CN108764777B - Electric logistics vehicle scheduling method and system with time window - Google Patents

Electric logistics vehicle scheduling method and system with time window Download PDF

Info

Publication number
CN108764777B
CN108764777B CN201810385664.2A CN201810385664A CN108764777B CN 108764777 B CN108764777 B CN 108764777B CN 201810385664 A CN201810385664 A CN 201810385664A CN 108764777 B CN108764777 B CN 108764777B
Authority
CN
China
Prior art keywords
path
planning
charging
electric logistics
logistics vehicle
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
CN201810385664.2A
Other languages
Chinese (zh)
Other versions
CN108764777A (en
Inventor
竹锦潇
李进
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.)
Zhejiang Gongshang University
Original Assignee
Zhejiang Gongshang 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 Zhejiang Gongshang University filed Critical Zhejiang Gongshang University
Priority to CN201810385664.2A priority Critical patent/CN108764777B/en
Publication of CN108764777A publication Critical patent/CN108764777A/en
Application granted granted Critical
Publication of CN108764777B publication Critical patent/CN108764777B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping
    • 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

Abstract

The invention discloses a method and a system for scheduling an electric logistics vehicle with a time window, wherein the method comprises the following steps: acquiring distribution parameters of the electric logistics vehicle, and establishing a mixed integer planning model according to the distribution parameters; acquiring planning constraint parameters of the electric logistics vehicle, and determining planning constraint conditions of the mixed integer planning model according to the planning constraint parameters; the planning constraint condition comprises a planning demand constraint condition and a charging constraint condition; performing optimization calculation on the mixed integer programming model by using a self-adaptive neighborhood search algorithm and a simulated annealing algorithm and combining a programming demand constraint condition and a charging constraint condition to obtain distribution path information; and finishing the optimization of the dispatching of the electric logistics vehicles according to the distribution path information. The invention can reasonably arrange the distribution path of the electric logistics vehicle to complete the dispatching, improve the efficiency of cargo distribution service by means of the electric logistics vehicle, and save electric energy so as to further reduce the pollution to the environment.

Description

Electric logistics vehicle scheduling method and system with time window
Technical Field
The invention relates to the technical field of vehicle scheduling, in particular to a method and a system for scheduling an electric logistics vehicle with a time window.
Background
In recent decades, the distribution problem of the green supply chain is always the focus of social attention and is also the primary problem facing energy conservation and emission reduction in China. The green supply chain management not only can obviously improve social benefits, but also is an effective means for obtaining economic benefits. The green supply chain can avoid resource waste, enhance the social responsibility of enterprises, bring good reputation and brand image of green products to the enterprises and is beneficial to developing product markets. The supply chain is a flow-through network of suppliers, manufacturers, retailers and distribution centers, etc., with the purpose of being centered on customer service. The most important part in a supply chain system is the transport of material between different centers, such as supplier to manufacturer, manufacturer to distribution center, and distribution center to customer. With the attention of customers to low carbon and environmental protection, the nation takes the popularization and application of the electric logistics vehicle as a long-term strategic policy. A scheduling method and a scheduling technology in distribution and transportation of a supply chain based on an electric logistics vehicle become a key problem in a green supply chain.
The problem of scheduling electric logistics vehicles is one of the important problems in the field of transportation in a plurality of practical applications. An Electric Logistics Vehicle Scheduling Problem (EL-VSPTW) with a Time window is an extension of a Vehicle Scheduling Problem (VSP), and is a delivery optimization Scheduling performed by taking an Electric Logistics Vehicle as a carrying tool and considering the characteristics of insufficient endurance, small number of charging stations and slow charging of the Electric Logistics Vehicle in order to meet the Time and environmental requirements of customers on Logistics service. The problem plays a crucial role in improving customer service satisfaction, realizing sustainable development and reducing energy utilization.
In the logistics transportation technology, most enterprises neglect the balanced use of the electric logistics vehicle electric energy and the influence on the environment. Recently, many enterprises have begun to adopt different technologies to try to save the electric energy consumption of the electric logistics vehicles. The driving distance is one of the main factors affecting electric power, and the consumption of electric power is proportional to the distance traveled by the electric logistics vehicles. The electric logistics vehicle scheduling problem is one of main problems in a transportation and supply chain management system, and is an integer programming problem in a type of combination optimization. The electric logistics vehicles transport goods from a distribution center to customers, supply power to nearby charging stations during transportation according to power consumption, and aim to minimize total travel distance to save cost. In the problem of dispatching the electric logistics vehicles with the time window, the requirement of customers on service time is also considered, and the electric energy consumption is minimized.
