CN115577886A - Combined distribution method and system for multiple unmanned aerial vehicles - Google Patents
Combined distribution method and system for multiple unmanned aerial vehicles Download PDFInfo
- Publication number
- CN115577886A CN115577886A CN202211248044.7A CN202211248044A CN115577886A CN 115577886 A CN115577886 A CN 115577886A CN 202211248044 A CN202211248044 A CN 202211248044A CN 115577886 A CN115577886 A CN 115577886A
- Authority
- CN
- China
- Prior art keywords
- operator
- solution
- unmanned aerial
- score
- cooperation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/083—Shipping
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a combined distribution method and a system of multiple unmanned aerial vehicles, wherein the method comprises the following steps: collecting the traveling salesman data of a plurality of unmanned aerial vehicles, constructing a mathematical integer programming model according to the traveling salesman data, and verifying the correctness of the data integer programming model by using a gurobi tool; processing multi-unmanned aerial vehicle traveling salesman data by using a cooperation-based self-adaptive large-field search algorithm to obtain an optimal scheduling scheme; and (3) processing according to a mathematical integer programming model to obtain an optimal solution based on a cooperative adaptive large-field search algorithm, and circularly optimizing the path of the vehicle by using a random disturbance and 2-opt disturbance method according to the optimal solution as an optimal scheduling scheme, wherein a preset number of destruction operators and repair operators are set in a third strategy. The invention solves the technical problems of low delivery and delivery efficiency caused by limited unmanned aerial vehicle station simulation and path planning scale and more constraint limitation.
Description
Technical Field
The invention relates to the field of computational intelligence and combined distribution model optimization, in particular to a combined distribution method and a combined distribution system for multiple unmanned aerial vehicles.
Background
The rapid development of electronic commerce drives the development of express delivery industry and also promotes the continuous change of logistics distribution modes. The logistics of the last mile of courier has become a popular area of interest to retailers. Companies are always looking for fast and cost effective ways to deliver goods to their customers, and drones are receiving widespread attention as emerging delivery tools. Compared with the traditional goods delivery of trucks, the unmanned aerial vehicle delivery has more advantages. The flight of the unmanned aerial vehicle is not limited by road conditions and geographical conditions, and the space utilization efficiency can be effectively expanded; the unmanned aerial vehicle has good flexibility and convenient operation; unmanned aerial vehicle uses the battery as the power supply, and is more energy-concerving and environment-protective. However, the unmanned aerial vehicle still has the defects of small cargo capacity, limited range, easy interference and the like, and cannot meet the requirement of modern logistics when being used alone. Unmanned aerial vehicle accomplishes the delivery with ground vehicle jointly, enables unmanned aerial vehicle and reduces the delivery task of freight train in its ability within range, makes logistics distribution more high-efficient, and economic advantages is more obvious.
The vehicle and unmanned aerial vehicle are delivered in parallel, the vehicle and unmanned aerial vehicle all participate in delivery, but the two are two independent transportation units, and the operation is independent, and the mutual influence is not. Unmanned aerial vehicles individually come and go to the distribution center, and the customers around the distribution center are served in a point-to-point mode, and the vehicles are matched with cargos at one time and serve customers far away. This distribution mode is well suited for distribution to suburban warehouses.
The problem of vehicle and drone parallel delivery was first addressed by Murray and Chu, they studied the case of delivery of one truck and multiple drones and called PDSTSP (parallel hydrocarbon scheduling and sampling plan) the problem can be broken down into two classical operational raising problems, TSP and parallel machine scheduling Problem (PMS).
Murray,Chase C.,and Amanda G.Chu."The flying sidekick traveling salesman problem:Optimization of drone-assisted parcel delivery."Transportation Research Part C:Emerging Technologies 54(2015):86-109。
Considering that the flight range of the unmanned aerial vehicle in the PDSTSP model is limited, kim and Moon add an unmanned aerial vehicle station (crane station) which is close to a customer and far from a distribution center in the model and specially provide goods storage and charging service for the unmanned aerial vehicle; they assumed that the location of the drone station was known, constructed an MILP model of TSP-DS (translating and purifying protocol with a hydrocarbon station), and found that TSP-DS was more efficient than PDSTSP. Kim S, moon I.traveling Salesman Problem With aDrone Station [ J ]. IEEE Transactions on Systems, mancybernetics-Systems,2019,49 (1): 42-52.
In the prior art, for example, the prior invention patent application document CN113671985A entitled "a method for task allocation and track planning of a multi-phase and multi-base unmanned aerial vehicle" includes three steps of area setting, task allocation and track planning. The method takes heterogeneous multi-unmanned aerial vehicles which are distributed in a plurality of bases in an evacuation manner and cooperatively execute multi-target detection tasks as an application background, and realizes the task planning of the multi-base multi-unmanned aerial vehicles through a plurality of stages of regional setting, A-algorithm pre-estimation of a flight distance matrix, multi-base multi-unmanned aerial vehicle task allocation, single-unmanned aerial vehicle time sequence task allocation in the bases, flight path planning and the like. In the multi-base multi-unmanned aerial vehicle task allocation stage, when the number of tasks is large and the areas are distributed in a centralized manner, the solution is carried out based on an improved K-means algorithm, and when the number of tasks is small and the areas are distributed in an evacuation manner, the solution is carried out by adopting a depth traversal method; a single unmanned aerial vehicle time sequence task allocation stage, which is based on a TSP model and a genetic algorithm for solving; and in the flight path planning stage, optimizing the flight path based on an improved A-star algorithm and a cubic B-spline curve method. As can be seen from the description of the prior art document, the prior art uses modified a-star algorithm to obtain the estimated distance matrix, and uses the TSP traveler model to solve. For another example, patent document CN109857143B of the prior invention, "a method for planning trajectories of unmanned aerial vehicles with maximized throughput", includes the following steps: s1: establishing an unmanned aerial vehicle ground communication system model, and determining a throughput optimization objective function according to the track and the transmission power of the unmanned aerial vehicle; s2: setting a distance threshold, grouping a plurality of randomly distributed ground nodes according to the threshold, and analyzing the influence of different distance thresholds on grouping; s3: after grouping, calculating the geometric center of each group to determine the flying circle center of the unmanned aerial vehicle, solving the problem of the shortest flying path of the unmanned aerial vehicle by solving the problem of a traveling salesman and determining the communication sequence of the unmanned aerial vehicle to the grouped ground nodes; s4: determining the optimal flight radius, the flight speed and the number of turns of the unmanned aerial vehicle; s5: during optimization, firstly, the track is optimized under the condition of a certain track, secondly, the track is optimized under the condition of a certain power, and finally, the two are optimized in a combined manner. Known by the concrete realization content of this current scheme, this prior art constructs new TSP problem through predetermineeing and adds initial and final position, through solving TSP problem and figuring out unmanned aerial vehicle's shortest flight path to acquire shortest flight distance. However, in the TSP-DS model in the conventional art, there is only one unmanned station, and the location of the station is random, and cannot satisfy the needs of most customers. Simultaneously, customers in the range of the unmanned aerial vehicle station can only be served by the unmanned aerial vehicle station, and the unmanned aerial vehicle station can stop working after the vehicle returns to the warehouse, so that the assumptions do not conform to the real scene.
