Method for dispatching vehicles in logistics site by considering road limitation

Abstract

The invention provides a method for dispatching vehicles in a logistics site considering road limitation. The method comprises the following steps: abstracting a road network in the logistics site into a topological road network based on a space-time network; constructing a vehicle path planning model in the logistics site based on the space-time network; solving the vehicle path planning model; carrying out vehicle-cargo matching on the idle vehicle and the to-be-transported cargo; a loading area in the logistics site sends a request, and the idle vehicles go to the loading area to carry out loading work after receiving the request; proposing an active real-time scheduling method based on a dynamic relocation strategy for the working process of the vehicle in the loading area; and after all loading and unloading work is completed, a shortest departure path is planned for the vehicle to leave the logistics site. According to the invention, vehicle path planning can be dynamically adjusted, the operation process of the vehicle in the unloading area and the loading area is optimized, meanwhile, efficient matching of the vehicle and goods is realized, the waiting time and the empty driving distance of the vehicle are reduced to the maximum extent, the overall operation efficiency of a logistics site is improved, and the logistics transportation cost is reduced.

Description

Technical Field

[0001] The invention relates to a vehicle dispatching method in a logistics site taking road restrictions into consideration, and belongs to the technical field of logistics. Background Art

[0002] With the rapid development of e-commerce and logistics industries, the operational efficiency and management capabilities of logistics sites have become key factors affecting the response speed and cost control of the entire supply chain. In logistics sites, vehicle scheduling is the core link connecting cargo loading and unloading, storage and transportation. Its efficiency and flexibility are directly related to the quality of logistics services and customer satisfaction. However, in actual operations, logistics sites often face complex road restrictions, such as one-way driving, height restrictions, and no-driving periods, which greatly increase the difficulty of vehicle scheduling.

[0003] Traditional vehicle scheduling methods are mostly based on static road network information and fixed vehicle route planning, ignoring the dynamically changing traffic conditions and real-time road restrictions within the logistics site, resulting in inefficient vehicle scheduling, high vehicle idle rates, and even traffic congestion and delays.

[0004] In order to solve the above problems, a new vehicle scheduling method is needed that can comprehensively consider the road restrictions, real-time traffic conditions and vehicle loading and unloading operation requirements in the logistics site. This method should be able to dynamically adjust vehicle path planning, optimize the operation process of vehicles in the unloading and loading areas, and achieve efficient matching of vehicles and goods, minimize vehicle waiting time and empty driving distance, and improve the overall operational efficiency of the logistics site.

[0005] Therefore, a vehicle scheduling method considering road restrictions in logistics sites is proposed. This method can accurately reflect the road restrictions and real-time traffic conditions in logistics sites by constructing a topological road network model based on a spatiotemporal network; by constructing a vehicle path planning model in the unloading area and solving the optimal solution, the vehicle can operate quickly and efficiently in the unloading area; through vehicle-cargo matching and dynamic relocation strategies, the vehicle operation process in the loading area is optimized; and finally the shortest departure path is planned for the vehicle to ensure the smooth progress of logistics operations. This comprehensive scheduling method not only improves the operational efficiency of logistics sites, but also reduces operating costs, which is of great significance for promoting the intelligent and efficient development of the logistics industry. Summary of the invention

[0006] The purpose of the present invention is to provide a vehicle scheduling method within a logistics site that takes road restrictions into consideration, to solve the problem of insufficient operational efficiency and management capabilities of the logistics site that affects the response speed and cost control of the entire supply chain, to improve overall logistics transportation efficiency, to reduce operating costs, and to promote the efficient development of the logistics industry.

[0007] In order to achieve the above object, the present invention provides a method for dispatching vehicles in a logistics site taking into account road restrictions, which is used to reduce transportation costs and improve operational efficiency in the logistics site, and mainly includes the following steps:

[0008] Step 1: Abstract the road network in the logistics site into a topological road network based on the space-time network;

[0009] Step 2: Construct a vehicle path planning model within the logistics site based on the spatiotemporal network to plan the shortest path for incoming vehicles to unload goods;

[0010] Step 3: Solve the vehicle path planning model to obtain the optimal solution so that the vehicle stays in the unloading area for the shortest time;

[0011] Step 4: Match idle vehicles with goods to be transported;

[0012] Step 5: A request is sent from the loading area in the logistics yard, and after receiving the request, the idle vehicle goes to the loading area to load the cargo;

[0013] Step 6: Propose an active real-time scheduling method based on dynamic relocation strategy for the working process of vehicles in the loading area;

[0014] Step 7: After completing all loading and unloading work, plan the shortest departure route for the vehicle to leave the logistics site.

[0015] As a further limitation of the present invention, in step 1, the road network in the logistics site is abstracted into a topological network G = (S, L), where L (l ∈ L) is the set of edges in the topological network, and S (s ∈ S) is the set of nodes in the topological network. Nodes include intersection nodes and meeting points, S * Gather at the meeting point.

[0016] The attributes of a point in a space-time network include the corresponding node in the topological network and the corresponding discrete time point, so the set of points is in, is the node in the topological network corresponding to point i, is the discrete time point corresponding to point i, and T represents the set of discrete time points within the study period.

[0017] The arcs in the space-time network describe the activities of vehicles, which are divided into waiting arcs and driving arcs. The waiting arc represents the waiting activity of the vehicle to avoid the meeting point, which only occurs at the meeting point. The starting and ending points of the waiting arc correspond to the same nodes in the topological network, and the corresponding discrete time points differ by 1, that is, Where A represents the set of arcs in the space-time network, (i,j)∈A, A w represents the set of waiting arcs in the space-time network, is the node in the topological network corresponding to point j, is the discrete time point corresponding to point j.

[0018] A driving arc represents the driving activity of a vehicle on an edge in a topological network, that is, driving from the starting node of the edge to its end node. The driving arc is constructed based on each edge in the topological network. The starting and ending points of the driving arc correspond to the starting and ending nodes of an edge in the topological network, and the corresponding discrete time point difference is the vehicle driving time of the edge, that is, The set of travel arcs corresponding to the edges in the topological network is Among them, A τ is the set of travel arcs in the space-time network, s(l) and e(l) are the start / end nodes of edge l in the topological network, s(l), e(l)∈S, represents the running time of the vehicle on edge l in the topological network.

[0019] As a further limitation of the present invention, in step 2, based on the constructed spatiotemporal network, a binary variable is set When its value is 1, it means that vehicle k passes through the arc (i, j) in the space-time network, otherwise it is 0. The goal of the model is to minimize the total weighted time of all vehicles in the unloading area of ​​the logistics site, including waiting time and driving time. The weight is measured by the priority of the vehicle to ensure that vehicles with low priority avoid waiting when meeting each other. The objective function is Among them, q k Indicates the priority of vehicle k. The larger the value, the higher the priority.

[0020] The constraints of the model are as follows: Formula (1) indicates that the vehicle enters the site at a given starting point and time; Formula (3) indicates that the vehicle eventually arrives at a given destination; Formulas (2) and (4) prohibit the vehicle from returning to the starting point and exiting the destination, thereby eliminating sub-loops on the vehicle path; Formula (5) indicates the flow balance constraint of the vehicle at other points.

[0021]

[0022] in, represents the entry time of vehicle k, o k ,d k represents the starting point and end point of vehicle k, o k ,d k ∈S.

[0023] The vehicle collision avoidance constraint is shown in formula (6). For a conflicting edge pair (l,l') in a topological network, formula (6) indicates that when vehicle k' passes through a driving arc (i',j') corresponding to edge l, the discrete time point Before and after Δt follow Within time (i.e., interval ) No other vehicles enter road section l.

[0024]

[0025] Among them, Δt pass It represents the time required for two transport vehicles traveling in opposite directions to complete the meeting at the meeting point, and M represents a sufficiently large positive number.

