals.

[0134] ; In the formula, This refers to any of the explored individuals after the initial update. This refers to any of the exploring individuals. This represents the third random number, whose value range is [0,1]. These represent the leader individual and the position difference vector, respectively.

[0135] As can be seen from the aforementioned position update formula, the updated elite individuals generated by the aforementioned update strategy based on the combination of neighborhood individuals and global individuals provide an evolutionary direction for the exploration individuals (i.e., randomly selecting one as the leader individual); at the same time, the difference vector between exploration individuals introduces differential perturbation during the evolutionary process, which in turn helps to explore new areas; thus, the combination of the two realizes the evolutionary mechanism of elite guidance + exploration perturbation, thereby ensuring that the evolutionary direction moves toward the optimal direction (i.e., ensuring convergence) and avoiding the problem of the population getting trapped in local optima due to over-reliance on elite individuals.

[0136] Therefore, after updating the location of the explored individuals using the aforementioned method, the location of inferior individuals can be updated, as shown in the following steps.

[0137] Step 1: For any inferior individual, construct the reproductive vector of that inferior individual; In this embodiment, the position of inferior individuals is usually far from the global optimum. Therefore, diversified evolution is required to guide them back to the vicinity of the global optimum, thereby transforming low-quality individuals into new individuals with potential to produce diversified offspring.

[0138] Additionally, the reproductive vector can be calculated using, for example, but not limited to, the following formula.

[0139] ; In the formula, Represents the reproduction vector. Indicates reproductive intensity (value is 0.1). This represents a standard Gaussian distributed random matrix.

[0140] As can be seen from the above formula, this embodiment constructs a breeding vector near the globally optimal individual based on a Gaussian random matrix. In this way, multiple different but similar candidate points can be generated around the optimal solution, which maintains diversity and avoids excessive dispersion. Moreover, if the population concentrates too early in a local optimum, the Gaussian perturbation may still produce individuals that jump out slightly, providing new evolutionary directions for subsequent iterations and avoiding the problem of getting trapped in local optima.

[0141] After calculating the reproduction vector, boundary constraints are applied to it to obtain the constraint vector, as shown in step 2 below.

[0142] Step 2: Apply boundary constraints to the breeding vector to obtain the constraint vector; in specific implementation, the following formula can be used, but is not limited to, to perform boundary constraints.

[0143] ; In the formula, Represents the constraint vector. These represent the lower bound and the upper bound of the search space, respectively.

[0144] Thus, after obtaining the constraint vector based on the aforementioned formula, the position update of inferior individuals can be performed by combining the globally optimal individual, as shown in step 3 below.

[0145] Step 3: Using the constraint vector and the globally optimal individual at the qth iteration, update any inferior individual to obtain the updated inferior individual.

[0146] In this embodiment, any inferior individual is updated according to the following formula; ; In the formula, This refers to any of the aforementioned inferior individuals after the update. This refers to any of the inferior individuals mentioned above. These represent the fourth and fifth random numbers (both taking values ​​between (0,1)).

[0147] The aforementioned method can be used to update each inferior individual. Then, by combining the updated exploration individuals and the updated elite individuals, the individual population at the (q+1)th iteration can be formed. Next, the population is updated in the aforementioned manner until the iteration stopping condition is met, and the optimal logistics and distribution plan can be obtained. The process is shown in step S47 below.

[0148] S47. Increment q by 1 and reacquire the individual population at the q-th iteration until the iteration stopping condition is met. Based on the globally optimal individual at the iteration stopping condition, determine the optimal logistics distribution plan.

[0149] Thus, by enhancing exploration diversity through the aforementioned neighborhood individuals, guiding the evolution of elite individuals in stages (i.e., elite individuals are guided by neighborhood individuals in the early stage and by the globally optimal individuals in the later stage), and through the synergy of multiple strategies such as elite guidance-differential exploration, reproduction and elite local refinement, the problems of premature convergence, insufficient population diversity and imbalance between exploration and development in traditional population optimization algorithms are solved, and the unity of convergence accuracy, convergence speed and population diversity is achieved.