In summary, the prior art has the technical problems of low delivery and delivery efficiency caused by limited unmanned aerial vehicle station simulation and path planning scale and more constraint limitations.
Disclosure of Invention
The invention aims to solve the technical problem of low delivery and delivery efficiency caused by limited scale of unmanned aerial vehicle station simulation and path planning and more constraint limitations in the prior art.
The invention adopts the following technical scheme to solve the technical problems: a combined distribution method of multiple unmanned aerial vehicle stations comprises the following steps:
the first strategy is as follows: collecting the data of the traveling salesmen of the multiple unmanned aerial vehicles, constructing a mathematical integer programming model according to the data, and verifying the correctness of the data integer programming model by using a gurobi tool;
the second strategy is as follows: processing the multi-UAV traveler data by using a cooperation-based adaptive large-area search algorithm to obtain an optimal scheduling scheme, wherein the second strategy further comprises:
s1, processing preset scale problem data by using a heuristic algorithm;
s2, designing heuristic information according to preset scale problem data;
s3, constructing all initial solutions according to heuristic information, and selecting cooperative operators by using a roulette selection mechanism according to the weight of each group of operators in the initial solutions;
s4, utilizing cooperation operator U ij Carrying out destruction operation and repair operation on the current solution to generate a new field solution;
s5, if the new domain solution is better than the global optimal solution, taking the current new domain solution as the global optimal solution, and generating a cooperative operator score of the new domain solution;
s6, if the new domain solution is better than the current solution, updating the current solution by the new domain solution, and generating a cooperative operator score of the new domain solution;
s7, if the new field solution is not superior to the current solution, generating a random number R (0, 1) belonging to the field, and judging whether the random number R is larger than a set value or not;
s8, updating the using times of the cooperation operator, and updating the weight of the cooperation operator according to the fraction of the cooperation operator;
s9, processing according to the weight of the cooperative operator to obtain an optimal scheduling scheme;
the third strategy is as follows: and (3) processing according to a mathematical integer programming model to obtain an optimal solution based on a cooperative adaptive large-field search algorithm, and circularly optimizing the path of the vehicle by using a random disturbance and 2-opt disturbance method according to the optimal solution as an optimal scheduling scheme, wherein a preset number of destruction operators and repair operators are set in a third strategy.
The invention discloses a mixed integer planning model for a vehicle combined distribution problem of multiple unmanned aerial vehicle stations, and a cooperative-based adaptive large-field search algorithm for solving the problem. Combine a plurality of unmanned aerial vehicle websites and vehicle, realize the delivery goods for customer jointly. The present application addresses some of the deficiencies in the existing TSP-DS models by providing for the addition of multiple drone sites into a vehicle combination delivery model while allowing customers within service range of the drone sites to be serviced by the vehicle, and then each drone site can continue to service customers after the vehicle returns to the warehouse.
In a more specific solution, a mathematical integer programming model is constructed in a first strategy using the following logic:
Minimize z#(1)
x ij ∈{0,1},(i∈V,j∈V)#(15)
in the formula, x i,j Indicating that when the decision variable is one, the vehicle is moving from customer point i to customer point j, serving customer j,
indicating that when the decision variable is one, customer i is served by the kth drone of drone station s,
indicating that if arc (i, j) is part of a truck to drone station route, the binary variable equals 1, otherwise 0,u i ∈R:u i Represents the ith customer on the vehicle path, z represents the objective function of the mathematical integer programming model, and the final delivery time.
This patent is to TSP-DS model's shortcoming, through having expanded a plurality of unmanned aerial vehicle stations to two restraint restrictions have been relaxed, avoided single unmanned aerial vehicle website can not satisfy customer's needs, stand alone unmanned aerial vehicle station simultaneously, after the vehicle got back to the warehouse, the unmanned aerial vehicle website can continue to release unmanned aerial vehicle service customer, and established integer planning model MILP, verified the correctness of this model on some small-scale data sets in the column.
In a more specific solution, in a first strategy, the final delivery time is minimized using an objective function defined by equation (1);
considering the operating time of the truck using the constraints defined by equation (2);
the constraint defined by the formula (3) is used for ensuring that the vehicle passes through the path of the unmanned aerial vehicle station and ensuring that the departure of the vehicle section is always a part of each station-to-station route;
considering, for each site s ∈ s, the time of activation taken by the truck to reach the site, and the service time of the unmanned aerial vehicle to service the customer, using the constraint defined by equation (4);
specifying a flow rate of the truck using constraints defined by equations (5), (6), and (7);
representing that the truck is opened from the warehouse once by using the constraint defined by the formula (5);
the constraint defined by equation (6) indicates that the truck must return to the distribution center once;
the constraint defined by equation (7) represents a classical flow conservation constraint for truck tours;
the constraint defined by the formula (8) and the formula (9) is utilized to ensure that each client only visits once;
the constraint defined by equations (10, 11, 12, and 13) ensures that only customers using drones are accessed by drones and only customers using trucks are accessed by trucks;
limiting the route of the truck before the truck reaches the unmanned aerial vehicle station s epsilon s by using the constraint defined by the formula (14);
eliminating the secondary tour of the truck by using the formula (15) and the formula (16);
the decision variable domain is specified by equations (17), (18), (19) and (20).
