Dynamic path optimization method for unmanned aerial vehicle combined multi-mode traffic in emergency logistics

Abstract

The invention relates to an unmanned aerial vehicle combined multi-mode traffic dynamic path optimization method in emergency logistics. The method comprises the steps that firstly, an initial transportation plan is acquired, and the initial transportation plan is a combined emergency logistics transportation path plan jointly composed of aviation, a road, a railway and an unmanned aerial vehicle; then, a mixed integer programming model is constructed based on the initial transportation plan, and an objective function and constraint conditions associated with unmanned aerial vehicle combined multi-mode traffic are defined; and finally, when the initial transportation plan is interrupted, solving based on the objective function and the constraint condition by adopting a self-adaptive large neighborhood search algorithm to obtain a new path plan of emergency logistics transportation. A mixed integer linear programming model is developed for dynamic multimodal transport of the unmanned aerial vehicle with double roles, the long-distance efficiency of a traditional transport mode and the accuracy, flexibility and adaptability of the unmanned aerial vehicle in a complex environment are combined, an ALNS algorithm is introduced, the calculation time is remarkably shortened, and the practicability of the model is enhanced.

Description

Technical Field

[0001] This application relates to the technical field of emergency logistics, and particularly to a dynamic path optimization method for drones combined with multi-modal transportation in emergency logistics. Background Art

[0002] More and more technologies have explicitly adopted collaborative optimization techniques for multi-transportation modes. For example, helicopters and vehicles are regarded as interdependent subsystems, first optimized separately, and then the results are integrated into a coherent and globally optimized solution; or the optimization process is customized according to the specific requirements of each transportation mode, so as to utilize their respective advantages to improve the overall system performance. Drones have also been explored for various applications in the prior art. Some technologies focus on using drones as tools for monitoring and mapping, while others use them for air transportation, similar to cargo planes. However, most of these technologies treat drones in isolation and fail to integrate them into the multimodal transportation system. This lack of integration limits the full play of the complementary advantages of drones and traditional transportation modes (such as bypassing inaccessible areas or shortening delivery times in emergencies). Filling this gap is crucial for improving the efficiency and adaptability of multi-modal transportation planning in emergency logistics.

[0003] Dynamic optimization is another area with significant gaps in the prior art. Despite progress in the development of optimization models, many models are still static and fail to consider real-time adjustments in complex scenarios. For example, early technologies introduced a dynamic model for the distribution and transportation of relief supplies, and subsequent technologies made adaptive improvements to real-time routing based on this, as well as optimization in dynamic environments. However, there is still a lack of models in the prior art that can dynamically integrate drones and traditional transportation modes (especially in emergencies). In addition, in the prior art of multimodal transportation, the heterogeneity of goods has received relatively little attention. Although many technologies involve multi-commodity scenarios, few technologies explicitly consider the diversity of goods and still have not solved the challenges brought by the heterogeneity of goods in emergency logistics, such as adapting to different cargo sizes, sensitivities, and priorities.

[0004] Therefore, in the related technologies, there is an urgent need for a way to improve the accuracy, flexibility, and adaptability of dynamic multimodal transportation in emergency logistics. Summary of the Invention

[0005] Based on this, in view of the above technical problems, it is necessary to provide a dynamic path optimization method for drones combined with multi-modal transportation in emergency logistics, which can improve the accuracy, flexibility, and adaptability of dynamic multimodal transportation in emergency logistics.

[0006] In a first aspect, this application provides a dynamic path optimization method for drones combined with multi-modal transportation in emergency logistics. The method includes:

[0007] Obtain an initial transportation plan, which is a joint emergency logistics transportation route plan composed of aviation, road, railway, and drones;

[0008] Based on the initial transportation plan, construct a mixed-integer programming model, and define the objective function and constraints associated with the joint multi-modal transportation of drones;

[0009] When the initial transportation plan is interrupted, use the adaptive large neighborhood search algorithm to solve based on the objective function and constraints to obtain a new path plan for emergency logistics transportation.

[0010] Optionally, in an embodiment of the present application, the initial transportation plan includes a preset take-off point for the drone, and the preset take-off point for the drone is determined based on the Thiessen polygon theory.

[0011] Optionally, in an embodiment of the present application, the objective function is:

[0012] min F = F 1 + F 2 + F 3 + F 4 + F 5

[0013]

[0014] where F is the total cost, F 1 is the transportation cost, F 2 is the transshipment cost, F 3 is the storage cost, F 4 is the waiting cost, F 5 is the delay penalty, K is the vehicle set, including cargo planes, trucks, trains, and drones; A is the arc set; O is the order set, including the ordinary rescue cargo order set, the emergency rescue cargo order set, and the drones in the order state; N is the location set, including the supply point, the demand point, the transshipment point, the preset drone take-off point, and the road interruption point; S is the supply point, D is the demand point, T is the transshipment point, T′ is the preset drone take-off point, B is the road interruption point; is the unit cost of different items, n ∈ {1, 1′, 2, 3, 4, 5}, is the transportation cost per ton per hour, is the transportation cost per ton per kilometer, is the loading and unloading cost per ton, is the storage cost per ton per hour, is the waiting cost per hour, is the delay penalty per ton per hour; is the travel time of vehicle k on arc (i, j), is the distance between vehicle k at locations i and j, q o is the number of order o; is a binary variable, which is 1 if order o is transported by vehicle k and uses arc (i, j), otherwise 0; is a binary variable, which is 1 if order o is transferred from vehicle k to vehicle l at transfer location i, otherwise 0; and are the start times of service for vehicle k and vehicle l at location i for order o, is the end time of service for vehicle k at location i for order o, a s(o) is the start time of pick-up for order o at the supply point, is the waiting time of vehicle k at location i, is the delay time of order o at the delivery location.

[0015] Optionally, in an embodiment of the present application, the constraint conditions include rescue cargo delivery guarantee constraints, multi-modal transfer constraints, vehicle and order flow conservation constraints, multi-modal transport characteristic constraints, service time constraints, and emergency constraints.

[0016] Optionally, in an embodiment of the present application, the rescue cargo delivery guarantee constraints include vehicle departure and return constraints, sub-tour route constraints, cargo pick-up and delivery constraints, and vehicle capacity constraints. The multi-modal transfer constraints include transfer process constraints and transfer operation timing constraints. The vehicle and order flow conservation constraints include vehicle flow conservation constraints, order flow conservation constraints, and vehicle transport arc constraints. The multi-modal transport characteristic constraints include vehicle route selection constraints and cargo heterogeneity constraints. The service time constraints include service start time constraints, service end time constraints, service completion constraints, order arrival time constraints, last service constraints, time and driving distance and speed constraints, and truck arrival time constraints. The emergency constraints include waiting time constraints and delay time constraints.

[0017] Optionally, in an embodiment of the present application, the solving using the adaptive large neighborhood search algorithm based on the objective function and constraint conditions includes:

[0018] Selecting an insertion operator and a removal operator according to the dependencies of the drone joint multi-modal transport, and adjusting the multi-modal transport plan based on the insertion operator and the removal operator.

[0019] Optionally, in an embodiment of the present application, the insertion operators include a greedy insertion operator, a transfer insertion operator, a random insertion operator, a regret insertion operator, and a most constrained first insertion operator. The removal operators include a worst removal operator, a random removal operator, a related removal operator, a route removal operator, and a node removal operator.