[0026] Considering the road restrictions in the logistics site, additional road width and turning radius constraints are added. Formula (7) represents the road width constraint. In practical applications, since vehicles usually do not occupy the entire road width when driving, it means that the maximum lateral offset of the vehicle (relative to the center line of the road) will not exceed the actual road width. Formula (8) represents the turning radius constraint, which ensures that the vehicle will not scratch or collide when turning due to a small turning radius.

[0027] num×Offset_max≤W(7)

[0028]

[0029] Where num is the number of lanes in the site, Offset_max represents the maximum lateral offset of the vehicle relative to the center line of the road during driving, and W is the road width. min Indicates the minimum turning radius of the vehicle.

[0030] It represents the minimum turning radius required by the road, WB is the wheelbase of the vehicle, ξ is the maximum steering angle of the vehicle, and M is an additional safety margin distance added to ensure safety.

[0031] As a further limitation of the present invention, when multiple vehicles gather in a certain area of ​​the road network in step 3, congestion may occur. At this time, if the subsequent impact is not taken into account when adjusting the vehicle's driving path, the waiting of the first node of the conflicting section or the readjustment of the path may cause a secondary conflict of the vehicle trajectory. In order to solve the above-mentioned secondary vehicle conflict and the problem of multiple vehicles meeting each other on the same road section, a conflict resolution method based on a space-time network is proposed. Assume that it is necessary to plan the path for a transport vehicle C to reach D at time t after n time scales. If there is no C in the space-time network, t To D t+n The path (arcs and points in the path are occupied by other transport vehicles), at this time C needs to reach D t+n Wait at the meeting point t Until the congestion is eliminated, the network node at the time when the transport vehicle C arrives is updated to

[0032] The main process of the algorithm is as follows:

[0033] Step 3.1, the first vehicle enters the road network, numbered i = 1, and the Dijkstra algorithm is used to calculate the shortest path for vehicle 1, and the shortest segment set of vehicle 1 is recorded {[S p ,S q ]}, the time period of the road section in From the starting point S o To S p distance, v is the speed of the vehicle.

[0034] Step 3.2: Whenever a new vehicle enters the road network, record i=i+1. If i>I, proceed to step 3.5. Otherwise, initialize the K short-circuit set P. i , candidate set X i , determine the starting point s i , end point t i , relaxation coefficient δ (the ratio of the length of the K shortest path to the length of the shortest path), using the Dijkstra algorithm to calculate the shortest path for vehicle i Put into set P i In the calculation Length make Let k = 1, record Road section collection and road segment time period collection in

[0035] Step 3.3: If k>K, go to step 3.2, otherwise From the point closest to t to point s i , point r swings to all possible connected points m, rm satisfies the point not in the candidate set X i Use Dijkstra's method to search for points m to t i The shortest path is denoted by p m , put p ksr , rm and p m The deviating paths composed of i In which p ksr express From s i to the subpath of r.

[0036] Step 3.4: If the candidate set X i If it is empty, go to step 3.2; if it is not empty, calculate all the paths f(p) in the candidate set and arrange them in ascending order. The minimum path is recorded as Only keep the first K paths, if Go to step 3.5, otherwise Move into P i, k=k+1, go to step 3.3.

[0037] Step 3.5: Compare the queues of the sections that all vehicles in the field pass through and the queues of the occupied time of the sections they pass through. If there are sections where two vehicles travel in opposite directions and have conflicting occupied time, extract the section where the conflict occurs first [S i ,S j ], and record the occupancy time of the two vehicles when they pass the above road section It is determined that the two vehicles will have a collision while driving in the field, and then go to step 3.6 to compare the priority P of the two vehicles. m , P n ; If there is no road section with overlapping occupancy time [S i ,S j ], the vehicle will travel according to the initial shortest path spatiotemporal information.

[0038] Step 3.6: Compare the priority attribute values ​​of the vehicles on the field, set the vehicle with high priority as m1 and the vehicle with low priority as m2. Guide the two vehicles to meet according to the vehicle meeting avoidance criterion. First, traverse the K short-circuit set of vehicle m2 to determine whether there is a secondary short-circuit that is not occupied by a vehicle with higher priority, and calculate the total travel time t on the path. k If it exists, let vehicle m2 travel along this short path and update the road section passed by vehicle m2 {[S p ,S q ]} and time period collection If it does not exist, then traverse the intersections between the low priority vehicles from the entrance to the conflicting road section. If there is a meeting point, let the vehicle wait at the meeting point and update the road section where vehicle m2 passes {[S p ,S q ]} and time period collection If there is no meeting point, vehicle m2 is made to wait at the vehicle entrance for a waiting time of And recalculate the total travel time t of the shortest route traveled by the vehicle at this time, where Update the avoidance vehicle path information and return to step 3.5.

[0039] Step 3.7, compare the queues of the sections that all vehicles in the field have passed through and the occupied time of the sections that they have passed through, and terminate the overlapping part of the queues.

[0040] As a further limitation of the present invention, step 4 adopts a clustering-based vehicle-cargo matching method. Assume that there are n vehicles to be matched in the logistics site, and record the vehicle set K = {k1, k2, ..., k n}, the maximum load of each truck is There are c goods to be matched, and the goods set G = {G1, G2, ... Gb ,...G c}, let the distance between the delivery destinations of goods b and c be d bc , record vehicle k n A matching relationship between and goods b is

[0041] By all matching relationships The matrix of n rows and c columns is a matching scheme for the vehicle-cargo matching problem, that is, a solution to the problem, denoted as The nth row vector of Y represents vehicle k n The c-th column vector represents the matching solution for item c.

[0042] In the process of vehicle-cargo matching, in order to make the goods in the matching scheme as concentrated as possible, the transportation points of the goods can be clustered, and DBSCAN clustering can be adopted. After clustering, the original goods set G can be expressed as g = {g1, g2, ..., g c}. Among them, g c Represents a cluster after clustering. According to the DBSCDN clustering rules, each noise point output is regarded as a cluster. Then there may be multiple transportation points or only one transportation point in the cluster, and the center coordinates of the cluster containing multiple cargo points are calculated. For a vehicle, its loading rate Here, full load is defined as 1, and the calculation formula is: in, Indicates whether cargo b is transported by vehicle k n For delivery, k n ∈K,b∈G;W b represents the weight of cargo b, Indicates the maximum load; V b represents the volume of cargo b, Indicates the maximum volume.

[0043] The loading rate (SZR) of a matching scheme Y is the ratio of the sum of the loading rates of all trucks to the sum of the matching relationships. The calculation formula is:

[0044] For a certain transport vehicle, because the cargo it transports is highly concentrated, the impact of its loading and unloading time on the vehicle's transportation time can be ignored during the transportation process, and the delivery time problem can be simplified to only consider the vehicle's driving time. Then, the driving time tk of a certain vehicle is n The formula is The driving time of a certain plan is Stk n , whose formula is in, Indicates the vehicle's travel distance; Indicates the vehicle's speed.

[0045] Taking the minimum number of vehicles dispatched from the site, the shortest vehicle driving time and the maximum vehicle loading rate as the goal, assuming that there are u options in total, the optimization goal is to minimize the number of vehicles Z1, the shortest delivery time Z2 and the maximum loading rate Z3. The objective function is as follows:

[0046] Z1=min{Y1,Y2,...,Y u}(9)

[0047] Z2=max{SZR1,SZR2,...,SZR u}(10)

[0048] Z3=min{Stk n1 ,Stk n2 ,...,Stk nu}(11)

[0049] The decision variables are set to k n ∈K,b∈G.

[0050] Assume that we need to use total trips to complete the delivery, y is the trip number, y∈total, y∈total,b∈G.