The electric logistics vehicle scheduling problem with the time window is a vehicle scheduling problem which is difficult to solve so far. At present, an algorithm for solving the scheduling of the electric logistics vehicles with the time windows is mainly based on the premise that once the electric logistics vehicles are charged in a charging station, batteries must be fully charged, so that the scheduling result of the electric logistics vehicles is not optimized, the efficiency in the aspect of electric energy saving is low, the transportation cost of the electric logistics vehicles is high, and the development of logistics companies is not facilitated. When the position of the charging station is considered and the specific charging quantity is determined, the state space of the solution is rapidly expanded, so that the solution efficiency is obviously reduced and even stagnates.
Disclosure of Invention
The invention aims to provide a method and a system for dispatching an electric logistics vehicle with a time window, which can reasonably arrange a distribution path of the electric logistics vehicle to finish dispatching, improve the efficiency of goods distribution service by means of the electric logistics vehicle, and save electric energy so as to further reduce the pollution to the environment.
The invention provides a method for dispatching an electric logistics vehicle with a time window, which comprises the following steps of;
acquiring distribution parameters of the electric logistics vehicle, and establishing a mixed integer planning model according to the distribution parameters;
acquiring planning constraint parameters of the electric logistics vehicle, and determining planning constraint conditions of the mixed integer planning model according to the planning constraint parameters; the planning constraint condition comprises a planning demand constraint condition and a charging constraint condition;
performing optimization calculation on the mixed integer programming model by using a self-adaptive neighborhood search algorithm and a simulated annealing algorithm and combining the programming demand constraint condition and the charging constraint condition to obtain distribution path information;
and finishing the optimization of the dispatching of the electric logistics vehicles according to the distribution path information.
As an implementable manner, the determining the planning constraint condition of the mixed integer planning model according to the planning constraint parameter includes the following steps;
and performing correlation screening on the planning constraint parameters and the distribution parameters in the mixed integer planning model, and establishing planning constraint conditions according to the planning constraint parameters after the correlation screening.
As an implementable embodiment, the charging constraint includes a state-of-charge constraint and a charging time window constraint;
the mathematical formula for the state of charge constraint is represented as:
Figure BDA0001642157210000021
wherein m represents the mth charging station; bmIs shown in an electric objectWhen the electric logistics vehicle arrives at the charging station m, the electric quantity state of the battery of the electric logistics vehicle is obtained; z represents the capacity of the battery of the electric logistics vehicle, BmA decision variable representing the battery charging state of the battery of the electric logistics vehicle when the electric logistics vehicle leaves the charging station m; f represents a set of charging stations, F' represents a virtual network point set generated according to F, and each power conversion station in the set F is allowed to be accessed for multiple times; in the set subscripts 0 and I +1, 0 represents the distribution center as a starting point, I +1 represents the distribution center as an end point, and each path starts from 0 and ends at I + 1; f'0F 'is a union of a distribution center and a virtual network point set'0＝F′Y{0}；
The mathematical formula of the charging time window constraint is expressed as:
Figure BDA0001642157210000031
where θ represents the start time of customer service, θmRepresents the start time, theta, of customer service at the mth charging stationnIndicating a start time of customer service at the nth charging station; p is a radical ofmnThe decision variables are binary decision variables, when the electric logistics vehicle passes through a road section from a client point m to n, the value is 1, otherwise, the value is 0; t is tmnRepresenting the time for the electric logistics vehicle to travel between the customer point m and the customer point n; r ismRepresents the service time at customer site m; l0Represents an upper limit of a time window when the distribution center point is used as a starting point; q represents a charge rate of the battery; d ═ 1.. times, I } represents a set of customer points, D0Representing the union of the starting point and the set of customer points, i.e. D, with the distribution centre as the starting point0DY {0 }; d 'represents a set of F' and D; d'I+1Indicates the end point, the union of the F 'and D sets, that is, D'I+1＝D′Y{I+1}。
As an implementable embodiment, the optimization calculation of the mixed integer programming model by using the adaptive neighborhood search algorithm and the simulated annealing algorithm and combining the programming demand constraint condition and the charging constraint condition to obtain the distribution path information includes the following steps;
constructing an initial path solution of the mixed integer programming model according to a preset path selection rule;
taking an adaptive neighborhood search algorithm as a basic frame, and combining the planning demand constraint condition and the charging constraint condition to process the mixed integer planning model so as to define the initial iteration value of each preset operator; sequentially executing initial iteration values corresponding to all preset operators, and optimizing the initial path solution to obtain a first path solution;
and carrying out probability random optimization on the first path solution by using a simulated annealing algorithm, and obtaining distribution path information according to a probability random optimization result and a preset receiving or rejecting criterion.
