CN117094629B - Dangerous goods full-load distribution path optimization and vehicle dispatching method - Google Patents

Dangerous goods full-load distribution path optimization and vehicle dispatching method Download PDF

Info

Publication number
CN117094629B
CN117094629B CN202311199824.1A CN202311199824A CN117094629B CN 117094629 B CN117094629 B CN 117094629B CN 202311199824 A CN202311199824 A CN 202311199824A CN 117094629 B CN117094629 B CN 117094629B
Authority
CN
China
Prior art keywords
vehicle
demand
distribution center
transportation
path
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
CN202311199824.1A
Other languages
Chinese (zh)
Other versions
CN117094629A (en
Inventor
柴获
何瑞春
韩桢铖
熊一辉
董凯凯
张会茹
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.)
Lanzhou Jiaotong University
Original Assignee
Lanzhou Jiaotong 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 Lanzhou Jiaotong University filed Critical Lanzhou Jiaotong University
Priority to CN202311199824.1A priority Critical patent/CN117094629B/en
Publication of CN117094629A publication Critical patent/CN117094629A/en
Application granted granted Critical
Publication of CN117094629B publication Critical patent/CN117094629B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G06Q10/0835—Relationships between shipper or supplier and carriers
    • G06Q10/08355—Routing methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • 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
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06315—Needs-based resource requirements planning or analysis
    • 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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention belongs to the technical field of distribution path planning, and discloses a dangerous goods full-load distribution path optimization and vehicle scheduling method, which comprises the steps of establishing a model of vehicle quantity, total vehicle running cost and total risk and distribution time, establishing a vehicle path optimization mathematical model P1, establishing a vehicle scheduling model P2, and carrying out two-stage method solution (comprising a path optimization solution method and a vehicle scheduling optimization solution method). The invention can minimize the transportation cost and transportation risk in the transportation process by optimizing the driving route and scientifically dispatching the vehicles, and simultaneously, the invention can minimize the number of the distributed vehicles and ensure the fairness of the transportation task quantity of the vehicles.

Description

Dangerous goods full-load distribution path optimization and vehicle dispatching method
Technical Field
The invention relates to the technical field of distribution path planning, in particular to a dangerous goods full-load distribution path optimization and vehicle scheduling method.
Background
The problem of full-load vehicle dispatch, also known as multiple transport dispatch, was proposed by Ball, bodin, golden et al in 1983, which requires multiple times of full-load execution of vehicles between the dispatch center and the point of demand. In the dangerous goods transportation process, if the demand of the task point is not less than the capacity of the transportation vehicle, more than one vehicle can execute each transportation task, and the vehicles need to run in full load to complete the distribution task. For example, the oil transportation problem of an automobile filling station is that the storage tank capacity of the filling station is usually 30-120m 3, the nuclear capacity of a common tank truck is 3-30m 3, and one filling station can meet the requirement only by fully transporting a plurality of tank trucks.
The optimization of the dangerous goods transportation path of the fully loaded vehicle is a multi-objective optimization problem, a plurality of non-dominant paths possibly exist between the distribution center and the demand points, the vehicles need to be selected to be transported according to the risk preference, after the transportation path is selected, fewer vehicles participate in the distribution task by dispatching the transportation vehicles, namely, each vehicle needs to bear as many transportation tasks as possible, namely, the idle time from the beginning of the whole transportation task to the completion of all tasks is minimum. Only if the driving route is optimized and vehicles are scientifically scheduled, the driving cost and the total risk of completing all transportation tasks can be minimized, meanwhile, the number of the distributed vehicles is minimized, the fairness of the transportation tasks of the vehicles is guaranteed, and then the technical problems to be solved are how to optimize the route and how to optimize the vehicle scheduling.
Disclosure of Invention
The invention aims to solve the technical problem in the prior art of dangerous goods full-load distribution, and provides a dangerous goods full-load distribution path optimization and vehicle dispatching method.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
The dangerous goods full-load distribution path optimizing and vehicle dispatching method is characterized by comprising the following steps of:
Step 1), establishing a model of the number of vehicles, the total running cost and total risk of the vehicles and the distribution time;
1.1 The demand quantity at the demand point d is q d, the nuclear capacity of each vehicle is g, q d is more than or equal to g, namely, one vehicle can only complete all or part of one task, and the number of vehicles required for completing the distribution task of the task point d is a d, if The total number of vehicles required to complete the distribution tasks of all the demand points is Σ dad;
1.2 Transportation cost of dangerous goods transported by transportation vehicles on road sections (i, j) is divided into two cases, and the full-load transportation cost is The empty driving cost is/>The transportation cost c 1 od for completing the arrival of the demand point d from the distribution center o is the sum of the transportation costs of the sections passed by, namely:
Variable(s) Indicating that the road segment (i, j) E is on the way from the delivery center o to the demand point d, otherwise
The route is the least cost route from the demand point d to the distribution center o when the empty car returns, and is usedThe delivery task of the demand point d is completed, and the transportation cost is/>
1.3 The risk of the transportation vehicle transporting dangerous goods on the road section (i, j) is r ij, and the total transportation risk of completing the delivery task from the delivery center o to the demand point d is as follows:
1.4 Setting the travel time of the vehicle on the road section (i, j) as The average loading time of the vehicle is delta t 1, the unloading time is delta t 2, and the running time of the vehicle from the distribution center o to the demand point d is
The travel time of the vehicle from the demand point d to the distribution center o is
The total time of returning the delivery tasks from the delivery center o to the completion demand point d to the delivery center is as follows:
Assume that the moment when the vehicle starts from the distribution center is The moment of finishing the return of the demand point d to the distribution center is
1.5 If the earliest time window for the vehicle to start from the distribution center and the latest time window for the vehicle to return to the distribution center is [ b d,ed ]
The moment when the vehicle reaches the demand point d must satisfy the following condition:
the moment when the vehicle starts from the distribution center o The following conditions must be met:
At the same time, it is also necessary to meet the time window of the distribution center [ b 0,e0 ], i.e.
Step 2), establishing a vehicle path optimization mathematical model P1;
The P1 model is as follows:
min f＝(f1,f2) (10)
s.t.
Wherein:
Equation (10) represents a minimized target vector composed of two target functions;
equation (11) is a transportation cost function expression;
formula (12) is a transport risk function expression;
The formula (13) ensures that a complete transportation path is formed between the distribution center o and the demand node d;
formula (14) is a decision variable;
N represents a set of N nodes, E represents a set of inter-node road segments;
Step 3), a vehicle scheduling model P2 is established, so that x ijk(xijk epsilon [0,1 ]) is whether the operation j is executed after the operation i is completed by the vehicle k, if yes, x ijk =1, otherwise x ijk＝0;yik(yik epsilon [0,1 ]) is whether the operation i is executed by the vehicle k, if yes, y ik =1, otherwise y ik =0; for job i, if i ε J d its time window is The service time is deltat 0d;
The P2 model is:
s.t.
Wherein: the expression (15) represents an objective function, In the function/>Representing the minimum number of vehicles, M being an integer, guaranteeing the priority of the vehicle number target,/>The standard deviation of running time of all vehicles is minimum under the condition of the same number of vehicles, so that the intensity of transportation tasks born by each vehicle is guaranteed to be fair, formulas (16) - (19) ensure that each vehicle can simultaneously carry out only one operation and sequentially execute the operation, each operation is completed by only one vehicle, formula (20) represents the moment when a vehicle k reaches a node j, and formula (21) ensures that the moment when the vehicle k reaches the node j must meet the time window constraint; equations (22) and (23) are time window requirements of earliest departure and latest return of the vehicle, and equations (24) and (25) are decision variables; k is a vehicle set, and C is a distribution node set;
step 4), carrying out two-stage solving;
4.1 P1 model, i.e. path optimization solving method: the first stage adopts a pulse algorithm to obtain Pareto solutions of all paths from a distribution center to each demand node; the second stage is to obtain a Pareto path between the distribution center and each demand node obtained in the first stage to encode, and solve by adopting NSGA-II algorithm;
4.2 P2 model, i.e. vehicle dispatch optimization solving method:
First, initializing requirement information;
secondly, adopting a pulse algorithm to calculate a Pareto path from the distribution center o to the demand node d;
thirdly, solving Pareto paths from the distribution center o to all demand nodes by adopting a multi-objective optimization method based on NSGA-II;
Fourthly, decision-making staff selects one path scheme according to the risk preference;
Fifthly, a VRP solving method based on UMDA is adopted to obtain a vehicle scheduling schedule, namely a vehicle scheduling scheme, under the path scheme.
Further, the paths from the first stage distribution center to each demand node in step 4.1) are respectively Representing the Pareto path set from the distribution center node o to the demand node d. 3. The method for optimizing the full-load distribution path of dangerous goods and dispatching the vehicles according to claim 1, wherein the method is characterized by comprising the following steps: in the step 4.1), the second stage NSGA-II algorithm is performed according to the coding mode, the fitness function and the population updating strategy.
Further, the coding mode is individual coding, natural number coding is adopted, the coding length is m, and the format is n 1,n2...nm, wherein: The i-th value in the code is n i, which indicates that the distribution center o to the demand node d i adopts the/> And (5) transporting along a path.
Further, fitness function represents fitness of individual indi using [ c_value, r_value ] =f (indi), where c_value and r_value are cost and risk of row transportation according to the path selection scheme in individual indi, respectively.
Further, in the population updating strategy, the population obtains new individuals through crossing, and the crossing operation adopts an integer crossing method; firstly, randomly selecting two individuals from a population, randomly generating two positions pos1 and pos2 (pos 1 is less than pos 2), and exchanging the positions pos1 to pos2 of the two individuals to generate two new individuals;
mutation operation using integer mutation method to obtain new individuals, randomly selecting one individual, randomly generating two positions pos1 and pos2 (pos 1 < pos 2), and combining the positions pos1 to pos2 of the selected individuals one by one And performing difference operation and taking an absolute value to generate a new individual.
Compared with the prior art, the invention has the following beneficial effects:
according to the invention, the vehicle driving route from the distribution center to each demand point and returning to the distribution center after distribution is completed can be obtained by solving the model P1, and after the vehicle driving route is selected according to the risk preference, a decision maker still has to solve the problem how to determine the vehicle dispatching plan. In the case of a limited number of dispatchable vehicles, the number of dispatches is an optimization objective that needs to be considered.
When the decision maker selects a certain path scheme according to the risk preference, in order to achieve the purpose of fewer vehicles to participate in the delivery task, each vehicle is required to bear as many transportation tasks as possible, that is, the idle time in the process of completing all tasks is minimum at the beginning of the whole transportation task. The invention can convert the vehicle dispatching problem into the VRP problem with the time window, and can obtain the vehicle dispatching optimization scheme by solving the P2 model.
According to the invention, by arranging the vehicle transportation scheme and optimizing the path and the vehicle dispatching scheme, the transportation cost and the transportation risk can be minimized in the transportation process, the number of the distributed vehicles is minimized, and the fairness of the vehicle transportation task amount is ensured.
Drawings
Fig. 1 is a vehicle departure arrival timeline.
FIG. 2 is a vehicle departure arrival time line after the adjustment of the working time according to the present invention.
Fig. 3 is a schematic diagram of a job scheduling transition to VRP problem.
FIG. 4 is a schematic diagram of total cost and total risk after the distribution center has been connected to the paths between the demand nodes.
FIG. 5 shows the individual encoding scheme in NSGA-II algorithm.
FIG. 6 is a flow chart for solving a scheduling problem for a single truck type hazardous materials transportation vehicle.
FIG. 7 is a diagram of a small-scale test network in accordance with an embodiment of the present invention.
Fig. 8 is a path selection Pareto solution of the present invention.
FIG. 9 is a graph comparing the two-stage algorithm (TSA) with the general NSGA-II algorithm (GGA).
Detailed Description
The invention is further described below with reference to the drawings and the detailed description.
In this distribution process, in order to simplify the operation flow, the following assumptions are made:
(1) When the dangerous goods transport vehicle is fully transported, the risk is only related to the risk of the driving road section, and is irrelevant to the vehicle loading capacity, the vehicle condition and the like, namely the risk calculation is calculated according to the risk measurement value of the driving road section.
(2) When the dangerous goods transport vehicle is in idle transport, the risk value is 0, namely, when the vehicle returns to the distribution center after the dangerous goods are unloaded at the demand point, the vehicle runs along a path with small cost (shortest distance or shortest time).
Step 1), establishing a model of the number of vehicles, the total running cost and total risk of the vehicles and the distribution time;
1.1 Assuming m demand points are to be completed) The dangerous goods distribution task of (1) is characterized in that the demand quantity of each task is q 1,q2,…,qm, the distribution time window of a demand point d is [ b d,ed ], a plurality of paths P od exist between a distribution center o and the demand point d, the demand quantity of each vehicle is q d, the nuclear quantity of each vehicle is g, q d is more than or equal to g, namely, one vehicle can only complete all or part of one task, the number of vehicles required for completing the distribution task of the task point d is a d, a d＝qd/g is a 3962/g, and a d＝qd/g is an integer when q d/g is an integer; when q d/g is not an integer,/>The total number of vehicles required to complete the distribution task at all the demand points is Σ dad, which is constant where the demand at the demand point is known. Although the total number of vehicles is constant, the same vehicle can participate in the delivery tasks at a plurality of demand points when the time window condition is satisfied, and studies are made on the total number of delivery vehicles, the arrival time of departure of the vehicles, and the like in the optimization of vehicle scheduling.
1.2 Transportation cost of dangerous goods transported by transportation vehicles on road sections (i, j) is divided into two cases, and the full-load transportation cost isThe empty driving cost is/>The transportation cost c 1 od for completing the arrival of the demand point d from the distribution center o is the sum of the transportation costs of the sections passed by, namely:
Variable(s) Indicating that the road segment (i, j) E is on the way from the delivery center o to the demand point d, otherwise
The route is the least cost route from the demand point d to the distribution center o when the empty car returns, and is usedThe value of the requirement point d can be obtained by adopting a shortest path solving algorithm, so that the distribution task of the requirement point d is completed, and the transportation cost is/>
1.3 The risk of the transportation vehicle transporting dangerous goods on the road section (i, j) is r ij, and the total transportation risk of completing the delivery task from the delivery center o to the demand point d is as follows:
1.4 Setting the travel time of the vehicle on the road section (i, j) as The average loading time of the vehicle is delta t 1, the unloading time is delta t 2, and the running time of the vehicle from the distribution center o to the demand point d is
The travel time of the vehicle from the demand point d to the distribution center o is
The total time of returning the delivery tasks from the delivery center o to the completion demand point d to the delivery center is as follows:
Assume that the moment when the vehicle starts from the distribution center is The moment of finishing the return of the demand point d to the distribution center is
1.5 If the earliest time window for the vehicle to start from the distribution center and the latest time window for the vehicle to return to the distribution center is [ b d,ed ]
The moment when the vehicle reaches the demand point d must satisfy the following condition:
the moment when the vehicle starts from the distribution center o The following conditions must be met:
At the same time, it is also necessary to meet the time window of the distribution center [ b 0,e0 ], i.e.
Step 2), establishing a vehicle path optimization mathematical model P1;
P1:
minf＝(f1,f2) (10)
s.t.
Wherein:
Equation (10) represents a minimized target vector composed of two target functions;
equation (11) is a transportation cost function expression;
formula (12) is a transport risk function expression;
The formula (13) ensures that a complete transportation path is formed between the distribution center o and the demand node d;
formula (14) is a decision variable;
n represents a set of N nodes and E represents a set of inter-node road segments.
The vehicle driving route from the distribution center to each demand point and back to the distribution center after the distribution is completed can be obtained by solving the model P1, and after the vehicle driving route is selected according to the risk preference, a decision maker determines how to determine the vehicle dispatching plan is still a problem to be solved. In the case of a limited number of dispatchable vehicles, the number of dispatches is an optimization objective that needs to be considered.
When the decision maker selects a certain path scheme according to the risk preference, in order to achieve the purpose of fewer vehicles to participate in the delivery task, each vehicle is required to bear as many transportation tasks as possible, that is, the idle time in the process of completing all tasks is minimum at the beginning of the whole transportation task.
For analysis, each distribution task of each demand point is decomposed into a plurality of jobs, that is, one distribution of each vehicle is divided into one job, and for any one job, the tasks need to be completed within a range from the earliest start time to the latest end time, and all the tasks of the distribution center can be represented in a time line manner as shown in fig. 1. It can be seen that to minimize the number of vehicles used throughout the distribution cycle, i.e., the jobs projected onto the overall timeline (dark bars) do not overlap as much as possible, because the dark bars overlap on the timeline, their corresponding jobs must be assigned different vehicles. Thus, the optimization objective for the minimum number of vehicles can be seen as making the jobs projected onto the overall time line as non-overlapping as possible by sliding the dark blocks on the respective task time lines, so that the minimum number of vehicles used can be achieved.
As shown in fig. 2, which is a working time line after the working start time of the part is adjusted, all the current delivery tasks only need two vehicles in view of the working overlapping condition on the total time line in the figure, so as to achieve the goal of optimizing the number of vehicles. The method provides a thought for optimizing the vehicle dispatching, namely, the number of vehicles can be optimized by adjusting the starting time of the operation on the premise that the operation process meets the time window requirement. However, the scheduling result of fig. 2 is not a best result for the balance of the tasks undertaken by the vehicles (the 1 st vehicle completes 8 jobs and the 2 nd vehicle completes only 1 job), so the vehicle job distribution balance is also considered in the job scheduling optimization process.
Through the analysis, each job can be executed by only one vehicle, the vehicle shifts to the next job after executing one job, if each job is regarded as a node of a two-dimensional plane, all the demand points are arranged according to the node numbers, and then each job in the distribution tasks is numbered according to the number, for example, the job of the distribution task reaching the demand point d 1 isThe job of d 2 is/>D 3 is operated asBy analogy, j=j d1∪Jd2∪...∪Jdm denotes all job sets. The vehicle scheduling problem may be translated into a time windowed VRP problem (fig. 3) with an objective function of minimum number of vehicles. For job i, if i ε J d its time window is/>The service duration is Δt 0d. Since the node represents the job information, the demand information is not included, and can be regarded as 0, and there is no capacity constraint on the transportation vehicle.
Step 3), a vehicle scheduling model P2 is established, so that x ijk(xijk epsilon [0,1 ]) is whether the operation j is executed after the operation i is completed by the vehicle k, if yes, x ijk =1, otherwise x ijk＝0;yik(yik epsilon [0,1 ]) is whether the operation i is executed by the vehicle k, if yes, y ik =1, otherwise y ik =0; for job i, if i ε J d its time window isThe service time is deltat 0d;
The P2 model is:
s.t.
Wherein: the expression (15) represents an objective function, In the function/>Representing the minimum number of vehicles, M being an integer, guaranteeing the priority of the vehicle number target,/>The standard deviation of running time of all vehicles is minimum under the condition of the same number of vehicles, so that the intensity of transportation tasks born by each vehicle is guaranteed to be fair, formulas (16) - (19) ensure that each vehicle can simultaneously carry out only one operation and sequentially execute the operation, each operation is completed by only one vehicle, formula (20) represents the moment when a vehicle k reaches a node j, and formula (21) ensures that the moment when the vehicle k reaches the node j must meet the time window constraint; equations (22) and (23) are time window requirements of earliest departure and latest return of the vehicle, and equations (24) and (25) are decision variables; k is a vehicle set, and C is a distribution node set;
step 4), carrying out two-stage solving;
4.1 P1 model, i.e. path optimization solving method: in the first stage, a pulse algorithm is adopted to obtain Pareto solutions of all paths from a distribution center to each demand node, the paths among all nodes need to be completely included in calculation by directly adopting NSGA-II, when the network scale is large, an evolutionary algorithm is difficult to obtain an optimal solution, and the fact that partial paths which cannot become the optimal solution can be eliminated can be considered. The optimization objective of model P1 is to obtain Pareto solutions of the total cost and risk between the distribution center and all the demand nodes, if this calculation process is considered as the path connection between the distribution center and the respective demand nodes, then find the cost and risk of transportation on the connected path (fig. 4). By using Representing the cost and risk of a path from the distribution center to the demand node d,/>Representing the cost and risk, respectively, after returning from the demand node d to the distribution center. Assume thatFor one of the Pareto solutions, if (c d,rd)＜(c′d,r′d) (< represent dominant), then/>Does not appear in the Pareto solution set. This conclusion is readily demonstrated by countercheck if/>Is one of the Pareto solutions, at least satisfiesOr/>One of the two conditions, i.e. c 'd≤cd or r' d≤rd, must be met, which is in contradiction with (c d,rd)＜(c′d,r′d).
From the above analysis, to obtain the Pareto solution set of the model P1, the Pareto solution may be first found for the path from the distribution center to each demand node, and only the partial path may appear in the Pareto path set of the model. And in the second stage, a Pareto path between the distribution center obtained in the first stage and each demand node is encoded, and the NSGA-II algorithm is adopted for solving. The paths from the first-stage distribution center to each demand node are respectively Representing the Pareto path set from the distribution center node o to the demand node d. In the two-stage method, other paths except the Pareto path are not needed to be considered in the second-stage calculation, so that the calculated amount can be greatly reduced compared with the method for directly adopting NSGA-II algorithm to solve, and the opportunity for obtaining the optimal solution is obviously increased under the condition of the same population scale and iteration times.
The coding mode is individual coding, natural number coding is adopted, the coding length is m, the format is n 1,n2...nm,
Wherein: The i-th value in the code is n i, which indicates that the distribution center o to the demand node d i adopts the/> And (5) transporting along a path. For example, the number of Pareto paths from the distribution center 0 to the demand node d 1-d4 is 2,3,5, and 4, respectively, and the code [1,4,3,5] indicates that in the individual, the 1 st path from the distribution center to the Pareto path set d 1 is selected, the 1 st path from the distribution center to the Pareto path set d 2 is selected, the 3 rd path from the distribution center to the Pareto path set d 3 is selected, and the 1 st path from the distribution center to the Pareto path set d 4 is selected for transportation.
Fitness function represents fitness of individual indi using [ c_value, r_value ] =f (indi), where c_value and r_value are cost and risk of transportation in accordance with the path selection scheme in individual indi, respectively.
In the population updating strategy, the population acquires new individuals through crossing, and the crossing operation adopts an integer crossing method; firstly, randomly selecting two individuals from a population, randomly generating two positions pos1 and pos2 (pos 1 is less than pos 2), and exchanging the positions pos1 to pos2 of the two individuals to generate two new individuals; mutation operation using integer mutation method to obtain new individuals, randomly selecting one individual, randomly generating two positions pos1 and pos2 (pos 1 < pos 2), and combining the positions pos1 to pos2 of the selected individuals one by oneAnd performing difference operation and taking an absolute value to generate a new individual.
The model P2 can be converted into a VRP problem which aims at achieving fairness of the number of vehicles and the intensity of the transportation tasks born by each vehicle through analysis, and besides the goal function is changed from the transportation distance and the number of vehicles to the number of vehicles and the fairness of the transportation tasks, other calculation steps can be completely solved by adopting a VRPHTW solving method based on a distribution estimation algorithm.
4.2 P2 model, i.e. vehicle dispatch optimization solving method:
First, initialization requirement information d=1;
secondly, adopting a pulse algorithm to calculate a Pareto path from the distribution center o to the demand node d, wherein d=d+1;
thirdly, if d is larger than m, solving a Pareto path between the distribution center o and all the demand nodes by adopting a multi-objective optimization method based on NSGA-II.
Fourth, decision-making personnel select one of the path schemes according to the risk preference.
Fifthly, a VRP solving method based on UMDA is adopted to obtain a vehicle scheduling schedule, namely a vehicle scheduling scheme, under the path scheme.
Examples:
The transport network is tested, node 0 in fig. 7 being the distribution center and node 1,2,4,7,8 being the point of demand. The distance and risk of each road section in the network are shown in table 1, and the demand and time window of each demand point are shown in table 2. The nuclear load of the transport vehicle is g=13.5m 3, the average loading time is deltat 1 =0.75 h, the unloading time is deltat 2 =0.75 h, the average running speed of the vehicle is 45km/h, the full-load transport cost is 50 yuan/km, and the empty-vehicle transport cost is 10 yuan/km. By arranging the vehicle transportation scheme, the transportation process achieves the transportation cost and the transportation risk are minimized.
Table 1 lengths of road segments and risk of transport vehicles traveling on the road segments
Table 2 is a window of demand and time to accept offloading for all demand nodes, a hard window of time, must arrive within a specified time, and wait if it arrives earlier than the earliest start time. The earliest time the vehicle starts from the delivery center and the latest time it returns to the delivery center are 08:00 and 20:00, respectively.
TABLE 2 demand Point demand and time Window
After step 4.1, pareto shortest path (departure path) and shortest distance path (return path) from the distribution center to each demand point can be obtained (see table 3).
TABLE 3 delivery center to demand Point outgoing Path return alternative Path
After obtaining the Pareto path set from the distribution center to each of the demand nodes, a Pareto solution of total transportation cost and total risk from the distribution center to all of the demand nodes is obtained by adopting an improved NSGA-II algorithm (fig. 8 is a Pareto solution distribution case, and table 4 is a path selection scheme after decoding the Pareto solution corresponding to the individual).
TABLE 4 Path selection scheme (Pareto solution, scheme 1-scheme 14)
When a certain transportation scheme is selected, the model P2 solving method is adopted for the route in the scheme, so that the vehicle scheduling schedule under the route scheme can be obtained, the vehicle scheduling scheme (table 5) in the case of the route scheme 1 in the table 4 is selected, and the total number of 17 delivery tasks is 17, and 5 vehicles are required to execute. The node array represents delivery nodes reached by the transport vehicle, loading is started in a schedule, the vehicle starts from the delivery center, unloading is started when the vehicle reaches a demand node, and the unloading is completed when the vehicle returns to the delivery center and reaches the delivery center.
TABLE 5 vehicle schedule with minimum total cost of transportation (scheme 1)
Table 6 is a vehicle scheduling scheme in the case of option 14, where the distribution tasks are still 17, but 6 vehicles are required to perform. It can be seen that the number of vehicles required when selecting different path schemes, and the number of required nodes served by each vehicle, are also different. However, it can be found through calculation that the scheduling scheme is not unique under the same path selection scheme, but the number of used vehicles and the standard deviation of the running time of each vehicle are the same. This is because when the demand time window overlap of two different demand nodes is large, the order of the transportation tasks to these nodes is not violated by the time window constraint, which also coincides with the actual situation in transportation.
TABLE 6 vehicle dispatch schedules with minimum total risk of transportation (scenario 14)
Verification example:
To verify the effectiveness of the method of the present invention, for the path planning problem, a two-stage algorithm (TSA) is proposed in addition to the two-stage algorithm (TSA), and a general NSGA-II algorithm (GGA) is designed, unlike the two-stage algorithm, in the general NSGA-II algorithm, a priority-based coding method PriorityBased Encoding is adopted, where the coding length is |n| m, and in the two-stage algorithm of the present invention, the coding length is m.
Two algorithm parameter settings are as follows: the population size is 100, and the iteration number is 50 generations. The Pareto optimal fronts obtained at iterations 5, 10, 20 and 50 in fig. 9 (a) to (d), respectively, can be found by comparison: when iterating to passage 10, the final solution can be obtained using the TSA algorithm, while when iterating to passage 50, the GGA algorithm still does not obtain the optimal solution. However, this is only for a 9-node network, and for a large-scale network, this gap will become larger. The two-stage algorithm designed for path planning has no direct relation between the coding length and the network scale, so that the algorithm is still effective for a large-scale network. In the second stage, the problem size of the VRPTW is only related to the number of the demand nodes and is not related to the size of the transportation network, so that the vehicle scheduling method has higher efficiency, and is applicable to small-scale or large-scale transportation networks.

Claims (2)

1. The dangerous goods full-load distribution path optimizing and vehicle dispatching method is characterized by comprising the following steps of: step 1), establishing a model of the number of vehicles, the total running cost and total risk of the vehicles and the distribution time;
1.1 Set the demand at the demand point d as q d, the nuclear capacity of each vehicle as g, set q d equal to or greater than g, and the number of vehicles needed for completing the delivery task of the task point d as a d, then The total number of vehicles required for completing the distribution tasks of all the demand points is Sigma dad;
1.2 Transportation cost of dangerous goods transported by transportation vehicles on road sections (i, j) is divided into two cases, and the full-load transportation cost is The empty driving cost is/>The transportation cost c 1 od for completing the arrival of the demand point d from the distribution center o is the sum of the transportation costs of the sections passed by, namely:
Variable(s) Representing the road segment (i, j) E on the way from the delivery center o to the demand point d, otherwise/>
The route is the least cost route from the demand point d to the distribution center o when the empty car returns, and is usedThe delivery task of the demand point d is completed, and the transportation cost is/>
1.3 The risk of the transportation vehicle transporting dangerous goods on the road section (i, j) is r ij, and the total transportation risk of completing the delivery task from the delivery center o to the demand point d is as follows:
1.4 Setting the travel time of the vehicle on the road section (i, j) as The average loading time of the vehicle is delta t 1, the unloading time is delta t 2, and the running time of the vehicle from the distribution center o to the demand point d is
The travel time of the vehicle from the demand point d to the distribution center o is
The total time of returning the delivery tasks from the delivery center o to the completion demand point d to the delivery center is as follows:
Assume that the moment when the vehicle starts from the distribution center is The moment of completion of the return of the demand point d to the distribution center is/>
1.5 If the earliest time window for the vehicle to start from the distribution center and the latest time window for the vehicle to return to the distribution center is [ b d,ed ]
The moment when the vehicle reaches the demand point d must satisfy the following condition:
the moment when the vehicle starts from the distribution center o The following conditions must be met:
At the same time, it is also necessary to meet the time window of the distribution center [ b 0,e0 ], i.e.
Step 2), establishing a vehicle path optimization mathematical model P1;
The P1 model is as follows:
minf＝(f1,f2) (10)
s.t.
Wherein:
Equation (10) represents a minimized target vector composed of two target functions;
equation (11) is a transportation cost function expression;
formula (12) is a transport risk function expression;
The formula (13) ensures that a complete transportation path is formed between the distribution center o and the demand node d;
formula (14) is a decision variable;
N represents a set of N nodes, E represents a set of inter-node road segments;
Step 3), a vehicle scheduling model P2 is established, so that x ijk(xijk epsilon [0,1 ]) is whether the operation j is executed after the operation i is completed by the vehicle k, if yes, x ijk =1, otherwise x ijk＝0;yik(yik epsilon [0,1 ]) is whether the operation i is executed by the vehicle k, if yes, y ik =1, otherwise y ik =0; for job i, if i ε J d its time window is The service time is deltat 0d;
The P2 model is:
s.t.
Wherein: the expression (15) represents an objective function, In the function/>Representing the minimum number of vehicles, M being an integer, guaranteeing the priority of the vehicle number target,/>The standard deviation of running time of all vehicles is minimum under the condition of the same number of vehicles, so that the intensity of transportation tasks born by each vehicle is guaranteed to be fair, formulas (16) - (19) ensure that each vehicle can simultaneously carry out only one operation, each operation is sequentially carried out, each operation is completed by only one vehicle, and formula (20) represents the moment when a vehicle k reaches a node j; equation (21) ensures that the time of arrival at node j must meet the time window constraint; equations (22) and (23) are time window requirements of earliest departure and latest return of the vehicle, and equations (24) and (25) are decision variables; k is a vehicle set, and C is a distribution node set;
step 4), carrying out two-stage solving;
4.1 P1 model, i.e. path optimization solving method:
the first stage adopts a pulse algorithm to obtain Pareto solutions of all paths from a distribution center to each demand node;
The second stage is to obtain a Pareto path between the distribution center and each demand node obtained in the first stage to encode, and solve by adopting NSGA-II algorithm; in the second stage NSGA-II algorithm, according to the coding mode, the fitness function and the population updating strategy;
The coding mode is individual coding, natural number coding is adopted, the coding length is m, the format is n 1,n2...nm,
Wherein: The i-th value in the code is n i, which indicates that the distribution center o to the demand node d i adopts the/> Transporting along a path;
the fitness function represents fitness of the individual indi by [ c_value, r_value ] =f (indi), wherein c_value and r_value are respectively the cost and risk of line transportation according to the path selection scheme in the individual indi;
The population in the population updating strategy obtains new individuals through crossing, and the crossing operation adopts an integer crossing method; firstly, randomly selecting two individuals from a population, randomly generating two positions pos1 and pos2 (pos 1 is less than pos 2), and exchanging the positions pos1 to pos2 of the two individuals to generate two new individuals;
mutation operation using integer mutation method to obtain new individuals, randomly selecting one individual, randomly generating two positions pos1 and pos2 (pos 1 < pos 2), and combining the positions pos1 to pos2 of the selected individuals one by one Performing difference operation and taking an absolute value to generate a new individual;
4.2 P2 model, i.e. vehicle dispatch optimization solving method:
First, initializing requirement information;
secondly, adopting a pulse algorithm to calculate a Pareto path from the distribution center o to the demand node d;
thirdly, solving Pareto paths from the distribution center o to all demand nodes by adopting a multi-objective optimization method based on NSGA-II;
Fourthly, decision-making staff selects one path scheme according to the risk preference;
Fifthly, a VRP solving method based on UMDA is adopted to obtain a vehicle scheduling schedule, namely a vehicle scheduling scheme, under the path scheme.
2. The method for optimizing the full-load distribution path of dangerous goods and dispatching the vehicles according to claim 1, wherein the method is characterized by comprising the following steps: the paths from the first stage distribution center to the demand nodes in the step 4.1) are respectively as follows Representing the Pareto path set from the distribution center node o to the demand node d.
CN202311199824.1A 2023-09-18 2023-09-18 Dangerous goods full-load distribution path optimization and vehicle dispatching method Active CN117094629B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311199824.1A CN117094629B (en) 2023-09-18 2023-09-18 Dangerous goods full-load distribution path optimization and vehicle dispatching method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311199824.1A CN117094629B (en) 2023-09-18 2023-09-18 Dangerous goods full-load distribution path optimization and vehicle dispatching method

Publications (2)

Publication Number Publication Date
CN117094629A CN117094629A (en) 2023-11-21
CN117094629B true CN117094629B (en) 2024-05-10

Family

ID=88775089

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311199824.1A Active CN117094629B (en) 2023-09-18 2023-09-18 Dangerous goods full-load distribution path optimization and vehicle dispatching method

Country Status (1)

Country Link
CN (1) CN117094629B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102236825A (en) * 2010-04-26 2011-11-09 兰州交通大学 Method for scheduling tasks of conveyer of airport freight logistics system
WO2015011661A2 (en) * 2013-07-26 2015-01-29 Elettric 80 S.P.A. Device and method for optimising the movement of automated-guided vehicles, and the like

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102236825A (en) * 2010-04-26 2011-11-09 兰州交通大学 Method for scheduling tasks of conveyer of airport freight logistics system
WO2015011661A2 (en) * 2013-07-26 2015-01-29 Elettric 80 S.P.A. Device and method for optimising the movement of automated-guided vehicles, and the like

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
关于危险品运输安全性路径选择仿真研究;殷勇等;计算机仿真;20170831;34(08);184-189, 259 *
危险品运输车辆调度的多目标优化;周愉峰等;中国安全科学学报;20130430;23(04);113-119 *
危险品道路运输车辆时空协调调度优化研究;柴获;《中国博士学位论文全文数据库（工程科技II辑）》;20181215(第02期);C034-64 *
基于区域风险分布公平性的危险品运输路径优化;柴获等;兰州交通大学学报;20161231;35(06);106-111 *
考虑载货量的危险品运输车辆路径优化;张圣忠等;武汉理工大学学报(信息与管理工程版);20200831;42(04);290-297 *

Also Published As

Publication number Publication date
CN117094629A (en) 2023-11-21

Similar Documents

Publication Publication Date Title
CN109583650B (en) Electric vehicle battery replacement station site selection and logistics distribution joint scheduling method
CN109919541B (en) Modeling solving method for multi-stage positioning inventory path problem
CN109002902B (en) Regional multi-stage fresh agricultural product dynamic vehicle path optimization method
CN111553507B (en) Multi-commodity-flow-based optimization method for Chinese and European container transportation scheme
CN105976030A (en) Multi-agent-based platform scheduling intelligent sorting model structure
Liu et al. Integrated scheduling of ready-mixed concrete production and delivery
Lv et al. Operational optimization of transit consolidation in multimodal transport
CN107578119A (en) A kind of resource allocation global optimization method of intelligent dispatching system
CN111091238A (en) Automatic container terminal AGV intelligent scheduling method
CN115689247A (en) Complete vehicle logistics scheduling optimization method based on improved genetic algorithm
CN114925995A (en) Road trunk vehicle and cargo matching and path planning method based on single parent genetic algorithm
CN101840200A (en) Adaptive processing method for optimizing dynamic data in dispatching control
CN113408988A (en) Scheduling optimization method of warehousing system, electronic equipment and storage medium
CN114358233A (en) Multi-AGV path planning problem optimization method and system based on double-hybrid particle swarm
CN117094629B (en) Dangerous goods full-load distribution path optimization and vehicle dispatching method
Mnif et al. A multi-objective formulation for multimodal transportation network's planning problems
CN108416482A (en) One kind is based on regional shifty logistics distribution paths planning method
CN115358455A (en) Electric vehicle path optimization method considering battery replacement and hybrid time window constraints
CN116882883A (en) Power grid material distribution path planning method utilizing big data
CN114537479A (en) Motor train unit operation plan compiling method considering fluctuation within week of train running scheme
Al Theeb et al. Optimization of logistic plans with adopting the green technology considerations by utilizing electric vehicle routing problem
CN113393111B (en) Cross-border transportation double-side connection vehicle scheduling method based on variable neighborhood tabu search algorithm
CN113420926A (en) Method for compiling multi-traffic mode transportation capacity collaborative optimization scheme between urban hubs
CN113743788A (en) Industrial vehicle scheduling method based on improved genetic algorithm
Savushkin Models of forming and processing packages of transport services

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