[0051] In order to meet the above requirements and the actual situation of vehicle-cargo matching in the distribution center, the constraints of the model are mainly based on the following factors:

[0052] Constraint 1: When a vehicle is loaded according to clusters, the weight of the cargo in the cluster cannot exceed the rated load of the vehicle, expressed as

[0053] Constraint 2: When vehicles are loaded according to clusters, the volume of cargo in the cluster cannot exceed the rated capacity of the vehicle, expressed as

[0054] Constraint 3: After the vehicle is loaded, the weight of the loaded cargo cannot exceed the rated load of the vehicle, expressed as

[0055] Constraint 4: After the vehicle is loaded, the volume of the loaded cargo cannot exceed the rated capacity of the vehicle, expressed as

[0056] Constraint 5: All goods in the site must be loaded, expressed as

[0057] Constraint 6: The vehicles included in the vehicle-cargo matching must be within the dispatchable range of the logistics site, expressed as Among them, D represents the dispatchable range of logistics sites; Indicates the distance between the vehicle and the center of the logistics site.

[0058] Constraint 7: Each cargo can only be loaded into one vehicle, expressed as

[0059] Next, we measure the satisfaction of both parties with the vehicle-cargo matching results. Assume that CO γ Indicates the γth cargo owner, VO in the vehicle owner set λ represents the λth car owner, S sn Indicates the snth index of the consignor, E sn Indicates the owner's snth indicator; For cargo owner CO γ VO to the owner λ About the indicator S sn satisfaction, VO for the owner λ VO to cargo owner λ About Indicator E sn satisfaction.

[0060] The first is the satisfaction of the cargo owners.

[0061] (1) For the demand date S 1 Satisfaction

[0062] Shipowner CO γ Give the required date S 1 The interval Owner VO λ Give the working date E 1 The specific value of The calculation formula for cargo owner satisfaction under this indicator is as follows:

[0063]

[0064] Among them, 0<σ γ <1, M represents a sufficiently large positive number.

[0065] (2) For the transport quotation S 2 Satisfaction

[0066] Shipowner CO γ Give a shipping quote 2 The specific value of Owner VO λ Given its delivery pricing E 2 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows:

[0067]

[0068] in, 0<θ γ ≤1.

[0069] (3) For Model S 3 Satisfaction

[0070] Shipowner CO γ Given the required model S 3 The specific value of Owner VO λ Given its model E 3 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows:

[0071]

[0072] in,

[0073] Next is the owner's satisfaction.

[0074] (1) For delivery pricing E 2 Satisfaction

[0075] Owner VO λ Given its delivery pricing E 2 The specific value of Shipowner CO γ Give a shipping quote 2 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows:

[0076]

[0077] in, 0<θ λ ≤1.

[0078] (2) For the place where delivery is possible 4 Satisfaction

[0079] Owner VO λ Give its deliverable place E 4 Preference order Shipowner CO γ Give the place of delivery S 4 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows:

[0080]

[0081] Among them, rankγλ Indicates the cargo owner CO γ The actual value of the indicator given In the car owner VO λ The ranking in the preference order, 0<τ λ ≤1,0<ξ λ <1.

[0082] The coefficient of variation method is used to determine the indicator weights, and the calculation process is as follows:

[0083] First calculate in, represents the coefficient of variation of the κth indicator; μ κ represents the standard deviation of the κth indicator; x κ Represents the average of the κth indicator.

[0084] Assume that the weight of each indicator is ω κ , and its calculation formula is

[0085] Based on the above weight determination method, the weight values ​​of each indicator involved in the satisfaction measurement are determined. The attribute indicator does not participate in the satisfaction aggregation, and the weight is set to 0. Thus, the overall satisfaction of the cargo owner and the car owner for the potential matching object is obtained:

[0086] Among them, α γλ Indicates the overall satisfaction of cargo owners with vehicle owners; β γλ Indicates the overall satisfaction of the vehicle owner with the cargo owner; Indicates the owner's index S sn The weight of Indicates the owner's sn The weight of

[0087]

[0088] When the overall satisfaction of either party is too low, the vehicle-cargo matching plan needs to be replanned.

[0089] As a further limitation of the present invention, the dynamic relocation strategy in step 6 is intended to actively guide idle vehicles into areas where future service requests may be concentrated, thereby reducing the response time to the requests. The dynamic relocation strategy is only executed for idle vehicles.

[0090] The dynamic relocation strategy aims to take proactive actions based on the information of future requests, focusing on the arrival time of virtual requests in typical transportation plans. The dynamic relocation strategy consists of three parts: trigger mechanism, implementation mechanism and remediation mechanism.

[0091] The trigger object of the dynamic relocation strategy is the idle vehicle located in the parking area p∈P at time t The trigger time for these vehicles is any time t that they are within the parking area.

[0092] For parking areas (a represents the current partition of vehicle k) The vehicle must meet the following conditions to trigger the implementation mechanism:

[0093] In the parking area There should be at least one vehicle other than vehicle k in k and k' park at the same time. In addition, there is no dispatch service request for vehicle k', and vehicle k' is not in the set of vehicles for which the implementation mechanism of the relocation strategy can be executed. In, or is the set of vehicles traveling at time t, i.e., there is at least one vehicle Its destination is the parking area The above conditions ensure that there is at least one vehicle parked in the current parking zone within time t, so as to quickly respond to newly arrived requests in the partition at any time.

[0094] RT sea is the search time range, in [t+T lead ,t+T lead +RT sea ], the number of virtual requests for vehicles arriving in the typical transportation plan partition should be less than the number of vehicles currently in the parking area This condition ensures that after the relocation of vehicle k, the remaining vehicles in the partition can quickly respond to new requests that may arrive densely in a short period of time.

[0095] RT int is the minimum time interval between two consecutive relocations. There should be a minimum time interval RT between the moment and the current time t int , to prevent the vehicle from being repositioned continuously within a short period of time.

[0096] The destination of the vehicle after relocation should be the area with the highest request arrival rate in the specified future period. The implementation time of vehicle migration should be earlier than the specified latest migration time τ. The specific steps are as follows:

[0097] Step 6.1: If If the current time t≤τ, go to step 6.2; otherwise, jump to step 6.7.

[0098] Step 6.2, RT sta represents the time interval between cargo arrivals. Model a typical cargo transportation plan and calculate the time interval between [t+T lead,t+T lead +RT sta ]The number of goods arriving in each partition a∈A during the period Q a If the highest arrival collected from a partition satisfies the minimum relocation criterion Q for the vehicle min ,Right now, Then remove the partitions in the set that do not meet the minimum relocation criteria and go to step 6.3; otherwise, jump to step 6.7.

[0099] Step 6.3: Update set A and set the arrival quantity Q a The highest partition is defined as a partition. If there are partitions with the same number of arrivals, the partitions are selected in the order of the partition numbers and the set Vehicles currently in the parking area of ​​this zone Removed.

[0100] Step 6.4: Update the collection if If it flashes, go to step 6.5; otherwise, go to step 6.7.

[0101] Step 6.5: Determine the current location and move to partition a * Whether the number of vehicles exceeds the capacity limit, if it does not exceed the capacity limit, go to step 6.6; otherwise, partition a * Remove from set A and return to step 6.3.

[0102] Step 6.6: In the collection The vehicle closest to the parking area in the partition is found and the vehicle will be relocated. The destination of the vehicle relocation is the parking area in the above partition. At the same time, the vehicle will be removed from the collection Remove and return to step 6.4.

[0103] Step 6.7, end the process.

[0104] For vehicles that cannot be repositioned Vehicles located in a parking area will continue to be parked in the current parking area until dynamic repositioning conditions are met or until they receive a service appointment request (e.g., cannot be repositioned due to limitations on the maximum number of vehicles that can be accommodated in a partition).

