CN111160690A

CN111160690A - Vehicle loading planning method and device and storage medium

Info

Publication number: CN111160690A
Application number: CN201911120115.3A
Authority: CN
Inventors: 张庆梅
Original assignee: Hangzhou Pinjie Network Technology Co Ltd
Current assignee: Hangzhou Pinjie Network Technology Co Ltd
Priority date: 2019-11-15
Filing date: 2019-11-15
Publication date: 2020-05-15
Anticipated expiration: 2039-11-15
Also published as: CN111160690B

Abstract

The invention discloses a vehicle loading planning method, which clusters all orders needing to be delivered in a system according to delivery addresses by combining a clustering algorithm, distributes loaded available vehicles for each order according to a clustering result, then respectively judges whether the orders loaded by the selected available vehicles meet system requirements, and simultaneously loads each available vehicle respectively based on the remaining orders until all the orders in the system are loaded into the corresponding available vehicles, so that vehicle loading planning in an ordering platform is realized, single logistics cost and overall rate can be greatly reduced, logistics cost is controlled in a reasonable range, and cost is saved. The invention also provides a vehicle loading planning device and a storage medium.

Description

Vehicle loading planning method and device and storage medium

Technical Field

The present invention relates to the field of logistics, and in particular, to a method for planning cargo loading in a logistics system, an electronic device, and a storage medium.

Background

Vehicle loading planning is an important part of the logistics system, namely, the problem of how to distribute delivery orders to corresponding logistics vehicles. The loading plan is to sort and load the goods onto the delivery vehicles according to the order demands of the customers and aiming at completing the delivery tasks with the mileage, the cost, the time and the vehicle number as few as possible, thereby realizing the delivery of the delivery orders.

However, current load planning, which involves a small amount of data, generally employs dynamic planning to obtain an optimal solution within a limited time. However, when a large amount of data is involved, such as uniform loading of vehicles for a city, the optimal solution is obtained by NP (non-deterministic Polynomial) problem, and the pure dynamic programming cannot adapt to the data of a city volume. The existing loading planning method cannot be combined with vehicle type analysis distribution of vehicles, meanwhile, the vehicles are considered to be infinitely available, and distribution areas are not limited. However, for the delivery orders nowadays, there are various types of regions, scenes and vehicle types, and the inventory is limited.

Disclosure of Invention

In order to overcome the defects of the prior art, an object of the present invention is to provide a vehicle loading planning method, which can solve the problem of high logistics cost caused by unreasonable vehicle loading planning in a distribution system in an order platform in the prior art.

The second objective of the present invention is to provide a vehicle loading planning device, which can solve the problem of high logistics cost caused by unreasonable vehicle loading planning in the distribution system of the order platform in the prior art.

The invention further aims to provide a storage medium, which can solve the problem of high logistics cost caused by unreasonable vehicle loading planning in a distribution system in an ordering platform in the prior art.

One of the purposes of the invention is realized by adopting the following technical scheme:

a vehicle loading planning method, the loading planning method comprising:

an acquisition step: acquiring all orders and available vehicles which need to be distributed in the system; each available vehicle comprises a model of the available vehicle; each vehicle type is provided with available quantity; each order includes a delivery address;

determining available vehicles: determining the number of available vehicles and vehicle types required by the delivery of all orders according to the minimum vehicle principle and available vehicles in the system;

an initial distribution step: clustering all orders to be delivered in the system according to a first clustering algorithm to obtain a first clustering result according to the delivery address of each order, and distributing a current available vehicle for each cluster in the first clustering result according to the vehicle type of the available vehicle required by the fact that all the orders are delivered and the available quantity of the corresponding vehicle type; the number of clusters in the first clustering result is the same as the number of available vehicles required by all the orders after the orders are distributed;

selecting: selecting one of the currently available vehicles and recording as a vehicle X to be completed; the order loaded by the vehicle X to be completed is marked as an order to be delivered;

a judging step: judging whether the vehicle X to be finished meets the system requirements, if so, executing a filling step; if not, executing a removing step;

filling: selecting a vehicle to be loaded from the vehicle to be loaded order set to load a vehicle X to be completed; when the loading of the vehicle X is finished, executing a driver matching step; the order to be loaded is an order of the vehicle X to be finished which is not loaded; the order set to be loaded is a set of orders to be loaded; suppose the collection of orders to be loaded Y: y is_j,j∈{1,n}，y_jOrdering for loading; set of orders to be delivered X loaded by vehicle to be fulfilled X_i,i∈{1,m}，x_iIs an order to be delivered; wherein m and n are both natural numbers larger than zero;

removing: sequentially removing orders to be distributed in the vehicles X to be completed, adding the orders to be distributed into the order set Y to be loaded simultaneously after removing one order to be distributed from the vehicles X to be completed each time, updating the orders to be distributed in the vehicles X, and executing the judging step;

a driver allocation step: acquiring available drivers of the system, distributing current available drivers for the vehicle X to be completed, removing the order to be distributed loaded by the vehicle X to be completed from the order to be distributed in the system, removing the vehicle X to be completed from the available vehicles of the system, removing the current available drivers from the available drivers of the system, and returning to the step of determining available vehicles; until the order to be delivered is empty in the system.

Further, the initial assigning step further comprises: the method comprises the steps of firstly calculating the total volume of goods of all orders of each cluster, sequencing the orders according to the total volume of the goods from large to small, and then sequentially distributing the orders to each cluster after sequencing according to the available quantity corresponding to the vehicle types of available vehicles required by all the orders after distribution.

Further, the first clustering algorithm is a kmean clustering algorithm, and when the kmean clustering algorithm clusters all orders required to be delivered in the system, the clustering number is equal to the number of available vehicles required to deliver all orders.

Further, the selecting step includes: firstly, calculating the distance between each cluster and all order center points in a first clustering result, and then selecting one cluster as an off-cluster according to the distance between each cluster and all order center points, wherein the current available vehicle distributed by the off-cluster is a vehicle X to be finished; the distance cluster is the cluster with the largest distance from all the cluster center points of all the orders; all orders are all orders needing to be delivered in the system;

wherein, the number of all orders needing to be delivered in the system is assumed to be A, and the number of the orders in the cluster is assumed to be k:

the location of the center point for all orders is expressed as:

the position of the cluster is represented as:

the distance between the cluster and the center point of all orders is as follows:

wherein the longitude and latitude of the order are derived from the shipping address of the order.