The invention relates to a method for selecting the address of an unmanned aerial vehicle station, which is characterized in that a clustering method is used for dividing customers into a plurality of clustering clusters according to the distribution condition of the customers, the clustering center of each clustering cluster is regarded as an unmanned aerial vehicle station, and the unmanned aerial vehicle station can serve the customers in the clustering cluster range. When the vehicle starts from the warehouse, partial customers are served, then all the unmanned aerial vehicles are activated in sequence, and the customers in the respective areas are served, so that the pressure of the vehicle can be greatly reduced, a large amount of manpower and material resources are reduced, the distribution time is shortened, and the problem of the last kilometer is effectively solved.
In a more specific technical solution, in step S3, the cooperation operator includes: a destruction operator and a repair operator.
In a more specific technical solution, in step S5, the cooperative operator score of the new domain solution is generated with the following logic:
Score(U ij )＝Score(U ij )+1.6。
in a more specific technical solution, in step S6, the cooperative operator score of the new domain solution is generated with the following logic:
Score(U ij )＝Score(U ij )+1.2。
in a more specific technical solution, in step S7, the cooperative operator score of the new domain solution is generated with the following logic:
Score(U ij )＝Score(U ij )+0.8；
and when the current solution is not updated, generating a co-operator score for the new domain solution with the following logic: score (U) ij )＝Score(U ij )+0.2。
In a more specific technical solution, step S8 includes:
s81, updating the weight of the cooperation operator by the following logic to obtain the weight of the first cooperation operator:
weight(U ij )＝weight(U ij )*r+(1-γ)score(U ij )/num(U ij )
in the formula, r is an updating parameter;
s82, updating the weight of the cooperation operator by the following logic according to the weight of the first cooperation operator:
the invention also considers that the integral programming model can not solve the problem in reasonable time along with the increase of the scale of the customer, so the invention introduces An Adaptive large-scale search algorithm (An Collaboration-based Adaptive large-scale neighbor search algorithm for TSP with Drone states) based on cooperation to solve the large-scale problem, verifies the superiority of the Adaptive large-scale search algorithm based on cooperation provided by the invention on a large-scale problem calculation example, and ensures the combined distribution planning effect of the invention applied to a large-scale application scene.
In a more specific aspect, the destroy operator includes: a random-based destruction operator, a neighborhood-based destruction operator, a region destruction operator based on drone station scanning, a maximum distance-based destruction operator, a maximum time-saving-based destruction operator, a tabu-based destruction operator.
The repair operator includes:
random repair operators, greedy repair operators, and historical knowledge repair operators.
In a more specific aspect, a multiple-drone station combined delivery system includes:
the system comprises a first subsystem, a second subsystem and a third subsystem, wherein the first subsystem is used for collecting traveling salesman data of a plurality of unmanned planes, constructing a mathematical integer programming model according to the traveling salesman data, and verifying the correctness of the data integer programming model by using a gurobi tool;
the second subsystem is used for processing the multi-unmanned aerial vehicle traveler data by using a cooperation-based self-adaptive large-field search algorithm to obtain an optimal scheduling scheme, and is connected with the first subsystem, wherein the second subsystem further comprises:
the scale problem processing module is used for processing preset scale problem data by utilizing a heuristic algorithm;
the heuristic information module is used for designing heuristic information according to preset scale problem data and is connected with the scale problem processing module;
the cooperative operator selection module is used for constructing all initial solutions according to heuristic information, selecting cooperative operators by using a roulette selection mechanism according to the weight of each group of operators in the initial solutions, and is connected with the heuristic information module;
a new domain solution generation module to utilize the cooperation operator U ij Carrying out destruction operation and repair operation on the current solution to generate a new domain solution, wherein the new domain solution generation module is connected with the cooperation operator selection module;
the global optimal solution updating module is used for taking the current new field solution as the global optimal solution and generating a cooperative operator score of the new field solution when the new field solution is superior to the global optimal solution, and the global optimal solution updating module is connected with the new field solution generating module;
the current solution updating module is used for updating the current solution by the new field solution and generating a cooperative operator score of the new field solution when the new field solution is superior to the current solution, and the current solution updating module is connected with the new field solution generating module;
the random number judging module is used for generating a random number R epsilon (0, 1) when the new field solution is not superior to the current solution, judging whether the random number R is larger than a set value or not, and connecting the random number judging module with the new field solution generating module;
the weight updating module is used for updating the using times of the cooperative operator and updating the weight of the cooperative operator according to the cooperative operator score, and the weight updating module is connected with the random number judging module;
s9, processing according to the weight of the cooperative operator to obtain an optimal scheduling scheme;
and the third subsystem is used for processing by utilizing a cooperative self-adaptive large-field search algorithm according to the mathematical integer programming model to obtain an optimal solution, and circularly optimizing the path of the vehicle by utilizing a random disturbance and 2-opt disturbance method according to the optimal scheduling scheme, wherein a preset number of destruction operators and repair operators are set in a third strategy, and the third subsystem is connected with the first subsystem.
Compared with the prior art, the invention has the following advantages: the invention discloses a mixed integer planning model for a vehicle combined distribution problem of multiple unmanned aerial vehicle stations, and a cooperative-based adaptive large-field search algorithm for solving the problem. Combine a plurality of unmanned aerial vehicle websites and vehicle, realize the delivery goods for customer jointly. The present application addresses some of the deficiencies in the existing TSP-DS models by providing for the addition of multiple drone sites into a vehicle combined delivery model while allowing customers within the service range of the drone sites to be serviced by the vehicle, and then each drone site can continue to service the customer after the vehicle returns to the warehouse.