[0150] After obtaining the globally optimal individual that satisfies the iteration stopping condition based on the aforementioned step S4 and its sub-steps, its corresponding logistics and distribution scheme is the optimal logistics and distribution scheme that minimizes the logistics and distribution time. Then, the logistics tasks of the UAV can be allocated based on this optimal logistics and distribution scheme, thereby achieving high-efficiency logistics and distribution.

[0151] Therefore, through the logistics delivery route optimization method described in detail in steps S1 to S4 above, this invention first constructs a flight speed correction model and a payload increase model reflecting the impact of severe weather such as rainfall on the added weight of the UAV, and further establishes a flight power correction model, so that the energy consumption constraint is closer to the actual operating environment. At the same time, combining the flight speed correction model and the flight power correction model, and taking minimizing logistics delivery time as the optimization objective, a route optimization function and a corresponding delivery energy consumption constraint function are constructed. Then, based on the aforementioned constructed objective function and constraint function, a logistics delivery optimization model is generated. Finally, by solving the constructed logistics delivery optimization model, the optimal delivery plan that can still guarantee timeliness under complex weather conditions is obtained. Thus, this invention solves the problem of ignoring the correlation between weather conditions and UAV operating performance and energy consumption in the prior art, significantly improves the rationality and adaptability of UAV route planning under complex weather conditions, effectively reduces flight risks and energy waste caused by weather factors, and provides reliable technical support for intelligent and refined scheduling of logistics delivery.

[0152] like Figure 2 As shown, the second aspect of this embodiment provides a hardware system for implementing the logistics distribution route optimization method described in the first aspect of the embodiment, including: The acquisition unit is used to acquire logistics and delivery information, wherein the logistics and delivery information includes the number of delivery stations and the number of drones.

[0153] The model building unit is used to build a flight speed correction model and a flight load increase model based on weather factors, and to establish a flight power correction model based on the flight load increase model. The flight load increase model is used to calculate the additional weight caused by rainfall.

[0154] The path optimization unit is used to construct a path optimization function with the goal of minimizing logistics delivery time based on the number of delivery stations and the number of drones in the logistics delivery information and the flight speed correction model, and to establish a delivery energy consumption constraint function for the path optimization function based on the flight power correction model and the flight speed correction model.

[0155] The path optimization unit is also used to construct a logistics delivery optimization model using the path optimization function and the delivery energy consumption constraint function, and solve the logistics delivery optimization model to obtain the optimal logistics delivery scheme that minimizes the logistics delivery time. The optimal logistics delivery scheme includes the optimal delivery station sequence for each drone.

[0156] The working process, working details and technical effects of the system provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0157] like Figure 3 As shown, the third aspect of this embodiment provides a logistics delivery route optimization device. Taking the device as an electronic device as an example, it includes: a memory, a processor, and a transceiver that are connected in sequence. The memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the logistics delivery route optimization method as described in the first aspect of the embodiment.

[0158] For specific examples, the memory may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; specifically, the processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor may be implemented using at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), PLA (Programmable Logic Array). The processor may also include a main processor and a coprocessor. The main processor, also known as the CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state.

[0159] In some embodiments, the processor may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. For example, the processor may not be limited to microprocessors of the STM32F105 series, reduced instruction set computer (RISC) microprocessors, x86 architecture processors, or processors with integrated neural network processing units (NPUs). The transceiver may be, but is not limited to, a Wi-Fi transceiver, a Bluetooth transceiver, a General Packet Radio Service (GPRS) transceiver, a ZigBee (a low-power LAN protocol based on the IEEE 802.15.4 standard) transceiver, a 3G transceiver, a 4G transceiver, and / or a 5G transceiver. Furthermore, the device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0160] The working process, working details and technical effects of the electronic device provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0161] The fourth aspect of this embodiment provides a storage medium for storing instructions containing the logistics distribution route optimization method described in the first aspect of the embodiment. That is, the storage medium stores instructions that, when the instructions are run on a computer, execute the logistics distribution route optimization method as described in the first aspect of the embodiment.

[0162] The storage medium refers to a carrier for storing data, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or memory sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0163] The working process, working details and technical effects of the storage medium provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0164] The fifth aspect of this embodiment provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the logistics distribution route optimization method as described in the first aspect of this embodiment, wherein the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0165] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A logistics distribution path optimization method characterized by, include: Obtain logistics and delivery information, including the number of delivery stations and the number of drones; A flight speed correction model and a flight load increase model based on weather factors were constructed. Based on the flight load increase model, a flight power correction model was established. The flight load increase model was used to calculate the additional weight caused by rainfall. Based on the number of delivery stations and drones in the logistics delivery information, and the flight speed correction model, a path optimization function is constructed with minimizing logistics delivery time as the optimization objective. Based on the flight power correction model and the flight speed correction model, a delivery energy consumption constraint function for the path optimization function is established. By utilizing the path optimization function and the delivery energy consumption constraint function, a logistics delivery optimization model is constructed, and the logistics delivery optimization model is solved to obtain the optimal logistics delivery scheme that minimizes the logistics delivery time. The optimal logistics delivery scheme includes the optimal delivery station sequence for each drone.

2. The method of claim 1, wherein, Logistics delivery information also includes: the drone's flight speed, of which a flight speed correction model based on weather factors has been constructed, including: For any two delivery stations, obtain the weather information of the airspace between the two delivery stations, wherein the weather information includes airspace wind speed and airspace wind direction angle; Based on the airspace wind direction angle in the weather information, the influence angle of the airspace wind speed on the flight heading of the UAV when flying between any two delivery stations is calculated. Based on the influence angle, the flight speed, and the airspace wind speed, a weather-based flight speed correction model is constructed between any two delivery stations.

3. The method of claim 2, wherein, Based on the airspace wind direction angle in the weather information, calculate the angle by which the airspace wind speed affects the flight heading of the UAV when flying between any two delivery stations, including: Obtain the heading angle of the drone when it flies between any two delivery stations; Calculate the angle difference between the heading angle and the airspace wind direction angle, and map the angle difference to the range [-π, π) to obtain the influence angle; Accordingly, based on the influence angle, the flight speed, and the airspace wind speed, a weather-based flight speed correction model is constructed between any two delivery stations, which includes: The flight speed correction model is constructed according to the following formula; ； In the formula, denotes the corrected flight speed, denote the flight speed of the drone and the airspace wind speed, respectively, denotes the influence angle.

4. The method of claim 1, wherein, A flight payload weight gain model was constructed, including: For any two delivery stations, based on the flight speed correction model, the corrected flight speed of the drone between the two delivery stations is obtained, and the flight time of the drone between the two delivery stations is determined based on the corrected flight speed. Obtain the windward area of ​​the drone, and the rainfall in the airspace between any two delivery stations; Based on the corrected flight speed, flight time, rainfall, and windward area, and using the following formula, a flight load increase model between any two delivery stations is constructed. ； wherein, represents an additional weight of the drone due to rainfall, is a water accumulation efficiency of the drone, represents a rainfall amount, represents a precipitation coefficient, represents a corrected flight speed, represents the windward area, represents an attack angle of the drone when flying between any two distribution sites, represents the flight time, represents a set maximum additional weight.

5. The method according to claim 1, characterized in that, Establish a flight power correction model, including: Establish a power influence function based on flight altitude and ambient temperature; The actual payload of the UAV is determined using the aforementioned flight payload weight-increasing model; Based on the power influence function and the actual load, the flight power correction model is constructed.

6. The method according to claim 5, characterized in that, Establish a power influence function based on flight altitude and ambient temperature, including: The power influence function is established according to the following formula; ； In the formula, This represents the power influence function. Indicates standard air density, Let h be the atmospheric pressure at which the drone is at its altitude. Represents the gas constant of air. The ambient temperature at flight altitude h. This indicates the factor affecting sensitivity. Indicates the temperature effect coefficient. Indicates reference temperature; Accordingly, based on the power influence function and the actual load, the flight power correction model is constructed, which includes: The flight power correction model is constructed according to the following formula; ； In the formula, This represents the corrected flight power of the drone when it is in the nth state. This represents the base power of the drone in the nth state. This represents the payload coefficient of the drone when it is in the nth state. This represents the actual payload of the drone in the nth state, where n=1 indicates the drone is in delivery mode, n=2 indicates the drone is in return mode, and n=3 indicates the drone is in unloading hovering mode. Furthermore, when the drone is in either delivery or unloading hovering mode... This is the sum of the total cargo volume and the additional weight of the drone due to rainfall, when the drone is on its return journey. This refers to the additional weight of the drone due to rainfall.

7. The method according to claim 1, characterized in that, Based on the flight power correction model and the flight speed correction model, a delivery energy consumption constraint function for the path optimization function is established, including: Establish a power penalty function based on rainfall; Based on the flight speed correction model and power penalty function, a delivery energy consumption constraint function is constructed.

8. The method according to claim 7, characterized in that, A power penalty function based on rainfall is constructed, including: The power penalty function is constructed according to the following formula; ； In the formula, Denotes the power penalty function, This represents the set minimum power penalty value. Indicates rainfall amount, Indicates the impact coefficient of rainfall; Accordingly, based on the flight speed correction model and power penalty function, the delivery energy consumption constraint function is constructed, which includes: The delivery energy consumption constraint function is constructed according to the following formula; ； ； ； ； In the formula, This represents the total energy consumption for delivery by the k-th drone. This represents the delivery energy consumption of the k-th drone. This represents the return trip energy consumption of the k-th drone. This represents the energy consumption of the k-th drone during unloading and hovering. Let be the maximum available energy of the k-th drone, where This represents the corrected flight power of the k-th drone between delivery station i and delivery station j when the drone is in delivery mode. This represents the power penalty coefficient between delivery station i and delivery station j when the k-th drone is in delivery mode. This represents the flight time of the k-th drone from delivery station i to delivery station j. Let $k$ be the first decision variable, representing whether the kth drone flies from delivery station i to delivery station j. This represents the corrected flight power of the k-th drone between the last delivery station and the delivery center when it is on its return journey. This represents the power penalty coefficient for the drone between the last delivery station and the distribution center. This represents the flight time of the k-th drone from the last delivery station to the delivery center. This represents the corrected flight power of the k-th drone at delivery station i when it is in a loading and unloading hovering state. This represents the power penalty coefficient for the k-th drone at delivery station i. This represents the service time of the k-th drone at delivery station i. Let be the second decision variable, representing whether the k-th drone docks at delivery station i; The power penalty coefficient is calculated based on the power penalty function, and each flight time is calculated based on the corrected flight speed output by the flight speed correction model.

9. The method according to claim 1, characterized in that, Solve the logistics and distribution optimization model to obtain the optimal logistics and distribution solution that minimizes the logistics and distribution time, including: Obtain the individual population at the q-th iteration, where when q is 1, the individual population at the q-th iteration is the initial population, and each initial individual in the initial population corresponds to a logistics delivery scheme. Based on the logistics and distribution optimization model, the fitness of each individual in the population at the qth iteration is calculated. Based on the fitness of each individual, the globally optimal individual at the q-th iteration is determined; Determine if the iteration stopping condition is met; If not, then the individual population at the qth iteration will be divided into elite individuals, inferior individuals, and exploratory individuals; A multi-strategy update approach is adopted to update elite individuals, inferior individuals, and exploration individuals separately, so as to obtain the individual population at the (q+1)th iteration after the individual update; Increment q by 1 and reacquire the individual population at the q-th iteration until the iteration stopping condition is met. Based on the globally optimal individual at the iteration stopping condition, determine the optimal logistics and distribution plan.

10. A logistics distribution route optimization system, characterized in that, include: An acquisition unit is used to acquire logistics and delivery information, wherein the logistics and delivery information includes the number of delivery stations and the number of drones; The model building unit is used to build a flight speed correction model and a flight load increase model based on weather factors, and to establish a flight power correction model based on the flight load increase model. The flight load increase model is used to calculate the additional weight caused by rainfall. The path optimization unit is used to construct a path optimization function with the goal of minimizing logistics delivery time based on the number of delivery stations and the number of drones in the logistics delivery information and the flight speed correction model, and to establish a delivery energy consumption constraint function for the path optimization function based on the flight power correction model and the flight speed correction model. The path optimization unit is also used to construct a logistics delivery optimization model using the path optimization function and the delivery energy consumption constraint function, and solve the logistics delivery optimization model to obtain the optimal logistics delivery scheme that minimizes the logistics delivery time. The optimal logistics delivery scheme includes the optimal delivery station sequence for each drone.

Images