[0105] The beneficial effects of the present invention are: it can dynamically adjust vehicle path planning, optimize the vehicle's operating procedures in the unloading area and the loading area, and at the same time achieve efficient matching of vehicles and goods, minimize vehicle waiting time and empty driving distance, improve the overall operating efficiency of the logistics site, and reduce logistics transportation costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0106] Figure 1 It is a flow chart of the present invention.

[0107] Figure 2 This is the path planning based on the space-time network in the present invention.

[0108] Figure 3 Comparison of vehicle trajectories with and without the dynamic relocalization strategy. DETAILED DESCRIPTION

[0109] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0110] like Figure 1 As shown, the present invention provides a method for dispatching vehicles in a logistics site taking into account road restrictions, comprising the following steps:

[0111] Step 1: Abstract the road network in the logistics site into a topological road network based on the space-time network. The specific process is as follows:

[0112] The road network in the logistics site is abstracted into a topological network G = (S, L), where L (l ∈ L) is the set of edges in the topological network, and S (s ∈ S) is the set of nodes in the topological network. Nodes include intersection nodes and meeting points. * Gather at the meeting point.

[0113] The attributes of a point in a space-time network include the corresponding node in the topological network and the corresponding discrete time point, so the set of points is in, is the node in the topological network corresponding to point i, is the discrete time point corresponding to point i, and T represents the set of discrete time points within the study period.

[0114] The arcs in the space-time network describe the activities of vehicles, which are divided into waiting arcs and driving arcs. The waiting arc represents the waiting activity of the vehicle to avoid the meeting point, which only occurs at the meeting point. The starting and ending points of the waiting arc correspond to the same nodes in the topological network, and the corresponding discrete time points differ by 1, that is, Where A represents the set of arcs in the space-time network, (i,j)∈A, A w represents the set of waiting arcs in the space-time network, is the node in the topological network corresponding to point j, is the discrete time point corresponding to point j.

[0115] A driving arc represents the driving activity of a vehicle on an edge in a topological network, that is, driving from the starting node of the edge to its end node. The driving arc is constructed based on each edge in the topological network. The starting and ending points of the driving arc correspond to the starting and ending nodes of an edge in the topological network, and the corresponding discrete time point difference is the vehicle driving time of the edge, that is, The set of travel arcs corresponding to the edges in the topological network is Among them, A τ is the set of travel arcs in the space-time network, s(l) and e(l) are the start / end nodes of edge l in the topological network, s(l), e(l)∈S, represents the running time of the vehicle on edge l in the topological network.

[0116] Step 2: Construct a vehicle path planning model within the logistics site based on the spatiotemporal network to plan the shortest path for incoming vehicles to unload. The specific process is as follows:

[0117] Based on the constructed spatiotemporal network, set binary variables When its value is 1, it means that vehicle k passes through the arc (i, j) in the space-time network, otherwise it is 0. The goal of the model is to minimize the total weighted time of all vehicles in the unloading area of ​​the logistics site, including waiting time and driving time. The weight is measured by the priority of the vehicle to ensure that vehicles with low priority avoid waiting when meeting each other. The objective function is Among them, q k Indicates the priority of vehicle k. The larger the value, the higher the priority.

[0118] The constraints of the model are as follows: Formula (1) indicates that the vehicle enters the site at a given starting point and time; Formula (3) indicates that the vehicle eventually arrives at a given destination; Formulas (2) and (4) prohibit the vehicle from returning to the starting point and exiting the destination, thereby eliminating sub-loops on the vehicle path; Formula (5) indicates the flow balance constraint of the vehicle at other points.

[0119]

[0120] in, represents the entry time of vehicle k, o k ,d k represents the starting point and end point of vehicle k, o k ,d k ∈S.

[0121] The vehicle collision avoidance constraint is shown in formula (6). For a conflicting edge pair (l,l') in a topological network, formula (6) indicates that when vehicle k' passes through a driving arc (i',j') corresponding to edge l, the discrete time point Before and after Δt follow Within time (i.e., interval ) No other vehicles enter road section l.

[0122]

[0123] Among them, Δt pass It represents the time required for two transport vehicles traveling in opposite directions to complete the meeting at the meeting point, and M represents a sufficiently large positive number.

[0124] Considering the road restrictions in the logistics site, additional road width and turning radius constraints are added. Formula (7) represents the road width constraint. In practical applications, since vehicles usually do not occupy the entire road width when driving, it means that the maximum lateral offset of the vehicle (relative to the center line of the road) will not exceed the actual road width. Formula (8) represents the turning radius constraint, which ensures that the vehicle will not scratch or collide when turning due to a small turning radius.

[0125] num×Offset_max≤W(7)

[0126]

[0127] Where num is the number of lanes in the site, Offset_max represents the maximum lateral offset of the vehicle relative to the center line of the road during driving, and W is the road width. min Indicates the minimum turning radius of the vehicle. It represents the minimum turning radius required by the road, WB is the wheelbase of the vehicle, ξ is the maximum steering angle of the vehicle, and M is an additional safety margin distance added to ensure safety.

[0128] Step 3: Solve the vehicle path planning model to obtain the optimal solution so that the vehicle stays in the unloading area for the shortest time. The specific process is as follows:

[0129] When multiple vehicles gather in a certain area of ​​the road network, congestion may occur. At this time, if the subsequent impact is not taken into account when adjusting the vehicle's driving path, waiting or path readjustment before the conflicting node may lead to secondary conflict of vehicle trajectories. In order to solve the above-mentioned secondary conflict of vehicles and the problem of multiple vehicles meeting each other on the same road section, a conflict resolution method based on spatiotemporal network is proposed.

[0130] like Figure 2 As shown in the figure, it is assumed that it is necessary to plan the path for transport vehicle C to reach D at time t after n time scales. If C does not exist in the space-time network t To D t+n The path (arcs and points in the path are occupied by other transport vehicles), at this time C needs to reach D t+n Wait at the meeting point tUntil the congestion is eliminated, the network node at the time when the transport vehicle C arrives is updated to

[0131] The main process of the algorithm is as follows:

[0132] Step 3.1, the first vehicle enters the road network, numbered i = 1, and the Dijkstra algorithm is used to calculate the shortest path for vehicle 1, and the shortest segment set of vehicle 1 is recorded {[S p ,S q ]}, the time period of the road section in From the starting point S o To S p distance, v is the speed of the vehicle.

[0133] Step 3.2: Whenever a new vehicle enters the road network, record i=i+1. If i>I, proceed to step 3.5. Otherwise, initialize the K short-circuit set P. i , candidate set X i , determine the starting point s i , end point t i , relaxation coefficient δ (the ratio of the length of the K shortest path to the length of the shortest path), using the Dijkstra algorithm to calculate the shortest path for vehicle i Put into set P i In the calculation Length make Let k = 1, record Road segment collection {[S p ,S q ]} and road section time period collection in

[0134] Step 3.3: If k>K, go to step 3.2, otherwise From the point closest to t to point s i , point r swings to all possible connected points m, rm satisfies the point not in the candidate set X i Use Dijkstra's method to search for points m to t i The shortest path is denoted by p m , put p ksr , rm and p m The deviating paths composed of i In which p ksr express From s i to the subpath of r.

[0135] Step 3.4: If the candidate set Xi If it is empty, go to step 3.2; if it is not empty, calculate all the paths f(p) in the candidate set and arrange them in ascending order. The minimum path is recorded as Only keep the first K paths, if Go to step 3.5, otherwise Move into P i , k=k+1, go to step 3.3.

[0136] Step 3.5: Compare the queues of the sections that all vehicles in the field pass through and the queues of the occupied time of the sections they pass through. If there are sections where two vehicles travel in opposite directions and have conflicting occupied time, extract the section where the conflict occurs first [S i ,S j ], and record the occupancy time of the two vehicles when they pass the above road section It is determined that the two vehicles will have a collision while driving in the field, and then go to step 3.6 to compare the priority P of the two vehicles. m , P n ; If there is no road section with overlapping occupancy time [S i ,S j ], the vehicle will travel according to the initial shortest path spatiotemporal information.