This patent is to TSP-DS model's shortcoming, through having expanded a plurality of unmanned aerial vehicle stations to two restraint restrictions have been relaxed, avoided single unmanned aerial vehicle website can not satisfy customer's needs, stand alone unmanned aerial vehicle station simultaneously, after the vehicle got back to the warehouse, the unmanned aerial vehicle website can continue to release unmanned aerial vehicle service customer, and established integer planning model MILP, verified the correctness of this model on some small-scale data sets in the column.
The invention relates to a method for selecting the address of an unmanned aerial vehicle station, which is characterized in that a clustering method is used for dividing customers into a plurality of clustering clusters according to the distribution condition of the customers, the clustering center of each clustering cluster is regarded as an unmanned aerial vehicle station, and the unmanned aerial vehicle station can serve the customers in the clustering cluster range. When the vehicle starts from the warehouse, partial customers are served, then all the unmanned aerial vehicles are activated in sequence, and the customers in the respective areas are served, so that the pressure of the vehicle can be greatly reduced, a large amount of manpower and material resources are reduced, the distribution time is shortened, and the problem of the last kilometer is effectively solved.
The invention also considers that the integral planning model can not solve the problem in reasonable time along with the increase of the scale of the customers, so the invention introduces a cooperative-based Adaptive large area search algorithm (TSP with Dry states) to solve the large-scale problem, verifies the superiority of the cooperative-based Adaptive large area search algorithm on a large-scale problem calculation example, and ensures the combined distribution planning effect of the invention in the large-scale application scene.
The invention solves the technical problems of low delivery and delivery efficiency caused by limited unmanned aerial vehicle station simulation and path planning scale and more constraint limitations in the prior art.
Drawings
Fig. 1 is a conceptual diagram of a combined distribution model of multiple unmanned aerial vehicles in embodiment 1 of the present invention;
fig. 2 is a model coding structure diagram according to embodiment 1 of the present invention.
FIG. 3 is a schematic diagram of the steps of the cooperative-based adaptive large-scale search algorithm of embodiment 1 of the present invention;
FIG. 4 is a flowchart of the collaborative adaptive domain search algorithm according to embodiment 1 of the present invention;
FIG. 5 is a schematic diagram of a conventional ALNS operator selection mechanism according to embodiment 1 of the present invention;
fig. 6 is a schematic diagram of a cooperative ALNS-based operator selection mechanism according to embodiment 1 of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
Model assumptions
1. Considering the limitation that the drone does not receive the road surface, the invention uses euclidean metrics to calculate the flight path length of the drone while using manhattan distance to describe the truck travel distance to simulate city block distance. The time spent on the project in loading operations, battery exchange drones, and unloading packages is neglected.
2. The unmanned stations are determined based on the location distribution of the customers and all of the unmanned stations must be activated.
3. The capacity of the drone is limited and can only serve one customer at a time.
4. The speed of the drone is 1.5 times the speed of the truck.
5. Due to practical limitations, such as limited payload or flight range of drones, there are some customers that cannot be served by drones.
Solution X consists of two parts. First part X 1 It is determined whether the customer is serviced by a truck (denoted by 0) or a drone (denoted by 1) in index order. Second part X 2 A sequence of nodes representing truck visits (client nodes or drone stations).
As shown in fig. 1, a solution with 18 client nodes and 3 drone stations (each carrying a drone) is shown.
As shown in fig. 2, in the present embodiment, the client nodes 5 and 9 are drone sites. A set of customers that can be served by the unmanned aerial vehicle station s1 = {1,2 3 4}, a set of customers that can be served by the unmanned aerial vehicle station s1 = {7, 8, 9, 10, 11, 12}, and a set of customers that can be served by the unmanned aerial vehicle station s3 = [13,14,15,16,17 } ]]. The truck departs from warehouse 0, serves customer node 1, and arrives at unmanned station s1. Thus, X 1 The cell of nodes 1 and s1 in (1) is 0. Station s1 is activated by a truck and the drone takes off from the station, serving customer nodes 2, 3, 4, 5 and 6. Subsequent trips of the truck follow the same rules.
The present invention solves this problem in two ways.
First describing this problem in mathematical language, a mixed integer programming model is constructed.
Consider that the MILP model can only solve small scale problems in a reasonable time.
In order to solve the large-scale problem calculation, the invention provides a cooperative-based self-adaptive large-field search algorithm to solve the problem. Meanwhile, the invention provides a new damage operator and a repair operator according to the characteristics of the problem model.
The patent discloses a solution method of a vehicle combined distribution model based on multiple unmanned aerial vehicles:
the method comprises the following steps: establishing mathematical integer programming model
And establishing a mathematical model of the problem according to the characteristics of the traveler problem of the multiple unmanned aerial vehicles. Meanwhile, the integer programming model is constructed by using gurobi to verify the correctness of the model provided by the invention on a small-scale or medium-scale example.
The mathematical integer programming model is as follows, and the English name of the model is travel Salesman Problem with Multiple draw statons, TSP-MDS for short.
·x i,j When the decision variable is one, the vehicle moves from customer point i to customer point j, serving customer j.
·
When the decision variable is one, the customer i is served by the kth unmanned aerial vehicle of the unmanned aerial vehicle station s;
·
if arc (i, j) is part of a truck-to-drone station route, the binary variable equals 1, otherwise it is 0. U.u i ∈R:u i Is the ith customer on his vehicle path.
Z is the objective function of the model, the last delivery time.