[0020] In a second aspect, the present application also provides a dynamic path optimization device for the joint multi-mode transportation of drones in emergency logistics. The device includes:

[0021] An initial planning module, configured to obtain an initial transportation plan, where the initial transportation plan is a joint emergency logistics transportation path plan jointly composed of aviation, highway, railway, and drones;

[0022] A mixed integer programming model construction module, configured to construct a mixed integer programming model based on the initial transportation plan, and define an objective function and constraint conditions associated with the joint multi-mode transportation of drones;

[0023] A multimodal transportation replanning module, configured to, when the initial transportation plan is interrupted, solve based on the objective function and constraint conditions by using an adaptive large neighborhood search algorithm to obtain a new path plan for emergency logistics transportation.

[0024] In a third aspect, the present application also provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and the processor executes the steps of the methods in the above respective embodiments.

[0025] In a fourth aspect, the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the methods in the above respective embodiments are implemented.

[0026] For the above dynamic path optimization method for the joint multi-mode transportation of drones in emergency logistics, first, an initial transportation plan is obtained, where the initial transportation plan is a joint emergency logistics transportation path plan jointly composed of aviation, highway, railway, and drones; then, a mixed integer programming model is constructed based on the initial transportation plan, and an objective function and constraint conditions associated with the joint multi-mode transportation of drones are defined; finally, when the initial transportation plan is interrupted, an adaptive large neighborhood search algorithm is used to solve based on the objective function and constraint conditions to obtain a new path plan for emergency logistics transportation. That is to say, a mixed integer linear programming (MIP) model is developed for the dynamic multi-modal transportation (DMT-Drones) of drones with dual roles. This model combines the long-distance efficiency of traditional transportation methods and the precision, flexibility, and adaptability of drones in complex environments to highlight their complementary roles. An adaptive large neighborhood search (ALNS) heuristic algorithm is introduced to significantly reduce the calculation time and enhance the practicality of the model. At the same time, a rolling horizon technique is adopted to achieve real-time adaptability, allowing the flight path of the drone to be dynamically adjusted according to interruptions such as damaged infrastructure to ensure uninterrupted and efficient material delivery. The problem of cargo heterogeneity is solved by distinguishing between drone-specific materials and traditional materials to optimize the transportation strategy for emergency and diverse demands. Brief Description of the Drawings

[0027] Figure 1 It is an application environment diagram of the dynamic path optimization method for the joint use of drones and multi - mode transportation in emergency logistics in one embodiment;

[0028] Figure 2 It is a schematic flowchart of the dynamic path optimization method for the joint use of drones and multi - mode transportation in emergency logistics in one embodiment;

[0029] Figure 3 It is a schematic diagram of the dynamic multimodal transportation characteristics of drones with dual roles in one embodiment;

[0030] Figure 4 It is a schematic diagram of cargo heterogeneity and complex scheduling process in one embodiment;

[0031] Figure 5 It is a schematic diagram of the dynamic adjustment of the drone take - off point and route replanning in one embodiment;

[0032] Figure 6 It is a schematic diagram of the principle of the event - triggered rolling horizon replanning method in one embodiment;

[0033] Figure 7 It is a schematic flowchart of the principle of the adaptive large neighborhood search algorithm in one embodiment;

[0034] Figure 8 It is a schematic diagram of the optimization process of the adaptive large neighborhood search algorithm customized for the dynamic multimodal transportation of drones with dual roles in one embodiment;

[0035] Figure 9 It is a schematic diagram of the principle of the removal operator function in one embodiment;

[0036] Figure 10 It is a coordination network diagram of drones and multimodal transportation in one embodiment;

[0037] Figure 11 It is a structural block diagram of the dynamic path optimization device for the joint use of drones and multi - mode transportation in emergency logistics in one embodiment;

[0038] Figure 12 It is an internal structure diagram of a computer device in one embodiment. Detailed Description of the Embodiments

[0039] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0040] The dynamic path optimization method for the joint use of unmanned aerial vehicles (UAVs) and multi-modal transportation in emergency logistics provided by the embodiments of the present application can be applied to, for example, Figure 1 the application environment shown. Among them, the terminal communicates with the server through the network. The data storage system can store the data that the server needs to process. The data storage system can be integrated on the server, or placed in the cloud or other network servers. Among them, the terminal can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, Internet of Things (IoT) devices, and portable wearable devices. The IoT devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented by an independent server or a server cluster composed of multiple servers.

[0041] In one embodiment, as Figure 2 shown, a dynamic path optimization method for the joint use of UAVs and multi-modal transportation in emergency logistics is provided. Taking the server in Figure 1 as an example, the method includes the following steps:

[0042] S201: Obtain an initial transportation plan, where the initial transportation plan is a joint emergency logistics transportation path plan jointly composed of aviation, road, railway, and UAVs.

[0043] In the embodiments of the present application, three main transportation modes are used, including aviation, road, and railway, and are combined with UAVs for emergency logistics technology. These modes are represented by the mode set W = {f, h, r, u}, where f represents aviation, h represents road, r represents railway, and u represents UAVs. Since UAVs are initially transported as goods to various transfer points or take-off points before transforming into active carrier roles, they are regarded as an independent transportation mode rather than a pure air transportation mode. UAVs have a dual role, including UAVs as goods and UAVs as carriers. The goods transported by aviation, road, and railway include general relief supplies O n , emergency relief supplies O e and UAVs as goods O u, among which, general relief supplies include bulk items, such as staple foods (such as rice, flour), large medical equipment (such as ambulances, CT machines) and building materials (such as steel, cement). These items are usually large in size and heavy in weight. Long-term support to the disaster area requires a large amount of such supplies, and the requirements for rapid delivery are not high. Emergency relief supplies include high-priority items, such as medicines, compressed food and communication equipment, which are small in size and light in weight. They need to be delivered quickly to meet the immediate needs of the disaster area. At the transfer point or the preset drone take-off point, the drone as the carrier only transports emergency relief supplies. After the disaster occurs, the disaster-stricken area is the demand point, and the q neighboring cities around it are the supply points to provide relief supplies to the disaster-stricken area. These supply points are represented by the supply point set S = {S i |i=1,2,…,q} means that the set T′={T i ′|i=1,2,…,n} represents a transfer point.

[0044] Based on the node locations of supply points, demand points, and transfer points and their distances from each other, as well as the vehicles and parameters of each vehicle, the initial transportation plan is determined in combination with the adaptive large neighborhood search algorithm (ALNS). Figure 3 As shown in Figure 2, the real-life transportation network is decomposed into four levels, each of which represents the route planning and scheduling of a specific vehicle type: Railway level k r , Highway level k h , UAV level k u and aviation level k f . The horizontal axis represents route planning, and the vertical axis represents order scheduling. The scheduling of vehicle k and order o is carried out simultaneously. A single order can be transported by any combination of these four modes of transportation. Each mode has its own unique characteristics. The railway level provides the lowest cost and the largest capacity. The road level provides moderate capacity, speed and cost. The drone level is the most flexible and numerous, but the capacity is the smallest. The aviation level is the fastest, but the cost is significantly higher and the availability is limited. Any order o can be transferred between different vehicles at a designated transfer point (that is, a transfer point equipped with transfer facilities and temporary cargo storage sites). Typically, the transfer point also acts as a regular terminal, allowing it to be used as a pick-up or delivery location. Transfer between vehicles operating in the same mode but on different schedules is also feasible. Flexible transportation modes such as the road level and the drone level significantly increase the opportunities for transfer. In addition, such as Figure 3 As shown in , transit times at transfer points between different modes are also considered.