Further, the criterion that the vehicle X to be completed meets the system requirement is as follows: the total volume of the cargos of all orders to be dispensed loaded by the vehicle X to be finished does not exceed the available quantity of the vehicle X to be finished;

or the total volume of the cargos of all orders to be delivered loaded by the vehicle X to be completed does not exceed the available volume of the vehicle X to be completed, and the delivery distance of all orders to be delivered loaded by the vehicle X to be completed does not exceed the preset value of the system.

Further, the filling step specifically includes:

step S11, selecting one or more orders to be loaded from the order set Y to be loaded, and recording the orders as candidate orders respectively; the set of all candidate orders is marked as a candidate order set;

step S12, clustering all candidate orders according to the distribution address of each candidate order in the candidate order set by a second clustering algorithm to obtain a second clustering result;

and step S13, selecting an order filled in the vehicle X to be completed from the candidate order set according to the second classification result, and removing the order filled in the vehicle X to be completed from the order set Y to be loaded.

Further, the step S11 includes: firstly, each order Y to be loaded in the order set Y to be loaded is calculated respectively_jDistance from vehicle X to be completed, and then according to each order y to be loaded_jSelecting one or more to-be-loaded orders from the to-be-loaded order set Y according to the principle that the distance between the to-be-loaded orders and the to-be-loaded vehicle X is from small to large; wherein the content of the first and second substances,

each order Y to be loaded in order set Y to be loaded_jWith each order X to be delivered of the vehicle X to be fulfilled_iThe distance of (a) is:

then each order Y to be loaded in the order set Y to be loaded_jThe distance from the vehicle X to be completed is:

y_jx distance is min (y)_jAnd x₁Distance of order, y_jAnd x₂Distance of order, once_jAnd x_mOrder distance).

Further, the step S12 further includes: step S121, clustering all candidate orders in the candidate order set through multiple clustering algorithms;

s122, determining a second clustering algorithm according to the turnaround distance of the clustering result of each clustering algorithm; the turnaround distance is the sum of the distribution distances of all candidate orders in each cluster after distribution in the clustering result of each clustering algorithm;

and S123, clustering all candidate orders in the candidate order set according to a second clustering algorithm to obtain a second clustering result.

Further, the second clustering algorithm is any one of the following: a Kmeans-based clustering algorithm, an EM-based clustering algorithm, a FARTHEST-based clustering algorithm, and a HIERARCHICAL-based clustering algorithm.

Further, the step S13 includes:

s131, selecting a cluster with the minimum distance to the vehicle X to be finished as a current cluster according to a second clustering result; wherein, assume that the order set of the current cluster C is: c. C_i,i∈{1,n1}：

The distance between the vehicle X to be completed and the current cluster C is:

step S132, judging whether all candidate orders of the current cluster can be completely loaded into the vehicle X to be finished, if so, executing step S133; if not, go to step S134;

step S133, loading all candidate orders of the current cluster into the vehicle X to be finished, updating the order to be delivered loaded by the vehicle X to be finished, removing all candidate orders of the current cluster from the order set Y to be loaded, and then executing step S135;

step S134, selecting one or more candidate orders from the current cluster according to the current loading space of the vehicle X to be finished to load the candidate orders into the vehicle X to be finished; when the loading of the vehicle X is finished, executing a driver allocation step; meanwhile, removing the candidate orders loaded into the to-be-finished vehicle X from the to-be-loaded order set Y, and updating the to-be-dispensed orders loaded by the to-be-finished vehicle X; the current loading space of the vehicle X to be finished is obtained by subtracting the total volume of the goods of all the orders to be distributed currently loaded by the vehicle X to be finished from the available quantity of the vehicle X to be finished;

step S135: judging whether the vehicle X to be finished has a loading space, if so, executing the step S11; if not, the driver allocation step is executed after the vehicle X loading is finished.

Further, step S134 includes selecting one or more candidate orders from the current cluster to be loaded into the to-be-completed vehicle X according to the sequential loading method and the swap loading method until the to-be-completed vehicle loading X is completed when all candidate orders in the current cluster are completely traversed or the to-be-completed vehicle has no loading space.

Further, the sequential loading method comprises: firstly, respectively calculating the distance between each candidate order in the current cluster and a vehicle X to be finished, sequencing according to the distance between each candidate order in the current cluster and the vehicle X to be finished from large to small, sequentially traversing each candidate order in the current cluster according to a sequencing result, simultaneously judging whether each candidate order can be loaded into the vehicle X to be finished, if so, loading the corresponding candidate order into the vehicle X to be finished, removing the corresponding candidate order from the current cluster, updating the order to be delivered of the vehicle X to be finished, and continuously traversing the next candidate order; if not, stopping traversing.

Further, the exchange loading method comprises a first exchange loading method and a second exchange loading method;

when the sequential loading method traverses to the current candidate order and the current candidate order cannot be loaded into the vehicle X to be completed, stopping traversing, and at this time, continuing loading the vehicle X to be completed according to the first exchange loading method or the second exchange loading method:

the first exchange loading method comprises the following steps: sequentially removing the corresponding orders to be delivered from the vehicle X to be completed according to the sequence that the distance between the orders to be delivered in the vehicle X to be completed and the vehicle X to be completed is from small to large until the current candidate orders can be loaded into the vehicle X to be completed; at the moment, adding the to-be-dispensed order removed from the to-be-completed vehicle X into the to-be-loaded order, loading the current candidate order into the to-be-completed vehicle X, removing the current candidate order from the current cluster, and updating the to-be-dispensed order in the to-be-completed vehicle X;

the second exchange loading method comprises the following steps: and continuously traversing other candidate orders in the current cluster, finding a new candidate order capable of being loaded into the vehicle X to be completed, loading the new candidate order into the vehicle X to be completed, updating the order to be delivered in the vehicle X to be completed, and removing the new candidate order from the current cluster.