The TSP-MDS can be formulated as follows:
The model is:
Minimize z#(1)
subject to:
x ij ∈{0,1},(i∈V,j∈V)#(15)
The objective function (1) minimizes the final delivery time. The constraint (2) takes into account the operating time of the truck. Constraints (3) ensure that the vehicle follows the path of the unmanned aerial vehicle station and that the departure from the vehicle section is always part of each arrival route. In the constraint (4), for each site s e s, the invention considers the time of activation that the truck takes to reach the site, and the service time that the drone station services the customer. The flow of the designated truck is restricted (5) to (7). Constraint (5) indicates that the truck is driven from the warehouse once and constraint (6) indicates that the truck must be returned to the distribution center once. Constraint (7) is the classical flow conservation constraint for truck tours. Constraints (8) - (9) mean that each customer is only visited once, either truck or drone. Constraints (10) - (13) ensure that customers using only drones must be accessed by drones and customers using only trucks must be accessed by trucks. The constraint (14) limits the route of the truck before reaching the unmanned station s e s. Constraints (15) - (16) eliminate the secondary tour of the truck. Finally, constraints (17) through (20) specify decision variable domains.
As shown in fig. 3, the cooperative adaptive large-area search algorithm adopted in the cooperative adaptive large-area search based multi-unmanned aerial vehicle station combined distribution method provided by the invention comprises the following steps:
as shown in fig. 4, step S1: the large scale problem is solved by a heuristic algorithm. Initializing a maximum number of iterations T max Initial temperature
Termination temperature
And (5) destroying the operator set D and repairing the operator set R. Initial combination operator set U = { U = ij ＝(D i ,R j ),1<＝i<＝|D|,1<＝j<= R | }, initialize the weight of each set of cooperative operators
1, the current iteration number t =0;
step S2: designing heuristic information according to the problem, and constructing all initial solutions through the heuristic information, referring to fig. 2;
and step S3: t = T0, a cooperative operator is first selected using the roulette selection mechanism based on the weights of each set of operators, the cooperative operator including a destruction operator and a repair operator.
And step S4: using selected co-operators U ij Carrying out destruction operation and repair operation on the current solution to generate a new domain solution;
step S5: if the newly generated solution is better than the globally optimal solution, it is changed to the globally optimal solution, while the scores of the co-operators of the new solution are generated, score (U) ij )＝Score(U ij )+1.6；
Step S6: if the newly generated solution is better than the current solution, the current solution is updated with the new solution, and at the same time, the Score of the co-operator of the new solution, score (U), will be generated ij )＝Score(U ij )+1.2；
Step S7: if the newly generated solution is different from the current solution, simultaneously generating a random number R epsilon (0, 1), and judging whether R is larger than a set value; in the embodiment, according to the experiment and the working experience, the value e- ((f (new solution) -f (current solution))/T) is set; the current solution is updated with the new solution and the Score of the co-operator (U) of the new solution will be generated ij )＝Score(U ij ) +0.8, otherwise the current solution is not updated, while the Score of the co-operator Score (U) for the new solution will be generated ij )＝Score(U ij )+0.2；
Step S8: updating the number of uses num (U) of the cooperation operator ij )＝num(U ij ) +1, updating the weight of the cooperation operator according to the fraction of the cooperation operator, wherein the updating formula is as follows: weight (U) ij )＝weight(U ij )*r+(1-γ)score(U ij )/num(U ij ) R is an updating parameter, r =0.3 according to experiment and working experience, and the weight of the cooperation operator is updated according to the weight of the cooperation operator
Step S9: t = α × T0, α is a temperature cooling factor, α =0.97
Step S10: let t = t +1 and,
step S11: t > = Tmax, and the optimal scheduling scheme is output.
In this embodiment, the second solution method is a collaborative based adaptive large domain search algorithm. The adaptive large neighborhood search algorithm is a meta-heuristic proposed by Ropke and Pisinger in 2006, which is an extension of the LNS heuristic proposed by Shaw (1998), except that it allows the user to use various destruction and insertion operators. By designing a plurality of groups of destruction operators and repair operators, the solution space search range is expanded, the current solution is improved, the well-represented destruction and repair method correspondingly obtains high scores, and the weight is higher. In each iteration, each damage and repair operator is selected and weight adjusted according to the previous performance, and an efficient combination method is used to improve the optimization capability of the algorithm, so that the optimal solution is found.
As shown in fig. 6, in the embodiment, the adaptive domain search algorithm based on cooperation is provided, in which, based on the conventional adaptive domain search algorithm, the cooperative relationship between the destruction operators and the repair operators is concerned, the destruction operators and the repair operators are combined one by one, each destruction operator and each repair operator form a team, each team is assigned with a score, and the destruction operators and the repair operators are selected according to the score of the combined operator of each team.
As shown in fig. 5, in the present embodiment, a score is assigned to the destroy operator and the repair operator respectively in the conventional adaptive large-scale domain search algorithm, and in the evolutionary process, the destroy operator and the repair operator are selected according to their scores respectively, and the selection processes of the destroy operator and the repair operator are independent.
Meanwhile, according to the characteristics of a combined distribution model of the unmanned aerial vehicle stations, six destruction operators and three repair operators are provided.
For the destruction operator, the present invention assumes that when a customer node i is destroyed, the delivery mode of the customer node becomes a vehicle and is added to the destruction set L.
Random-based destruction operator: phi customers are randomly selected from a set of customers that can be serviced by both the vehicle and the drone for destruction, and a destruction set L is added.
Neighborhood-based destruction operator: a customer is first randomly selected from a set of customers that can be serviced by both the vehicle and the drone, then the euclidean distance of the remaining customers to the customer is calculated, the phi customers closest to the customer are deleted, and they are added to the destruction set L.
Region destruction operator based on unmanned aerial vehicle station scanning: first, a drone station s is randomly selected from a set of drone stations. The drone station si will then be the pole with the pole axis pointing to the right of the pole. Seed angles e are randomly selected from the continuously and uniformly distributed U (0 degrees and 360 degrees), and the polar axis is taken as the direction. The selected seed angle draws a straight line I from sy. Then, the φ client nodes nearest to the first line L are removed from the customer set and added to the destruction set L.
Maximum distance based destroy operator: for each drone station, the invention calculates the service distance of customers within the service range of the drone station from the drone station, deletes φ/n customers, n being the number of drone stations, and adds to the destruction combination L.