[0045] In one embodiment of the present application, the initial transportation plan includes a preset take-off point for the drone, and the preset take-off point for the drone is determined based on Thiessen polygon theory.

[0046] In one embodiment of the present application, the set T = {Ti|i = 1, 2, …, n} represents the preset take-off points of the unmanned aerial vehicles (UAVs), which are essentially the transfer points T′ between other transportation modes and UAVs. The positions of these preset take-off points are calculated using the Thiessen polygon theory to ensure full coverage and fairness in the disaster area. The process includes the following steps: (a) Divide the disaster area into several regions according to the geographical characteristics of the disaster area, the distribution of material demand points, and the accessibility of transportation vehicles, with each region corresponding to one or more material demand points; (b) Use the Thiessen polygon algorithm to construct a Thiessen polygon map with these material demand points as the generators, where the distance from any point within each polygon to its corresponding generator is less than the distance to other generators; (c) According to factors such as the flight range, endurance, and material delivery efficiency of the UAVs, select one or more suitable points within each Thiessen polygon as the preset take-off points of the UAVs. These points should be as close as possible to the center of the polygon or the area with intensive material demand to ensure that the UAVs can efficiently and accurately cover their corresponding material demand points; (d) Further optimize the positions and quantities of the preset take-off points through optimization algorithms (such as the shortest path algorithm, coverage algorithm, etc.) to maximize the overall benefit of emergency logistics transportation. In specific applications, the initial transportation plan sets 4 preset take-off points for the UAVs, which are determined in advance before departure. However, in case of emergency, their positions and quantities will be adjusted according to the actual situation.

[0047] S203: Construct a mixed-integer programming model based on the initial transportation plan, and define the objective function and constraints associated with the UAV joint multi-modal transportation.

[0048] In the embodiment of the present application, a corresponding mixed-integer programming model is constructed based on the initial transportation plan, converting the solution problem during path optimization into the corresponding objective function and corresponding constraints, and minimizing the total cost including monetary costs such as transportation and handling, and time costs such as transportation time, waiting time, and transfer time.

[0049] In one embodiment of the present application, the objective function is:

[0050] min F = F 1 + F 2 + F 3 + F 4 + F 5

[0051]

[0052] where F is the total cost, F 1 is the transportation cost, F 2 is the transfer cost, F 3 is the storage cost, F 4is the waiting cost, F 5 is the delay penalty, K is the set of vehicles, including cargo planes, trucks, trains, and drones; A is the set of arcs; O is the set of orders, including the set of ordinary rescue cargo orders, the set of emergency rescue cargo orders, and the drones in the order state; N is the set of locations, including supply points, demand points, transfer points, preset drone take-off points, and road interruption points; S is the supply point, D is the demand point, T is the transfer point, T′ is the preset drone take-off point, and B is the road interruption point; is the unit cost of different projects, n ∈ {1, 1′, 2, 3, 4, 5}, is the transportation cost per ton per hour, is the transportation cost per kilometer per ton, is the loading and unloading cost per ton, is the storage cost per ton per hour, is the waiting cost per hour, is the delay penalty per ton per hour; is the travel time of vehicle k on arc (i, j), is the distance between locations i and j for vehicle k, q o is the quantity of order o; is a binary variable, which is 1 if order o is transported by vehicle k and uses arc (i, j), otherwise 0; is a binary variable, which is 1 if order o is transferred from vehicle k to vehicle l at transfer location i, otherwise 0; and are the start times of service for vehicle k and vehicle l at location i for order o, is the end time of service for vehicle k at location i for order o, a s(o) is the pick-up start time of order o at the supply point, is the waiting time of vehicle k at location i, is the delay time of order o at the delivery location.

[0053] In an embodiment of the present application, the constraint conditions include rescue cargo delivery guarantee constraints, multimodal transfer constraints, flow conservation constraints of vehicles and orders, multimodal transport characteristic constraints, service time constraints, and emergency constraints.

[0054] In an embodiment of the present application, the defined constraints mainly include rescue cargo delivery guarantee constraints, multimodal transfer constraints, flow conservation constraints of vehicles and orders, multimodal transport characteristic constraints, service time constraints, and emergency constraints. Among them, the rescue cargo delivery guarantee constraint means that each rescue cargo can ensure that there is a corresponding vehicle for transportation. The multimodal transfer constraint means to ensure the normal progress of the rescue cargo transfer process. The flow conservation constraints of vehicles and orders mean to ensure that the flows of vehicles and orders during transportation are within the defined conditions. The multimodal transport characteristic constraints include vehicle route selection constraints and corresponding cargo heterogeneity constraints. The service time constraints include service start time constraints, service end time constraints, service completion constraints, order arrival time constraints, last service constraints, time, driving distance, and speed constraints, and truck arrival time constraints. The emergency constraints mean that the transportation time in case of emergencies should conform to the actual situation.

[0055] In an embodiment of the present application, the rescue cargo delivery guarantee constraints include vehicle departure and return constraints, sub-tour route constraints, cargo pickup and delivery constraints, and vehicle capacity constraints. The multimodal transfer constraints include transfer process constraints and transfer operation timing constraints. The flow conservation constraints of vehicles and orders include vehicle flow conservation constraints, order flow conservation constraints, and vehicle transportation arc constraints. The multimodal transport characteristic constraints include vehicle route selection constraints and cargo heterogeneity constraints. The service time constraints include service start time constraints, service end time constraints, service completion constraints, order arrival time constraints, last service constraints, time, driving distance, and speed constraints, and truck arrival time constraints. The emergency constraints include waiting time constraints and delay time constraints.

[0056] In an embodiment of the present application, the rescue cargo delivery guarantee constraints include vehicle departure and return constraints, sub-tour route constraints, cargo pickup and delivery constraints, and vehicle capacity constraints. Among them, the vehicle departure and return constraint means to ensure that the vehicle departs from and returns to its starting warehouse and ending warehouse respectively. In addition, the route of the truck service is not restricted because the fleet is regarded as multiple trucks and each truck may take a different route. The specific representation is as follows:

[0057]

[0058] Among them, represents the starting node, represents that among the arcs from the starting node to all other nodes j for all vehicles k except trucks, at most one can be used (i.e., the vehicle cannot fork from the starting node), represents the termination node, represents that for all vehicles k except trucks, the total number of arcs from its starting node to all other nodes j (i.e., the number of times the vehicle leaves the starting node) must be equal to the number of times from all other nodes j back to its termination node The total number of arcs (i.e., the number of times the vehicle returns to the termination node), which ensures that each vehicle departs from its starting node and returns to its termination node.

[0059] The sub-tour route constraints are as follows, used to eliminate sub-tours and provide strict bounds for several polynomial-sized versions of the sub-tour elimination constraints, expressed as follows:

[0060]

[0061] where, is a binary variable, which is 1 if vehicle k uses arc (i, j), and 0 otherwise; is a binary variable, which is 1 if location i is before location j (not necessarily adjacent) in the route of vehicle k, and 0 otherwise.