[0137] Step 3.6: Compare the priority attribute values ​​of the vehicles on the field, set the vehicle with high priority as m1 and the vehicle with low priority as m2. Guide the two vehicles to meet according to the vehicle meeting avoidance criterion. First, traverse the K short-circuit set of vehicle m2 to determine whether there is a secondary short-circuit that is not occupied by a vehicle with higher priority, and calculate the total travel time t on the path. k If it exists, let vehicle m2 travel along this short path and update the road section passed by vehicle m2 {[S p ,S q ]} and time period collection If it does not exist, then traverse the intersections between the low priority vehicles from the entrance to the conflicting road section. If there is a meeting point, let the vehicle wait at the meeting point and update the road section where vehicle m2 passes {[S p ,S q ]} and time period collection If there is no meeting point, vehicle m2 is made to wait at the vehicle entrance for a waiting time of And recalculate the total travel time t of the shortest route traveled by the vehicle at this time, where Update the avoidance vehicle path information and return to step 3.5.

[0138] Step 3.7, compare the queues of the sections that all vehicles in the field have passed through and the occupied time of the sections that they have passed through, and terminate the overlapping part of the queues.

[0139] Step 4: Match the idle vehicles with the goods to be transported. The specific process is as follows:

[0140] Using the clustering-based vehicle-cargo matching method, assuming that there are n vehicles to be matched in the logistics site, the vehicle set K = {k1, k2, ..., k n}, the maximum load of each truck is There are c goods to be matched, and the goods set G = {G1, G2, ... G b ,...G c}, let the distance between the delivery destinations of goods b and c be d bc , record vehicle k n A matching relationship between and goods b is

[0141] By all matching relationships The matrix of n rows and c columns is a matching scheme for the vehicle-cargo matching problem, that is, a solution to the problem, denoted as The nth row vector of Y represents vehicle k n The c-th column vector represents the matching solution for item c.

[0142] In the process of vehicle-cargo matching, in order to make the goods in the matching scheme as concentrated as possible, the transportation points of the goods can be clustered, and DBSCAN clustering can be adopted. After clustering, the original goods set G can be expressed as g = {g1, g2, ..., g c}. Among them, g c Represents a cluster after clustering. According to the DBSCDN clustering rules, each noise point output is regarded as a cluster. Then there may be multiple transportation points or only one transportation point in the cluster, and the center coordinates of the cluster containing multiple cargo points are calculated.

[0143] For a vehicle, its loading rate Here, full load is defined as 1, and the calculation formula is: in, Indicates whether cargo b is transported by vehicle k n For delivery, k n ∈K,b∈G;W b represents the weight of cargo b, Indicates the maximum load; V b represents the volume of cargo b, Indicates the maximum volume.

[0144] The loading rate (SZR) of a matching scheme Y is the ratio of the sum of the loading rates of all trucks to the sum of the matching relationships. The calculation formula is:

[0145] For a certain transport vehicle, because the cargo it transports is highly concentrated, the impact of its loading and unloading time on the vehicle's transportation time can be ignored during the transportation process, and the delivery time problem can be simplified to only consider the vehicle's driving time. Then, the driving time of a certain vehicle tk n The formula is The driving time of a certain plan is Stk n , whose formula is in, Indicates the vehicle's travel distance; Indicates the vehicle's speed.

[0146] Taking the minimum number of vehicles dispatched from the site, the shortest vehicle driving time and the maximum vehicle loading rate as the goal, assuming that there are u options in total, the optimization goal is to minimize the number of vehicles Z1, the shortest delivery time Z2 and the maximum loading rate Z3. The objective function is as follows:

[0147] Z1=min{Y1,Y2,...,Y u}(9)

[0148] Z2=max{SZR1,SZR2,...,SZR u}(10)

[0149] Z3=min{Stk n1 ,Stk n2 ,...,Stk nu}(11)

[0150] The decision variables are set to k n ∈K,b∈G.

[0151] Assume that we need to use total trips to complete the delivery, y is the trip number, y∈total, y∈total,b∈G.

[0152] In order to meet the above requirements and the actual situation of vehicle-cargo matching in the distribution center, the constraints of the model are mainly based on the following factors:

[0153] Constraint 1: When vehicles are loaded according to clusters, the weight of the cargo in the cluster cannot exceed the rated load of the vehicle, expressed as

[0154] Constraint 2: When vehicles are loaded according to clusters, the volume of cargo in the cluster cannot exceed the rated capacity of the vehicle, expressed as

[0155] Constraint 3: After the vehicle is loaded, the weight of the loaded cargo cannot exceed the rated load of the vehicle, expressed as

[0156] Constraint 4: After the vehicle is loaded, the volume of the loaded cargo cannot exceed the rated capacity of the vehicle, expressed as

[0157] Constraint 5: All goods in the site must be loaded, expressed as

[0158] Constraint 6: The vehicles included in the vehicle-cargo matching must be within the dispatchable range of the logistics site, expressed as Among them, D represents the dispatchable range of logistics sites; Indicates the distance between the vehicle and the center of the logistics site.

[0159] Constraint 7: Each cargo can only be loaded into one vehicle, expressed as

[0160] Next, we measure the satisfaction of both parties with the vehicle-cargo matching results. Assume that CO γ Indicates the γth cargo owner, VO in the vehicle owner set λ represents the λth car owner, S sn Indicates the snth index of the consignor, E sn Indicates the owner's snth indicator; For cargo owner CO γ VO to the owner λ About the indicator S sn satisfaction, VO for the owner λ VO to cargo owner λ About Indicator E sn satisfaction.

[0161] The first is the satisfaction of the cargo owners.

[0162] (1) For the demand date S 1 Satisfaction

[0163] Shipowner CO γ Give the required date S 1 The interval Owner VO λ Give the working date E 1 The specific value of The calculation formula for cargo owner satisfaction under this indicator is as follows:

[0164]

[0165] Among them, 0<σ γ <1, M represents a sufficiently large positive number.

[0166] (2) For the transport quotation S2 Satisfaction

[0167] Shipowner CO γ Give a shipping quote 2 The specific value of Owner VO λ Given its delivery pricing E 2 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows:

[0168]

[0169] in,

[0170] (3) For Model S 3 Satisfaction

[0171] Shipowner CO γ Given the required model S 3 The specific value of Owner VO λ Given its model E 3 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows:

[0172]

[0173] in,

[0174] Next is the owner's satisfaction.

[0175] (1) For delivery pricing E 2 Satisfaction

[0176] Owner VO λ Given its delivery pricing E 2 The specific value of Shipowner CO γ Give a shipping quote 2 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows:

[0177]

[0178] in, 0<θ λ ≤1.

[0179] (2) For the place where delivery is possible 4 Satisfaction

[0180] Owner VO λ Give its deliverable place E 4 Preference order Shipowner CO γ Give the place of delivery S 4 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows:

[0181]

[0182] Among them, rank γλ Indicates the cargo owner CO γ The actual value of the indicator given In the car owner VO λ The ranking in the preference order, 0<τ λ ≤1,0<ξ λ <1.

[0183] The coefficient of variation method is used to determine the indicator weights, and the calculation process is as follows:

[0184] First calculate in, represents the coefficient of variation of the κth indicator; μ κ represents the standard deviation of the κth indicator; x κ Represents the average of the κth indicator.