Destruction operators based on maximum time savings: one drone station is randomly selected. The service time for each client node is then recorded from the set of customers that the drone station may service, the first φ client nodes with the highest service times are deleted, and the deleted nodes are added to the destruction union L.
Destruction operator based on tabu: the operator will record the number of times each node was corrupted. The phi client nodes with the least number of damages are added into the damage set L.
For the repair operator, the customers in the damage set L are used as input, and new service modes are allocated to the customers again.
Random-based repair operator: the operator will then hand the customer in disruption set L to a vehicle or a drone station for service.
Greedy-based repair operator: the operator selects a better delivery mode for each customer in the disruption set L from the two service modes of truck and drone sites. If the customer is serviced with a vehicle, which would reduce the time to final delivery, the vehicle is used, otherwise the unmanned station service is used.
Historical knowledge-based repair operator: the operator records historical optimal solutions in an algorithm iteration process, and a knowledge matrix of a customer delivery mode is constructed. First, the number of times the client node uses the unmanned aerial vehicle service and the vehicle service is individually calculated for each client node using the historical optimal solution set, and if the number of times the vehicle is used is greater than the number of times the unmanned aerial vehicle is used, a vehicle will be selected to service the customer.
And (3) solution optimization:
considering that for a determined solution the service time of each drone station is determined, the target value affecting the solution is the activation time of each drone station, related to the path of the vehicle. For each new solution generated by the combination operator, the invention continuously optimizes the path of the vehicle by using a random disturbance method and a 2-opt disturbance method, so that the vehicle can more quickly activate the unmanned station, and the unmanned station can serve more customers, thereby reducing the final delivery time.
In the TSP-DS model, there is only one unmanned station, and the location of the station is random and cannot meet the needs of most customers. Simultaneously, customers in the range of the unmanned aerial vehicle station can only be served by the unmanned aerial vehicle station, and the unmanned aerial vehicle station can stop working after the vehicle returns to the warehouse, so that the assumptions do not conform to the real scene. This patent addresses some of the deficiencies in the existing TSP-DS model by providing for the addition of multiple drone sites into a vehicle combination delivery model while allowing customers within the service range of the drone sites to be serviced by the vehicle, and then after the vehicle returns to the warehouse, each drone site can continue to service the customer.
In the embodiment, the problem of address selection of unmanned aerial vehicle sites is crucial, the invention applies a clustering method to divide customers into a plurality of clustering clusters according to the distribution condition of the customers, the clustering center of each clustering cluster is regarded as an unmanned aerial vehicle station, and the unmanned aerial vehicle station can serve the customers in the clustering cluster range. When the vehicle starts from the warehouse, partial customers are served, then all the unmanned aerial vehicles are activated in sequence, and the customers in the respective areas are served, so that the pressure of the vehicle can be greatly reduced, a large amount of manpower and material resources are reduced, the distribution time is shortened, and the problem of the last kilometer is effectively solved.
The invention discloses a mixed integer planning model for a vehicle combined distribution problem of multiple unmanned aerial vehicle stations, and a cooperative-based adaptive large-field search algorithm for solving the problem. Combine a plurality of unmanned aerial vehicle websites and vehicle, realize the delivery goods for customer jointly. The present application addresses some of the deficiencies in the existing TSP-DS models by providing for the addition of multiple drone sites into a vehicle combined delivery model while allowing customers within the service range of the drone sites to be serviced by the vehicle, and then each drone site can continue to service the customer after the vehicle returns to the warehouse.
This patent is to TSP-DS model's shortcoming, through having expanded a plurality of unmanned aerial vehicle stations to two restraint restrictions have been relaxed, avoided single unmanned aerial vehicle website can not satisfy customer's needs, stand alone unmanned aerial vehicle station simultaneously, after the vehicle got back to the warehouse, the unmanned aerial vehicle website can continue to release unmanned aerial vehicle service customer, and established integer planning model MILP, verified the correctness of this model on some small-scale data sets in the column.
The invention relates to a method for selecting the address of an unmanned aerial vehicle station, which is characterized in that a clustering method is used for dividing customers into a plurality of clustering clusters according to the distribution condition of the customers, the clustering center of each clustering cluster is regarded as an unmanned aerial vehicle station, and the unmanned aerial vehicle station can serve the customers in the clustering cluster range. When the vehicle starts from the warehouse, partial customers are served, then all the unmanned aerial vehicles are activated in sequence, and the customers in the respective areas are served, so that the pressure of the vehicle can be greatly reduced, a large amount of manpower and material resources are reduced, the distribution time is shortened, and the problem of the last kilometer is effectively solved.
The invention also considers that the integral programming model can not solve the problem in reasonable time along with the increase of the scale of the customer, so the invention introduces An Adaptive large-scale search algorithm (An Collaboration-based Adaptive large-scale neighbor search algorithm for TSP with Drone states) based on cooperation to solve the large-scale problem, verifies the superiority of the Adaptive large-scale search algorithm based on cooperation provided by the invention on a large-scale problem calculation example, and ensures the combined distribution planning effect of the invention applied to a large-scale application scene.
The invention solves the technical problems of low delivery efficiency caused by limited unmanned aerial vehicle station simulation and path planning scale and more constraint limitation in the prior art.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A combined distribution method of a plurality of unmanned aerial vehicles is characterized by comprising the following steps:
the first strategy is as follows: collecting the traveling salesman data of a plurality of unmanned aerial vehicles, constructing a mathematical integer programming model according to the traveling salesman data, and verifying the correctness of the data integer programming model by using a gurobi tool;
the second strategy is as follows: processing the multi-UAV traveler data by using a cooperation-based adaptive large-area search algorithm to obtain an optimal scheduling scheme, wherein the second strategy further comprises:
s1, processing preset scale problem data by using a heuristic algorithm;
s2, designing heuristic information according to the preset scale problem data;
s3, constructing all initial solutions according to the heuristic information, and selecting cooperative operators by using a roulette selection mechanism according to the weight of each group of operators in the initial solutions;
s4, utilizing the cooperation operator U ij Carrying out destruction operation and repair operation on the current solution to generate a new field solution;
s5, if the new domain solution is better than the global optimal solution, taking the current new domain solution as the global optimal solution, and generating a cooperative operator score of the new domain solution;
s6, if the new domain solution is better than the current solution, updating the current solution by the new domain solution, and generating the cooperative operator score of the new domain solution;
s7, if the new field solution is not superior to the current solution, generating a random number R epsilon (0, 1), and judging whether the random number R is larger than a set value or not;
s8, updating the using times of the cooperation operator, and updating the weight of the cooperation operator according to the score of the cooperation operator;
s9, processing according to the weight of the cooperation operator to obtain an optimal scheduling scheme;
the third strategy is as follows: and processing according to the mathematical integer programming model to obtain an optimal solution based on a cooperative self-adaptive large-field search algorithm, and circularly optimizing the path of the vehicle by using a random disturbance and 2-opt disturbance method according to the optimal scheduling scheme, wherein a preset number of destruction operators and repair operators are set in the third strategy.