[0062] The cargo pickup and delivery constraints ensure that the cargo for each order must be picked up and delivered at its supply point and demand point respectively, expressed as follows:

[0063]

[0064] where s(o) represents the supply node and d(o) represents the demand node.

[0065] The vehicle capacity constraint is expressed as:

[0066]

[0067] where q o is the quantity of order o, is a binary variable, which is 1 if order o is transported by vehicle k and uses arc (i, j), and 0 otherwise, c′ k is the capacity of vehicle k, is a binary variable, which is 1 if vehicle k uses arc (i, j), and 0 otherwise.

[0068] The multi-modal transfer constraints include transfer process constraints and transfer operation timing constraints. Among them, the transfer process constraints ensure that at most one transfer is performed at a given location and prohibit transfers within the same vehicle k. The transfer operation timing constraints manage the timing of transfer operations. In the case where the cargo is transferred from vehicle k to vehicle l and vehicle l arrives before vehicle k finishes unloading, vehicle l is allowed to wait until unloading is complete, specifically expressed as follows:

[0069]

[0070] where, and are binary variables, which are 1 if order o is transported by vehicle k or vehicle l and uses arc (i, j), and 0 otherwise; is a binary variable, which is 1 if order o is transferred from vehicle k to vehicle l at transfer location i, and 0 otherwise. is the end time of service of vehicle k for order o at transfer location i. is the start time of service of vehicle l for order o at transfer location i, and M is a sufficiently large positive number.

[0071] The flow conservation constraints for vehicles and orders include vehicle flow conservation constraints, order flow conservation constraints, and vehicle transportation arc constraints. Among them, the vehicle flow conservation constraint is expressed as:

[0072]

[0073] The order flow conservation constraint is expressed as:

[0074]

[0075] The vehicle transportation arc constraint ensures that an order can only be transported along an arc when the corresponding vehicle passes through that arc, and can be expressed as:

[0076]

[0077] The multimodal transportation characteristic constraints include vehicle route selection constraints and cargo heterogeneity constraints. Among them, the vehicle route selection constraint ensures that the vehicle uses a suitable route. For example, it is not allowed for a truck to drive on a railway track, and can be expressed as:

[0078]

[0079] The cargo heterogeneity constraint ensures that transshipment occurs at appropriate locations because some transfer points only support transshipment between specific transportation modes. For example, when goods need to be transferred from a train to a truck, locations only used for transshipment between trains and airplanes are excluded. These constraint conditions are unique to this model and reflect the specific requirements of vehicle path planning in the multimodal transportation system, and can be expressed as:

[0080]

[0081] In addition, the cargo heterogeneity constraint is also used to handle the time windows at supply locations and scheduled locations, and can be expressed as:

[0082]

[0083] Among them, is the start time of service of vehicle k for order o at the supply point. is the end time of service of vehicle k for order o at the supply point, [a s(o) , b s(o) is the pickup time window of order o at the supply point. is the arrival time of vehicle k at location i for order o. is the opening time window of vehicle k at location i.

[0084] Service time constraints include service start time constraint, service end time constraint, service completion constraint, order arrival time constraint, last service constraint, time and driving distance and speed constraint, truck arrival time constraint. The service start time constraint ensures that the service start time occurs after the goods arrive. The service end time constraint defines the service end time as the sum of the service start time and the service duration. The service completion constraint requires that departure can only occur after all services are completed. The order arrival time constraint stipulates that the arrival time of the order cannot be earlier than the arrival time of the vehicle. The last service constraint specifies the latest allowed start time for the last service of the vehicle. The time and driving distance and speed constraint ensures that the time on the route is consistent with the driving distance and speed. The truck arrival time constraint takes into account the arrival time of the truck to ensure accurate time management within the model, and can be specifically expressed as:

[0085]

[0086] where, is the arrival time of vehicle k at location i, is the start time of the last service of vehicle k at location i, is the departure time of vehicle k at location i.

[0087] Emergency constraints include waiting time constraint and delay time constraint. The waiting time and delay time are calculated respectively, aiming to minimize the waiting time and related delay costs, and can be expressed as:

[0088]

[0089] where, is the delay time of order o at the delivery location, is the service end time of vehicle k for order o at the demand point, b d(o) is the delivery end time of order o at the demand point.

[0090] In an embodiment of the present application, the scheduling considerations include waiting time, storage time and delay. When vehicle k arrives at the pickup point s(o) or the transfer point T′ before order o, it may need to wait for the arrival of order o. On the contrary, if order o arrives before vehicle k, it must be temporarily stored until vehicle k arrives, thus generating storage costs. Delays at the delivery terminal will result in penalty costs for the vehicle. Transfers between vehicles will affect the waiting and delay times because the scheduling of one vehicle may depend on another vehicle. To ensure synchronization, the waiting and storage times will be adjusted as needed. As Figure 4As shown, it involves complex scheduling and diverse cargo characteristics. Railway level k r Load general rescue supplies O n and emergency rescue supplies O e and the time window for the drone as cargo O u is [1, 2], while for general rescue supplies O n and emergency rescue supplies O e and the time window for the drone as cargo O u is [0.5, 2]. Therefore, they will be picked up at time 1 and arrive at the transfer point T′ at time 4.5. At T′, the unloading time window of railway level kr and for general rescue supplies O n and emergency rescue supplies O e and the time window for the drone as cargo O u are all [4.5, 5.5]. Immediately afterwards, the drone as cargo O u transforms into the carrier role of drone level k u and loads the emergency rescue supplies O e within the delivery time window [5.5, 5.8] of the emergency rescue supplies O e . At time 5.8, the drone of drone level k u takes off and transports the emergency rescue supplies O e to the destination D. Meanwhile, road level k h loads the general rescue supplies O n , and its time window is [5.8, 6.5], while the pickup time window for the general rescue supplies O n is [5.8, 7]. Therefore, road level k h completes the loading of the general rescue supplies O n at time 6.5 and departs for D. Drone level k u arrives at D for unloading at time 7.2, and at this time the delivery time window for the emergency rescue supplies O e is [7.2, 7.4]. Road level k h arrives at D at time 10.5. Since the delivery time window for the general rescue supplies O n at D is [10, 10.9], and road level k h arrives after this window, plus the unloading time window for road level k h is [10.5, 11.6], this results in a delay of 0.7 hours.

[0091] S205: When the initial transportation plan is interrupted, use the adaptive large neighborhood search algorithm to solve based on the objective function and constraints to obtain a new path plan for emergency logistics transportation.

[0092] In the embodiments of the present application, when an emergency occurs and the initial transportation plan is interrupted, the plan is dynamically updated by changing the transportation mode of the affected requests, re-planning the vehicle route, and modifying the take-off point of the drone. In such emergencies as sudden damage to roads, railways, or terminals caused by disasters, the vehicle will be stalled at the nearest terminal or the current location, and the model will recalculate the feasible route to ensure service continuity. If the vehicle is loaded with drones as cargo and no feasible route can be determined, the drones will take off from the nearest terminal or the current location of the vehicle. These drones will fly to the designated disaster demand points for material distribution and then return to their original deployment locations to ensure an effective disaster response. As Figure 5 shown, it demonstrates the dynamic adjustment of the drone take-off point and the re-planning of the transportation route to cope with sudden interruptions. Initially, the planned route from the supply point (S) to the demand point (D) includes a designated drone take-off point (T 1 ). After the unexpected damage at the sudden interruption point (B), the initial route becomes invalid. At this time, a new take-off point (T 2 ) is generated, and the route is re-planned while discarding the damaged section. This process ensures that the material distribution can continue despite the interruption. An event-triggered rolling horizon re-planning method is adopted, and the adaptive large neighborhood search algorithm is used to solve based on the objective function and constraints to obtain a new path planning for emergency logistics transportation. As Figure 6 shown, the planning horizon is defined as the interval [t 0 , t n . Initially, the model optimizes the path L of vehicle k in the interval [t 0 , t 1 according to the current information. At time t e ∈ [t 0 , t 1 , if an external event occurs (for example, road damage caused by a secondary disaster), the remaining path L(t e , t 1 ) will be discarded, and a re-planning for the new time domain [t e , t 2 will be triggered, using the real-time information about the external event. If no further event occurs in the interval [t e , t 2 , the vehicle will continue to travel along the planned path. At time t 2 , the path optimization is extended to the next interval [t 2 , t 3 , and the path L(t 0 , t 2 ) is fixed. If at any time t i ′ ∈ [t e , t 2An event occurring at a certain moment will trigger replanning again. This rolling horizon approach will iterate until the final stage t f , completing the dynamic optimization for the 0 , t n interval to ensure real-time path updates throughout the planning horizon.

[0093] Specifically, in an embodiment of the present application, the use of the adaptive large neighborhood search algorithm to solve based on the objective function and constraints includes:

[0094] Selecting an insertion operator and a removal operator according to the dependencies of the UAV combined multimodal transportation, and adjusting the multimodal transportation plan based on the insertion operator and the removal operator.

[0095] In an embodiment of the present application, first, ensure compatibility with multimodal and UAV-specific constraints, such as restricted routes of UAVs or transfer restrictions between different transportation modes. Select appropriate insertion operators and removal operators according to the dependencies of the UAV combined multimodal transportation. Through iterative application of the insertion and removal operators, use an update mechanism based on simulated annealing to dynamically balance exploration and exploitation. Key modifications include ensuring that UAVs are assigned to shorter and higher-priority routes and enforcing the feasibility of multimodal transportation at transfer points. By continuously adjusting the multimodal transportation plan to minimize the objective function, i.e., delays and total costs. As Figure 7 shown, it is a principle flow chart of the adaptive large neighborhood search algorithm.

[0096] As Figure 8 shown, it is a schematic diagram of the optimization process of the adaptive large neighborhood search algorithm customized for the dynamic multimodal transportation of UAVs with dual roles. In Figure 8 (a), the initial solution s shows multiple intersections and overlaps between the routes of different transportation modes, and the paths of each mode are relatively long. Figure 8 (b) shows the solution after applying the removal operator, which eliminates specific orders by deleting unreasonable paths. Finally, Figure 8 (c) shows the solution after applying the insertion operator, where the originally unconnected demand and supply points are reconnected, solving the intersection and overlap problems, and significantly shortening the routes of each mode. If the optimized solution s' is worse than the current best solution, it will be discarded, and the ALNS algorithm will continue to iterate until the optimal solution s is found. best .

[0097] In an embodiment of the present application, the insertion operators include a greedy insertion operator, a transshipment insertion operator, a random insertion operator, a regret insertion operator, and a most-constrained-first insertion operator, and the removal operators include a worst removal operator, a random removal operator, a related removal operator, a route removal operator, and a node removal operator.

[0098] In one embodiment of the present application, the selected ALNS operator takes into account the interdependencies among drones, trucks, and railways, ensuring efficient transshipment, capacity management, and prioritized processing of high-priority orders. The operator aims to balance exploration and exploitation, leveraging historical data and predictive heuristic techniques to improve solution quality. The insertion operators include a greedy insertion operator, a transshipment insertion operator, a random insertion operator, a regret insertion operator, and a most-constrained-first insertion operator.

[0099] Among them, the greedy insertion operator evaluates all feasible solutions for inserting an order into a single-vehicle or multi-vehicle route. For drones, it prioritizes urgent deliveries by evaluating the minimum-time and shortest-distance routes, and the order is inserted into the route that generates the lowest additional cost: ΔF = F after -F before where F after and F before represent the objective function values before and after insertion, respectively. This operator is crucial for scenarios where drones must handle critical and time-sensitive deliveries in the intermodal transportation network.

[0100] The transshipment insertion operator focuses on multi-vehicle solutions by inserting orders that require transshipment at a specific terminal. For example, an order initially served by a truck is transferred to a drone for the last-mile delivery. This operator increases the probability of allocating a drone for the last leg of the intermodal journey.

[0101] The random insertion operator inserts an order by randomly selecting a vehicle and a location, introducing diversity. It ensures that drones and other vehicles are explored in various scenarios to avoid getting stuck in local optima. Feasibility checks ensure that the randomly inserted order complies with time and capacity constraints.

[0102] The regret insertion operator predicts future costs by considering the difference between the best and second-best insertion costs of an order: where c o represents the regret value, is the second-lowest insertion cost, is the lowest cost. This operator is particularly valuable in emergency logistics as it helps prioritize orders that may become more expensive or infeasible in future iterations.

[0103] The most-constrained-first insertion operator prioritizes orders based on the constraints of the order (such as distance, time window, and load). This operator is crucial for managing orders with strict delivery time windows and high load requirements, especially when drones are involved in emergency deliveries. The priority score of order o is given by the following formula:

[0104]

[0105] Among them, w 1 、w 2 、w 3 are weights, represents the truck distance, q o represents the load.

[0106] As Figure 9 shown, the removal operators include the worst removal operator, the random removal operator, the correlation removal operator, the route removal operator, and the node removal operator. Among them, the worst removal operator removes the order with the highest cost in the current route for better reallocation, which is crucial for the drone to handle the multimodal transportation of high-cost last-mile deliveries, because reallocating these orders can reduce the overall cost. The random removal operator randomly selects an order from the route for removal, and this operator helps to explore different solutions to ensure that the drones and other vehicles are not restricted by the predetermined pattern. The correlation removal operator removes clusters of similar orders according to distance, time, load, and vehicle compatibility. For example, orders that can be served by drones within a certain geographical radius can be removed together to evaluate alternative routing strategies. The route removal operator clears the entire route and returns all orders to the unserved pool, which is particularly beneficial in multimodal transportation scenarios because integrating routes can reduce vehicle usage and maximize drone efficiency. The node removal operator removes a specific terminal from the vehicle route. For example, removing the truck terminal may result in reallocating the order to the drone for direct delivery, thus reducing the overall transportation time.

[0107] In an embodiment of the present application, as Figure 10 shown, it is a coordination network diagram of drones and multimodal transportation. This network can be represented as (N, A), where N is the set of nodes (including the supply point S, the transfer point T′, the drone takeoff point T, the demand point D, and the road interruption point B), and A is the set of arcs. This network includes four typical transportation scenarios. (a) S 1 →T 1 →D 1 : This route is a dedicated transportation line for emergency relief supplies. Use trains and trucks to transport the supplies (O e and O u ) from S 1 to T 1 . Then, the drone takes off from T 1 and transports O e to D 1 . (b) S 2 &S 3 →T 2 →D 2 : Use trains and trucks to transport the supplies (O 2 and S 3 separately from S e and S n and O u) Transported to T 2 The truck continues to transport the supplies stored at T 2 (O e 、O n and O u ). When it enters point B within the disaster-stricken area AA and suddenly discovers that the road L ahead is damaged and it cannot proceed further. At point B at this time, the drone immediately changes from the cargo state to the vehicle state and carries O e and takes off. It uses the ALNS algorithm to re-plan a new drone transportation route L′ based on the real-time position of the drone, enabling the drone to accurately transport O e to D 2 . This process reflects the dynamic nature of the method. In a traditional multimodal transportation network lacking drones, when encountering a point where the road ahead is damaged, the transportation of supplies will be unable to continue and can only return along the original route.

[0108] (c) S 3 & S 4 → T 3 ′ → T 3 → D 3 : This route highlights the heterogeneity of the goods and demonstrates the coordinated transportation of the truck and the drone. Use trains and airplanes to transport the supplies (O 3 、O 4 and O e 、O n and O u ) from S 3 ′ (This point is one of the few transfer points that can provide a suitable landing space for large airplanes, while other points cannot provide air transfer). The truck continues to transport the supplies (O 3 ′ stored at T e 、O n and O u ) to the preset take-off point T 3 of the drone calculated using the Thiessen polygon technique. The drone takes off from this point carrying O e and accurately delivers it to D 3 . At the same time, the truck also continues to move forward from T 3 and transports O n to D 3 according to the planned route. (d) S 3 → D 1 : When D 1 urgently needs a large amount of O e , in order to solve the problem as soon as possible, the airplane loads O e from S3 and flies directly to D 1 for airdropping.

[0109] In an embodiment of the present application, a control experiment is conducted using a traditional multimodal transport model that completely excludes drones as goods and transportation tools. Among all transportation orders o, O e and O n each account for 50%. The results are shown in Table 1. It can be seen that in small-scale examples with 5 and 10 order quantities, the two methods show high similarity in terms of total cost, number of vehicles used, and order service rate. The delay penalty cost of the method of the present application is significantly lower than that of the traditional multimodal transport method without drones, which fully demonstrates the advantages of the method of the present application in terms of efficiency and flexibility.

[0110] In the context of emergency logistics, the total monetary cost is not the only factor to be concerned about. Time costs, including waiting time cost F4 and delay penalty F5, are equally crucial. In multiple order quantity scenarios (such as 20, 30, 50, and 400), its waiting cost is lower than or equal to that of the traditional multimodal transport method without drones. If the waiting cost is calculated based on the unit cost per ton per kilometer, the method of the present application shows obvious advantages. In addition, in the comparison of unit delay penalties, when the order quantity is 50, the unit delay penalty of the method of the present application is also lower than that of the traditional multimodal transport method without drones.

[0111] In terms of order service rate, the performance of the method of the present application is far superior to that of the traditional multimodal transport method without drones. Except that the total service rates of the two methods are the same when the order quantities are 5, 10, and 30, and the emergency order service rate of the traditional multimodal transport method without drones is better than that of the present application when the order quantity is 30, in other cases, the total service rate of the method of the present application leads by 6% - 18%. In terms of emergency material service rate and general material service rate, the method of the present application leads by 4% - 24% and 4% - 12% respectively. In disaster relief activities, it is crucial to ensure that orders at any demand point are satisfied as much as possible. Therefore, compared with the traditional multimodal transport method without drones, the method of the present application has significant advantages in handling emergency logistics demands.

[0112] In addition to comparing with the traditional multimodal transport method without drones, it can also be seen that among all test data, when the order quantity is 50, the method of the present application shows the best performance. However, when dealing with a large-scale order scenario with 400 order quantities, the order service rate of the method of the present application drops significantly. The reason for this phenomenon is that the preset fleet size in the method of the present application is limited between 15 and 20 vehicles, and the load capacity of each fleet is about 150 tons. Therefore, the current capacity configuration of the method of the present application is difficult to meet the transportation needs of 400 orders, which can be regarded as an extreme test case for verifying the upper limit of capacity. The results show that insufficient transportation capacity leads to a significant drop in the order service rate.

[0113] Table 1

[0114]

[0115] O: Order quantity; F: Total cost, F1: Transportation cost, F2: Transshipment cost, F3: Storage cost, F4: Waiting cost, F5: Delay penalty; Air: Proportion of orders served by air transportation; Road: Proportion of orders served by road transportation; Rail: Proportion of orders served by rail transportation; Drone: Proportion of orders served by drone transportation; Initial: Running time of the initial solution, Best:

[0116] Running time of the best solution, the percentage in parentheses equals the running time of the best solution divided by the total running time; Order service rate: Proportion of orders delivered to the demand points among the total number of orders; Total orders: Proportion of all orders delivered to the demand points among the total number of orders; Emergency orders: Proportion of all emergency supply orders delivered to the demand points among the total number of emergency supply orders; Regular orders: Proportion of all regular orders delivered to the demand points among the total number of regular orders; Segment size: 20 iterations; Cooling rate c = 0.9; 200 iterations; Time limit: 72 hours.

[0117] In an embodiment of the present application, during the operation process, the attributes of the orders may change. For example, in the case where the carrier cannot move forward due to road damage ahead (resulting in an infinite increase in the transportation cost of that section), this may render the original solution infeasible. To solve this problem, the method of the present application can be extended to dynamic settings for re-planning the route and time. When the information related to the order changes, the original plan will be adjusted accordingly. To avoid having a significant impact on other orders and meet the real-time requirements of multimodal transportation, only the orders that have changed will be re-planned.

[0118] The Adaptive Large Neighborhood Search (ALNS) algorithm only formulates plans for all orders at the beginning of the planning period to obtain the original solution. When new order information appears, ALNS is used to update the current solution. If the change occurs before pick-up, the original route will be deleted from the current solution and optimized based on the newly revealed information. If the change occurs after pick-up, the corresponding constraints will be added to ensure that order o is transported by drone.

[0119] In ALNS, only the un-traveled part of the current route is deleted, and the insertion operator is used to optimize the route and schedule according to the original itinerary. When the vehicle has flexibility, the route, departure time, waiting time, and storage time can be adjusted to adapt to order changes. As shown in Table 2 below, a comparison between static planning and dynamic planning is presented. It can be seen that static planning shows a relatively homogeneous approach in vehicle usage, and compared with dynamic planning, it relies on a significantly smaller number of vehicles. Specifically, the proportion of trucks in static planning is extremely high, undertaking almost all transportation tasks. In contrast, dynamic planning demonstrates a more diverse vehicle usage strategy, in which drones and trucks are jointly regarded as the main transportation methods, each undertaking approximately 40% of the total transportation volume. This balanced vehicle allocation not only improves the overall efficiency of the transportation system but also enhances its ability to cope with complex environments and emergencies.

[0120] Meanwhile, the order service rate is used to evaluate the effectiveness of the two types of planning. The order service rate is defined as the ratio of the number of completed orders to the total number of orders. Static planning only shows a certain order fulfillment ability in the case of involving five orders, while in other scenarios, the order service rate is generally low, and approximately 25% of the orders are not fulfilled. Such a low order service rate cannot meet the urgent needs of emergency material supply in the disaster area during actual transportation. In contrast, the order service rate of dynamic planning is close to 100%, almost completely fulfilling all orders.

[0121] In summary, dynamic planning shows significant advantages in terms of vehicle utilization efficiency, transportation capacity allocation, and order fulfillment rate. Its flexible vehicle allocation strategy, intelligent scheduling ability, and excellent order fulfillment ability indicate that the application of dynamic planning is crucial for improving the overall performance of the logistics system.

[0122] Table 2

[0123]

[0124] In an embodiment of the present application, as shown in Table 3 below, to study the impact of cargo heterogeneity on the method of the present application, 10, 30, 100, and 400 orders are randomly selected from the instances, and each group of orders is further divided into three scenarios: (a) 100% are ordinary supply orders o n ; (b) 20% are emergency supply orders o e and 80% are ordinary supply orders o n ; and (c) 50% are emergency supply orders o e and 50% are ordinary supply orders o n .

[0125] In the analysis of vehicle usage, it can be seen that: in scenario (a), except for the small-scale calculation example of 10 orders (where the proportion is relatively low), the proportion of trucks is the highest in all scenarios. In all the surveyed scenarios, the proportion of trains also ranks first. Since drones are specifically responsible for transporting emergency supplies, when the proportion of emergency supplies is 0, the usage proportion of drones also drops to 0. In the other two scenarios except for the 30-order scenario, the proportion of large cargo planes is at the highest level. In addition, except for the 30-order scenario, the total number of vehicles in all scenarios is the lowest because drones are not needed in this scenario, and the carrying capacity of the other three transportation methods is much higher than that of drones. In scenario (c), the total number of vehicles in all surveyed cases is the highest. However, the proportion of large cargo planes is the smallest, and the proportion of trains is also at the lowest level. Among all orders, except for the 400-order scenario, the usage proportion of drones is the highest, which reflects the strategy of using drones to transport emergency supplies in this scenario. As for scenario (b), except that the proportion of trucks is the lowest in the 100-order scenario, the proportions of various transportation methods under different order volumes are generally between those of scenarios (a) and (c).

[0126] In terms of order service rate, it can be seen that: scenario (b) shows the best overall order service rate. Only in the case of 100 orders, the total order service rate of scenario (c) exceeds that of scenario (a), and in the remaining order examples, the total order service rates of scenarios (c) and (a) are equal. In addition, the service rate of general supplies in scenario (c) is the highest, while the service rate of general supplies in scenario (a) is relatively low.

[0127] Table 3

[0128]

[0129] In the above dynamic path optimization method for the joint use of drones and multi-modal transportation in emergency logistics, first, an initial transportation plan is obtained. The initial transportation plan is a joint emergency logistics transportation path plan composed of aviation, highway, railway, and drones. Then, a mixed-integer programming model is constructed based on the initial transportation plan, and objective functions and constraints associated with the joint use of drones and multi-modal transportation are defined. Finally, when the initial transportation plan is interrupted, an adaptive large neighborhood search algorithm is used to solve based on the objective functions and constraints to obtain a new path plan for emergency logistics transportation. That is to say, a mixed-integer linear programming (MIP) model is developed for the dynamic multi-modal transportation with dual-role drones (DMT-Drones). This model combines the long-distance efficiency of traditional transportation methods and the precision, flexibility, and adaptability of drones in complex environments to highlight their complementary roles. An adaptive large neighborhood search (ALNS) heuristic algorithm is introduced to significantly reduce the calculation time and enhance the practicality of the model. At the same time, a rolling horizon technique is adopted to achieve real-time adaptability, allowing the flight path of the drone to be dynamically adjusted according to interruptions such as damaged infrastructure to ensure uninterrupted and efficient material delivery. The problem of cargo heterogeneity is solved by distinguishing between drone-specific materials and traditional materials to optimize transportation strategies for emergency and diverse demands.

[0130] It should be understood that although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the indication of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same moment, but can be executed at different moments. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least some of the steps or stages in other steps or other steps.

[0131] Based on the same inventive concept, an embodiment of the present application also provides a dynamic path optimization device for the joint use of drones and multi-modal transportation in emergency logistics for implementing the above-mentioned dynamic path optimization method. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the following dynamic path optimization device for the joint use of drones and multi-modal transportation in emergency logistics can refer to the limitations on the dynamic path optimization method for the joint use of drones and multi-modal transportation in emergency logistics in the above text, and will not be repeated here.

[0132] In one embodiment, as Figure 11As shown, a dynamic path optimization device 1100 for unmanned aerial vehicle (UAV) combined with multi-modal transportation in emergency logistics is provided, including: an initial planning module 1101, a mixed integer programming model construction module 1103, and a multi-modal transportation re-planning module 1105, where:

[0133] The initial planning module 1101 is configured to obtain an initial transportation plan, where the initial transportation plan is a combined emergency logistics transportation path plan composed of aviation, road, railway, and UAV.

[0134] The mixed integer programming model construction module 1103 is configured to construct a mixed integer programming model based on the initial transportation plan, and define an objective function and constraint conditions associated with UAV combined with multi-modal transportation.

[0135] The multi-modal transportation re-planning module 1105 is configured to, when the initial transportation plan is interrupted, solve based on the objective function and constraint conditions by using an adaptive large neighborhood search algorithm to obtain a new path plan for emergency logistics transportation.

[0136] In an embodiment of the present application, the initial transportation plan includes a preset take-off point of the UAV, and the preset take-off point of the UAV is determined based on the Thiessen polygon theory.

[0137] In an embodiment of the present application, the objective function is:

[0138] min F = F 1 + F 2 + F 3 + F 4 + F 5

[0139]

[0140] where F is the total cost, F 1 is the transportation cost, F 2 is the transshipment cost, F 3 is the storage cost, F 4 is the waiting cost, F 5 is the delay penalty, K is a set of vehicles, including cargo planes, trucks, trains, and UAVs; A is a set of arcs; O is a set of orders, including an ordinary rescue cargo order set, an emergency rescue cargo order set, and UAVs in the order state; n is a set of locations, including supply points, demand points, transshipment points, preset UAV take-off points, and road interruption points; S is the supply point, D is the demand point, T is the transshipment point, T′ is the preset UAV take-off point, and B is the road interruption point; is the unit cost of different items, n ∈ {1, 1′, 2, 3, 4, 5}, is the transportation cost per ton per hour, is the transportation cost per kilometer per ton, is the loading and unloading cost per ton, is the storage cost per ton per hour, is the waiting cost per hour, is the delay fine per ton per hour; is the travel time of vehicle k on arc (i, j), is the distance between locations i and j for vehicle k, q o is the quantity of order o; is a binary variable, which is 1 if order o is transported by vehicle k and uses arc (i, j), otherwise 0; is a binary variable, which is 1 if order o is transferred from vehicle k to vehicle l at transfer location i, otherwise 0; and are the start times of service for vehicle k and vehicle l at location i for order o, is the end time of service for vehicle k at location i for order o, a s(o) is the start time of picking up goods for order o at the supply point, is the waiting time of vehicle k at location i, is the delay time of order o at the delivery location.

[0141] In an embodiment of the present application, the constraint conditions include rescue goods delivery guarantee constraints, multi-mode transfer constraints, vehicle and order flow conservation constraints, multimodal transport characteristic constraints, service time constraints, and emergency constraints.

[0142] In an embodiment of the present application, the rescue goods delivery guarantee constraints include vehicle departure and return constraints, sub-tour route constraints, goods pickup and delivery constraints, and vehicle capacity constraints. The multi-mode transfer constraints include transfer process constraints and transfer operation timing constraints. The vehicle and order flow conservation constraints include vehicle flow conservation constraints, order flow conservation constraints, and vehicle transportation arc constraints. The multimodal transport characteristic constraints include vehicle route selection constraints and goods heterogeneity constraints. The service time constraints include service start time constraints, service end time constraints, service completion constraints, order arrival time constraints, last service constraint, time and travel distance and speed constraints, and truck arrival time constraints. The emergency constraints include waiting time constraints and delay time constraints.

[0143] In an embodiment of the present application, the solution of the adaptive large neighborhood search algorithm based on the objective function and constraint conditions includes:

[0144] Select the insertion operator and removal operator according to the dependencies of the drone joint multimodal transport, and adjust the multimodal transport plan based on the insertion operator and removal operator.

[0145] In one embodiment of the present application, the insertion operator includes a greedy insertion operator, a transshipment insertion operator, a random insertion operator, a regret insertion operator, and a most-constrained-first insertion operator, and the removal operator includes a worst-removal operator, a random removal operator, a correlation removal operator, a route removal operator, and a node removal operator.

[0146] Each module in the above dynamic path optimization device for unmanned aerial vehicle combined with multi-modal transportation in emergency logistics can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0147] In one embodiment, a computer device is provided. The computer device can be a terminal, and its internal structure diagram can be as Figure 12 shown. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be implemented through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it realizes the dynamic path optimization method for unmanned aerial vehicle combined with multi-modal transportation in emergency logistics. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0148] Those skilled in the art can understand that Figure 12 the structure shown in

[0149] is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0150] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0151] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0152] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data that have been authorized by the user or fully authorized by all parties.

[0153] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above method embodiments. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., and are not limited thereto. The processors involved in the embodiments provided in this application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., and are not limited thereto.

[0154] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0155] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A dynamic path optimization method for unmanned aerial vehicles combined with multi-modal transportation in emergency logistics, characterized in that: The method comprises: Obtaining an initial transportation plan, wherein the initial transportation plan is a joint emergency logistics transportation route plan consisting of aviation, highway, railway, and drone; constructing a mixed integer programming model based on the initial transportation plan, defining objective functions and constraints associated with UAV joint multimodal transportation; When the initial transportation plan is interrupted, an adaptive large neighborhood search algorithm is used to solve the objective function and constraint conditions to obtain a new path plan for emergency logistics transportation.

2. The method for dynamic path optimization of unmanned aerial vehicles combined with multi-modal transportation in emergency logistics according to claim 1 is characterized in that: The initial transportation plan includes a preset take-off point for the drone, and the preset take-off point for the drone is determined based on Thiessen polygon theory.

3. The method for dynamic path optimization of unmanned aerial vehicles combined with multi-modal transportation in emergency logistics according to claim 1 is characterized in that: The objective function is: minF=F1+F2+F3+F4+F5 Among them, F is the total cost, F1 is the transportation cost, F2 is the transfer cost, F3 is the storage cost, F4 is the waiting cost, F5 is the delay penalty, K is the vehicle set, including cargo planes, trucks, trains, and drones; A is the arc set; O is the order set, including the ordinary rescue cargo order set, the emergency rescue cargo order set, and the drones in the order state; N is the location set, including the supply point, demand point, transfer point, preset drone take-off point, and road interruption point; S is the supply point, D is the demand point, T is the transfer point, T′ is the preset drone take-off point, and B is the road interruption point; is the unit cost of different items, n∈{1,1′,2,3,4,5}, is the transportation cost per ton per hour, is the transportation cost per ton per kilometer, is the loading and unloading cost per ton, is the storage cost per ton per hour, is the waiting cost per hour, Penalty for each tonne per hour of delay; is the travel time of vehicle k on arc (i, j), is the distance between vehicle k at locations i and j, q o is the quantity of order o; is a binary variable, which is 1 if order o is transported by vehicle k and uses arc (i, j), otherwise it is 0; is a binary variable, which is 1 if order o is transferred from vehicle k to vehicle l at transshipment location i, otherwise it is 0; and is the service start time of vehicle k and vehicle l at location i for order o, is the service end time of vehicle k at location i for order o, a s(o) is the pickup start time for order o at the supply point, is the waiting time of vehicle k at location i, The delay time of order o at the delivery location.

4. The method for dynamic path optimization of unmanned aerial vehicles combined with multi-modal transportation in emergency logistics according to claim 1 is characterized in that: The constraints include rescue cargo delivery guarantee constraints, multimodal transport constraints, vehicle and order flow conservation constraints, multimodal transport feature constraints, service time constraints, and emergency constraints.

5. The method for dynamic path optimization of unmanned aerial vehicles combined with multi-modal transportation in emergency logistics according to claim 4 is characterized in that: The rescue cargo delivery guarantee constraints include vehicle departure and return constraints, sub-circuit route constraints, cargo pickup and delivery constraints, and vehicle capacity constraints. The multimodal transfer constraints include transfer process constraints and transfer operation timing constraints. The vehicle and order flow conservation constraints include vehicle flow conservation constraints, order flow conservation constraints, and vehicle transportation arc constraints. The multimodal transport feature constraints include vehicle route selection constraints and cargo heterogeneity constraints. The service time constraints include service start time constraints, service end time constraints, service completion constraints, order arrival time constraints, last service constraints, time and driving distance and speed constraints, and truck arrival time constraints. The emergency constraints include waiting time constraints and delay time constraints.

6. The method for dynamic path optimization of unmanned aerial vehicles combined with multi-modal transportation in emergency logistics according to claim 1 is characterized in that: The use of the adaptive large neighborhood search algorithm to solve the objective function and the constraint conditions includes: Insertion operators and removal operators are selected according to the dependency relationship of UAV joint multimodal transport, and the multimodal transport plan is adjusted based on the insertion operators and removal operators.

7. The method for dynamic path optimization of unmanned aerial vehicles combined with multi-modal transportation in emergency logistics according to claim 6 is characterized in that: The insertion operators include greedy insertion operator, transit insertion operator, random insertion operator, regretful insertion operator, and most constrained first insertion operator; the removal operators include worst removal operator, random removal operator, related removal operator, route removal operator, and node removal operator.

8. A dynamic path optimization device for unmanned aerial vehicles combined with multi-modal transportation in emergency logistics, characterized in that: The device comprises: An initial planning module is used to obtain an initial transportation plan, which is a joint emergency logistics transportation route plan consisting of aviation, highway, railway, and drones; A mixed integer programming model building module, used to build a mixed integer programming model based on the initial transportation plan, and define objective functions and constraints associated with UAV joint multi-modal transportation; The multimodal transport re-planning module is used to adopt an adaptive large neighborhood search algorithm to solve the objective function and constraint conditions when the initial transport plan is interrupted to obtain a new path plan for emergency logistics transportation.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images