[0185] Assume that the weight of each indicator is ω κ , and its calculation formula is

[0186] Based on the above weight determination method, the weight values ​​of each indicator involved in the satisfaction measurement are determined. The attribute indicator does not participate in the satisfaction aggregation, and the weight is set to 0. Thus, the overall satisfaction of the cargo owner and the car owner for the potential matching object is obtained:

[0187] Among them, α γλ Indicates the overall satisfaction of cargo owners with vehicle owners; β γλ Indicates the overall satisfaction of the vehicle owner with the cargo owner; Indicates the owner's index S sn The weight of Indicates the owner's sn The weight of

[0188]

[0189] When the overall satisfaction of either party is too low, the vehicle-cargo matching plan needs to be replanned.

[0190] Step 5: A request is sent from the loading area in the logistics yard, and after receiving the request, the idle vehicle goes to the loading area to load the cargo;

[0191] Step 6: Propose an active real-time scheduling method based on dynamic relocation strategy for the working process of vehicles in the loading area. The specific process is as follows:

[0192] like Figure 3 The figure shows a comparison of vehicle trajectories with and without the dynamic relocalization strategy.

[0193] The dynamic relocation strategy aims to proactively guide idle vehicles into areas where future service requests are likely to be concentrated, thereby reducing the response time to requests. The dynamic relocation strategy is only executed for idle vehicles.

[0194] The dynamic relocation strategy aims to take proactive actions based on the information of future requests, focusing on the arrival time of virtual requests in typical transportation plans. The dynamic relocation strategy consists of three parts: trigger mechanism, implementation mechanism and remediation mechanism.

[0195] The trigger object of the dynamic relocation strategy is the idle vehicle located in the parking area p∈P at time t The trigger time for these vehicles is any time t that they are within the parking area.

[0196] For parking areas (a represents the current partition of vehicle k) The vehicle must meet the following conditions to trigger the implementation mechanism:

[0197] In the parking area There should be at least one vehicle other than vehicle k in k and k' park at the same time. In addition, there is no dispatch service request for vehicle k', and vehicle k' is not in the set of vehicles for which the implementation mechanism of the relocation strategy can be executed. In, or is the set of vehicles traveling at time t, i.e., there is at least one vehicle Its destination is the parking area The above conditions ensure that there is at least one vehicle parked in the current parking zone within time t, so as to quickly respond to newly arrived requests in the partition at any time.

[0198] RT sea is the search time range, During this period, the number of virtual requests for vehicles arriving in the typical transportation plan partition should be less than the number of vehicles currently in the parking area. This condition ensures that after the relocation of vehicle k, the remaining vehicles in the partition can quickly respond to new requests that may arrive densely in a short period of time.

[0199] RT intis the minimum time interval between two consecutive relocations. There should be a minimum time interval RT between the moment and the current time t int , to prevent the vehicle from being repositioned continuously within a short period of time.

[0200] The destination of the vehicle after relocation should be the area with the highest request arrival rate in the specified future period. The implementation time of vehicle migration should be earlier than the specified latest migration time τ. The specific steps are as follows:

[0201] Step 6.1: If If the current time t≤τ, go to step 6.2; otherwise, jump to step 6.7.

[0202] Step 6.2, RT sta represents the time interval between cargo arrivals. Model a typical cargo transportation plan and calculate the time interval between [t+T lead ,t+T lead +RT sta ]The number of goods arriving in each partition a∈A during the period Q a If the highest arrival collected from a partition satisfies the minimum relocation criterion Q for the vehicle min ,Right now, Then remove the partitions in the set that do not meet the minimum relocation criteria and go to step 6.3; otherwise, jump to step 6.7.

[0203] Step 6.3: Update set A and set the arrival quantity Q a The highest partition is defined as a partition. If there are partitions with the same number of arrivals, the partitions are selected in the order of the partition numbers and the set Vehicles currently in the parking area of ​​this zone Removed.

[0204] Step 6.4: Update the collection if If it flashes, go to step 6.5; otherwise, go to step 6.7.

[0205] Step 6.5: Determine the current location and move to partition a * Whether the number of vehicles exceeds the capacity limit, if it does not exceed the capacity limit, go to step 6.6; otherwise, partition a * Remove from set A and return to step 6.3.

[0206] Step 6.6: In the collection The vehicle closest to the parking area in the partition is found and the vehicle will be relocated. The destination of the vehicle relocation is the parking area in the above partition. At the same time, the vehicle will be removed from the collection Remove and return to step 6.4.

[0207] Step 6.7, end the process.

[0208] For vehicles that cannot be repositioned Vehicles located in a parking area will continue to be parked in the current parking area until dynamic repositioning conditions are met or until they receive a service appointment request (e.g., cannot be repositioned due to limitations on the maximum number of vehicles that can be accommodated in a partition).

[0209] Step 7: After completing all loading and unloading work, plan the shortest departure route for the vehicle to leave the logistics site.

[0210] The above description describes the present invention in detail, and those skilled in the art should be able to understand that appropriate changes and modifications can be made without departing from the essential scope.

Claims

1. A method for dispatching vehicles in a logistics site taking into account road restrictions, characterized in that: The following steps are involved: Step 1: Abstract the road network in the logistics site into a topological road network based on the space-time network; Step 2: Construct a vehicle path planning model within the logistics site based on the spatiotemporal network to plan the shortest path for incoming vehicles to unload goods; Step 3: Solve the vehicle path planning model to obtain the optimal solution so that the vehicle stays in the unloading area for the shortest time; Step 4: Match idle vehicles with goods to be transported; Step 5: A request is sent from the loading area in the logistics yard, and after receiving the request, the idle vehicle goes to the loading area to load the cargo; Step 6: Propose an active real-time scheduling method based on dynamic relocation strategy for the working process of vehicles in the loading area; Step 7: After completing all loading and unloading work, plan the shortest departure route for the vehicle to leave the logistics site.

2. The method for dispatching vehicles in a logistics site considering road restrictions according to claim 1, characterized in that: Step 1 includes the following: The road network in the logistics site is abstracted into a topological network G = (S, L), where L (l ∈ L) is the set of edges in the topological network, S (s ∈ S) is the set of nodes in the topological network; nodes include intersection nodes and meeting points, S * Gather for the meeting point; The attributes of a point in a space-time network include the corresponding node in the topological network and the corresponding discrete time point, so the set of points is in, is the node in the topological network corresponding to point i, is the discrete time point corresponding to point i, and T represents the set of discrete time points in the research period; The arcs in the space-time network describe the activities of vehicles, which are divided into waiting arcs and driving arcs. The waiting arc represents the waiting activity of the vehicle to avoid the meeting point, which only occurs at the meeting point. The starting and ending points of the waiting arc correspond to the same nodes in the topological network, and the corresponding discrete time points differ by 1, that is, Where A represents the set of arcs in the space-time network, (i,j)∈A, A w represents the set of waiting arcs in the space-time network, is the node in the topological network corresponding to point j, is the discrete time point corresponding to point j; A driving arc represents the driving activity of a vehicle on an edge in a topological network, that is, driving from the starting node of the edge to its end node; the driving arc is constructed based on each edge in the topological network, and the starting and ending points of the driving arc correspond to the starting and ending nodes of an edge in the topological network, and the corresponding discrete time point difference is the vehicle driving time of the edge, that is, The set of travel arcs corresponding to the edges in the topological network is Among them, A τ is the set of travel arcs in the space-time network, s(l) and e(l) are the start / end nodes of edge l in the topological network, s(l), e(l)∈S, represents the running time of the vehicle on edge l in the topological network.