2. The method of claim 1, wherein the mathematical integer programming model is constructed in the first strategy using the following logic:
Minimize z#(1)
x ij ∈{0，1}，(i∈V，j∈V)#(15)
in the formula, x i，j Indicating that when the decision variable is one, the vehicle is moving from customer point i to customer point j, serving customer j,
indicating that when the decision variable is one, customer i is served by the kth drone of drone station s,
indicating that if arc (i, j) is part of the truck-to-drone station route, the binary variable equals 1, otherwise 0,u i ∈R:u i Represents the ith customer on the vehicle path, z represents the objective function of the mathematical integer programming model, and the final delivery time.
3. A multi-robot station combined distribution method according to claim 2, characterized in that in said first strategy,
minimizing a final delivery time using an objective function defined by said equation (1);
considering the operating time of the truck using the constraints defined by said equation (2);
ensuring that the vehicle passes through the path of the unmanned aerial vehicle station by using the constraint defined by the formula (3), and ensuring that the departure of the vehicle section is always a part of each station-arriving route;
considering, for each site s e s, the time of activation taken by the truck to reach the site, and the service time of the unmanned aerial vehicle to service the customer, using the constraint defined by said formula (4);
specifying a flow rate of the truck using constraints defined by the equation (5), the equation (6), and the equation (7);
representing that the truck is opened from the warehouse once by using the constraint defined by the formula (5);
representing that the truck must return to the distribution center once using the constraint defined by the equation (6);
representing a classical flow conservation constraint of the truck tour with the constraint defined by said equation (7);
using the constraints defined by said equations (8) and (9) to make each of said clients access only once;
the constraint conditions defined by the equations (10, 11, 12, and 13) ensure that only customers using drones are accessed by drones and only customers using trucks are accessed by trucks;
limiting a route of the truck until the truck reaches the unmanned aerial vehicle station s e s using the constraint defined by the equation (14);
eliminating the secondary tour of the truck by using the formula (15) and the formula (16);
a decision variable domain is specified using equation (17), equation (18), equation (19), and equation (20).
4. A method for delivering in combination of multiple unmanned aerial vehicles according to claim 1, wherein in step S3, said cooperation operator comprises: a destruction operator and a repair operator.
5. The method for delivering in combination of multiple unmanned aerial vehicles according to claim 1, wherein in step S5, the cooperative operator score of the new domain solution is generated with the following logic:
Score(U ij )＝Score(U ij )+1.6。
6. the method for dispatching in combination of multiple unmanned aerial vehicles stations of claim 1, wherein in step S6, the cooperative operator score of the new domain solution is generated with the following logic:
Score(U ij )＝Score(U ij )+1.2。
7. the method for delivering in combination of multiple unmanned aerial vehicles according to claim 1, wherein in step S7, the cooperative operator score of the new domain solution is generated with the following logic:
Score(U ij )＝Score(U ij )+0.8；
and when the current solution is not updated, generating the partner operator score for the new domain solution with the logic: score (U) ij )＝Score(U ij )+0.2。
8. The method for the combined distribution of the plurality of the unmanned aerial vehicles as claimed in claim 1, wherein the step S8 comprises:
s81, updating the weight of the cooperation operator by the following logic to obtain the weight of the first cooperation operator:
weight(U ij )＝weight(U ij )*r+(1-γ)score(U ij )/num(U ij )
in the formula, r is an updating parameter;
s82, updating the weight of the cooperation operator by the following logic according to the weight of the first cooperation operator:
9. a method for combined distribution of multiple unmanned aerial vehicles stations as claimed in claim 1, wherein said destruction operator comprises: a random-based destruction operator, a neighborhood-based destruction operator, a region destruction operator based on unmanned aerial vehicle station scanning, a maximum distance-based destruction operator, a maximum time-saving-based destruction operator, and a taboo-based destruction operator;
the repair operator includes:
random repair operators, greedy repair operators, and historical knowledge repair operators.