As an implementation manner, the obtaining of the distribution path information according to the probability random optimizing result and the preset receiving or rejecting criteria comprises the following steps;
obtaining a second path solution according to the probability random optimization result, and judging whether the second path solution meets the termination condition or not to evaluate and judge;
if the second path solution does not meet the termination condition, taking an evaluation judgment result as a new termination condition, and changing the adaptive weight of each operator by using an adaptive neighborhood search algorithm according to the evaluation judgment result so as to adjust the initial iteration value; sequentially executing the adjusted initial iteration values corresponding to the preset operators, and optimizing the second path solution to obtain a third path solution; carrying out probability random optimization on the third path solution by using a simulated annealing algorithm to obtain a fourth path solution; continuously and circularly evaluating and judging until the probability is randomly optimized to obtain the current path solution which meets the last termination condition;
and if the second path solution meets the termination condition, taking the second path solution as the distribution path information.
As an implementation, the mixed integer programming model is:
Figure BDA0001642157210000041
wherein m represents the mth charging station; n represents an nth charging station; x is the number ofmnRepresents the distance from the mth customer point to the nth customer point; p is a radical ofmnThe decision variables are binary decision variables, when the electric logistics vehicle passes through a road section from a customer point m to a customer point n, the decision variables are 1, otherwise, the decision variables are 0; d'0Represents a union of node sets of origin, F 'and D, i.e., D'0＝D′Y{0}。
Correspondingly, the invention also provides an electric logistics vehicle dispatching system with a time window, which comprises a model establishing module, a constraint condition setting module, an optimizing module and a dispatching module;
the model establishing module is used for acquiring distribution parameters of the electric logistics vehicle and establishing a mixed integer planning model according to the distribution parameters;
the constraint condition setting module is used for acquiring planning constraint parameters of the electric logistics vehicle and determining planning constraint conditions of the mixed integer planning model according to the planning constraint parameters; the planning constraint condition comprises a planning demand constraint condition and a charging constraint condition;
the optimization module is used for carrying out optimization calculation on the mixed integer programming model by utilizing a self-adaptive neighborhood search algorithm and a simulated annealing algorithm and combining the programming demand constraint condition and the charging constraint condition to obtain distribution path information;
and the scheduling module is used for finishing the optimization of the scheduling of the electric logistics vehicles according to the distribution path information.
As an implementable embodiment, the constraint condition setting module is configured to: the charging constraint condition comprises a charging state constraint condition and a charging time window constraint condition;
the mathematical formula for the state of charge constraint is represented as:
Figure BDA0001642157210000042
wherein m represents the mth charging station; bmIndicating arrival charging at an electric logistics vehicleWhen the electric logistics vehicle stands for m, the electric quantity state of the battery of the electric logistics vehicle is judged; z represents the capacity of the battery of the electric logistics vehicle, BmA decision variable representing the battery charging state of the battery of the electric logistics vehicle when the electric logistics vehicle leaves the charging station m; f represents a set of charging stations, F' represents a virtual network point set generated according to F, and each power conversion station in the set F is allowed to be accessed for multiple times; in the set subscripts 0 and I +1, 0 represents the distribution center as a starting point, I +1 represents the distribution center as an end point, and each path starts from 0 and ends at I + 1; f'0F 'is a union of a distribution center and a virtual network point set'0＝F′Y{0}；
The mathematical formula of the charging time window constraint is expressed as:
Figure BDA0001642157210000043
where θ represents the start time of customer service, θmRepresents the start time, theta, of customer service at the mth charging stationnIndicating a start time of customer service at the nth charging station; p is a radical ofmnThe decision variables are binary decision variables, when the electric logistics vehicle passes through a road section from a client point m to n, the value is 1, otherwise, the value is 0; t is tmnRepresenting the time for the electric logistics vehicle to travel between the customer point m and the customer point n; r ismRepresents the service time at customer site m; l0Represents an upper limit of a time window when the distribution center point is used as a starting point; q represents a charge rate of the battery; d ═ 1.. times, I } represents a set of customer points, D0Representing the union of the starting point and the set of customer points, i.e. D, with the distribution centre as the starting point0DY {0 }; d 'represents a set of F' and D; d'I+1Indicates the end point, the union of the F 'and D sets, that is, D'I+1＝D′Y{I+1}。
As an implementation, the optimization module includes a construction unit, a definition unit, and a path optimization unit;
the construction unit is used for constructing an initial path solution of the mixed integer programming model according to a preset path selection rule;
the defining unit is used for processing the mixed integer programming model by taking the adaptive neighborhood search algorithm as a basic frame and combining the programming demand constraint condition and the charging constraint condition so as to define the initial iteration value of each preset operator; sequentially executing initial iteration values corresponding to all preset operators, and optimizing the initial path solution to obtain a first path solution;
and the path optimization unit is used for carrying out probability random optimization on the first path solution by using a simulated annealing algorithm and obtaining the distribution path information according to the probability random optimization result and a preset receiving or rejecting criterion.