3. The method for dispatching vehicles in a logistics site considering road restrictions according to claim 2, characterized in that: Step 2 includes the following: Based on the constructed spatiotemporal network, set binary variables When its value is 1, it means that vehicle k passes through the arc (i, j) in the space-time network, otherwise it is 0; the goal of the model is to minimize the total weighted time of all vehicles in the unloading area of ​​the logistics site, including waiting time and driving time; the weight is measured by the priority of the vehicle to ensure that vehicles with low priority avoid waiting when meeting each other, then the objective function is Among them, q k Indicates the priority of vehicle k. The larger the value, the higher the priority. The constraints of the model are as follows: Formula (1) indicates that the vehicle enters the site at a given starting point and time; Formula (3) indicates that the vehicle eventually arrives at a given destination; Formulas (2) and (4) prohibit the vehicle from returning to the starting point and exiting the destination, thereby eliminating sub-loops on the vehicle path; Formula (5) indicates the flow balance constraint of the vehicle at other points; in, represents the entry time of vehicle k, o k ,d k represents the starting point and end point of vehicle k, o k ,d k ∈S; The vehicle meeting avoidance constraint is shown in formula (6); for a conflicting edge pair (l,l') in the topological network, formula (6) means that when vehicle k' passes through a driving arc (i',j') corresponding to edge l, at the discrete time point Before and after Δt follow Within time, that is, interval No other vehicles enter road section l; Among them, Δt pass It represents the time required for two transport vehicles traveling in opposite directions to complete the meeting at the meeting point, and M represents a sufficiently large positive number; Considering the road restrictions in the logistics site, additional road width and turning radius constraints are added; Formula (7) represents the road width constraint. In practical applications, since the vehicle usually does not occupy the entire road width when driving, it means that the maximum lateral offset of the vehicle will not exceed the actual road width; Formula (8) represents the turning radius constraint, which ensures that the vehicle will not scratch or collide when turning due to a small turning radius; num×Offset_max≤W (7) Where num is the number of lanes in the site, Offset_max represents the maximum lateral offset of the vehicle relative to the center line of the road during driving, and W is the road width; R min Indicates the minimum turning radius of the vehicle. It represents the minimum turning radius required by the road, WB is the wheelbase of the vehicle, ξ is the maximum steering angle of the vehicle, and M is an additional safety margin distance added to ensure safety.

4. The method for dispatching vehicles in a logistics site considering road restrictions according to claim 3, characterized in that: Step 3 includes the following: When multiple vehicles gather in a certain area of ​​the road network, congestion may occur. At this time, if the subsequent impact is not taken into account when adjusting the vehicle's driving path, the waiting of the first node of the conflicting section or the readjustment of the path may cause secondary conflict of vehicle trajectories. In order to solve the above-mentioned secondary conflict of vehicles and the problem of multiple vehicles meeting each other in the same section, a conflict resolution method based on space-time network is proposed. Assuming that it is necessary to plan the path for transport vehicle C to reach D at time t after n time scales, if C does not exist in the space-time network t To D t+n The path, that is, the arcs and points in the path are occupied by other transport vehicles. In this case, C needs to reach D t+n Wait at the meeting point t Until the congestion is eliminated, the network node at the time when the transport vehicle C arrives is updated to The algorithm flow is as follows: Step 3.1, the first vehicle enters the road network, numbered i = 1, and the Dijkstra algorithm is used to calculate the shortest path for vehicle 1, and the shortest segment set of vehicle 1 is recorded {[S p ,S q ]}, the time period of the road section in From the starting point S o To S p The distance, v is the speed of the vehicle; Step 3.2: Whenever a new vehicle enters the road network, record i=i+1. If i>I, proceed to step 3.

5. Otherwise, initialize the K short-circuit set P. i , candidate set X i , determine the starting point s i , end point t i , relaxation coefficient δ, which is the ratio of the length of the K shortest path to the length of the shortest path, using the Dijkstra algorithm to calculate the shortest path for vehicle i Put into set P i In the calculation Length make Let k = 1, record Road segment collection {[S p ,S q ]} and road section time period collection in Step 3.3: If k>K, go to step 3.2, otherwise From the point closest to t to point s i , point r swings to all possible connected points m, rm satisfies the point not in the candidate set X i conditions; Use Dijkstra's method to search for points m to t i The shortest path is denoted by p m , put p ksr , rm and p m The deviating paths composed of i In which p ksr express From s i subpath to r; Step 3.4: If the candidate set X i If it is empty, go to step 3.2; if it is not empty, calculate all the paths f(p) in the candidate set and arrange them in ascending order. The minimum path is recorded as Only keep the first K paths, if Go to step 3.5, otherwise Move into P i , go to step 3.3; Step 3.5: Compare the queues of the sections that all vehicles in the field pass through and the queues of the occupied time of the sections they pass through. If there are sections where two vehicles travel in opposite directions and have conflicting occupied time, extract the section where the conflict occurs first [S i ,S j ], and record the occupancy time of the two vehicles when they pass the above road section It is determined that the two vehicles will have a collision while driving in the field, and then go to step 3.6 to compare the priority P of the two vehicles. m , P n ; If there is no road section with overlapping occupancy time [S i ,S j ], the vehicle travels according to the initial shortest path time-space information; Step 3.6, compare the priority attribute values ​​of the vehicles on the field, set the vehicle with high priority as m1 and the vehicle with low priority as m2; guide the two vehicles to meet according to the vehicle meeting avoidance criterion, first traverse the K short-circuit set of vehicle m2, determine whether there is a secondary short-circuit that is not occupied by a vehicle with higher priority, and calculate the total travel time t on the path k ; If it exists, let vehicle m2 drive along this short path, and update the road section {[Sp, Sq]} and time period set passed by vehicle m2 If it does not exist, then traverse the intersections between the entrance and the conflicting road section of the low-priority vehicle. If there is a meeting point, let the vehicle wait at the meeting point and update the road section {[Sp, Sq]} and time period set passed by vehicle m2 If there is no meeting point, vehicle m2 is made to wait at the vehicle entrance for a waiting time of And recalculate the total travel time t of the shortest route traveled by the vehicle at this time, where Update the avoidance vehicle path information and return to step 3.5; Step 3.7, compare the queues of the sections that all vehicles in the field have passed through and the occupied time of the sections that they have passed through, and terminate the overlapping part of the queues.

5. The method for dispatching vehicles in a logistics site based on road restrictions according to claim 4, characterized in that: Step 4 includes the following: Using the clustering-based vehicle-cargo matching method, assuming that there are n vehicles to be matched in the logistics site, the vehicle set K = {k1, k2, ..., k n }, the maximum load of each truck is There are c goods to be matched, and the goods set G = {G1, G2, ... G b ,...G c }, let the distance between the delivery destinations of goods b and c be d bc , record vehicle k n A matching relationship between and goods b is By all matching relationships The matrix of n rows and c columns is a matching scheme for the vehicle-cargo matching problem, that is, a solution to the problem, denoted as The nth row vector of Y represents vehicle k n The c-th column vector represents the matching solution for item c; In the process of vehicle-cargo matching, the goods in the matching scheme should be concentrated as much as possible, and the transportation points of the goods should be clustered. DBSCAN clustering is adopted. After clustering, the original goods set G is expressed as g = {g1, g2, ..., gc}; where g c Represents a cluster after clustering. According to the DBSCDN clustering rules, each noise point output is regarded as a cluster. Then there may be multiple transportation points or only one transportation point in the cluster, and the center coordinates of the cluster containing multiple cargo points are calculated. For a vehicle, its loading rate Here, full load is defined as 1, and the calculation formula is: in, Indicates whether cargo b is transported by vehicle k n For delivery, k n ∈K,b∈G;W b represents the weight of cargo b, Indicates the maximum load; V b represents the volume of cargo b, Indicates the maximum volume; The loading rate (SZR) of a matching scheme Y is the ratio of the sum of the loading rates of all trucks to the sum of the matching relationships. The calculation formula is: For a certain transport vehicle, because the cargo it transports is highly concentrated, the impact of its loading and unloading time on the vehicle's transportation time can be ignored during the transportation process, and the delivery time problem is simplified to only consider the vehicle's driving time; then, the driving time of a certain vehicle tk n The formula is The driving time of a certain plan is Stk n , whose formula is in, Indicates the vehicle's travel distance; Indicates the vehicle's speed; Taking the minimum number of vehicles dispatched from the site, the shortest vehicle driving time and the maximum vehicle loading rate as the goal, assuming that there are u options in total, the optimization goal is to minimize the number of vehicles Z1, the shortest delivery time Z2 and the maximum loading rate Z3. The objective function is as follows: Z1=min{Y1,Y2,...,Y u } (9) Z2=max{SZR1,SZR2,...,SZR u } (10) Z3=min{Pcs n1 ,Piece n2 ,...,Piece nu } (11) The decision variables are set to Assume that we need to use total trips to complete the delivery, y is the trip number, y∈total, In order to meet the above requirements and the actual situation of vehicle-cargo matching in the distribution center, the constraints of the model are mainly based on the following factors: Constraint 1: When vehicles are loaded according to clusters, the weight of the cargo in the cluster cannot exceed the rated load of the vehicle, expressed as Constraint 2: When vehicles are loaded according to clusters, the volume of cargo in the cluster cannot exceed the rated capacity of the vehicle, expressed as Constraint 3: After the vehicle is loaded, the weight of the loaded cargo cannot exceed the rated load of the vehicle, expressed as Constraint 4: After the vehicle is loaded, the volume of the loaded cargo cannot exceed the rated capacity of the vehicle, expressed as Constraint 5: All goods in the site must be loaded, expressed as Constraint 6: The vehicles included in the vehicle-cargo matching must be within the dispatchable range of the logistics site, expressed as Among them, D represents the dispatchable range of logistics sites; Indicates the distance between the vehicle and the center of the logistics site; Constraint 7: Each cargo can only be loaded into one vehicle, expressed as Next, we measure the satisfaction of both parties with the vehicle-cargo matching results; assuming that CO γ Indicates the γth cargo owner, VO in the vehicle owner set λ represents the λth car owner, S sn Indicates the snth index of the consignor, E sn Indicates the owner's snth indicator; For cargo owner CO γ VO to the owner λ About the indicator S sn satisfaction, VO for the owner λ VO to cargo owner λ About Indicator E sn satisfaction; The first is the satisfaction of the cargo owner; (1) For the demand date S 1 Satisfaction Shipowner CO γ Give the required date S 1 The interval Owner VO λ Give the working date E 1 The specific value of The calculation formula for cargo owner satisfaction under this indicator is as follows: Among them, 0<σ γ <1, M represents a sufficiently large positive number; (2) For the transport quotation S 2 Satisfaction Shipowner CO γ Give a shipping quote 2 The specific value of Owner VO λ Given its delivery pricing E 2 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows: in, (3) For Model S 3 Satisfaction Shipowner CO γ Given the required model S 3 The specific value of Owner VO λ Given its model E 3 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows: in, Next is the owner’s satisfaction; (1) For delivery pricing E 2 Satisfaction Owner VO λ Given its delivery pricing E 2 The specific value of Shipowner CO γ Give a shipping quote 2 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows: in, (2) For the place where delivery is possible 4 Satisfaction Owner VO λ Give its deliverable place E 4 Preference order Shipowner CO γ Give the place of delivery S 4 The specific value of The calculation formula of cargo owners' satisfaction under this indicator is as follows: Among them, rank γλ Indicates the cargo owner CO γ The actual value of the indicator given In the car owner VO λ The ranking in the preference order, 0<τ λ ≤1,0<ξ λ <1; The coefficient of variation method is used to determine the indicator weights, and the calculation process is as follows: First calculate in, represents the coefficient of variation of the κth indicator; μ κ represents the standard deviation of the κth indicator; x κ represents the average of the κth index; Assume that the weight of each indicator is ω κ , and its calculation formula is Based on the above weight determination method, the weight values ​​of each indicator involved in the satisfaction measurement are determined. The attribute indicator does not participate in the satisfaction aggregation and its weight is set to 0. Thus, the overall satisfaction of the cargo owner and the vehicle owner for the potential matching object is obtained: Among them, α γλ Indicates the overall satisfaction of cargo owners with vehicle owners; β γλ Indicates the overall satisfaction of the vehicle owner with the cargo owner; Indicates the owner's index S sn The weight of Indicates the owner's sn The weight of When the overall satisfaction of either party is too low, the vehicle-cargo matching plan needs to be replanned.

6. The method for dispatching vehicles in a logistics site based on road restrictions according to claim 5, characterized in that: Step 6 includes the following: Dynamic relocation strategies aim to proactively guide idle vehicles into areas where future service requests are likely to be concentrated, thereby reducing response time to requests; The dynamic relocation strategy is only executed for idle vehicles; The dynamic relocation strategy aims to take proactive actions based on the information of future requests, focusing on the arrival time of virtual requests in typical transportation plans; the dynamic relocation strategy consists of three parts: trigger mechanism, implementation mechanism and remediation mechanism; The trigger object of the dynamic relocation strategy is the idle vehicle located in the parking area p∈P at time t The trigger time for these vehicles is any time t when they are located in the parking area; For parking areas Idle vehicles in The vehicle must meet the following conditions to trigger the implementation mechanism: In the parking area There should be at least one vehicle other than vehicle k in k and k' park at the same time; in addition, there is no dispatch service request for vehicle k', and vehicle k' is not in the set of vehicles for which the implementation mechanism of the relocation strategy can be executed In, or is the set of vehicles traveling at time t, i.e., there is at least one vehicle Its destination is the parking area The above conditions ensure that at least one vehicle is parked in the current parking area within time t, so as to quickly respond to the newly arrived requests in the partition at any time; RT sea is the search time range, in [t+T lead ,t+T lead +RT sea ], the number of virtual requests for vehicles arriving in the typical transport plan partition should be less than the number of vehicles currently in the parking area p k a The number of vehicles in the partition; this condition ensures that after the relocation of vehicle k, the remaining vehicles in the partition can quickly respond to new requests that may arrive densely in a short period of time; RT int is the minimum time interval between two consecutive relocalizations; when vehicle k reaches the parking area There should be a minimum time interval RT between the moment and the current time t int , to prevent the vehicle from being repositioned continuously within a short period of time; The destination of the vehicle after relocation should be the area with the highest request arrival rate in the specified future period; the implementation time of vehicle migration should be earlier than the specified latest migration time τ; the specific steps are as follows: Step 6.1: If If the current time t≤τ, go to step 6.2; otherwise, go to step 6.7; Step 6.2, RT sta represents the calculation of the time interval between cargo arrivals; a typical cargo transportation plan is modeled and the time interval between [t+T lead ,t+T lead +RT sta ]The number of goods arriving in each partition a∈A during the period Q a ; If the highest arrival collected from a partition satisfies the minimum relocation criterion Q of the vehicle min ,Right now, Then remove the partitions in the set that do not meet the minimum relocation criteria and go to step 6.3; otherwise, jump to step 6.7; Step 6.3: Update set A and set the arrival quantity Q a The highest partition is defined as a partition. If there are partitions with the same number of arrivals, the partitions are selected in the order of the partition numbers and the set Vehicles currently in the parking area of ​​this zone Remove; Step 6.4: Update the collection if If it flashes, go to step 6.5; otherwise, go to step 6.7; Step 6.5: Determine the current location and move to partition a * Whether the number of vehicles exceeds the capacity limit, if it does not exceed the capacity limit, go to step 6.6; otherwise, partition a * Remove from set A and return to step 6.3; Step 6.6: In the collection The vehicle closest to the parking area in the partition is found, and the vehicle will be relocated; the destination of the vehicle relocation is the parking area in the above partition, and at the same time, the vehicle will be removed from the collection Remove it and return to step 6.4; Step 6.7, end the process; For vehicles that cannot be repositioned Vehicles located in a parking area will continue to be parked in the current parking area until dynamic repositioning conditions are met or until they receive a service appointment request.

Images