CN114757394A - Logistics vehicle path optimization method, system and medium based on workload balance - Google Patents

Logistics vehicle path optimization method, system and medium based on workload balance Download PDF

Info

Publication number
CN114757394A
CN114757394A CN202210275223.3A CN202210275223A CN114757394A CN 114757394 A CN114757394 A CN 114757394A CN 202210275223 A CN202210275223 A CN 202210275223A CN 114757394 A CN114757394 A CN 114757394A
Authority
CN
China
Prior art keywords
cluster
small
preset
logistics
target
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.)
Granted
Application number
CN202210275223.3A
Other languages
Chinese (zh)
Other versions
CN114757394B (en
Inventor
李晶晶
汤娜
方姚惠琼
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.)
South China Normal University
Original Assignee
South China Normal University
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 South China Normal University filed Critical South China Normal University
Priority to CN202210275223.3A priority Critical patent/CN114757394B/en
Publication of CN114757394A publication Critical patent/CN114757394A/en
Application granted granted Critical
Publication of CN114757394B publication Critical patent/CN114757394B/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
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a method, a system and a medium for optimizing a logistics vehicle path based on workload balance, which can be applied to the technical field of vehicle paths. The method comprises the steps of constructing an objective function and a constraint condition of the objective function through a logistics scene, simultaneously setting the logistics scene as a first preset logistics scene, dividing clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to vehicle capacity, and fusing the small clusters into a plurality of large clusters according to a preset rule; and determining the logistics scene as the second preset logistics scene, fusing a plurality of small clusters into a plurality of large clusters according to a preset rule, then searching vehicle paths in the large clusters through an ant colony search algorithm according to the constraint conditions, and optimizing the searched paths according to the objective function, so that the working efficiency of workers during logistics distribution according to the optimized paths is improved, and the logistics transportation cost is reduced.

Description

Logistics vehicle path optimization method, system and medium based on workload balance
Technical Field
The invention relates to the technical field of vehicle paths, in particular to a method, a system and a medium for optimizing a logistics vehicle path based on workload balance.
Background
In logistics companies, the work hours of drivers are measured by driving time and service time. The driving time is related to the length of the path traveled, while the service time is mainly related to the amount of cargo being handled. In the existing vehicle path optimization method for ensuring the workload balance, the workload is simply defined as the length of a path for the vehicle to travel, or the load of the vehicle, or a profit hook, and the length and the load of the path for traveling are not comprehensively considered in the workload, so that the workload cannot be effectively balanced, and the service quality is improved.
The logistics chain generally comprises three links: the first is the movement of goods from the retailer to the large warehouse or distribution center, which is typically long distance transportation across regions, the second is the transportation from the large warehouse or distribution center to the community distribution point, and the third is the link from the distribution to the final customer. The first link uses airplane, ship and rail trunk transportation, and the route is fixed, so the current vehicle path optimization model only carries out independent optimization on the second link or the third link, and a method which can simultaneously optimize the two links and balance the workload is lacked.
Disclosure of Invention
The present invention is directed to solving at least one of the problems of the prior art. Therefore, the invention provides a method, a system and a medium for optimizing the logistics vehicle path based on workload balance, which can effectively improve the service quality and reduce the logistics transportation cost.
In one aspect, an embodiment of the present invention provides a method for optimizing a logistics vehicle path based on workload balancing, including the following steps:
determining a logistics scene, wherein the logistics scene comprises a first preset logistics scene and a second preset logistics scene;
constructing an objective function and a constraint condition of the objective function according to the logistics scene;
determining the logistics scene as the first preset logistics scene, dividing clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to vehicle capacity, and fusing the small clusters into a plurality of large clusters according to a preset rule;
determining the logistics scene as the second preset logistics scene, and fusing a plurality of small clusters into a plurality of large clusters according to a preset rule;
and searching the vehicle path in the large cluster through an ant colony search algorithm according to the constraint condition, and optimizing the searched path according to the objective function.
In some embodiments, when the logistics scenario is a first preset logistics scenario,
the objective function includes minimizing travel costs and minimizing a ratio of vehicle maximum workload to vehicle minimum workload;
the constraints include that the node obeys the flow conservation constraints, that each customer is visited only once, that at least one customer is served per vehicle, that each customer is served by only one vehicle, that all vehicles must leave the same warehouse, and that the vehicles are not overloaded.
In some embodiments, the dividing the clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to vehicle capacity includes:
calculating the radius and the first density of a first neighbor, and initializing an initial core point state and a node state;
determining a first clustering center point meeting a first preset requirement according to the first neighbor radius and the first density;
obtaining client nodes which can reach the first clustering center and meet the vehicle-mounted capacity to form a small cluster;
determining a second clustering center point meeting a second preset requirement;
obtaining client nodes which can reach the second cluster center and meet the vehicle-mounted capacity to form other small clusters;
and determining that the number of the small clusters is equal to a first preset number, and stopping the small cluster forming process.
In some embodiments, after the dividing the clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to the vehicle capacity, the method further comprises the following steps:
calculating the average distance from a first noise node to each small cluster, wherein the first noise node comprises a client node which is not accessed in the small cluster forming process;
acquiring a small cluster and a first noise node corresponding to the average distance smaller than a preset distance as a first target small cluster and a target node respectively;
and determining that the sum of the number of the target nodes and the number of the nodes in the first target small cluster is less than or equal to a second preset number, adding the target nodes to the first target small cluster, and marking the target nodes as an accessed state.
In some embodiments, said fusing said several small clusters into several large clusters according to a preset rule includes:
setting the states of all small clusters as usable states;
determining that a second target small cluster is not fused and the second target small cluster has an adjacent small cluster or an adjacent large cluster, and fusing the second target small cluster and a target cluster meeting a third preset requirement; the target cluster meeting the third preset requirement is a small cluster or a large cluster which has the maximum adjacency value and the number of small clusters in the fused large cluster is less than or equal to the third preset number; the second target small cluster is one of all the small clusters;
Determining that the second target small cluster is not fused, the second target small cluster has an adjacent small cluster or an adjacent large cluster, and the required node quantity of the second target small cluster is less than the required quantity of a preset small cluster, and fusing the second target small cluster and a target cluster meeting a fourth preset requirement; the target cluster meeting the fourth preset requirement is a nearby small cluster or a nearby large cluster which is closest to the second target small cluster and the number of the small clusters in the fused large cluster is less than or equal to the third preset number;
determining that the second small target cluster is fused into a first large target cluster, wherein the first large target cluster has an adjacent small cluster or an adjacent large cluster, and fusing the first large target cluster with a target cluster meeting a fifth preset requirement; the target cluster meeting the fifth preset requirement is a large cluster or a small cluster, and the number of small clusters in the fused large cluster is less than or equal to a third preset number.
In some embodiments, after said fusing said several small clusters into several large clusters according to a preset rule, said method further comprises the steps of:
calculating the average distance from a second noise node to each large cluster, wherein the second noise node comprises the remaining client nodes which are not accessed in the small cluster forming process;
Acquiring a large cluster meeting a sixth preset requirement as a second target large cluster; the sixth preset required large cluster is a large cluster which does not access the second noise node within a preset time period and the total demand of the second noise node and one of the large clusters is less than or equal to the upper limit of the total demand of the one of the large clusters;
merging the second noisy node into the second target large cluster, and marking the second noisy node as a visited state.
In some embodiments, the searching the vehicle path in the large cluster by the ant colony search algorithm according to the constraint condition and optimizing the searched path according to the objective function includes:
initializing a pareto optimal solution set to be empty and initializing pheromones on each path;
initializing each solution in the temporary pareto optimal solution set by a greedy algorithm;
selecting a next access node of an ant according to a next node selection algorithm, adding the node accessed by the ant to the solution, and updating the current vehicle load and the local pheromone concentration;
determining that all nodes in the current large cluster are added into the solution, and traversing the next large cluster;
Determining that the traversal of all the large clusters is completed, and updating the temporary pareto optimal solution set according to the quality of the solution;
determining a plurality of ants to obtain a final temporary pareto optimal solution set, and updating the pareto optimal solution set according to the quality of all solutions in the final temporary pareto optimal solution set;
determining that the pareto optimal solution set does not change in preset iteration, and obtaining a required pareto optimal solution set;
removing a cross path and a repeat path of a second target optimal solution in the required pareto optimal solution set;
updating the global pheromone concentration;
and determining that the iteration times are larger than the preset iteration times to obtain a final pareto optimal solution set.
In another aspect, an embodiment of the present invention provides a system for optimizing a logistics vehicle path based on workload balancing, including:
the system comprises a determining module, a judging module and a judging module, wherein the determining module is used for determining a logistics scene, and the logistics scene comprises a first preset logistics scene and a second preset logistics scene;
the construction module is used for constructing an objective function and a constraint condition of the objective function according to the logistics scene;
the clustering module is used for determining the logistics scene as the first preset logistics scene and dividing the clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to the vehicle capacity;
A fusion module for determining the logistics scene as the first preset logistics scene or the second preset logistics scene, and fusing a plurality of small clusters into a plurality of large clusters according to a preset rule
And the path generating module is used for searching the vehicle path in the large cluster through an ant colony search algorithm according to the constraint condition and optimizing the searched path according to the target function.
In another aspect, an embodiment of the present invention provides a system for optimizing a logistics vehicle path based on workload balancing, including:
at least one memory for storing a program;
at least one processor for loading the program to execute the method for optimizing the logistics vehicle path based on the workload balance.
In another aspect, an embodiment of the present invention provides a storage medium, in which a computer-executable program is stored, and the computer-executable program is executed by a processor to implement the method for optimizing a logistic vehicle path based on workload balancing.
The logistics vehicle path optimization method based on the workload balance has the following beneficial effects that:
according to the method, an objective function and constraint conditions of the objective function are constructed through a logistics scene, the logistics scene is determined to be the first preset logistics scene, clients in the logistics scene are divided into a plurality of small clusters based on a preset density clustering algorithm according to vehicle capacity, and the small clusters are fused into a plurality of large clusters according to a preset rule; and determining the logistics scene as the second preset logistics scene, fusing a plurality of small clusters into a plurality of large clusters according to preset rules, then performing vehicle path search in the large clusters through an ant colony search algorithm according to the constraint conditions, and optimizing the searched paths according to the objective function, so that the working efficiency of staff in logistics distribution according to the optimized paths is improved, and the logistics transportation cost is reduced.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The invention is further described with reference to the following figures and examples, in which:
FIG. 1 is a flow chart of a method for optimizing a logistics vehicle path based on workload balancing according to an embodiment of the present invention;
fig. 2 is a schematic diagram illustrating generation of a small cluster in a first preset logistics scene according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a large cluster fusion according to an embodiment of the present invention;
fig. 4 is a schematic diagram of a path optimization process according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention and are not to be construed as limiting the present invention.
In the description of the present invention, it should be understood that the orientation or positional relationship referred to in the description of the orientation, such as the upper, lower, front, rear, left, right, etc., is based on the orientation or positional relationship shown in the drawings, and is only for convenience of description and simplification of description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus, should not be construed as limiting the present invention.
In the description of the present invention, the meaning of a plurality is one or more, the meaning of a plurality is two or more, and larger, smaller, larger, etc. are understood as excluding the present numbers, and larger, smaller, inner, etc. are understood as including the present numbers. If there is a description of first and second for the purpose of distinguishing technical features only, this is not to be understood as indicating or implying a relative importance or implicitly indicating the number of technical features indicated or implicitly indicating the precedence of technical features indicated.
In the description of the present invention, unless otherwise specifically limited, terms such as set, installation, connection and the like should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above terms in the present invention by combining the specific contents of the technical solutions.
In the description of the present invention, reference to the description of "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples", etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Before specifically describing embodiments of the present invention, the following explanations are made for the terms used in the present embodiment:
optimizing the vehicle path: the system is characterized in that a certain number of customers respectively have different goods demands, a distribution center provides goods for the customers, a fleet is responsible for distributing the goods and organizing a proper driving route, and the aim is to meet the demands of the customers and achieve the aims of shortest route, minimum cost, minimum time consumption and the like under certain constraints.
The density is direct: if P is the core point and Q is in the Eps neighborhood of P, then the P-to-Q density is said to be through. Any core point is density-neutral to itself, which is not symmetrical, and if P is density-neutral to Q, then Q is not necessarily density-neutral.
The density can be achieved: if core points P2, P3, … …, Pn exist, and P1 through P2 density is direct, P2 through P3 density is direct, … …, P (n-1) through Pn density is direct, and Pn through Q density is direct, then P1 through Q density is direct. The density may be asymmetrical.
The logistics vehicle path optimization is beneficial to shortening the overtime time of the driver, so that the service quality of the driver is improved. The logistics chain comprises three links in total: the first is the movement of goods from retailers to large warehouses or distribution centers, which is typically a long distance transport across regions; the second is the delivery from a large warehouse or distribution center to a community distribution point; the third is the delivery to the end customer. The first link is transported by using airplanes, ships and railway trunk lines, and the route is fixed, so that the current path optimization method only carries out independent optimization on the second link or the third link, but the current path optimization method lacks a technical means which can simultaneously optimize the two links and balance the workload, and therefore the working efficiency of workers cannot be effectively improved, and the transportation cost is reduced.
Based on this, referring to fig. 1, an embodiment of the present invention provides a method for optimizing a logistics vehicle path based on workload balance, where the method of the embodiment includes, but is not limited to, step S110 to step S150:
s110, determining a logistics scene, wherein the logistics scene comprises a first preset logistics scene and a second preset logistics scene;
s120, constructing an objective function and a constraint condition of the objective function according to the logistics scene;
s130, determining that the logistics scene is the first preset logistics scene, dividing clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to vehicle capacity, and fusing the small clusters into a plurality of large clusters according to a preset rule; the preset density clustering algorithm is an algorithm which is improved based on the existing density clustering algorithm and is applicable to the embodiment.
S140, determining the logistics scene as the second preset logistics scene, and fusing a plurality of small clusters into a plurality of large clusters according to a preset rule;
s150, vehicle path searching is carried out in the large cluster through an ant colony search algorithm according to the constraint conditions, and the searched path is optimized according to the objective function.
In the embodiment of the application, a first preset logistics scene can be set to enable a vehicle to start from a distribution center and distribute goods to community distribution points; in this process, the community distribution point can be viewed as a customer node. The second predetermined logistics scenario can be configured such that the vehicle is dispatched to the customer from the community dispatch point.
When the logistics scene is a first preset logistics scene, the objective function comprises a minimized travel cost and a ratio of a minimized vehicle maximum workload to a minimized vehicle workload; the constraints include that the node obeys the flow conservation constraints, that each customer is visited only once, that at least one customer is served per vehicle, that each customer is served by only one vehicle, that all vehicles must leave the same warehouse, and that the vehicles are not overloaded. The vehicle workload is the sum of the demand of the node accessed by the vehicle and the vehicle travel distance. The cargo demand of a node is abstracted to the number of basic transport units.
In a first preset logistics scene, the step of dividing the clients in the logistics scene into a plurality of small clusters according to the vehicle capacity and based on a preset density clustering algorithm can be realized by the following steps:
Calculating the radius and the first density of a first neighbor, and initializing an initial core point state and a node state;
determining a first clustering center point meeting a first preset requirement according to the first neighbor radius and the first density;
obtaining client nodes which can reach the first clustering center and meet the vehicle-mounted capacity to form a small cluster;
determining a second clustering center point meeting a second preset requirement;
obtaining client nodes which can reach the second cluster center and meet the vehicle-mounted capacity to form other small clusters;
and determining that the number of the small clusters is equal to a first preset number, and stopping the small cluster forming process.
It is understood that, as shown in fig. 2, the present embodiment is illustrated by the following steps:
step 210, calculating a first neighbor radius Eps and a first density MinPts, and initializing an initial core point state and a node state;
step 220, selecting the client node which is farthest from the warehouse node and meets the first neighbor radius Eps and the first density MinPts as a first small cluster mCr1Cluster center point X of1(ii) a That is, the step is to find out the first clustering center point X meeting the first preset requirement1；
Step 230, determine all secondary cluster center points X1The density of the accessible client nodes forms a small cluster of mCr within the vehicle-mounted capacity limit 1；
Step 240, selecting the non-clustered node which has the maximum minimum distance median to each currently known cluster center point and meets the first neighbor radius Eps and the first density MinPts as the nth (n is more than or equal to 2) small cluster mCrnCluster center point X of (2)n(ii) a The step determines a second cluster center point X meeting a second preset requirementn；
Step 250, determining all secondary cluster center points XnCustomer with reachable densityNodes, and within the vehicle capacity limit, form a small cluster Xn；
Step 260, repeating step 240 and step 250 until m small clusters are generated; namely, the step is to determine that the number of the small clusters is equal to a first preset number, and stop the small cluster forming process.
In the embodiment of the present application, after the steps shown in fig. 2 are performed, the embodiment further needs to perform noise node processing. Specifically, the present embodiment may perform noise node processing by:
calculating the average distance from a first noise node to each small cluster, wherein the first noise node comprises a client node which is not accessed in the small cluster forming process; in this step, the average distance is an average value of the sum of distances from the noise node to all client nodes in the designated small cluster, and after the average distances of all small clusters are obtained, all the obtained average distances may be combined into a two-dimensional array.
Acquiring a small cluster and a first noise node corresponding to the average distance smaller than a preset distance as a first target small cluster and a target node respectively;
and determining that the sum of the number of the target nodes and the number of the nodes in the first target small cluster is less than or equal to a second preset number, adding the target nodes to the first target small cluster, and marking the target nodes as an accessed state. In this step, the second preset number may be adjusted according to an actual situation.
The noise node processing of the embodiment can cluster the client nodes which are not clustered, thereby reducing the probability of occurrence of missing situations of the client nodes.
When the logistics scene is the second preset logistics scene, the client nodes are already grouped into a plurality of small clusters in advance, for example, the client nodes are grouped in advance based on the geographic location. The objective function of the present embodiment includes minimizing the cost of travel and minimizing the ratio of the vehicle maximum workload to the vehicle minimum workload; the constraint conditions include that the nodes obey the stream conservation constraint, each client can be accessed only once, each vehicle can serve at least one client, each client can be served only by one vehicle, all vehicles have to leave the same warehouse, the vehicles are not overloaded, the vehicles are operated for multiple times to enter and leave small clusters, the small clusters are disjoint, the vehicles have to leave the small clusters after entering the small clusters, and each vehicle has to return to the warehouse nodes after leaving the small clusters.
After the small clusters of the first preset logistics scene or the second preset logistics scene are obtained, the small clusters can be fused, and therefore the accuracy of the path optimization result is improved. Specifically, the fusing the small clusters into the large clusters according to the preset rule, as shown in fig. 3, includes, but is not limited to, the following steps:
step 310, setting the states of all the small clusters as available states; also, in this implementation, the state may indicate that these small clusters may be fused into a large cluster;
step 320, determining that a second target small cluster is not fused and the second target small cluster has an adjacent small cluster or an adjacent large cluster, and fusing the second target small cluster and a target cluster meeting a third preset requirement; the target cluster meeting the third preset requirement is a small cluster or a large cluster which has the maximum adjacency value and the number of small clusters in the fused large cluster is less than or equal to the third preset number; the second target small cluster is one of all the small clusters. In the first traversal process, the sequence number n of the second target small cluster is 1;
step 330, determining that the second target small cluster is not fused, the second target small cluster has an adjacent small cluster or an adjacent large cluster, the quantity of required nodes of the second target small cluster is less than the quantity of required nodes of a preset small cluster, and fusing the second target small cluster and a target cluster meeting a fourth preset requirement; the target cluster meeting the fourth preset requirement is a nearby small cluster or a nearby large cluster which is closest to the second target small cluster and the number of the small clusters in the fused large cluster is less than or equal to the third preset number; the preset small cluster demand of the implementation can be set to be 50% of the demand of all small clusters;
Step 340, determining that the second target small cluster is fused into a first target large cluster, and the first target large cluster has an adjacent small cluster or an adjacent large cluster, and fusing the first target large cluster and a target cluster meeting a fifth preset requirement; the target cluster meeting the fifth preset requirement is a large cluster or a small cluster, and the number of small clusters in the fused large cluster is less than or equal to a third preset number.
Step 350, after the fusion is executed, if the number of the small clusters in the fused large cluster is larger than a third preset number, setting the states of all the small clusters in the large cluster as unavailable states;
and step 360, adding 1 to the serial number n of the second target small cluster, and then repeating the steps 320 to 350 until all the small cluster states are checked. If there are also small clusters that are not fused into any large cluster at this time, the small cluster becomes a large cluster directly.
In this embodiment, the adjacent large clusters or adjacent small clusters can be understood as follows:
two clusters are said to be contiguous small or large clusters if two nodes belonging to different small or large clusters are directly reachable (node Q is within the Eps of node P, then P is said to be directly reachable to Q).
Adjacency value A of the two adjacent clustersijAs shown in equation (1):
Aij＝Ki×KjFormula (1)
Wherein, KiIs the number of nodes which belong to the large cluster i or the small cluster i and can be reached directly from the large cluster i or the small cluster i, KjIs the number of nodes that belong to either the large cluster j or the small cluster j and that can be reached directly from either the large cluster j or the small cluster j.
The nearby large clusters or nearby small clusters can be understood as follows:
if there are nodes in IcoreAnd if the center is a circle with a radius of 3L, the small cluster or the large cluster containing the node is called as a nearby large cluster or a nearby small cluster of the current small cluster. I of the present examplecoreAs shown in equation (2), and L as shown in equation (3):
Figure BDA0003555601540000091
Figure BDA0003555601540000092
where i is the current small cluster mCrkNode n of (1)kIs mCrkThe number of nodes of the array MCDist records a small cluster mCrkNode in to IcoreThe distance of (c).
In the embodiment of the present application, after completing the fusion of small clusters into a large cluster, the small noise cluster needs to be processed in the following manner. Specifically, the small noise cluster of the present embodiment can be regarded as a noise node. The specific treatment process includes but is not limited to the following steps:
calculating the average distance from a second noise node to each large cluster, wherein the second noise node comprises the remaining client nodes which are not accessed in the small cluster forming process; the un-accessed residual client nodes in the step are the client nodes which are not accessed after the noise nodes after the small cluster fusion are processed;
Acquiring a large cluster meeting a sixth preset requirement as a second target large cluster; the sixth preset required large cluster is a large cluster which does not access the second noise node within a preset time period and the total demand of the second noise node and one of the large clusters is less than or equal to the upper limit of the total demand of the one of the large clusters; wherein the upper limit of the total demand quantity of the large clusters is the product of the number of the small clusters contained in the large clusters and the vehicle load quantity;
merging the second noisy node into the second target large cluster, and marking the second noisy node as a visited state.
In the embodiment, all the non-accessed client nodes are subjected to noise processing again, so that all the non-accessed nodes are absorbed into a large cluster, and the accuracy of the subsequent path optimization result is improved.
In the embodiment of the application, after all the small clusters are fused into the large cluster, the optimization processing of the logistics path can be performed based on the large cluster. Specifically, as shown in fig. 4, the path optimization process may include, but is not limited to, the following steps:
step 410, initializing a pareto optimal solution set P to be empty, and initializing pheromones on each path;
step 420 of initializing a temporary pareto optimal solution set S by a greedy algorithm PEach solution in (a);
step 430, for each ant a, setting the ant a to start from a warehouse, then selecting the next access node of the ant according to the next node selection algorithm, adding the node accessed by the ant to the solution, and updating the current vehicle load and the local pheromone concentration;
440, repeating the step 430 until the next node selected by the ant to access is the warehouse, and ending the access in the round;
step 450, if the number of access wheels is less than the number of vehicles, the number of access wheels ri+1, ant starts the next round of access;
step 460, repeating the steps 430 to 450, determining that all nodes in the current large cluster are added into the solution, and traversing the next large cluster; determining the traversal completion of all the large clusters, and updating the temporary pareto optimal solution set S according to the quality of the solutionP；
Step 470, determining that all ants have performed steps 430 to 460, that is, all ants obtain the final temporary pareto optimal solution set, and updating the pareto optimal solution set P according to the quality of all solutions in the final temporary pareto optimal solution set;
step 480, determining that the pareto optimal solution set does not change in the preset R0 iteration, and jumping out of the iteration loop to obtain a required pareto optimal solution set; removing a cross path and a repeat path of a second target optimal solution in the required pareto optimal solution set; and updating the global pheromone concentration;
And 490, regarding the steps 430 to 480 as one iteration, repeating the iteration, and ending the loop if the iteration number is determined to be greater than a preset iteration number R to obtain a final pareto optimal solution set.
In this embodiment, the next node selection algorithm includes the following contents:
calculating the current vehicle workload and the average vehicle workload of the current large cluster in the pareto optimal solution set P, and calculating the current workload degree according to the two workloads;
if the current vehicle load reaches the upper capacity limit, directly returning to the warehouse and exiting the program;
if the workload degree is greater than a given value and less than or equal to 1, comparing a random number from 0 to 1 with a certain value, and if the workload degree is less than the given value, returning to the warehouse and exiting the program; if the workload degree is more than 1, comparing the random number from 0 to 1 with another value, if the workload degree is less than 1, returning to the warehouse and exiting the program;
if the workload degree is more than a given value, comparing a random number from 0 to 1 with a certain value, if the workload degree is less than or equal to the given value, selecting nodes with the maximum visibility in the current large cluster and warehouse, and if the workload degree is more than the given value, selecting nodes in the current large cluster and warehouse according to a roulette method;
If the workload level is not greater than this given value, then nodes in the current large cluster must be selected according to roulette, i.e., no warehouse nodes must be selected.
For example, taking delivery data of a certain delivery company from a certain delivery center for one month as a first preset logistics scene as an example, the logistics vehicle path optimization method based on workload balance provided by the embodiment of the application is applied to the explanation of an actual process. Specifically, the distribution center is responsible for the transportation of goods at 57 community distribution points in the area, and in this embodiment, data of a certain day in a month is arbitrarily selected, where the goods are transported from the distribution center to all 57 distribution points, that is, N is a set including all nodes of the warehouse, N is {0,1, 2., N } N is 57, d is 57, andiis the quantity of goods to be delivered of service Point i, demand for short, d0The average demand per service point is 540 units, 0. There are 5 trucks responsible for the transport, i.e. V is the collection of vehicles, V ═ 1,2,3,4, 5. Load of each vehicle Cv72000 units, v 1.
When the data of the express company is processed in this embodiment, the method includes, but is not limited to, the following steps:
step M1, constructing a vehicle path optimization model considering workload balance, wherein the vehicle path optimization model comprises an objective function and constraint conditions;
The objective function is shown in equation (4):
Figure BDA0003555601540000111
the constraint conditions are as shown in equations (5) to (12):
Figure BDA0003555601540000112
Figure BDA0003555601540000121
Figure BDA0003555601540000122
Figure BDA0003555601540000123
Figure BDA0003555601540000124
Figure BDA0003555601540000125
Figure BDA0003555601540000126
Figure BDA0003555601540000127
wherein b isvIs the amount of work of the vehicle v,
Figure BDA0003555601540000128
cijis the cost of travel from customer i to customer j,
Figure BDA0003555601540000129
representing a binary variable that is equal to 1 when the vehicle v can go directly from client i to client j, otherwise equal to 0,
Figure BDA00035556015400001210
is a binary variable, i is equal to 1 when the vehicle v serves the customer i, and d is equal to 0 otherwiseiIs a requirement of a customer i, CvIs the available load of the vehicle v,
Figure BDA00035556015400001211
is the remaining available load of the vehicle v before serving the customer i. Where α is 1 and β is 3.5.
Step M2, divide the client node N ═ {0,1, 2., N }, N ═ 57 into small clusters and process noisy nodes.
Wherein the small cluster forming process comprises the following steps:
step M2.1, executing a preset density clustering algorithm:
step M2.1.1, comprising:
the density MinPts was calculated using equation (12):
Figure BDA00035556015400001212
the neighbor radius Eps is calculated using equation (13):
Figure BDA00035556015400001213
wherein n is the number of the client nodes, Total is the sum of values in Dist _ i [ MinPts ], and Dist _ i records a one-dimensional array of the distances from the node i to other nodes and arranged from small to large according to the values.
Initializing an initial core point: if a client node has MinPts un-visited client nodes in the radius of Eps, the node is the starting core point.
Step M2.1.2, including:
selecting the client node which is farthest away from the warehouse node and meets the parameters Eps and MinPts as the 1 st small cluster mCr1Cluster center X of (2)1(ii) a All nodes are initialized to a state without access; the number of small clusters Cnum ═ 1; k is 1;
step M2.1.3, comprising:
Ndirectto be at node XkSet of nodes in the Eps range, for NdirectIf the node X is a core point that has not been accessed and the current small cluster demand is present
Figure BDA0003555601540000131
Adding the node set with the distance from the node X as the Eps into N if the sum of the node requirements in the Eps of the node X is less than the load of the corresponding vehicledirect. If small cluster demand
Figure BDA0003555601540000132
And node X demand dXIf the sum is less than the load of the corresponding vehicle, adding the node X into the small cluster mCrkMedium and small cluster demand
Figure BDA0003555601540000133
And sets X as accessed.
Wherein the core points are: if a client node has MinPts client nodes in the radius of Eps, the node is a core node.
Step M2.1.4, comprising:
the number of small clusters Cnum plus 1.
Step M2.1.5, comprising:
if Cnum is more than 1 and less than or equal to m, calculating the distance from all nodes which are not accessed to the clustering starting point of the current known small cluster, and storing the distance as a one-dimensional array D; and if the initial core point exists in all the nodes which are not visited currently, selecting the node with the maximum minimum value in each array D as the clustering starting point of the next cluster. Otherwise, recalculating the clustering start point of the next cluster from the nodes which are not accessed and meet the core point requirement, and if all the nodes which are not accessed do not meet the core point requirement, selecting the node with the maximum minimum value in each array D as the clustering start point of the next cluster.
Step M2.1.6, including:
steps M2.1.3, M2.1.4, M2.1.5 are repeated until the number of clusters Cnum ═ m small clusters are generated.
Step M2.2, processing noise nodes:
step M2.2.1, inputting the MicroCluster ═ mCr1,mCr2,...,mCrkM and a series of unaccessed nodes, and compute an unaccessed node X to each small cluster mCrkAverage distance of inner nodes
Figure BDA0003555601540000134
And stores these distance values in an array DCXIn (1). DC (direct current)XThe minimum value of (1) is denoted as IfMax.
Step M2.2.2, gather MicroCluster ═ { mCr from the small cluster1,mCr2,...,mCrkSelecting a small cluster mCr which has the minimum average distance from the node not accessed and meets a certain conditionkAdding node X to the small cluster mCrkMedium and small cluster demand
Figure BDA0003555601540000141
And sets the state of the node X as visited.
Wherein the conditions of step M2.2 are shown in equation (14) and equation (15):
Figure BDA0003555601540000142
Figure BDA0003555601540000143
wherein, γ1CkIs mCrkThe upper limit of the load. Gamma ray1And gamma2Is a parameter. Gamma ray1And gamma2It can be adaptively adjusted based on the load of the small clusters and the distribution of the noise nodes. Gamma ray1The definition is shown in formula (16):
Figure BDA0003555601540000144
wherein N isvisitedRepresenting a set of nodes that have been clustered.
γ2The definition is shown as formula (17):
Figure BDA0003555601540000145
wherein
Figure BDA0003555601540000146
Representing the minimum average distance between noise point i and the node in the small cluster.
Figure BDA0003555601540000147
Is shown except that
Figure BDA0003555601540000148
In addition, the minimum average distance between noise point i and a node in a small cluster. N is a radical of hydrogennoiseRepresenting a set of noise nodes.
Step M2.2.3, repeat step M2.2.1 and step M2.2.2 until all the unvisited nodes have been checked once.
Step M3, obtaining a small cluster microcouster ═ { mCr) based on step M21,mCr2,...,mCrkMerge into large clusters and process noisy nodes.
Step M3.1, changing all small clusters MicroCluster to { mCr ═ mCr1,mCr2,...,mCrkThe status of is set to available.
Step M3.2, checking the small cluster mCrkThe state of (2): first point, if the small cluster mCrkAvailable and not fused, and checking whether it is available or notAdjacent small/large clusters. If there is a small cluster and the small cluster is fused with one of the clusters with the maximum adjacency value, the number of the small clusters in the cluster does not reach the upper limit MCr,
Figure BDA0003555601540000149
The fusion is performed. Otherwise, if the small cluster mCrkIs less than 50% of the average demand of all the small clusters, and the small cluster mCrkAfter fusing with the nearby small/large cluster closest to it, the number of small clusters in the cluster does not reach the upper limit MCrThen fusion is performed. Second point, if the small cluster mCrkIf it is available but has already merged into a certain large cluster, it is checked whether there is a neighboring large/small cluster for the large cluster. The number of small clusters in the cluster, if any, after fusing the large cluster with the neighboring cluster having the largest adjacency value does not reach the upper limit M CrThen fusion is performed.
Step M3.3, after the fusion is executed, if the number of the small clusters in the fused large cluster reaches the upper limit MCrThe status of all small clusters within the large cluster is set to unavailable.
And step M3.4, repeating the step M3.2 and the step M3.3 until all the small cluster states are checked. If there are still small clusters that are not fused into any large cluster at this time, the small cluster becomes a large cluster directly.
Step M3.5, processing noise nodes: the input is a large Cluster ═ Cr1,Cr2,...,CrlL ≦ m and a small number of nodes that have not been accessed.
Wherein step M3.5 comprises:
step M3.5.1, for each unaccessed node, calculating the node and each large cluster CrkAverage distance of middle nodes
Figure BDA0003555601540000151
And records it in an array DCXIn (1).
Step M3.5.2, select the nearest node X to be visited and the node X plus the total amount of demand in the cluster
Figure BDA0003555601540000152
Not exceeding the upper limit of Vnumk×CrkLarge cluster of CrkTo absorb the node X. Wherein
Figure BDA0003555601540000153
Is the demand of node X and a large cluster of CrkSum of the amount of load of, VnumkIs CrkThe number of small clusters contained in (a).
Step M3.5.3, the demand for the big cluster is
Figure BDA0003555601540000154
Node X is marked as visited.
Step M4, based on large Cluster set Cluster ═ { Cr }1,Cr2,...,CrlAnd l is less than or equal to m, and searching an optimal path meeting the workload balance for each vehicle. Given with n aOnly ants search for paths and the number of iterations is R.
Step M4.1, initializing the pareto optimal solution set P to be null, and initializing pheromones on each path to be 1/KP。
Step M4.2 of initializing a temporary pareto optimal solution set S by means of a greedy algorithmPEach solution of Sa。
Step M4.3, for each ant a, add depot0 to solution SaThe number of rounds ri is 0, and the requirement of the ant a on the round ri
Figure BDA0003555601540000155
Is 0.
M4.4, ant a selects the next accessed node Point according to the algorithm of selecting the next nodejPoint ofjAdding to solution SaIn the ri th round, the requirement of the ant a is equal to
Figure BDA0003555601540000156
Wherein pointjIs node PointjThe requirements of (a). The local pheromone concentration is simultaneously updated by equation (18):
pheij＝VP×pheij+(1-VP) (18)
wherein, VPIs 0.8.
Step M4.5, if the next access node PointjFor warehouse, the number of rounds ri is increased by 1.
Step M4.6, repeating steps M4.4 and M4.5 until all nodes in the current large cluster are added into the solution SaIn (1).
Step M4.7, repeating step M4.6, when ants traverse all the large clusters, updating the rule according to the pareto optimal solution set, and according to the solution SaQuality update temporary pareto solution set SP；
Step M4.8, repeating steps M4.3 to M4.7, when n isaObtaining final temporary pareto solution set S by ants only PAccording to the pareto optimal solution set updating rule and the temporary pareto solution set SPUpdating the pareto optimal solution set P by the quality of all the solutions;
step M4.9, if the pareto optimal solution set P has no change in R0 iterations, jumping out of an iteration loop to obtain a required pareto optimal solution set P;
m4.10, removing a cross path and a repeat path of the second target optimal solution in the Patot optimal solution set P;
step M4.11, update the global pheromone concentration by equation (19):
Figure BDA0003555601540000161
wherein, VPAnd KPIs a parameter, lengthpIs the average path length of the pareto optimal solution set P. In addition, each edge has an upper limit Maxphe and a lower limit Minphe, where the upper limit Maxphe is calculated by formula (20) and the lower limit Minphe is calculated by formula (21):
Figure BDA0003555601540000162
Figure BDA0003555601540000163
and step M4.12, regarding the step M4.2 to the step M4.11 as one iteration, repeating the iteration, and ending the loop when the upper limit R of the iteration times is reached to obtain the final pareto optimal solution set P.
Specifically, the algorithm for selecting the next node is as follows:
initializing a workload indicator by equation (22)
Figure BDA0003555601540000171
Workload of ant a on the ri wheel
Figure BDA0003555601540000172
It can be calculated by the formula (23) that all vehicles are in the large cluster CriAverage workload in
Figure BDA0003555601540000173
Can be calculated by equation (24):
Figure BDA0003555601540000174
Figure BDA0003555601540000175
Figure BDA0003555601540000176
Wherein, ClulenkExpressed in the large cluster CrkThe total length of the path of all vehicles in the vehicle.
If the ant a's demand on the ri' th wheel equals the total load of the vehicle, it returns to the warehouse.
If it is used
Figure BDA0003555601540000177
Greater than a given value, e equal to 0.7, andif less than 1, a Random number Random is generated1If it is determined that
Figure BDA0003555601540000178
Then return to the warehouse. Otherwise, if
Figure BDA0003555601540000179
If greater than 1, a Random number Random is generated2If, if
Figure BDA00035556015400001710
Then return to the warehouse, where ε1＝0.3，ε2＝0.7。
If it is not
Figure BDA00035556015400001711
When the value is larger than zeta, zeta is 0.85, a Random number Random is generated3If Random is detected3≤ζ1，ζ10, then in the current large cluster CrkAnd selecting Tp with maximum visibility in warehouseijUsing equation (25) to calculate Tpij：
Figure BDA00035556015400001712
Where K1, K2 and K3 are constant parameters, where K1 equals 100, K2 equals 2, and K3 equals 5. If Random3>ζ1Then according to roulette method, the current large cluster of CrkAnd selecting nodes in the warehouse;
if it is not
Figure BDA00035556015400001713
Less than ζ, ζ equal to 0.85, then the current large cluster of Cr is played according to roulettekNode, but not the warehouse node.
Specifically, for the first preset logistics scenario, specific information when the method of the present embodiment is used to perform path optimization on 5 vehicles is shown in table 1:
TABLE 1
Vehicle with a steering wheel Distance traveled Total amount of demand for transportation Total work volume
v1 30212.40 5976.00 51128.40
v2 29707.30 6048.00 50875.30
v3 28003.70 6480.00 50683.70
v4 32197.90 5400.00 51097.90
v5 26994.60 6876.00 51060.60
The total travel distance of 5 vehicles was 147116 meters. The float between vehicle workloads was within a small range [50683.7, 51128.4], and a ratio of vehicle max to min workload of 1.00877 indicated that the workloads were substantially equal for each vehicle.
Taking the delivery data of one month from one express service point of a certain express company as a second preset logistics scene as an example, the logistics vehicle path optimization method based on the workload balance provided by the embodiment of the application is applied to the explanation of the actual process. Specifically, this embodiment arbitrarily selects data of a day in a month, where goods are transported from a service point to 67 customers distributed in five areas in the community, that is, N is a set including all nodes of the warehouse, N ═ 67, d ═ 0,1,2iIs to meet the needs of customer i, d00, with an average demand of 10 units per customer. And 5 dollies are responsible for transportation, V is the set of dollies, and V is {1,2,3,4,5 }. Load of each vehicle Cv200 units, v 1, 5, and each vehicle is responsible for the transportation of one area, where α 1 and β 3.5.
When the present embodiment processes the data of the express service point, the method includes, but is not limited to, the following steps:
step N1, constructing a cluster-based workload balance vehicle path optimization model: wherein the vehicle path optimization model comprises an objective function and a constraint condition;
the objective function is shown in equation (26):
Figure BDA0003555601540000181
the constraints are as shown in equations (5) to (12) and equations (26) to (30):
Figure BDA0003555601540000182
Figure BDA0003555601540000183
Figure BDA0003555601540000184
Figure BDA0003555601540000191
Figure BDA0003555601540000192
Figure BDA0003555601540000193
Figure BDA0003555601540000194
Figure BDA0003555601540000195
Figure BDA0003555601540000196
Figure BDA0003555601540000197
Figure BDA0003555601540000198
Figure BDA0003555601540000199
Figure BDA00035556015400001910
Step N2, pre-dividing the client node N ═ 0,1, 2., N ═ 67 into 5 small clusters { mCr ═ according to the area where the client node N ═ 0,1, 2., N ═ 67 is located1,mCr2,...,mCr5Then according to step M3, the small clusters are fused into a large Cluster, Cluster ═ Cr1,Cr2,...,CrnH, l ≦ m and process noise nodes.
And step N3, searching an optimal path meeting the workload balance for each vehicle through an ant colony algorithm based on the large cluster set according to the step M4.
Specifically, for the second preset logistics scene, specific information obtained when the path optimization is performed by using the conventional ant colony algorithm and the method of the embodiment is shown in table 2:
TABLE 2
Figure BDA00035556015400001911
Figure BDA0003555601540000201
As can be seen from Table 2, the results obtained using the method of the present embodiment, in contrast to the experimental results obtained using the conventional ant colony algorithm to search for an optimized path in each given area, are for the vehicle v2And v4Is increased while the vehicle v1And v3The workload of (2) is reduced. Although the overall travel distance for the optimized frame is slightly more than that for the traditional ant colony algorithm, the vehicle workload float control in the optimized frame is in a small range [4294.93,4780.32 ]]In between. To achieve equilibrium, the method of the present embodiment schedules customers outside the given zone to leave vehicles v2，v3，v4To service, but the total additional service customers is only 14, 80% of the service time per vehicle is still in a given area, which reduces the number of additional service customers The cost of the courier needing additional knowledge of the new delivery area conditions is reduced.
The embodiment of the invention provides a logistics vehicle path optimization system based on workload balance, which comprises:
the system comprises a determining module, a judging module and a judging module, wherein the determining module is used for determining a logistics scene, and the logistics scene comprises a first preset logistics scene and a second preset logistics scene;
the construction module is used for constructing an objective function and a constraint condition of the objective function according to the logistics scene;
the clustering module is used for determining the logistics scene as the first preset logistics scene and dividing the clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to the capacity of the vehicles;
the fusion module is used for determining that the logistics scene is the first preset logistics scene or the second preset logistics scene and fusing a plurality of small clusters into a plurality of large clusters according to a preset rule;
and the path generating module is used for searching the vehicle path in the large cluster through an ant colony search algorithm according to the constraint condition and optimizing the searched path according to the target function.
The contents of the embodiment of the method of the invention are all applicable to the embodiment of the system, the functions specifically realized by the embodiment of the system are the same as those of the embodiment of the method, and the beneficial effects achieved by the embodiment of the system are also the same as those achieved by the method.
The embodiment of the invention provides a logistics vehicle path optimization system based on workload balance, which comprises:
at least one memory for storing a program;
at least one processor for loading the program to perform the method for optimizing a logistic vehicle path based on workload balancing as shown in fig. 1.
The contents of the embodiment of the method of the invention are all applicable to the embodiment of the system, the functions specifically realized by the embodiment of the system are the same as those of the embodiment of the method, and the beneficial effects achieved by the embodiment of the system are also the same as those achieved by the method.
An embodiment of the present invention provides a storage medium, in which a computer-executable program is stored, and the computer-executable program is executed by a processor to implement the method for optimizing a logistics vehicle path based on workload balancing shown in fig. 1.
Embodiments of the present invention also provide a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The computer instructions may be read by a processor of a computer device from a computer-readable storage medium, and the computer instructions executed by the processor cause the computer device to perform the method for logistics vehicle path optimization based on workload balancing as shown in fig. 1.
The embodiments of the present invention have been described in detail with reference to the drawings, but the present invention is not limited to the embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the gist of the present invention. Furthermore, the embodiments of the present invention and features of the embodiments may be combined with each other without conflict.

Claims (10)

1. A logistics vehicle path optimization method based on workload balance is characterized by comprising the following steps:
determining a logistics scene, wherein the logistics scene comprises a first preset logistics scene and a second preset logistics scene;
constructing an objective function and a constraint condition of the objective function according to the logistics scene;
determining the logistics scene as the first preset logistics scene, dividing clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to vehicle capacity, and fusing the small clusters into a plurality of large clusters according to a preset rule;
determining the logistics scene as the second preset logistics scene, and fusing a plurality of small clusters into a plurality of large clusters according to a preset rule;
and searching the vehicle path in the large cluster through an ant colony search algorithm according to the constraint condition, and optimizing the searched path according to the objective function.
2. The method for optimizing the logistics vehicle path based on workload balance as claimed in claim 1, wherein when the logistics scenario is a first preset logistics scenario;
the objective function includes minimizing travel costs and minimizing a ratio of vehicle maximum workload to vehicle minimum workload;
the constraints include that the node obeys the flow conservation constraint, each customer is only visited once, each vehicle serves at least one customer, each customer is only served by one vehicle, all vehicles must leave the same warehouse, and the vehicles are not overloaded.
3. The method for optimizing logistics vehicle path based on workload balancing according to claim 2, wherein the dividing the clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to vehicle capacity comprises:
calculating the radius and the first density of a first neighbor, and initializing an initial core point state and a node state;
determining a first clustering center point meeting a first preset requirement according to the first neighbor radius and the first density;
obtaining client nodes which can reach the first clustering center and meet the vehicle-mounted capacity to form a small cluster;
Determining a second clustering center point meeting a second preset requirement;
obtaining client nodes which can reach the second cluster center and meet the vehicle-mounted capacity to form other small clusters;
and determining that the number of the small clusters is equal to a first preset number, and stopping the small cluster forming process.
4. The method as claimed in claim 3, wherein after the customers in the logistics scene are divided into several small clusters based on the pre-set density clustering algorithm according to the vehicle capacity, the method further comprises the following steps:
calculating the average distance from a first noise node to each small cluster, wherein the first noise node comprises a client node which is not accessed in the small cluster forming process;
acquiring a small cluster and a first noise node corresponding to the average distance smaller than a preset distance as a first target small cluster and a target node respectively;
and determining that the sum of the number of the target nodes and the number of the nodes in the first target small cluster is less than or equal to a second preset number, adding the target nodes to the first target small cluster, and marking the target nodes as an accessed state.
5. The method for optimizing logistics vehicle path based on workload balancing according to claim 1, wherein said fusing the small clusters into big clusters according to a preset rule comprises:
Setting the states of all small clusters as usable states;
determining that a second small target cluster is not fused and the second small target cluster has an adjacent small cluster or an adjacent large cluster, and fusing the second small target cluster and a target cluster meeting a third preset requirement; the target cluster meeting the third preset requirement is a small cluster or a large cluster which has the maximum adjacency value and the number of small clusters in the fused large cluster is less than or equal to a third preset number; the second target small cluster is one of all the small clusters;
determining that the second target small cluster is not fused, the second target small cluster has an adjacent small cluster or an adjacent large cluster, the quantity of required nodes of the second target small cluster is smaller than the quantity of required nodes of a preset small cluster, and fusing the second target small cluster and a target cluster meeting a fourth preset requirement; the target cluster meeting the fourth preset requirement is a nearby small cluster or a nearby large cluster, the distance between the target cluster and the second target small cluster is the shortest, and the number of the small clusters in the fused large cluster is less than or equal to the third preset number;
determining that the second small target cluster is fused into a first large target cluster, wherein the first large target cluster has an adjacent small cluster or an adjacent large cluster, and fusing the first large target cluster with a target cluster meeting a fifth preset requirement; the target clusters meeting the fifth preset requirement are large clusters or small clusters, and the number of the small clusters in the fused large clusters is less than or equal to the third preset number.
6. The method for logistics vehicle path optimization based on workload balancing according to claim 5, wherein after said merging of the small clusters into a large cluster according to a preset rule, the method further comprises the following steps:
calculating the average distance from a second noise node to each large cluster, wherein the second noise node comprises the residual client nodes which are not accessed in the small cluster forming process;
acquiring a large cluster meeting a sixth preset requirement as a second target large cluster; the sixth preset required large cluster is a large cluster which does not access the second noise node within a preset time period and the total demand of the second noise node and one of the large clusters is less than or equal to the upper limit of the total demand of the one of the large clusters;
merging the second noisy node into the second target large cluster, and marking the second noisy node as a visited state.
7. The method for optimizing logistics vehicle path based on workload balance according to claim 1, wherein the searching vehicle paths in the large cluster through the ant colony search algorithm according to the constraint condition and optimizing the searched paths according to the objective function comprises:
Initializing a pareto optimal solution set to be empty and initializing pheromones on each path;
initializing each solution in the temporary pareto optimal solution set through a greedy algorithm;
selecting a next access node of an ant according to a next node selection algorithm, adding the node accessed by the ant to the solution, and updating the current vehicle load and the local pheromone concentration;
determining that all nodes in the current large cluster are added into the solution, and traversing the next large cluster;
determining that the traversal of all the large clusters is completed, and updating the temporary pareto optimal solution set according to the quality of the solution;
determining a plurality of ants to obtain a final temporary pareto optimal solution set, and updating the pareto optimal solution set according to the quality of all solutions in the final temporary pareto optimal solution set;
determining that the pareto optimal solution set does not change in preset iteration, and obtaining a required pareto optimal solution set;
removing a cross path and a repeat path of a second target optimal solution in the required pareto optimal solution set;
updating the global pheromone concentration;
and determining that the iteration times are larger than the preset iteration times to obtain a final pareto optimal solution set.
8. A system for optimizing a logistic vehicle path based on workload balancing, comprising:
The system comprises a determining module, a judging module and a judging module, wherein the determining module is used for determining a logistics scene, and the logistics scene comprises a first preset logistics scene and a second preset logistics scene;
the construction module is used for constructing an objective function and a constraint condition of the objective function according to the logistics scene;
the clustering module is used for determining the logistics scene as the first preset logistics scene and dividing the clients in the logistics scene into a plurality of small clusters based on a preset density clustering algorithm according to the capacity of the vehicles;
the fusion module is used for determining that the logistics scene is the first preset logistics scene or the second preset logistics scene and fusing a plurality of small clusters into a plurality of large clusters according to a preset rule;
and the path generating module is used for searching the vehicle path in the large cluster through an ant colony search algorithm according to the constraint condition and optimizing the searched path according to the target function.
9. A system for optimizing a logistic vehicle path based on workload balancing, comprising:
at least one memory for storing a program;
at least one processor configured to load the program to perform the method for workload balance based logistics vehicle path optimization of any of claims 1-7.
10. A storage medium having stored therein a computer-executable program which when executed by a processor is adapted to implement the method for logistics vehicle path optimization based upon workload balancing of any one of claims 1 to 7.
CN202210275223.3A 2022-03-21 2022-03-21 Logistics vehicle path optimization method, system and medium based on workload balance Active CN114757394B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210275223.3A CN114757394B (en) 2022-03-21 2022-03-21 Logistics vehicle path optimization method, system and medium based on workload balance

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210275223.3A CN114757394B (en) 2022-03-21 2022-03-21 Logistics vehicle path optimization method, system and medium based on workload balance

Publications (2)

Publication Number Publication Date
CN114757394A true CN114757394A (en) 2022-07-15
CN114757394B CN114757394B (en) 2024-08-13

Family

ID=82326552

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210275223.3A Active CN114757394B (en) 2022-03-21 2022-03-21 Logistics vehicle path optimization method, system and medium based on workload balance

Country Status (1)

Country Link
CN (1) CN114757394B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116645027A (en) * 2023-04-26 2023-08-25 中国民用航空飞行学院 ALNS frame-based distribution vehicle-unmanned aerial vehicle collaborative distribution path planning method

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103927643A (en) * 2014-04-30 2014-07-16 洪剑 Optimization method for large-scale order processing and distributing route
CN111222705A (en) * 2020-01-10 2020-06-02 聊城大学 Nonlinear charging vehicle path optimization method
CN111428991A (en) * 2020-03-20 2020-07-17 北京百度网讯科技有限公司 Method and device for determining delivery vehicles

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103927643A (en) * 2014-04-30 2014-07-16 洪剑 Optimization method for large-scale order processing and distributing route
CN111222705A (en) * 2020-01-10 2020-06-02 聊城大学 Nonlinear charging vehicle path optimization method
CN111428991A (en) * 2020-03-20 2020-07-17 北京百度网讯科技有限公司 Method and device for determining delivery vehicles

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
王旭坪;李新宇;张;: "考虑时空距离的异车型同时集送车辆路径优化", 管理学报, no. 06, 30 May 2018 (2018-05-30), pages 133 - 141 *
申艳光;张玲玉;刘永红;: "基于混合遗传算法的物流路径优化方法研究", 计算机技术与发展, no. 03, 4 December 2017 (2017-12-04), pages 198 - 202 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116645027A (en) * 2023-04-26 2023-08-25 中国民用航空飞行学院 ALNS frame-based distribution vehicle-unmanned aerial vehicle collaborative distribution path planning method
CN116645027B (en) * 2023-04-26 2024-01-30 中国民用航空飞行学院 ALNS frame-based distribution vehicle-unmanned aerial vehicle collaborative distribution path planning method

Also Published As

Publication number Publication date
CN114757394B (en) 2024-08-13

Similar Documents

Publication Publication Date Title
CN113393040B (en) Industrial park logistics scheduling method and system based on game theory
CN111860754B (en) AGV scheduling method based on ant colony and genetic algorithm
CN110782086B (en) Rescue vehicle distribution path optimization method and system with unmanned aerial vehicle
CN109345091B (en) Ant colony algorithm-based whole vehicle logistics scheduling method and device, storage medium and terminal
CN106355291B (en) Logistics path planning method based on store point group bisector
CN107798423A (en) Vehicle path planning Simulation Experimental Platform based on multi-intelligence algorithm
CN113359702B (en) Intelligent warehouse AGV operation optimization scheduling method based on water wave optimization-tabu search
CN107292440B (en) Container liner route optimization method based on BA (building block) scale-free network theory
CN114626794B (en) Internet of things-based warehousing vertical warehouse shuttle vehicle path planning method and system
CN113673922A (en) Fishbone type warehouse layout-based multi-vehicle picking path problem optimization method and system
CN111860957B (en) Multi-vehicle-type vehicle path planning method considering secondary distribution and balancing
CN115860613A (en) Part load and goods matching and vehicle scheduling method considering reservation mechanism
CN113848970A (en) Multi-target collaborative path planning method for vehicle and unmanned aerial vehicle
CN113780676A (en) Method for optimizing distribution path of bottled liquefied gas vehicle
CN114757394B (en) Logistics vehicle path optimization method, system and medium based on workload balance
CN115759917A (en) Logistics path planning method based on improved mixed ant colony algorithm
CN112418475A (en) Logistics path planning method and device, electronic equipment and storage medium
CN112613701A (en) Finished cigarette logistics scheduling method
CN114757591B (en) Multi-vehicle type collaborative sorting scheduling method based on behavior dependency graph
CN111507662B (en) Method for planning logistics vehicle path
CN111445100B (en) Vehicle and goods matching method based on self-adaptive time window under limited transport capacity
CN113570187A (en) Vehicle scheduling method and device, computer equipment and medium
CN111652550A (en) Method, system and equipment for intelligently searching optimal loop set
CN112837001A (en) Logistics network planning method and device and computer readable storage medium
Gam et al. Optimization of a two-stage logistic system using Hybrid Filtered Beam Search algorithm

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