Further, the basis for selecting the first exchange loading method or the second exchange loading method is to select according to the turnover loss size of each exchange loading method;

wherein, the turnaround distance loss of the first exchange loading method is: the sum of the driving distance generated when each order to be delivered is removed by the vehicle X to be completed and the driving distance generated when the current candidate order is loaded by the vehicle X to be completed; the turnaround loss for the second swap loading method is: the sum of the travel distance generated when the vehicle X to be completed loads a new candidate order and the travel distance generated when the current candidate order is delivered.

Further, the removing step specifically includes: step S21, respectively calculating the distance between each order to be distributed loaded by the vehicle X to be completed and the order set Y to be loaded;

step S22, finding out the order to be delivered with the minimum distance between each order to be delivered loaded by the vehicle X to be completed and the order set Y to be loaded, removing the order to be delivered from the vehicle X to be completed, adding the removed order to be delivered into the order set Y to be loaded, updating the order to be delivered in the vehicle X to be completed, and then executing the determining step.

Further, the step S21 includes: order x to be delivered_iWith each order Y to be loaded in the order set Y to be loaded_jThe distance of (a) is:

then the order x is to be delivered_iThe distance from the order set Y to be loaded is as follows:

x_idistance Y ═ min (x)_iAnd y₁Distance of order, x_iAnd y₂Distance of order_iAnd y_nOrder distance).

The second purpose of the invention is realized by adopting the following technical scheme:

a vehicle load planning apparatus comprising a memory and a processor, the memory having stored thereon a load planning program operable on the processor, the load planning program being a computer program, the processor, when executing the load planning program, performing the steps of a vehicle load planning method according to one of the objects of the invention.

The third purpose of the invention is realized by adopting the following technical scheme:

a storage medium being a computer readable storage medium having stored thereon a load planning program being a computer program which, when executed by a processor, carries out the steps of a method of load planning for a vehicle as employed in one of the objects of the invention.

Compared with the prior art, the invention has the beneficial effects that:

the invention combines the clustering algorithm to divide all orders to be delivered by the system according to delivery addresses, initially distributes corresponding available vehicles according to the divided results, then sequentially judges whether each distributed available vehicle meets the system requirements, removes or continuously loads the orders loaded by the available vehicles according to the judgment results until the loading of the available vehicles and the order distribution of the system are finished, then distributes corresponding available drivers for each available vehicle, further realizes the vehicle loading planning of the orders to be delivered in the ordering platform, can greatly reduce the single-key logistics cost and the overall rate, controls the logistics cost in a reasonable range, and saves the cost.

Drawings

FIG. 1 is a flow chart of an order for an order platform according to the present invention;

FIG. 2 is a flow chart of a vehicle loading planning method provided by the present invention;

FIG. 3 is a flowchart of step S16 in FIG. 2;

FIG. 4 is a flowchart of step S163 of FIG. 3;

FIG. 5 is a diagram showing the results of the second polymerization;

FIG. 6 is a sequential loading initial view;

FIG. 7 is a sequential load traversal stop diagram;

FIG. 8 is a schematic diagram of loading results of the first swap loading method;

FIG. 9 is a schematic diagram of loading results of a second swap loading method;

FIG. 10 is a schematic diagram of the turn-around loss calculation for the exchange loading method;

FIG. 11 is a graph of average individual piece stream cost trend;

fig. 12 is a rate change trend graph.

Detailed Description

The present invention will be further described with reference to the accompanying drawings and the detailed description, and it should be noted that any combination of the embodiments or technical features described below can be used to form a new embodiment without conflict.

Example one

The invention provides a vehicle loading planning system applied to an order platform, which realizes the loading of delivery vehicles through orders generated by the order platform so as to complete the delivery of the orders. Generally, an order platform, which is a resource integration platform, is connected to a store at one end to provide an order platform for the store, is connected to a supplier at one end to provide a selling platform for the supplier at the other end, and is connected to a logistics system at the other end to perform distribution. The whole service process of the ordering platform comprises the following steps: the staff in the shop places orders through the ordering platform and forms orders to be delivered in the system, then corresponding logistics schemes and stock suppliers are generated for the orders to be delivered through intelligent logistics delivery, and finally the drivers deliver the commodities to the orders to be delivered according to the logistics schemes to complete the whole ordering process.

As shown in fig. 1, the order process in the order platform generally comprises:

and step A, generating an order after a client places the order, and then storing the order in an order pool of the system.

And B, regularly acquiring all order information in the order pool from the order pool, wherein the order information comprises order commodity demands and order placing shop positions. For example, all order information is taken from the order pool at 4 pm each day. For example, most order platforms today generally set a fixed time, for example, the current order before 4 pm will be scheduled for shipment, delivery, etc., and the current order after 4 pm will need to wait for the next day for scheduling for shipment, delivery, etc.

And C, dividing the orders into orders in different areas according to the distribution areas to which the orders belong. In general, when obtaining delivery, a driver, a vehicle, a route, and the like are divided according to a delivery area, and therefore, the driver, the vehicle, the route, and the like are first divided according to the delivery area to which an order belongs.

And D, acquiring online driver information and available vehicle information of different areas, loading vehicles according to the orders of the different areas and the online driver information and available vehicle information of the corresponding areas, and selecting drivers to complete the distribution of the orders.

And E, forming a corresponding detail list of the order commodities, the suppliers, the vehicles and the drivers after all the orders are distributed, and realizing the distribution of the orders according to the corresponding detail list.

The invention is applied to the order platform, and the problem solved by the provided loading planning system is that after a user places an order, the system divides all orders in the system according to different areas according to the areas, namely, the problem how to load vehicles for all orders in the corresponding areas according to the orders in the different areas, the online driver information and the available vehicle information of the corresponding areas and select the drivers to complete the distribution of the orders in the step D.

As shown in fig. 2, the present invention provides a preferred embodiment, a method for load planning, comprising:

step S11, various data required for load planning, such as orders to be delivered in the system, available vehicles in the system, are obtained.

Each order comprises an order ID, commodity information and required quantity of the order and an order placing shop address. The position information in the invention is represented by two-dimensional coordinates, namely, by longitude and latitude. That is, the information of the ordering store address is represented by latitude and longitude, the ordering store address is the delivery address of each order, and the delivery distance is calculated according to the delivery address.

Each available vehicle includes a vehicle ID, a vehicle type, and an available amount. Different models of vehicles have different loading spaces, such as small-sized minivans and large-sized minivans, and the total volume of goods which can be loaded by the vehicles is different due to different models of vehicles. Therefore, in order to meet the actual situation, the model and the available amount of the vehicle need to be acquired during the loading planning. In addition, all the vehicles involved in the invention are normal vehicles which can be used for loading.

Step S12, determining the number of available vehicles and vehicle types required by the delivery of all orders according to the least vehicle principle and the available vehicles in the system; and then clustering all orders to be delivered in the system according to a first clustering algorithm according to delivery addresses to obtain a first clustering result, and distributing a current available vehicle to each cluster in the first clustering result according to the determined vehicle types and available quantities of the available vehicles required by the delivery of all the orders. Of course, in the actual use process, clustering may be performed according to the delivery distance, wherein the delivery distance is calculated according to the delivery address.

The invention adopts a kmeans clustering algorithm to cluster all orders according to distribution distance, and the key of the kmeans clustering algorithm is to firstly determine the clustering number.

The invention determines the clustering number of the kmeans clustering algorithm by adopting the minimum vehicle principle. That is, the cluster number is the number of available vehicles required to complete delivery of all orders to be delivered. For example, the determined currently available vehicles include IDs, vehicle types, available amounts, and the like of the available vehicles.

The method specifically comprises the following steps: the minimum vehicle principle is the minimum number of vehicles that are required to find a vehicle that can load all the orders in the system. Assuming that 5 small bread cars, or 3 medium bread cars, or 1 large bread car and 1 medium bread car are required to finish loading after a batch of orders is delivered, the minimum vehicle required when the batch of orders is loaded is determined to be 2 according to the minimum vehicle principle, namely 1 large bread car and 1 small bread car are adopted. Therefore, the number of clusters of the kmean algorithm is 2. Since the number of clusters is 2, the first clustering result obtained by clustering all orders through the kmeans clustering algorithm also includes 2 clusters.

That is, when all orders are clustered according to the distribution distance by using the kmean clustering algorithm, 2 clusters are obtained, and each cluster includes one or more orders, such as cluster 1 and cluster 2.

In connection with the above example, for the delivery of the batch of orders, a large flat car and a small flat car are required, each car will be loaded with one or more orders, and the delivery distance of the orders in each cluster is substantially consistent. Because the available quantity of the vehicles is different due to different vehicle types, namely the loaded goods have different volumes, in order to ensure that the vehicles with large available quantity can load more goods, the invention also needs to determine the available vehicles corresponding to each cluster.

In order to determine the available vehicles required by each cluster, the invention also determines the vehicle corresponding to each cluster obtained after clustering according to the available vehicles in the system.

When determining the cluster number by the minimum vehicle principle, the vehicle type and the available quantity of the selected currently available vehicles are determined. Therefore, the total volume of the goods of all the orders of each clustered cluster is calculated, then the order is carried out according to the total volume of the goods, then the order is carried out according to the determined available quantity of the vehicle types of the available vehicles, and finally the ordered available vehicles are matched with each cluster after the order is carried out one by one. At this point, each cluster is assigned an available vehicle and is recorded as the current available vehicle. The selected available vehicle is referred to herein as the currently available vehicle to facilitate differentiation from the available vehicles in the system.

For example, as described in the above example, when the total volume of the goods of all the orders in the cluster 1 is greater than the total volume of the goods of all the orders in the cluster 2, the available vehicle corresponding to the cluster 1 is a van, and the available vehicle corresponding to the cluster 2 is a minivan.

After the available vehicles determine the loaded orders to be delivered, in order to ensure that the delivery distance of the available vehicles meets the system constraint, or ensure the maximum loading capacity of the available vehicles, and further ensure the controllability of the logistics cost, further judgment needs to be performed on all the orders loaded by each current available vehicle, and whether each current available vehicle can completely load all the orders of the corresponding cluster, or whether other conditions of the system are met, and the like.

And step S13, determining the vehicle X to be completed, and marking the order loaded by the vehicle X to be completed as the order to be delivered. As can be seen from the above description, each cluster corresponds to one currently available vehicle, and since each cluster may include a plurality of orders, the total volume of the goods may exceed the available quantity of the allocated currently available vehicles, and thus if the total volume of the goods exceeds the available quantity of the allocated currently available vehicles, the currently available vehicles cannot fully load all the orders of the corresponding cluster, and therefore, it is necessary to determine whether the currently available vehicles of each cluster meet the system requirements.

When determining the vehicle X to be finished, the invention searches the vehicle X to be finished from the currently available vehicles through the concept of cluster leaving. The outlier cluster is the one of the clusters with the largest distance from all the order center points. That is, when a certain cluster is farthest away from all the order center points, the available vehicle corresponding to the cluster is also the vehicle X to be completed determined in this embodiment.

The calculation mode of the cluster is as follows: assuming that there are all a orders, the location of the center point of all orders (in the present invention, the location is expressed as latitude and longitude) is:

assuming that there are k orders in a cluster, the location of the cluster is specifically:

the distance between the cluster and the center point of all orders is:

and calculating the distance between each cluster and the center point of all orders according to the formulas (1) to (5). For example, the distance between cluster 1 and the center point of all orders is calculated as S1, and the distance between cluster 2 and the center point of all orders is calculated as S2, and then the distances between cluster S1 and cluster S2 are compared to determine the distance cluster. The available vehicle corresponding to the cluster is a vehicle to be completed and is marked as X. At this time, the order in the cluster corresponding to the vehicle X to be completed, that is, the order loaded by the vehicle X to be completed, is recorded as the order to be delivered.

Step S14, judging whether the vehicle X to be finished meets the system requirement, if so, executing step S16; if not, S15 is executed.

Wherein, the judgment basis is as follows:

A. the total volume of the goods loaded by the vehicle X to be finished and assigned orders does not exceed the loading space of the vehicle X to be finished. The loading space is also the available amount corresponding to the model of the vehicle X to be completed.

Generally, vehicles have corresponding carrying capacity, and when the total volume of the goods loaded by the vehicle-mounted order to be delivered exceeds the loading space of the vehicle, it indicates that the current vehicle cannot load all the goods of the order to be delivered.

B. The delivery distance satisfies the system constraint condition when the vehicle X to be completed delivers all the loaded orders to be delivered.

The delivery distance is the shortest mileage at which the delivery of the order to be delivered loaded on the vehicle to be completed X is completed. Such as the shortest mileage of all store vehicles involved in the delivery of the order to be delivered loaded on the vehicle to be delivered X. In order to guarantee the logistics cost, the system can stipulate the shortest mileage that a vehicle runs when distributing goods, and when the mileage is exceeded, the current order to be distributed by the vehicle is considered to be unreasonable and needs to be adjusted.

During actual application, whether a vehicle meets system requirements or not is judged, flexible adjustment can be performed according to actual requirements, and one or more conditions are set.

S16, when the vehicle X to be finished meets the system requirements, selecting an order to be loaded from the order set to be loaded to load the vehicle X to be finished; when the loading of the vehicle X is completed, step S17 is executed. For example, when the delivery distance of all the orders to be delivered loaded by the vehicle to be completed X is determined to satisfy the system constraint condition, and the total volume of the goods is smaller than the available volume of the vehicle to be completed X.

In order to save the logistics cost, since the total volume of the goods loaded by the vehicle to be completed X for all orders to be delivered is less than the available volume of the vehicle to be completed X, the order in the system can be continuously selected to load the vehicle to be completed X.

First, the invention specifies the following parameters:

and a to-be-loaded order set Y: y is_iI belongs to {1, n }, namely the order y to be loaded_jThe collection of orders to be loaded is the collection of all orders to be loaded for the order not loaded into the vehicle X to be completed.

To complete vehicle X: x is the number of_iI belongs to {1, m }, that is, the order loaded by the vehicle X to be completed is the order X to be delivered_i. Wherein m and n are both natural numbers larger than zero. In addition, all parameters referred to herein are natural numbers greater than zero.

As shown in fig. 3, the specific implementation process of selecting the to-be-loaded order from the to-be-loaded order set to load the to-be-loaded order into the to-be-completed vehicle X in step S16 is as follows:

step S161, selecting one or more to-be-loaded orders from the to-be-loaded order set Y, and respectively recording the one or more to-be-loaded orders as candidate orders; the set of all candidate orders is denoted as the set of candidate orders.

Specifically, the method comprises the following steps: the method comprises the steps of firstly sorting orders to be loaded in an order set to be loaded, and then selecting one or more orders to be loaded from the order set to be loaded according to a sorting result to serve as candidate orders accurately loaded into a vehicle X to be completed. The ordering of the to-be-loaded orders in the to-be-loaded order set Y is carried out according to the distance between each to-be-loaded order in the to-be-loaded order set and the to-be-completed vehicle X. And during selection, one or more orders to be loaded are sequentially selected according to the principle that the distance is from small to large and are respectively used as candidate orders. The number of the candidate orders is a value which is called by running batch of service data, and is a super parameter which is properly adjusted according to the service. According to the invention, part of orders in the to-be-loaded order set Y are selected as candidate orders loaded into the to-be-completed vehicle X, so that the number of algorithms in the traversal process can be reduced, and the operation speed is further improved.

And step S162, clustering all candidate orders in the candidate order set according to the distribution address of each candidate order to obtain a second clustering result. For example, the candidate order combinations are clustered to obtain M clusters, each cluster including one or more candidate orders. Wherein M is a natural number greater than or equal to 1. That is, the candidate orders are clustered to obtain one or more clusters, and each cluster includes one or more candidate orders.

In addition, in order to ensure the optimal clustering result, the invention clusters all candidate orders in the candidate order set by multiple algorithms, and then determines the optimal clustering algorithm according to the clustering result of the multiple algorithms.

The preferred clustering method in this embodiment includes clustering based on a Kmeans algorithm, clustering based on an EM algorithm, clustering based on a fatehest algorithm, and clustering based on a HIERARCHICAL algorithm.

Clustering all candidate orders in the candidate order set according to the distribution address of each candidate order by each clustering method to obtain corresponding clustering results, then respectively calculating the turnaround distance of the clustering results of each clustering method, and finally selecting the optimal clustering result according to the turnaround distance of the second clustering result, namely determining the clustering algorithm adopted by the embodiment. The optimal clustering algorithm is the clustering algorithm when the turnaround distance is the minimum among all the clustering results.

That is, step S162 specifically includes:

step S1621, clustering all candidate orders according to the distribution address of each candidate order by a plurality of clustering algorithms. For example, all candidate orders in the candidate order set are clustered respectively through the above clustering based on the Kmeans algorithm, the clustering based on the EM algorithm, the clustering based on the fathest algorithm, and the clustering based on the hiperlichical algorithm, and each clustering algorithm has one clustering result.

Step S1622, determining a second clustering algorithm according to the turnaround distance of the clustering result of each clustering algorithm; and the turnaround distance is the sum of the delivery distances of all candidate orders in each cluster after delivery in the clustering result of each clustering algorithm. When determining the clustering algorithm, that is, selecting the clustering algorithm with the smallest turnaround distance of the clustering result.

Step S1623, clustering all candidate orders in the candidate order set according to a second clustering algorithm to obtain a second clustering result. And after the optimal clustering algorithm is determined, clustering all candidate orders in the candidate order set to obtain one or more clusters, so that the candidate orders loaded into the vehicle X to be finished can be selected conveniently in the next step.

Preferably, this embodiment further provides a method for calculating a turnaround distance of a clustering result, as shown in fig. 5, the clustering result of the candidate order set is assumed that the number of clusters is 3, that is, 3 clusters are included, where cluster 1 includes M1 candidate orders, cluster 2 includes M2 candidate orders, and cluster 3 includes M3 candidate orders. Assume that each dot in FIG. 5 represents a delivery address for a candidate order.

The turnaround distance specified in this embodiment is the mileage traveled after all orders have been delivered. As shown in fig. 5, the turnaround distance is the sum of the miles traveled after the delivery of M1 candidate orders in cluster 1, the miles traveled after the delivery of M2 candidate orders in cluster 2, and the miles traveled after the delivery of M3 candidate orders in cluster 3.

And respectively calculating the turnaround distance of the clustering result of each clustering method according to the calculation method of the turnaround distance, and selecting the clustering result with the minimum turnaround distance as the optimal clustering result.

Step S163, selecting the to-be-loaded order filled in the to-be-completed vehicle X from the candidate orders according to the second clustering result of the determined clustering algorithm, removing the to-be-loaded order loaded in the to-be-completed vehicle X from the to-be-loaded order set Y, and updating the to-be-loaded order in the to-be-completed vehicle X at the same time until the to-be-completed vehicle X is completely loaded.

When a loading order to be filled in the to-be-completed vehicle X is selected from the candidate orders, as shown in fig. 4, the present invention further includes: step S1631, first, according to the second clustering result, a cluster with the smallest distance to the vehicle to be completed X is selected and recorded as the current cluster.

Wherein, assume that the order set of the current cluster C is: c. C_iI ∈ {1, n1}, the distance of the vehicle to be completed X from a certain cluster C is:

as shown in fig. 5, the distances of the cluster M1, the cluster M2, and the cluster M3 from the vehicle X to be completed are respectively determined according to the formula (8), and then the size of the distance of each cluster from the vehicle X to be completed is determined according to the calculation result, and the cluster having the smallest distance is selected.

Step S1632: judging whether all candidate orders of the current cluster can be completely loaded into the vehicle X to be completed or not, if so, executing S1633; if not, go to S1634.

Step S1633, load all candidate orders of the current cluster into the to-be-completed vehicle X, update the to-be-delivered orders loaded by the to-be-completed vehicle X, remove all candidate orders of the current cluster from the to-be-loaded order set Y, and then execute step S1635.

Step S1635, judging whether the vehicle X to be finished has a loading space, if so, executing the step S161, and continuing to select the order to be loaded to load the vehicle X to be finished; if not, the vehicle X loading is completed, and step S17 is executed.

And when all the candidate orders of the selected current cluster can be completely loaded into the vehicle X to be completed, directly loading all the candidate orders into the vehicle X to be completed. And then judging whether the vehicle X with the finished goods has a loading space, if so, continuously selecting a corresponding candidate order from the order set Y to be loaded to load the vehicle X with the finished goods into the vehicle X with the finished goods until the vehicle X with the finished goods is loaded completely.

Step S1634, selecting one or more candidate orders from the current cluster according to the current loading space of the vehicle X to be completed to load the candidate orders into the vehicle X to be completed, and executing step S17 until the vehicle X to be completed is loaded; and simultaneously, removing the candidate orders loaded into the to-be-finished vehicle X from the to-be-loaded order set Y, and updating the to-be-dispensed orders loaded by the to-be-finished vehicle X.

And when all the candidate orders of the selected current cluster cannot be completely loaded into the vehicle X to be completed, selecting partial candidate orders from the current cluster to fill the vehicle X to be completed.

That is, when the total volume of goods of all candidate orders in the current cluster is larger than the current available quantity of the vehicle X to be completed, a part of the candidate orders needs to be selected to fill the vehicle X to be completed. Wherein the current available quantity is the total available quantity of the to-be-completed vehicle X, divided by the total volume of the goods of all to-be-delivered orders already loaded in the to-be-completed vehicle X.

When part of candidate orders are selected from the current cluster to be filled in the vehicle X to be completed, the method also adopts two implementation modes of a sequential loading method and an exchange loading method to select part of candidate orders from the current cluster to be filled in the vehicle X to be completed, so that the vehicle X to be completed is loaded.

The sequential loading method comprises the following steps: firstly, respectively calculating the distance between each candidate order in the current cluster and the vehicle X to be finished, sequencing according to the distance between each candidate order in the current cluster and the vehicle X to be finished from large to small, sequentially traversing each candidate order in the current cluster according to a sequencing result, simultaneously judging whether each candidate order can be filled into the vehicle X to be finished, if so, loading the corresponding candidate order into the vehicle X to be finished, updating the order to be distributed of the vehicle X to be finished, removing the corresponding candidate order from the current cluster, and continuously traversing the next candidate order; if not, stopping traversing.

That is, the candidate orders in the cluster are sequentially loaded into the vehicle X to be completed according to the distance between each candidate order in the cluster and the vehicle X to be completed, and the traversing loading is stopped until the candidate orders cannot be completely filled into the vehicle X to be completed.

Suppose that: a cluster has 11 candidate orders and is ranked according to the distance between each candidate order and the vehicle X to be completed, and the initial state is shown in fig. 6.

Then, each candidate order in fig. 6 is traversed, each candidate order is sequentially filled into the vehicle X to be completed, and whether each candidate order can be filled into the vehicle X to be completed is determined at the same time. That is, it is determined whether the cargo volume of each candidate order is less than the current available volume of the to-be-fulfilled vehicle X.

For example, candidate orders 1, 2, 3, 4, 5, 6 are filled into the to-be-completed vehicle X.

When the candidate order 7 is traversed and the candidate order 7 is judged to be incapable of completely loading the vehicle X to be completed, the traversal is stopped. In this state after the sequential loading, the candidate orders 1, 2, 3, 4, 5, and 6 are moved from the cluster to the ready-to-complete vehicle X and changed to the ready-to-deliver order, as shown in fig. 7.

And after the traversal of the sequential loading is stopped, judging whether the vehicle X to be finished has the residual loading space, if not, indicating that the vehicle X to be finished has finished loading.

If the current vehicle X to be finished has the residual loading space, the current vehicle X to be finished can be loaded continuously, and the vehicle X to be finished is loaded continuously by exchanging the loading.

That is, when the sequential loading method traverses to the current candidate order and the current candidate order cannot be loaded into the vehicle X to be completed, the vehicle X to be completed continues to be loaded according to the first swap loading method or the second swap loading method.

The first loading method comprises the following steps: removing the corresponding orders to be delivered from the vehicle X to be completed according to the sequence that the distance between the orders to be delivered in the vehicle X to be completed and the vehicle X to be completed is from small to large until the current candidate orders can be loaded into the vehicle X to be completed; at this time, the to-be-dispensed order removed from the to-be-completed vehicle X is added to the to-be-loaded order, the current candidate order is loaded into the to-be-completed vehicle X and is removed from the current cluster, and the to-be-dispensed order in the to-be-completed vehicle X is updated at the same time.

For example, as shown in fig. 8, after sequentially removing the order to be delivered 5 and the order to be delivered 6 that are closest to the vehicle to be completed X from the vehicle to be completed X, the candidate order 7 that cannot be currently filled in the vehicle to be completed X when the sequential loading traversal is stopped is loaded in the vehicle to be completed X.

For example, as shown in fig. 9, traversing the remaining candidate orders in the cluster to obtain that the to-be-delivered order 9 can completely fill the to-be-completed vehicle X, and thus filling the candidate order 9 into the to-be-completed vehicle X.

According to the alternative implementation of the sequential loading method and the exchange loading method, the order to be loaded of the vehicle X to be completed is selectively loaded from the cluster, and the vehicle X to be completed is loaded when the order to be loaded which can be loaded into the vehicle X to be completed cannot be found from the cluster.

For example, sequential loading is performed firstly, after the sequential loading is stopped, the candidate order 9 is loaded into the vehicle X according to the second exchange loading method, then the candidate order 10 is continuously and sequentially loaded and traversed, the candidate order 10 cannot be loaded into the vehicle X to be completed, and traversing is stopped; and then, according to a first exchange loading method, sequentially removing one or more to-be-delivered orders in the to-be-completed vehicles X, further loading the candidate orders 10 into the to-be-completed vehicles, and then continuously and sequentially loading the traverse candidate orders until all the candidate orders in the cluster are completely traversed, and then finishing loading the to-be-completed vehicles X.

In addition, since the embodiment provides two exchange loading methods, the invention determines to select a proper exchange loading method to realize loading according to the principle of minimum turnover distance loss when in application. The turnaround distance loss here refers to a travel distance generated for order loading when loading the to-be-completed vehicle X is performed.

For the first swap loading method, the turnaround distance loss is the sum of the travel distance generated when each order to be delivered is removed by the vehicle to be completed X and the travel distance generated when the vehicle to be completed X loads the current candidate order.

As shown in fig. 10, the turnaround distance loss calculation: firstly, the vehicle X to be completed needs to get the order 7 to be delivered, and a distance d1 is generated at the moment; then, to-be-delivered order 5 and to-be-delivered order 6 are removed, and other vehicles are needed to fetch to-be-delivered order 5 and to-be-delivered order 6, so that distances d2 and d3 are generated. Then the turnaround distance loss at this time is: d1+ d2+ d 3.

For the second swap loading method, the turnaround distance loss is the sum of the travel distance generated when the vehicle X to be completed loads a new candidate order and the travel distance generated when the current candidate order is delivered.

As shown in fig. 10, the turnaround distance loss calculation: firstly, if the vehicle X to be completed needs to take the order 9 to be delivered, a distance d4 is generated; taking order to be delivered 9 then requires another vehicle to take order to be delivered 7, resulting in travel distance d 5. Then the turnaround distance loss at this time is: d4+ d 5.

Therefore, the scheme with the smallest turnaround distance loss value for swap loading is selected to select the next candidate order to fill in the to-be-completed vehicle X and remove the candidate order from the cluster while updating the to-be-delivered orders in the to-be-completed vehicle X.

That is, the present invention realizes the loading of the to-be-completed vehicle X according to the alternation of the sequential loading and the interchange loading until the to-be-delivered order that can be loaded into the to-be-completed vehicle X is not found from the cluster or the to-be-completed vehicle X has no loading space, at which time the loading of the to-be-completed vehicle X is completed. When the loading of the vehicle X to be completed is completed, step S17 is executed to match the currently available driver for the vehicle X to be completed from the available drivers of the system, and then the delivery is performed.

Further, when all the orders to be delivered loaded by the vehicle X to be completed cannot meet the system requirements, the orders to be delivered in the vehicle X to be completed need to be removed.

The vehicle loading planning method further comprises: step S15, removing the order to be distributed in the vehicle X to be completed, executing step S14, and continuously judging whether the order to be distributed loaded by the vehicle X to be completed meets the system requirement; when satisfied, step S16 is executed.

The removing method comprises the following steps: and removing each order to be delivered loaded by the vehicle X to be completed according to the order set to be loaded from near to far, and judging whether the vehicle X to be completed meets the system requirements or not when one order to be delivered is removed until all the orders to be delivered loaded by the vehicle X to be completed meet the system requirements. Meanwhile, when one order to be delivered is removed from the vehicle X to be completed, the order to be delivered is added into the order set Y to be loaded, and the order to be delivered in the vehicle X to be completed is updated. The embodiment further provides that the distance between order i and order j is:

then the order x is to be delivered_iY is the set of orders from waiting to load_iThe distance of i ∈ {1, n } is:

x_idistance Y ═ min (x)_iAnd y₁Distance, x_iAnd y₂Distance_iAnd y_nDistance) (8).

Step S17, when the loading of the vehicle X is completed, the available drivers in the system are acquired, and the available drivers are selected for the vehicle with the successful loading schedule.

When the available drivers are selected, all available drivers with matched vehicle types can be selected according to the requirements of the vehicle types, each available driver is scored, and the driver with the highest score of the driver seat for distributing the vehicle is selected. Wherein the score is weighted according to the available driver image score and the size of the intersection of the familiar area. The driver representation score may be determined in conjunction with a business such as credit rating, service attitude rating, etc., while the familiar area is described by a radius centered on the location where the available driver resides, and the size of the intersection is represented by the length of the distribution of the available orders and the overlap of the radii.

Step S18, when the to-be-delivered orders loaded by the vehicle with successful loading arrangement are determined, the vehicle type is determined, and the driver is determined, all the to-be-delivered orders loaded by the vehicle with successful loading arrangement are removed from the all-order set of the system, the current vehicle is removed from the available vehicles of the system, the current driver is removed from the available drivers of the system, the operation returns to step S12 until all the orders of the system are completely loaded and delivered, and then other operations are executed or quit.

As shown in fig. 11 and 12, after the present embodiment is applied, the average single-piece logistics cost change trend graph and the overall rate change trend graph are also shown in the present embodiment. Compared with the existing industries, the scheme has the advantages that the overall rate is greatly reduced, the logistics cost is better controlled within a reasonable range, and the cost is saved.

Example two

A vehicle load planning apparatus comprising a memory and a processor, the memory having a load planning program stored thereon, the load planning program being a computer program, the processor executing the load planning program to implement the steps of a vehicle load planning method as used in one embodiment.

EXAMPLE III

A storage medium, which is a computer-readable storage medium, on which a load planning program is stored, where the load planning program is a computer program and when executed by a processor, implements the steps of a vehicle load planning method according to an embodiment.

The above embodiments are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited thereby, and any insubstantial changes and substitutions made by those skilled in the art based on the present invention are within the protection scope of the present invention.

Claims

1. A vehicle loading planning method, characterized in that the loading planning method comprises:

filling: selecting a vehicle to be loaded from the vehicle to be loaded order set to load a vehicle X to be completed; when the loading of the vehicle X is finished, executing a driver matching step; the order to be loaded is an order of the vehicle X to be finished which is not loaded; the order to be loaded is collected as waitingA set of loading orders; suppose the collection of orders to be loaded Y: y is_j,j∈{1,n}，y_jOrdering for loading; set of orders to be delivered X loaded by vehicle to be fulfilled X_i,i∈{1,m}，x_iIs an order to be delivered; wherein m and n are both natural numbers larger than zero;

2. A vehicle load planning method according to claim 1, wherein the initial assigning step further comprises: the method comprises the steps of firstly calculating the total volume of goods of all orders of each cluster, sequencing the orders according to the total volume of the goods from large to small, and then sequentially distributing the orders to each cluster after sequencing according to the available quantity corresponding to the vehicle types of available vehicles required by all the orders after distribution.

3. The vehicle loading planning method according to claim 1, wherein the first clustering algorithm is a kmean clustering algorithm, and when the kmean clustering algorithm clusters all orders to be delivered in the system, the clustering number is equal to the number of available vehicles required for delivering all orders.

4. A vehicle load planning method according to claim 1, wherein the selecting step comprises: firstly, calculating the distance between each cluster and all order center points in a first clustering result, and then selecting one cluster as an off-cluster according to the distance between each cluster and all order center points, wherein the current available vehicle distributed by the off-cluster is a vehicle X to be finished; the distance cluster is the cluster with the largest distance from all the cluster center points of all the orders; all orders are all orders needing to be delivered in the system;

the location of the center point for all orders is expressed as:

the position of the cluster is represented as:

5. The vehicle loading planning method according to claim 1, wherein the criterion that the vehicle X to be completed meets the system requirement is as follows: the total volume of the cargos of all orders to be dispensed loaded by the vehicle X to be finished does not exceed the available quantity of the vehicle X to be finished;

6. The vehicle loading planning method according to claim 1, wherein the filling step specifically comprises:

7. The vehicle loading planning method according to claim 6, wherein the step S11 includes: firstly, each order Y to be loaded in the order set Y to be loaded is calculated respectively_jDistance from vehicle X to be completed, and then according to each order y to be loaded_jSelecting one or more to-be-loaded orders from the to-be-loaded order set Y according to the principle that the distance between the to-be-loaded orders and the to-be-loaded vehicle X is from small to large; wherein the content of the first and second substances,

8. The vehicle loading planning method according to claim 6, wherein the step S12 further includes: step S121, clustering all candidate orders in the candidate order set through multiple clustering algorithms;

9. A vehicle load planning method according to claim 8, wherein the second clustering algorithm is any one of the following: a Kmeans-based clustering algorithm, an EM-based clustering algorithm, a FARTHEST-based clustering algorithm, and a HIERARCHICAL-based clustering algorithm.

10. The vehicle loading planning method according to claim 6, wherein the step S13 includes:

11. The vehicle load planning method according to claim 10, wherein step S134 comprises selecting one or more candidate orders from the current cluster to be loaded into the to-be-completed vehicle X according to the sequential loading method and the swap loading method until the to-be-completed vehicle X is completed when all candidate orders in the current cluster are completely traversed or the to-be-completed vehicle has no loading space.

12. A vehicle loading planning method according to claim 11 in which the sequential loading method comprises: firstly, respectively calculating the distance between each candidate order in the current cluster and a vehicle X to be finished, sequencing according to the distance between each candidate order in the current cluster and the vehicle X to be finished from large to small, sequentially traversing each candidate order in the current cluster according to a sequencing result, simultaneously judging whether each candidate order can be loaded into the vehicle X to be finished, if so, loading the corresponding candidate order into the vehicle X to be finished, removing the corresponding candidate order from the current cluster, updating the order to be delivered of the vehicle X to be finished, and continuously traversing the next candidate order; if not, stopping traversing.

13. A vehicle loading planning method according to claim 12 in which the swap loading method comprises a first swap loading method and a second swap loading method;

14. A vehicle loading planning method according to claim 13 in which the first swap loading method or the second swap loading method is selected based on the magnitude of the turnaround loss for each swap loading method;

15. The vehicle loading planning method according to claim 1, wherein the removing step specifically comprises: step S21, respectively calculating the distance between each order to be distributed loaded by the vehicle X to be completed and the order set Y to be loaded;

16. A vehicle load planning method according to claim 15, wherein said step S21 includes: order x to be delivered_iWith each order Y to be loaded in the order set Y to be loaded_jThe distance of (a) is:

17. A vehicle load planning apparatus comprising a memory and a processor, the memory having a load planning program stored thereon that is executable on the processor, the load planning program being a computer program, characterized in that: the steps of a vehicle load planning method according to any one of claims 1-16 are implemented when the load planning program is executed by the processor.

18. A storage medium, the storage medium being a computer-readable storage medium having a load planning program stored thereon, the load planning program being a computer program, characterized in that: the load planning program when executed by a processor implements the steps of a vehicle load planning method according to any one of claims 1-16.