10. A multi-drone station combined delivery system, comprising:
the system comprises a first subsystem, a second subsystem and a third subsystem, wherein the first subsystem is used for collecting the traveler data of a plurality of unmanned aerial vehicles, so as to construct a mathematical integer programming model, and verifying the correctness of the data integer programming model by using a gurobi tool;
the second subsystem is used for processing multi-unmanned aerial vehicle traveling salesman data by using a cooperation-based adaptive large-area search algorithm to obtain an optimal scheduling scheme, and is connected with the first subsystem, wherein the second subsystem further comprises:
the scale problem processing module is used for processing preset scale problem data by utilizing a heuristic algorithm;
the heuristic information module is used for designing heuristic information according to the preset scale problem data and is connected with the scale problem processing module;
the cooperative operator selection module is used for constructing all initial solutions according to the heuristic information, selecting cooperative operators by using a roulette selection mechanism according to the weight of each group of operators in the initial solutions, and is connected with the heuristic information module;
a new domain solution generation module to utilize the cooperation operator U ij Carrying out destruction operation and repair operation on the current solution to generate a new domain solution, wherein the new domain solution generation module is connected with the cooperation operator selection module;
the global optimal solution updating module is used for taking the current new domain solution as the global optimal solution and generating a cooperative operator score of the new domain solution when the new domain solution is superior to the global optimal solution, and the global optimal solution updating module is connected with the new domain solution generating module;
a current solution update module configured to update the current solution with the new domain solution and generate the cooperative operator score of the new domain solution when the new domain solution is better than the current solution, the current solution update module being connected to the new domain solution generation module;
the random number judging module is used for generating a random number R epsilon (0, 1) when the new field solution is not superior to the current solution, judging whether the random number R is larger than a set value or not, and is connected with the new field solution generating module;
the weight updating module is used for updating the using times of the cooperation operator and updating the weight of the cooperation operator according to the score of the cooperation operator, and the weight updating module is connected with the random number judging module;
s9, processing according to the weight of the cooperation operator to obtain an optimal scheduling scheme;
and the third subsystem is used for processing by utilizing a cooperative self-adaptive large-field search algorithm according to the mathematical integer programming model to obtain an optimal solution, and circularly optimizing the path of the vehicle by utilizing a random disturbance and 2-opt disturbance method according to the optimal scheduling scheme, wherein a preset number of destructive operators and repair operators are set in the third strategy, and the third subsystem is connected with the first subsystem.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211248044.7A CN115577886A (en) | 2022-10-12 | 2022-10-12 | Combined distribution method and system for multiple unmanned aerial vehicles |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211248044.7A CN115577886A (en) | 2022-10-12 | 2022-10-12 | Combined distribution method and system for multiple unmanned aerial vehicles |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115577886A true CN115577886A (en) | 2023-01-06 |
Family
ID=84585905
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211248044.7A Pending CN115577886A (en) | 2022-10-12 | 2022-10-12 | Combined distribution method and system for multiple unmanned aerial vehicles |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115577886A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116341781A (en) * | 2023-03-28 | 2023-06-27 | 暨南大学 | Path planning method based on large-scale neighborhood search algorithm and storage medium |
| CN116757585A (en) * | 2023-08-22 | 2023-09-15 | 安徽大学 | Unmanned aerial vehicle and unmanned aerial vehicle collaborative distribution method based on mobile edge calculation |
| CN117726059A (en) * | 2024-02-08 | 2024-03-19 | 深圳大学 | Truck unmanned aerial vehicle task allocation method under time window constraint |
-
2022
- 2022-10-12 CN CN202211248044.7A patent/CN115577886A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116341781A (en) * | 2023-03-28 | 2023-06-27 | 暨南大学 | Path planning method based on large-scale neighborhood search algorithm and storage medium |
| CN116757585A (en) * | 2023-08-22 | 2023-09-15 | 安徽大学 | Unmanned aerial vehicle and unmanned aerial vehicle collaborative distribution method based on mobile edge calculation |
| CN116757585B (en) * | 2023-08-22 | 2023-10-31 | 安徽大学 | Unmanned aerial vehicle and unmanned aerial vehicle collaborative distribution method based on mobile edge calculation |
| CN117726059A (en) * | 2024-02-08 | 2024-03-19 | 深圳大学 | Truck unmanned aerial vehicle task allocation method under time window constraint |
| CN117726059B (en) * | 2024-02-08 | 2024-04-30 | 深圳大学 | Truck unmanned aerial vehicle task allocation method under time window constraint |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Raj et al. | The multiple flying sidekicks traveling salesman problem with variable drone speeds | |
| CN115577886A (en) | Combined distribution method and system for multiple unmanned aerial vehicles | |
| Psaraftis | An exact algorithm for the single vehicle many-to-many dial-a-ride problem with time windows | |
| Manchella et al. | Flexpool: A distributed model-free deep reinforcement learning algorithm for joint passengers and goods transportation | |
| Liu et al. | Mobility-aware dynamic taxi ridesharing | |
| CN112053117B (en) | Collaborative distribution path planning method and device | |
| Peng et al. | Wide-area vehicle-drone cooperative sensing: Opportunities and approaches | |
| CN113848970A (en) | Multi-target collaborative path planning method for vehicle and unmanned aerial vehicle | |
| Küçükoğlu et al. | A hybrid meta-heuristic algorithm for vehicle routing and packing problem with cross-docking | |
| CN114841582A (en) | Truck and unmanned aerial vehicle cooperative distribution method | |
| Luo et al. | Multi-objective optimization algorithm with adaptive resource allocation for truck-drone collaborative delivery and pick-Up services | |
| Zhang et al. | A review on the truck and drone cooperative delivery problem | |
| Hani et al. | Simulation based optimization of a train maintenance facility | |
| Montaña et al. | A novel mathematical approach for the Truck-and-Drone Location-Routing Problem | |
| Ren et al. | A dynamic routing optimization problem considering joint delivery of passengers and parcels | |
| Liu et al. | The optimization of the" UAV-vehicle" joint delivery route considering mountainous cities | |
| CN114611794A (en) | Vehicle-machine cooperative pick-and-place path optimization method and system based on sub-heuristic algorithm | |
| Meng et al. | The multi-visit drone-assisted pickup and delivery problem with time windows | |
| Lin et al. | Model and hybrid algorithm of collaborative distribution system with multiple drones and a truck | |
| Hou et al. | An Improved Particle Swarm Optimization Algorithm for the Distribution of Fresh Products. | |
| CN117032298A (en) | Unmanned aerial vehicle task allocation planning method under synchronous operation and cooperative distribution mode of truck unmanned aerial vehicle | |
| Ramírez-Villamil et al. | Integrating Clustering Methodologies and Routing Optimization Algorithms for Last-Mile Parcel Delivery | |
| Song et al. | Path Planning for Multi-Vehicle-Assisted Multi-UAVs in Mobile Crowdsensing | |
| CN113706081A (en) | Unmanned aerial vehicle goods taking and delivering system and method based on urban roof automatic express device | |
| Zhang et al. | Cooperative Route Planning for Fuel-constrained UGV-UAV Exploration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |