CN112001678B - Method for reducing tobacco delivery mileage - Google Patents

Method for reducing tobacco delivery mileage Download PDF

Info

Publication number
CN112001678B
CN112001678B CN202010870443.1A CN202010870443A CN112001678B CN 112001678 B CN112001678 B CN 112001678B CN 202010870443 A CN202010870443 A CN 202010870443A CN 112001678 B CN112001678 B CN 112001678B
Authority
CN
China
Prior art keywords
solution
delivery
algorithm
current
new
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010870443.1A
Other languages
Chinese (zh)
Other versions
CN112001678A (en
Inventor
陈是建
陈奋励
叶清云
陈益航
苏志欣
蔡文魁
张剑雄
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Fujian Tobacco Co Putian Branch
Original Assignee
Putian Tobacco Logistics Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Putian Tobacco Logistics Co ltd filed Critical Putian Tobacco Logistics Co ltd
Priority to CN202010870443.1A priority Critical patent/CN112001678B/en
Publication of CN112001678A publication Critical patent/CN112001678A/en
Application granted granted Critical
Publication of CN112001678B publication Critical patent/CN112001678B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Development Economics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention provides a method for reducing tobacco delivery mileage, which comprises the following steps: step S1, adjusting the delivery structure, namely acquiring the number of vehicle delivery times in a week and acquiring the number of delivery vehicles; step S2, constructing a minimum delivery unit of the client; step S3, rearrangement of the delivery route by the minimum delivery unit of the customer, the present invention can effectively reduce the repeat delivery distance.

Description

Method for reducing tobacco delivery mileage
Technical Field
The invention relates to the technical field of tobacco company distribution, in particular to a method for reducing tobacco distribution mileage.
Background
In recent years, a whole-province tobacco business system guides the high-quality development hoof to be steadily advanced according to a newly developed concept, source development, potential excavation and throttling are systematically performed, the important way of continuously improving the operation benefit and promoting the operation, quality and efficiency improvement of enterprises is formed, as an industry logistics enterprise of important links of cost reduction and efficiency improvement in the tobacco business system, the development concept of provincial authorities should be implemented in the cigarette distribution main industry and the non-cigarette distribution new industry, the delivery link with the largest cost occupation ratio is focused, the artificial intelligence algorithm is introduced, the route planning is more reasonable, the effective integration of operation resources is promoted, the total mileage of non-cigarette integrated distribution of cigarettes is continuously reduced, the distribution cost is saved, and the high-quality development and the quality of the enterprises are realized.
The existing logistics transportation capacity can only meet a part of non-smoke distribution requirements; namely, third-party logistics distribution influences the gross profit rate of non-tobacco products, and according to the win-win goal, how to effectively control the total logistics distribution mileage and simultaneously considering the development of the non-tobacco distribution is a new challenge of cigarette non-tobacco integration lower hem in front of logistics companies. The original delivery route arrangement of the tobacco company has the big problems that the total delivery mileage is relatively long, the delivery quantity is relatively small, and how to reduce the total delivery mileage becomes a problem to be solved under the comprehensive development of cigarette non-smoke integrated delivery work.
Disclosure of Invention
In order to overcome the problems, the invention aims to provide a method for reducing the tobacco delivery mileage, which effectively reduces the delivery mileage and improves the delivery efficiency.
The invention is realized by adopting the following scheme: a method of reducing the mileage since tobacco delivery, said method comprising the steps of:
step S1, adjusting the delivery structure, namely acquiring the number of vehicle deliveries in a week and acquiring the number of vehicle deliveries;
step S2, constructing a minimum delivery unit of the customer;
step S3, the delivery route is rearranged by the customer minimum delivery unit.
Further, the step S2 is further specifically: dividing the customers into a plurality of minimum distribution units according to the similarity degree of the longitude and the latitude through a clustering algorithm, and merging the similar customers to form respective minimum distribution units; setting the distributed client set as V ═ V 1,v2,...,vnThe longitude and latitude attribute of the client is xiAnd k cluster subsets c are constructed1,c2,...,ck-minimizing the client differentiation in k subsets of clusters; namely that
Figure GDA0003692878300000021
Wherein the content of the first and second substances,
Figure GDA0003692878300000022
for clustering subset { c1,c2,...,ckMiddle cluster cjCluster center of (j ═ 1, 2.., k), viRepresenting a cluster cjAll clients in; the selection method of the cluster center is as follows:
Figure GDA0003692878300000023
wherein, | cjI is the number of customers in cluster j, xjIs shown by cjLongitude and latitude data of the middle client;
clustering algorithm using Euclidean distance as client nodeSimilarity, so the Euclidean distance d (x) of any two client nodesi,xj) Expressed as:
Figure GDA0003692878300000024
further, the step S3 is further specifically: step 31: obtaining a minimum delivery unit of a customer;
step 32: setting an initialization parameter, i.e. setting an initial temperature T0End temperature Tf；
Step 33: generating an initial solution X0I.e. to solve for X initially0As the current optimal solution Xbest＝X0Obtaining an objective function f (X) of the simulated annealing algorithm0) (ii) a In the step, a customer node in the sub-path with the maximum order quantity is inserted into the sub-path with the minimum customer order quantity, and an initial feasible solution of the simulated annealing algorithm is obtained after continuous adjustment;
step 34: searching locally to obtain new solution, i.e. making a random change to current optimum point to produce a new solution X kAcquiring an objective function f (X) of a new simulated annealing algorithm, and calculating the increment delta f (f) (X) -f (X) of the objective function0) Let k be k +1, k being the number of iterations;
step 35: judging whether the new solution can become the current optimal solution, namely if delta f is less than 0, accepting the new solution X to become the current optimal solution, XbestX; otherwise, receiving X as a new current optimal solution by taking the probability p as exp (-delta f/T), wherein T is the temperature T corresponding to the optimal solution X;
step 36: judging whether the simulated annealing algorithm is finished or not, and if the current temperature is more than the termination temperature TfReturning to step 34, otherwise, outputting the current solution as the optimal solution, and ending the algorithm.
Further, the method further comprises: step S4, optimizing the delivery route through a tabu search algorithm; the step S4 further includes: based on the minimum delivery units of the customers, defining the positions of the customer nodes as the average longitude and latitude information of each minimum delivery unit, judging whether the distance between any two customer nodes is smaller than a preset Q value or not through a proximity KNN algorithm, if so, performing two neighborhood searching operations, and recording an accessed neighborhood solution; if not, abandoning the pair of nodes and carrying out the same operation on other nodes.
Further, the step S4 is further specifically: step 41: obtaining a minimum delivery unit of a customer;
step 42: setting initialization parameters, setting initial temperature T0Termination temperature TfAnd the value of G in the KNN algorithm; discarding the farthest nodes which do not conform to the KNN algorithm through a preset G value, and reducing the range of neighborhood searching;
step 43: generating an initial solution X0I.e. to solve for X initially0As the current optimal solution Xbest＝X0Obtaining an objective function f (X) of the simulated annealing algorithm0) (ii) a In the step, the customer node in the sub-path with the maximum order quantity is inserted into the sub-path with the minimum customer order quantity, and an initial feasible solution of the hybrid simulated annealing algorithm is obtained after continuous adjustment;
step 44: by searching locally to obtain a new solution, i.e. at the current solution X0Respectively carrying out local search operations of an exchange method and an interpolation method in the neighborhood of the target to generate a new feasible solution X, obtaining an objective function f (X) of a new simulated annealing algorithm, and calculating the increment delta f (f) (X) -f (X) of the objective function0) Let k be k +1, k being the number of iterations;
step 45: judging whether the new solution can become the current optimal solution, if delta f is less than 0, accepting the new solution X to become the current optimal solution, XbestX; otherwise, receiving X as a new current optimal solution by taking the probability p as exp (-delta f/T), wherein T is the temperature T corresponding to the optimal solution X;
Step 46: judging whether all neighborhood searching is finished, namely under the current temperature T, if not, turning to the step 47, otherwise, turning to the step 44;
step 47: judging whether the hybrid simulated annealing algorithm is finished or not, and if the current temperature is higher than the termination temperature TfGo back to step 44, otherwise output the current solution as the optimal solution, and end the algorithm.
The invention has the beneficial effects that: the bottleneck problem of manual line planning which is troubled for a long time is solved, a cigarette non-cigarette integrated 4D distribution new mode based on the ant colony working principle is preliminarily explored and formed, intelligent optimization model improvement methods such as a hybrid simulation annealing algorithm covering the minimum delivery unit and based on tabu search are built, the delivery mileage is effectively reduced, the delivery efficiency is improved, the characteristics of strong reproducibility and generalizability are achieved in the industry, and certain effect is achieved in the aspects of operation cost control and informatization construction.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention.
Fig. 2 is a detailed flowchart of the present invention for optimizing the delivery route by a tabu search algorithm.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to fig. 1, the present invention provides a method for reducing the mileage of tobacco distribution, the method comprising the steps of:
step S1, adjusting the delivery structure, namely acquiring the number of vehicle deliveries in a week and acquiring the number of vehicle deliveries;
step S2, constructing a minimum delivery unit of the client;
the step S2 further includes: dividing the customers into a plurality of minimum distribution units according to the similarity degree of the longitude and the latitude through a clustering algorithm, and merging the similar customers to form respective minimum distribution units; setting the distributed client set as V ═ V1,v2,...,vnThe longitude and latitude attribute of the client is xiConstructing k clustering subsets { c1,c2,...,ck}, minimizing client differentiation in k subsets of clusters; namely, it is
Figure GDA0003692878300000041
Wherein the content of the first and second substances,
Figure GDA0003692878300000042
for clustering subset { c1,c2,...,ckMiddle cluster cj(j ═ 1, 2.. times, k) cluster center, viRepresenting a cluster cjAll the clients; the selection method of the cluster center is as follows:
Figure GDA0003692878300000043
wherein, | cjI is the number of customers in cluster j, xjDenotes cjLongitude and latitude data of the middle client;
the clustering algorithm takes the Euclidean distance as the similarity of the client nodes, so the Euclidean distance d (x) of any two client nodesi,xj) Expressed as:
Figure GDA0003692878300000044
step S3, the delivery route is rearranged by the customer minimum delivery unit.
The step S3 further includes: step 31: obtaining a minimum delivery unit of a client;
step 32: setting initialization parameters, i.e. setting the initial temperature T0Termination temperature Tf；
Step 33: generating an initial solution X0I.e. to solve for X initially0As the current optimal solution Xbest＝X0Obtaining an objective function f (X) of the simulated annealing algorithm0) (ii) a In the step, the customer node in the sub-path with the maximum order quantity is inserted into the sub-path with the minimum customer order quantity, and an initial feasible solution of the simulated annealing algorithm is obtained after continuous adjustment;
step 34: obtaining a new solution by local search, i.e. randomly changing the current optimum point to generate a new solution XkAcquiring an objective function f (X) of a new simulated annealing algorithm, and calculating the increment delta f (f) (X) -f (X) of the objective function0) Making k equal to k +1, wherein k is the iteration number;
step 35: judging whether the new solution can become the current optimal solution, namely if delta f is less than 0, accepting the new solution X to become the current optimal solution, XbestX; otherwise, receiving X as a new current optimal solution by taking the probability p as exp (-delta f/T), wherein T is the temperature T corresponding to the optimal solution X;
step 36: judging whether the simulated annealing algorithm is finished or not, and if the current temperature is higher than the termination temperature TfReturning to step 34, otherwise, outputting the current solution as the optimal solution, and ending the algorithm.
Referring to fig. 2, the method further includes: step S4, optimizing the delivery route through a tabu search algorithm; the step S4 further includes: based on the minimum delivery units of the customers, defining the positions of the customer nodes as the average longitude and latitude information of each minimum delivery unit, judging whether the distance between any two customer nodes is smaller than a preset Q value or not through a proximity KNN algorithm, if so, performing two neighborhood searching operations, and recording an accessed neighborhood solution; and if not, abandoning the pair of nodes and performing the same operation on other nodes.
The step S4 further includes: step 41: obtaining a minimum delivery unit of a client;
step 42: setting initialization parameters, setting initial temperature T0Termination temperature TfAnd the value of G in the KNN algorithm; discarding the farthest nodes which do not conform to the KNN algorithm through a preset G value, and reducing the range of neighborhood searching;
step 43: generating an initial solution X0I.e. to solve for X initially0As the current optimal solution Xbest＝X0Obtaining an objective function f (X) of the simulated annealing algorithm0) (ii) a In the step, the customer node in the sub-path with the maximum order quantity is inserted into the sub-path with the minimum customer order quantity, and an initial feasible solution of the hybrid simulated annealing algorithm is obtained after continuous adjustment;
And step 44: searching locally to obtain a new solution, i.e. at the current solution X0Respectively carrying out local search operations of an exchange method and an interpolation method in the neighborhood of the target to generate a new feasible solution X, obtaining an objective function f (X) of a new simulated annealing algorithm, and calculating the increment delta f (f) (X) -f (X) of the objective function0) Making k equal to k +1, wherein k is the iteration number;
step 45: judging whether the new solution can become the current optimal solution, if delta f is less than 0, accepting the new solution X to become the current optimal solution, XbestX; otherwise with probabilityReceiving X as a new current optimal solution by p-exp (-delta f/T), wherein T is the temperature T corresponding to the optimal solution X;
step 46: judging whether all neighborhood searching is finished, namely under the current temperature T, if not, turning to the step 47, otherwise, turning to the step 44;
step 47: and judging whether the hybrid simulated annealing algorithm is finished, if the current temperature is higher than the termination temperature Tf, returning to the step 44, otherwise, outputting the current solution as the optimal solution, and finishing the algorithm.
The invention is further illustrated below with reference to a specific embodiment:
structure for implementing and adjusting delivery of countermeasure
(1) Calculating the number of laps of the turnover vehicle
In 2016 + 2018, the cigarette sales increase is relatively small, and the influence on the vehicle loading rate is mainly the number of trips. The average weekly sales is calculated by calling the total amount of the year 2018 in the whole year, and the weekly trips meeting the 75% loading rate index are calculated by taking the full load of 95 pieces and the provincial load rate index of 75% as basic parameters.
Calculating the formula: the number of vehicle-out passes is 75% of the average round-trip-rate/vehicle loading rate/95 of the number of single-vehicle full-load pieces. As shown in Table 1 below
TABLE 1
Figure GDA0003692878300000061
It is found by calculation that the weekly vehicle trips are controlled within 187 trips when the annual vehicle loading rate is 75%.
(2) Measuring and calculating distribution vehicle
After the delivery service data is accurately analyzed, the number of vehicles required to be delivered is determined by taking 187 weeks as reference conditions, the number of the vehicle passes and the proportion of clients on the day are respectively calculated according to 26, 27 and 28 vehicles, and the calculation results are as follows:
calculating the formula: the departure times are 4 days (tuesday to friday) on the day line, 5 days (tuesday to saturday) on the day line, 1 day (saturday) on the day delivery line
Customer proportion on the day: current day line, number of delivery customers (press 65 customers) day/total number of customers
Figure GDA0003692878300000071
Through data analysis, of the vehicle plans around the weekly trip 187, the daily delivery ratio of 27 vehicle plans and 26 vehicle plans is close to 30%, and the daily delivery customer ratio of 28 vehicle plans is only 22%. So the proposal of 28 vehicles is eliminated.
Through measurement and calculation, 27 distribution vehicles are adopted (namely one vehicle is reduced on the basis of the original delivery vehicle), the average daily average number of users of a single vehicle is increased by 4, the single vehicle distribution mileage is increased by about 8 kilometers, the total single vehicle distribution time is estimated to be increased by about 0.5 hour, the total average single vehicle distribution time is estimated to be 7.4 hours, the distribution time of a C25 line is only met, and the working time is 8.22 hours after 8 hours.
By adopting 26 delivery vehicles (namely, two vehicles are reduced on the basis of the original delivery vehicles), the average daily average number of users of a single vehicle is increased by 8, the delivery mileage of the single vehicle is increased by about 16 kilometers, the total delivery time of the single vehicle is expected to be increased by about 1 hour, the total delivery time of the single vehicle is expected to be 7.9 hours, and the working time of 13 delivery lines exceeds 8 hours.
By contrast, it is considered relatively reasonable to use 27 vehicles for distribution.
And implementing a second strategy: customer minimization unit
And dividing the customers into a plurality of minimum distribution units according to the similarity of the longitude and the latitude through a clustering algorithm, and merging the customers close to the city compartment customer service part to form respective minimum distribution units.
The K-means clustering algorithm for the pu tian tobacco stream can be described as: given a set V ═ V { V } containing 1420 city block customers1,v2,...,vn1420, where the latitude and longitude attribute of the client is xiConstructing k clustering subsets { c1,c2,...,ckAnd (4) minimizing the customer differentiation in the k clustering subsets. Target function of K-means clustering algorithmThe number is as in equation (9), and the right side also becomes the error sum of squares function for the class.
Namely, it is
Figure GDA0003692878300000081
Wherein the content of the first and second substances,
Figure GDA0003692878300000082
for clustering subset { c1,c2,...,ckMiddle cluster cj(j ═ 1, 2.. times, k) cluster center, vjRepresenting a cluster cjAll clients in; the selection method of the cluster center is formula (10):
Figure GDA0003692878300000083
Wherein, | cjI is the number of customers in cluster j, xjIs shown by cjLongitude and latitude data of the middle client;
the K-means clustering algorithm of the Pu Tian tobacco takes the Euclidean distance as the similarity of the client nodes, so that the Euclidean distance d (x) of any two client nodesi,xj) Expressed as:
Figure GDA0003692878300000084
the purpose of the customer base being divided into a number of minimum delivery units is to reduce the pressure on the optimization of the line, the number of minimum delivery units should not be too large. Meanwhile, if the number of customers in the minimum delivery unit is too large, the accuracy of the route optimization is affected, and therefore, the number of the minimum delivery unit is not too small.
And (3) implementing a strategy III: rearranging a delivery route:
the route planning of the present invention is done manually. The artificial line optimization can also be reluctantly completed in a mountainous area service department with sparser customer distribution, but for an urban compartment area service department with dense customers and large demand, the artificial line optimization takes time and has poor effect. The invention is based on the actual situation of the Pu Tian tobacco material flow, applies a simulated annealing algorithm, and comprises the following specific steps:
step 31: obtaining a minimum delivery unit of a client;
step 32: setting initialization parameters, i.e. setting the initial temperature T0Termination temperature T f；
Step 33: generating an initial solution X0I.e. to solve for X initially0As the current optimal solution Xbest＝X0Obtaining the objective function f (X) of the simulated annealing algorithm0) (ii) a In the step, the customer node in the sub-path with the maximum order quantity is inserted into the sub-path with the minimum customer order quantity, and an initial feasible solution of the simulated annealing algorithm is obtained after continuous adjustment;
step 34: obtaining a new solution by local search, i.e. randomly changing the current optimum point to generate a new solution XkAcquiring a new target function f (X) of the simulated annealing algorithm, and calculating the increment delta f (f) (X) -f (X) of the target function0) Let k be k +1, k being the number of iterations;
step 35: judging whether the new solution can become the current optimal solution, namely if delta f is less than 0, accepting the new solution X to become the current optimal solution, XbestX; otherwise, receiving X as a new current optimal solution by taking the probability p as exp (-delta f/T), wherein T is the temperature T corresponding to the optimal solution X;
step 36: judging whether the simulated annealing algorithm is finished or not, and if the current temperature is more than the termination temperature TfReturning to step 34, otherwise, outputting the current solution as the optimal solution, and ending the algorithm.
Through the intelligent line optimization of the subject group, nearly twenty thousand cigarettes are distributed by the service department of the urban compartment area every week, and account for 80.8 percent of all orders of the urban compartment area. After optimization, the number of customers 289 is sent to the current day, which accounts for 20.3% of the total number of customers in the urban area, and is nearly twice of that of the manual optimization line, so that the customer experience is greatly improved.
And implementing a strategy of four: taboo search optimization delivery line
According to the method, the distances from the client nodes to all the client nodes in the training set are calculated, K adjacent client nodes closest to the current client node are selected, and a KNN algorithm is adopted and combined with two neighborhood search algorithms, so that the calculation complexity of local search is reduced, the complexity of a hybrid simulated annealing algorithm is reduced, and the operation efficiency of the algorithm is improved.
Based on the minimum delivery units proposed above, the client node location is defined as the average longitude and latitude information of each minimum delivery unit, and there are 223 client nodes taking the urban area as an example. Judging whether the distance between any two nodes is smaller than a preset G value or not by a neighbor KNN algorithm for any two nodes, if so, performing two neighborhood searching operations, and recording an accessed neighborhood solution; if not, abandoning the pair of nodes and carrying out the same operation on other nodes. And performing secondary promotion on the basis of the re-simulation algorithm, and establishing a hybrid simulated annealing algorithm based on tabu search, wherein the specific steps are as follows, and a flow chart is shown in fig. 2.
The step S4 further includes: step 41: obtaining a minimum delivery unit of a client;
Step 42: setting initialization parameters, setting initial temperature T0End temperature TfAnd the value of G in the KNN algorithm; discarding the farthest nodes which do not conform to the KNN algorithm through a preset G value, and reducing the range of neighborhood searching;
step 43: producing an initial solution X0I.e. to solve for X initially0As the current optimal solution Xbest＝X0Calculating the objective function f (X) of the simulated annealing algorithm0) (ii) a In the step, the customer node in the sub-path with the maximum order quantity is inserted into the sub-path with the minimum customer order quantity, and an initial feasible solution of the hybrid simulated annealing algorithm is obtained after continuous adjustment;
step 44: by searching locally to obtain a new solution, i.e. at the current solution X0Respectively carrying out local search operations of an exchange method and an interpolation method in the neighborhood of the target to generate a new feasible solution X, calculating a new target function f (X) of the simulated annealing algorithm, and calculating the increment delta f (f) (X) -f (X) of the target function0) Let k be k +1, k being the number of iterations;
step 45: judging whether the new solution can become the current optimal solution, if delta f is less than 0, accepting the new solution X to become the current optimal solution, XbestX; otherwise, receiving X as a new current optimal solution according to the probability p ═ exp (-delta f/T);
step 46: judging whether all neighborhood searches are completed or not, namely, under the current temperature T, if all neighborhood searches in a tabu table in the KNN algorithm are not completed, turning to a step 47, and if not, turning to a step 44;
Step 47: and judging whether the hybrid simulated annealing algorithm is finished, if the current temperature is higher than the termination temperature Tf, returning to the step 44, otherwise, outputting the current solution as the optimal solution, and finishing the algorithm.
In a word, the bottleneck problem of manual line planning which is troubled for a long time is solved, a new cigarette non-smoke integration type 4D distribution mode based on the ant colony working principle is preliminarily explored and formed, intelligent optimization model improvement methods such as a mixed simulated annealing algorithm covering the minimum delivery unit of delivery and based on tabu search are built, the delivery mileage is effectively reduced, the delivery efficiency is improved, the characteristics of strong reproducibility and popularization are achieved in the industry, and certain effect is achieved in the aspects of operation cost control and information construction.
The above description is only a preferred embodiment of the present invention, and all equivalent changes and modifications made in accordance with the claims of the present invention should be covered by the present invention.

Claims (3)

1. A method of reducing tobacco delivery mileage, comprising: the method comprises the following steps:
step S1, adjusting the delivery structure, namely acquiring the number of vehicle delivery times in a week and acquiring the number of delivery vehicles;
step S2, constructing a minimum delivery unit of the client;
Step S3, rearranging the delivery route by the customer minimum delivery unit;
the step S2 further includes: dividing the customers into a plurality of minimum distribution units according to the similarity degree of the longitude and the latitude through a clustering algorithm, and merging the similar customers to form respective minimum distribution units; setting the distributed client set as V ═ V1,v2,...,vnThe longitude and latitude attribute of the client is xiConstructing k clustering subsets { c1,c2,...,ck-minimizing the client differentiation in k subsets of clusters; namely, it is
Figure FDA0003633551080000011
Wherein the content of the first and second substances,
Figure FDA0003633551080000012
for clustering subset { c1,c2,...,ckCluster in (c)jCluster center of (j ═ 1, 2.., k), viRepresenting a cluster cjAll clients in; the selection method of the cluster center is as follows:
Figure FDA0003633551080000013
wherein, | cjI is the number of customers in cluster j, xjDenotes cjLongitude and latitude data of the middle client;
the clustering algorithm takes the Euclidean distance as the similarity of the client nodes, so the Euclidean distance d (x) of any two client nodesi,xj) Expressed as:
Figure FDA0003633551080000014
the step S3 further includes: step 31: obtaining a minimum delivery unit of a client;
step 32: setting initialization parameters, i.e. setting the initial temperature T0Termination temperature Tf；
Step 33: generating an initial solution X0I.e. to solve for X initially0As the current optimal solution X best＝X0Obtaining an objective function f (X) of the simulated annealing algorithm0) (ii) a In the step, a customer node in the sub-path with the maximum order quantity is inserted into the sub-path with the minimum customer order quantity, and an initial feasible solution of the simulated annealing algorithm is obtained after continuous adjustment;
step 34: obtaining a new solution by local search, i.e. randomly changing the current optimum point to generate a new solution XkObtaining a new objective function f (X) of the simulated annealing algorithm, and obtaining the increment delta f (f) (X) -f (X) of the objective function0) Let k be k +1, k being the number of iterations;
step 35: judging whether the new solution can become the current optimal solution, namely if delta f is less than 0, accepting the new solution X to become the current optimal solution, XbestX; otherwise, receiving X as a new current optimal solution by taking the probability p as exp (-delta f/T), wherein T is the temperature T corresponding to the optimal solution X;
step 36: judging whether the simulated annealing algorithm is finished or not, and if the current temperature is more than the termination temperature TfReturning to step 34, otherwise, outputting the current solution as the optimal solution, and ending the algorithm.
2. A method of reducing tobacco delivery mileage as recited in claim 1, wherein: the method further comprises the following steps: step S4, optimizing the delivery route through a tabu search algorithm; the step S4 further includes: based on the minimum delivery units of the customers, defining the positions of the customer nodes as the average longitude and latitude information of each minimum delivery unit, judging whether the distance between any two customer nodes is smaller than a preset Q value or not through a proximity KNN algorithm, if so, performing two neighborhood searching operations, and recording an accessed neighborhood solution; if not, abandoning the pair of nodes and carrying out the same operation on other nodes.
3. A method of reducing tobacco delivery mileage as set forth in claim 2, wherein: the step S4 further includes: step 41: obtaining a minimum delivery unit of a client;
step 42: setting initialization parameters, setting initial temperature T0Termination temperature TfAnd the value of G in the KNN algorithm; discarding the farthest nodes which do not conform to the KNN algorithm through a preset G value, and reducing the range of neighborhood searching;
step 43: generating an initial solution X0I.e. to solve for X initially0As the current optimal solution Xbest＝X0Obtaining an objective function f (X) of the simulated annealing algorithm0) (ii) a In the step, the customer node in the sub-path with the maximum order quantity is inserted into the sub-path with the minimum customer order quantity, and an initial feasible solution of the hybrid simulated annealing algorithm is obtained after continuous adjustment;
step 44: by searching locally to obtain a new solution, i.e. at the current solution X0Respectively carrying out local search operations of an exchange method and an interpolation method in the neighborhood of the target to generate a new feasible solution X, obtaining an objective function f (X) of a new simulated annealing algorithm, and obtaining an increment delta f (f) (X) -f (X) of the objective function0) Let k be k +1, k being the number of iterations;
step 45: judging whether the new solution can become the current optimal solution, if delta f is less than 0, accepting the new solution X to become the current optimal solution, X bestX; otherwise, receiving X as a new current optimal solution by taking the probability p as exp (-delta f/T), wherein T is the temperature T corresponding to the optimal solution X;
step 46: judging whether all neighborhood searching is finished, namely under the current temperature T, if not, turning to the step 47, otherwise, turning to the step 44;
step 47: judging whether the hybrid simulated annealing algorithm is finished or not, and if the current temperature is more than the termination temperature TfGo back to step 44, otherwise output the current solution as the optimal solution, and end the algorithm.
CN202010870443.1A 2020-08-26 2020-08-26 Method for reducing tobacco delivery mileage Active CN112001678B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010870443.1A CN112001678B (en) 2020-08-26 2020-08-26 Method for reducing tobacco delivery mileage

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010870443.1A CN112001678B (en) 2020-08-26 2020-08-26 Method for reducing tobacco delivery mileage

Publications (2)

Publication Number Publication Date
CN112001678A CN112001678A (en) 2020-11-27
CN112001678B true CN112001678B (en) 2022-07-19

Family

ID=73471922

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010870443.1A Active CN112001678B (en) 2020-08-26 2020-08-26 Method for reducing tobacco delivery mileage

Country Status (1)

Country Link
CN (1) CN112001678B (en)

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000112925A (en) * 1998-10-07 2000-04-21 Ibm Japan Ltd Method/system for generating delivery route for miniaturizing number of routes
JP5612817B2 (en) * 2008-11-04 2014-10-22 インターナショナル・ビジネス・マシーンズ・コーポレーションＩｎｔｅｒｎａｔｉｏｎａｌ Ｂｕｓｉｎｅｓｓ Ｍａｃｈｉｎｅｓ Ｃｏｒｐｏｒａｔｉｏｎ Method, computer system and computer program for determining delivery routes of multiple vehicles
CN102880798A (en) * 2012-09-20 2013-01-16 浪潮电子信息产业股份有限公司 Variable neighborhood search algorithm for solving multi depot vehicle routing problem with time windows
CN103383756A (en) * 2013-07-22 2013-11-06 浙江省烟草公司绍兴市公司 Planning method for tobacco logistics distribution routes
CN107392539A (en) * 2017-07-23 2017-11-24 浙江省烟草公司温州市公司 A kind of cigarette delivery management system
CN107578178B (en) * 2017-09-11 2018-08-28 合肥工业大学 Based on the dispatching method and system for becoming neighborhood search and gravitation search hybrid algorithm
CN108764754A (en) * 2018-06-08 2018-11-06 浪潮软件股份有限公司 A kind of tobacco distribution vehicle dispatch system based on geographic area grouping
EP3820753B1 (en) * 2018-07-14 2023-08-02 Moove.AI Vehicle-data analytics
CN110930091B (en) * 2019-11-05 2023-08-18 哈尔滨工业大学（威海） Express delivery end net point optimization integration method based on neighborhood search simulated annealing algorithm
CN111178440B (en) * 2019-12-31 2023-06-20 哈尔滨智兀科技有限公司 Path optimization method based on efficient reduction of total transport cost of AGV

Also Published As

Publication number Publication date
CN112001678A (en) 2020-11-27

Similar Documents

Publication Publication Date Title
Chen et al. Solving the first‐mile ridesharing problem using autonomous vehicles
Pelzer et al. A partition-based match making algorithm for dynamic ridesharing
CN109117993B (en) Processing method for optimizing vehicle path
Morency et al. Typology of carsharing members
Rong et al. Mining efficient taxi operation strategies from large scale geo-location data
Yang et al. Dynamic pricing for integrated energy-traffic systems from a cyber-physical-human perspective
Asghari et al. Environmental and social implications of incorporating carpooling service on a customized bus system
Guo et al. Integrated location and routing planning of electric vehicle service stations based on users’ differentiated perception under a time-sharing leasing mode
Ren et al. Improved ant colony optimization for the vehicle routing problem with split pickup and split delivery
He et al. An adaptive variable neighborhood search ant colony algorithm for vehicle routing problem with soft time windows
TWI591492B (en) Energy consumption estimation system and method
CN114912740A (en) Intelligent decision-making method and system for travel on demand
CN113538067A (en) Inter-city network taxi appointment demand prediction method and system based on machine learning
Meng et al. Customer‐Oriented Vehicle Routing Problem with Environment Consideration: Two‐Phase Optimization Approach and Heuristic Solution
CN113935505A (en) Shared electric vehicle operation optimization method based on column generation and ant colony algorithm fusion
CN116415882A (en) Instant distribution order distribution system for collaborative distribution of riders and unmanned vehicles
CN115689431A (en) Tail end distribution path planning method for two-stage incoming line of new customer
Sun et al. A graphical game approach to electrical vehicle charging scheduling: Correlated equilibrium and latency minimization
CN114418213A (en) Urban electric vehicle scheduling method and system based on deep reinforcement learning
CN112001678B (en) Method for reducing tobacco delivery mileage
CN114936666A (en) Electric vehicle charging navigation method and system based on vehicle-station-platform system
CN112613701B (en) Finished cigarette logistics scheduling method
Guo et al. A prediction-based iterative Kuhn-Munkres approach for service vehicle reallocation in ride-hailing
Wang et al. An Affinity Propagation‐Based Clustering Method for the Temporal Dynamics Management of High‐Speed Railway Passenger Demand
CN115935820A (en) Large-scale road network traffic distribution model parameter calibration method based on improved C-SPSA

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right

Effective date of registration: 20230404

Address after: 351100 No. 185, Licheng South Avenue, Chengxiang District, Putian City, Fujian Province

Patentee after: Fujian tobacco company Putian Co.

Address before: 351131 No. 1459, Shaolin South Street, West Tianwei Town, Licheng District, Putian City, Fujian Province

Patentee before: PUTIAN TOBACCO LOGISTICS Co.,Ltd.

TR01 Transfer of patent right