As an implementation manner, the path optimization unit includes a judgment subunit, a loop optimization subunit and an acceptance subunit;
the judging subunit is configured to obtain a second path solution according to the probabilistic random optimization result, and evaluate and judge whether the second path solution satisfies a termination condition;
the loop optimization subunit is configured to, if the second path solution does not satisfy the termination condition, take an evaluation judgment result as a new termination condition, and change the adaptive weight of each operator by using an adaptive neighborhood search algorithm according to the evaluation judgment result to adjust the initial iteration value; sequentially executing the adjusted initial iteration values corresponding to the preset operators, and optimizing the second path solution to obtain a third path solution; carrying out probability random optimization on the third path solution by using a simulated annealing algorithm to obtain a fourth path solution; continuously and circularly evaluating and judging until the probability is randomly optimized to obtain the current path solution which meets the last termination condition;
and the receiving subunit is configured to, if the second path solution satisfies a termination condition, use the second path solution as the distribution path information.
Compared with the prior art, the technical scheme has the following advantages:
according to the method and the system for dispatching the electric logistics vehicles with the time windows, firstly, a mixed integer planning model is established according to the obtained distribution parameters of the electric logistics vehicles, then the planning constraint conditions of the mixed integer planning model including the planning requirement constraint conditions and the charging constraint conditions are determined, and the electric logistics vehicles are allowed to be partially charged in a charging station according to the electric quantity consumption of the electric logistics vehicles and the time window requirements of distribution service, so that the charging time can be saved, and the time window requirements of customer service are improved; and then, an adaptive neighborhood search algorithm and a simulated annealing algorithm are utilized, and a planning demand constraint condition and a charging constraint condition are combined to carry out optimization calculation on the mixed integer planning model, so that the time for obtaining an optimal feasible solution is effectively shortened, the efficiency of the whole algorithm flow is improved, the distribution path of the electric logistics vehicle can be reasonably arranged to complete scheduling, the efficiency of cargo distribution service by the electric logistics vehicle is improved, and the electric energy is saved, so that the pollution to the environment is further reduced.
Drawings
Fig. 1 is a schematic flowchart of a method for scheduling an electronic logistics vehicle with a time window according to an embodiment of the present invention;
FIG. 2 is a schematic flowchart of an optimization procedure using an adaptive neighborhood search algorithm and a simulated annealing algorithm according to a first embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating the dispatching of an electric logistics vehicle with a time window according to a first embodiment of the invention;
fig. 4a is a schematic diagram illustrating a random path before deleting operation according to an embodiment of the present invention;
FIG. 4b is a diagram illustrating a random path after a random path deleting operation according to an embodiment of the present invention;
FIG. 5a is a schematic view of a first charging station removal operation according to an embodiment of the present invention;
FIG. 5b is a schematic diagram illustrating a first time charging station removal operation according to an embodiment of the present invention;
FIG. 6a is a schematic diagram illustrating a second recharging station removal operation according to one embodiment of the present invention;
FIG. 6b is a schematic diagram illustrating a second charging station removal operation according to an embodiment of the present invention;
FIG. 7a is a schematic view of a charging station according to an embodiment of the present invention before a charging station removal and charging station insertion operation;
FIG. 7b is a schematic diagram of a charging station removal and charging station insertion operation according to an embodiment of the present invention;
fig. 8 is a schematic structural diagram of an electric logistics vehicle dispatching system with a time window according to a second embodiment of the present invention;
FIG. 9 is a schematic diagram of the structure of the optimization module of FIG. 8;
fig. 10 is a schematic structural diagram of the path optimizing unit in fig. 8.
In the figure: 100. a model building module; 200. a constraint condition setting module; 210. a screening unit; 300. an optimization module; 310. a construction unit; 320. a definition unit; 330. a path optimization unit; 331. a judgment subunit; 332. a loop optimization subunit; 333. an accepting subunit; 400. and a scheduling module.
Detailed Description
The above and further features and advantages of the present invention will be apparent from the following, complete description of the invention, taken in conjunction with the accompanying drawings, wherein the described embodiments are merely some, but not all embodiments of the invention.
The electric logistics vehicle scheduling problem with the time window is one of the standard vehicle scheduling problems. While the conventional method often considers that the battery of the electric logistics vehicle must be fully charged once the electric logistics vehicle is charged at the charging station, in the invention, the electric logistics vehicle is allowed to be partially charged at the charging station according to the electric quantity consumption and the time window requirement of distribution service, so that the charging time can be saved, and the time window requirement and the satisfaction degree of customer service can be improved. The invention models the problem as a 0-1 mixed integer linear programming problem and develops an adaptive neighborhood search and simulated annealing mixed intelligent algorithm to effectively solve the problem. The scheduling problem of the electric logistics vehicle with the time window has two aims: the first objective is to find the minimum travel distance, and the second objective is to find the path that uses the least amount of electricity and has the shortest accumulated charging time. The method comprises the following specific steps.
Referring to fig. 1, a method for dispatching an electric logistics car with a time window according to an embodiment of the present invention includes the following steps;
s100, obtaining distribution parameters of the electric logistics vehicle, and establishing a mixed integer planning model according to the distribution parameters;
s200, acquiring planning constraint parameters of the electric logistics vehicle, and determining planning constraint conditions of the mixed integer planning model according to the planning constraint parameters; the planning constraint condition comprises a planning demand constraint condition and a charging constraint condition;
s300, performing optimization calculation on the mixed integer programming model by using a self-adaptive neighborhood search algorithm and a simulated annealing algorithm and combining a programming demand constraint condition and a charging constraint condition to obtain distribution path information;
and S400, completing the optimization of the dispatching of the electric logistics vehicles according to the distribution path information.
It should be noted that the distribution parameters and the planning constraint parameters of the electric logistics vehicle may be directly obtained from the local database, or may be obtained from the cloud server by using big data. In this embodiment, the source of the acquired data is not limited. Moreover, according to the types of the electric logistics vehicles, the specific parameter types in the distribution parameters of each type of electric logistics vehicle are different, and the specific parameter data in the specific parameter types are different. And screening the acquired data according to the requirements of the mixed integer programming model, and establishing the mixed integer programming model by taking a 0-1 mixed integer linear programming as a problem according to the screened distribution parameters.
In this embodiment, the mixed integer programming model is:
Figure BDA0001642157210000071
wherein m represents the mth charging station; n represents an nth charging station; x is the number ofmnRepresents the distance from the mth customer point to the nth customer point; p is a radical ofmnThe decision variables are binary decision variables, when the electric logistics vehicle passes through a road section from a customer point m to a customer point n, the decision variables are 1, otherwise, the decision variables are 0; d'0Represents a union of a set of origin points, F 'and D, that is, D'0＝D′Y{0}，D′I+1D ' is a union of the distribution center as the terminal point and the set of F ' and D 'I+1＝D′Y{I+1}。
The mixed integer programming model can also be understood as an objective function for calculating the minimum value of the total travel distance and the total distance of the electric logistics vehicles. For the mixed integer programming model, various physical parameters in the model are delivery parameters.
In order to solve the problems that in the prior art, an algorithm for solving the scheduling of the electric logistics vehicles with the time windows mainly aims at the premise that the electric logistics vehicles need to be fully charged once being charged in a charging station, so that the scheduling results of the electric logistics vehicles are not optimized enough, the efficiency in the aspect of electric energy saving is low, the transportation cost of the electric logistics vehicles is high, and the development of logistics companies is not facilitated, the planning constraint conditions of the mixed integer planning model are determined according to the planning constraint parameters, and the planning constraint conditions comprise planning requirement constraint conditions and charging constraint conditions. The electric logistics vehicle is allowed to be partially charged in the charging station according to the electric quantity consumption and the time window requirement of distribution service by utilizing the charging constraint condition, so that the charging time can be saved, and the time window requirement of customer service is improved.
The planning constraints are explained in detail below by way of example.
The charging constraint condition comprises a charging state constraint condition and a charging time window constraint condition;
the mathematical formula for the state of charge constraint is expressed as:
Figure BDA0001642157210000081
wherein m represents the mth charging station; bmRepresenting the electric quantity state of the battery of the electric logistics vehicle when the electric logistics vehicle reaches the charging station m; z represents the capacity of the battery of the electric logistics vehicle, BmA decision variable representing the battery charging state of the battery of the electric logistics vehicle when the electric logistics vehicle leaves the charging station m; f represents a set of charging stations, F' represents a virtual network point set generated according to F, and each power conversion station in the set F is allowed to be accessed for multiple times; in the set subscripts 0 and I +1, 0 represents the distribution center as a starting point, I +1 represents the distribution center as an end point, and each path starts from 0 and ends at I + 1; f'0F 'is a union of a distribution center and a virtual network point set'0F' Y {0 }; thereby determining the battery state of charge after charging at a charging station and ensuring that the battery state of charge does not exceed its capacity;
the mathematical formula for the charging time window constraint is expressed as:
Figure BDA0001642157210000082
where θ represents the start time of customer service, θmRepresents the start time, theta, of customer service at the mth charging stationnIndicating a start time of customer service at the nth charging station; p is a radical ofmnThe decision variables are binary decision variables, when the electric logistics vehicle passes through a road section from a client point m to n, the value is 1, otherwise, the value is 0; t is tmnRepresenting the time for the electric logistics vehicle to travel between the customer point m and the customer point n; r ismRepresents the service time at customer site m; l0Represents an upper limit of a time window when the distribution center point is used as a starting point; q represents a charge rate of the battery; d ═ 1.. times, I } represents a set of customer points, D0Representing the union of the starting point and the set of customer points, i.e. D, with the distribution centre as the starting point0DY {0 }; d 'represents a set of F' and D; d'I+1Indicates the end point, the union of the F 'and D sets, that is, D'I+1D' Y { I +1 }. Thereby ensuring that the charging process can meet the time window requirements of all customers.
In the embodiment, the optimized distribution route information is limited by the charging state constraint condition and the charging time window constraint condition, and the electric logistics vehicles are allowed to partially charge at the charging stations according to the electric quantity consumption and the time window requirement of distribution service. And the planning demand constraints include, but are not limited to, connectivity constraints, access constraints, traffic conservation constraints, binary decision variable constraints, time window limit constraints, and cargo demand constraints.
The mathematical formulas of the above planning requirement constraints are expressed as follows.
Due to p in actual transportationmnWill change correspondingly with the change of the number of customer points in order to change pmnIn conjunction with the customer site, connectivity constraints and access constraints are defined to handle the customer's connectivity and access to the charging station, respectively.
The mathematical formula for the connectivity constraint is expressed as:
Figure BDA0001642157210000091
the mathematical formula for the access constraints is expressed as:
Figure BDA0001642157210000092
the traffic conservation constraint will be such that the number of segments entering a node is equal to the number of segments from which the segment departed. The mathematical formula for the flow conservation constraint is expressed as:
Figure BDA0001642157210000093
the mathematical formula for the binary decision variable constraints is represented as:
Figure BDA0001642157210000094
to meet the time window limits of the customer and the distribution center, the mathematical formula for the time window limit constraints is expressed as:
Figure BDA0001642157210000095
in the formula, thetanRepresents the start time, k, of customer service at the nth charging stationnAnd lnRespectively representing the upper and lower limits of a time window, i.e. the time window is kn,ln]。
In order to be able to guarantee the cargo demands of all customers, the mathematical formula of the cargo demand constraint is expressed as:
Figure BDA0001642157210000096
wherein v ismRepresenting the remaining cargo amount of the electric logistics vehicle at the mth customer point; v. ofnRepresenting the remaining cargo amount of the electric logistics vehicle at the nth customer point; v. of0The residual cargo quantity of the electric logistics vehicle at the starting point of the distribution path is represented; c represents the load capacity of the electric logistics vehicle, wmRepresenting the demand of the mth customer site.
For the planning constraint conditions, the physical parameters in the constraint conditions are delivery parameters.
In order to solve the problem, when the position of a charging station is considered and the specific electric quantity of charging is determined, the state space of a solution is rapidly expanded, so that the solution efficiency is obviously reduced and even stagnates. After a mixed integer planning model of an electric logistics vehicle scheduling problem with a time window is established and a planning constraint condition is determined, an adaptive neighborhood search algorithm and a simulated annealing algorithm are utilized, and the planning requirement constraint condition and a charging constraint condition are combined to carry out optimization calculation on the mixed integer planning model to obtain distribution path information; specifically, it can be understood that: defining initial iteration values of preset operators by taking a self-adaptive neighborhood search algorithm as a basic frame, and executing different operators to obtain a current path solution; and determining to accept or reject the current path solution through a simulated annealing algorithm, judging whether the predicted effect is achieved, changing the self-adaptive weight of each preset operator according to the algorithm searching effect, performing dynamic algorithm adjustment, and continuously and circularly executing the whole process until no more optimal solution is generated. If the battery charge level of the electric logistics vehicle is negative when it arrives at the customer site, the whole process will be repeated for the customer sites. If the charging station insertion does not produce a feasible solution, it will revert to the previous feasible solution.
In other embodiments, other meta heuristic algorithms may be used for solving, and one or more of a tabu search algorithm, a genetic algorithm, a variable neighborhood search algorithm, and a particle swarm algorithm may be used.
The scheduling method of the electric logistics vehicle with the time window, provided by the invention, comprises the steps of firstly establishing a mixed integer planning model according to the acquired distribution parameters of the electric logistics vehicle, then determining a planning constraint condition of the mixed integer planning model comprising a planning demand constraint condition and a charging constraint condition, and allowing the electric logistics vehicle to partially charge at a charging station according to the electric quantity consumption of the electric logistics vehicle and the time window requirement of distribution service, so that the charging time can be saved, and the time window requirement of customer service is improved; and then, an adaptive neighborhood search algorithm and a simulated annealing algorithm are utilized, and a planning demand constraint condition and a charging constraint condition are combined to carry out optimization calculation on the mixed integer planning model, so that the time for obtaining an optimal feasible solution is effectively shortened, the efficiency of the whole algorithm flow is improved, the distribution path of the electric logistics vehicle can be reasonably arranged to complete scheduling, the efficiency of cargo distribution service by the electric logistics vehicle is improved, and the electric energy is saved, so that the pollution to the environment is further reduced. Meanwhile, the invention can also be applied to the technical fields of unmanned aerial vehicle logistics distribution, robot distribution, artificial intelligence and the like.
Further, step S200 includes the following steps; and performing correlation screening on the planning constraint parameters and the distribution parameters in the mixed integer planning model, and establishing planning constraint conditions according to the planning constraint parameters after the correlation screening.
In order to improve the efficiency of establishing the planning constraint conditions, the relevance screening can be completed in a comparison screening mode. And comparing and screening all the acquired planning constraint parameters with distribution parameters in the mixed integer planning model, simultaneously completing correlation calculation, and establishing planning constraint conditions for the planning constraint parameters after the correlation screening. Unnecessary parameters are quickly eliminated, and the establishment process of planning constraint conditions is accelerated. It should be noted that, during the relevance screening, a relevance value may be pre-selected and set for the planning constraint parameter, and the relevance value is always used as a necessary planning constraint parameter, so as to establish a relevant planning constraint condition.
Further, step S300 includes the following steps;
s310, constructing an initial path solution of the mixed integer programming model according to a preset path selection rule;
s320, taking the self-adaptive neighborhood search algorithm as a basic frame, and combining a planning demand constraint condition and a charging constraint condition to process the mixed integer planning model so as to define the initial iteration value of each preset operator; sequentially executing initial iteration values corresponding to the preset operators, and optimizing the initial path solution to obtain a first path solution;
s330, carrying out probability random optimization on the first path solution by using a simulated annealing algorithm, and obtaining distribution path information according to a probability random optimization result and a preset receiving or rejecting criterion.
And when constructing an initial path solution of the mixed integer programming model, performing the initial path solution according to a preset path selection rule. In this embodiment, the preset path selection rule is to construct an initial path solution by first selecting a path, then calculating the insertion cost of the customer point, and then determining whether any customer point can be inserted, and finally determining whether a charging station is needed. Therefore, the method can quickly obtain the initial path solution and pave the path for the subsequent algorithm.
The operators can be set as desired, with each operator having its corresponding weight. That is, whether the random optimization result is the optimal value can be determined according to the preset operator weight and the evaluation probability of the operator iteration value. Specifically, the preset operators include, but are not limited to, a customer point insertion operator, a customer point removal operator, a charging station insertion operator, and a charging station removal operator. Firstly, constructing a plurality of destroying operators and rebuilding operators, namely inserting operators and removing operators, by utilizing the principle of destroying and rebuilding the established operators; then, the operators are put into an algorithm for iteration, and a current first path solution is searched. The destroy operator and the reconstruction operator are applied to the neighborhood expansion process in pairs, meanwhile, the operators are given certain use weight, and the selected probability is changed according to the weight. The method utilizes the self-adaptive neighborhood search algorithm to more effectively obtain the optimal path of the electric logistics vehicle, determines the use of an operator through an iteration value, and makes a corresponding solution for each condition appearing in the algorithm flow, so that the operation efficiency is greatly increased, and the time for obtaining the path solution is shortened. And an insertion and removal optimization mechanism of the self-adaptive neighborhood search algorithm is constructed, the electric logistics vehicle scheduling problem with a time window can be solved more effectively by adopting a destroying and rebuilding principle, the global optimization capability of an optimal path is improved, the optimal solution can effectively reduce energy consumption, and simultaneously the service time requirement of customers is better met, so that the logistics transportation cost is greatly reduced, and the requirement of a green supply chain is met.
After the first path solution is subjected to probability random optimization by using a simulated annealing algorithm, the distribution path information is obtained according to the probability random optimization result and a preset receiving or rejecting criterion. The receiving or rejecting criteria may be determined according to a preset threshold, for example, a threshold is set, and the algorithm is stopped only when both targets of the electric logistics vehicle scheduling problem with the time window meet the preset threshold. That is, the minimum travel distance is smaller than the preset travel distance value, and the route having the least amount of electricity and the shortest accumulated charging time is smaller than the preset route value having the least amount of electricity and the shortest accumulated charging time. But this approach tends to trap the algorithm into infinite loops.
Or the path solution obtained by determining the current probability random optimization result according to a preset threshold value, and if the path solution is rejected. The rejected path solution is used as a judgment condition and then optimized again. If the probability random optimizing result meets the preset receiving criterion, namely the second path solution meets the termination condition, taking the second path solution as the distribution path information; and if the probability random optimizing result meets the preset rejection criterion, namely the second path solution does not meet the termination condition, carrying out optimization calculation again. It can be understood that if the number of electronic logistics vehicles in the new solution is smaller than or the same as the number of electronic logistics vehicles in the current best solution, but the total distance of the new solution is shorter, the new solution is accepted. On the other hand, if more electric logistics car numbers are needed, the new solution is rejected. When the number of electric logistics vehicles is equal but the distance is long, the new solution is accepted with a certain probability. The specific process comprises the following steps;
obtaining a second path solution according to the probability random optimization result, and judging whether the second path solution meets the termination condition or not to carry out evaluation judgment;
if the second path solution does not meet the termination condition, taking the evaluation judgment result as a new termination condition, and changing the self-adaptive weight of each operator by using a self-adaptive neighborhood search algorithm according to the evaluation judgment result so as to adjust the initial iteration value; sequentially executing the adjusted initial iteration values corresponding to the preset operators, and optimizing the second path solution to obtain a third path solution; carrying out probability random optimization on the third path solution by using a simulated annealing algorithm to obtain a fourth path solution; continuously and circularly evaluating and judging until the probability is randomly optimized to obtain the current path solution which meets the last termination condition;
and if the second path solution meets the termination condition, taking the second path solution as the distribution path information.
The step of continuously and circularly evaluating and judging refers to that after evaluation and judgment, if the termination condition is not met all the time, new path solutions are obtained all the time through probability random optimization until the current path solutions obtained through probability random optimization meet the last termination condition. And determining to accept or reject the current solution by simulating an annealing algorithm, judging whether the predicted effect is achieved, changing the self-adaptive weight of each operator according to the algorithm searching effect, and performing dynamic algorithm adjustment. The method can carry out self-adaptive dynamic adjustment according to the historical performance of the first path, and can quickly obtain the optimal first path solution. The simulated annealing algorithm is used for accepting or rejecting the solution, and is used for searching the neighborhood structure in the main loop of the algorithm and further improving the best feasible solution, so that the time for obtaining the best feasible solution is effectively shortened, and the efficiency of the whole algorithm process is improved.
Firstly, a schematic diagram of electric logistics vehicle scheduling with a time window is described by way of example with reference to fig. 3; in the figure, an example involving eight customers (C1-C8), three charging stations (S1-S3) and a distribution center (D) that can also be used for charging is shown. The percentage values on the path show the battery charge status when the electric logistics vehicle arrives at the customer site or distribution center and leaves the charging station after the battery is recharged. The electric logistics vehicle B serves C6, C5 and C1 and returns to the distribution center without any charging. On the other hand, the electric logistics vehicle a visits S1 after servicing C2, C8, C3 and charges it before visiting C7 and C4, then charges it once again at S2, then visits C4 again, and finally returns to the distribution center. Wherein one charging station may pass through the same vehicle or a different vehicle and may be accessed multiple times, not necessarily each charging station.
The following describes in detail the optimization steps using the adaptive neighborhood search algorithm and the simulated annealing algorithm by way of example with reference to fig. 2, and may specifically include the following steps.
And S001, constructing an initial path solution.
And S011 and selecting a path.
A path is selected that is the shortest path from the distribution center to any customer site. It is determined whether all customer sites are serviced. If yes, the process is ended.
And S012, calculating the insertion cost.
The insertion cost of all customer points not served to the current path is calculated.
And S013, judging whether a client point can be inserted. If not, selecting a new path of the non-service customer points closest to the distribution center, and if so, selecting the customer point with the minimum distance added by the path and inserting the customer point into the current path.
And S014, judging whether a charging station is needed. If necessary, a power station swapping insertion algorithm that minimizes costs is performed. Jumping to S011.
And S002, carrying out a self-adaptive variable neighborhood searching process.
S021, defining an initial iteration value i ← 1, and removing the iteration value N of an operator by the initial charging stationSROid No. 1, iteration value N of the client point removal operatorRROid ← 1, sum N of numbers of operator iterations for client point insertion and removalC← 1, sum N of iterative values of charge station insertion and removal operatorsS←1。
S022, determining i and N at the momentSRIf the remainder of the division is constantly equal to 0, if yes, executing a charging station removal operator and removing the corresponding charging station, and enabling N to beSR←NSR+1，NS←NS+1, the removal operators include invalid charging station removal operators and full charging station removal operators, and the specific algorithm operations are respectivelyThe following were used:
invalid charging station removal: the main purpose of this operator is the electric quantity of make full use of electronic commodity circulation car, improves the availability factor of charging station. The present invention attempts to remove charging stations that contain higher amounts of power for electric logistics vehicle access. The charging stations are sequenced from large to small according to the electric quantity of the electric logistics vehicles. The present invention will remove theta charging stations in this sequence starting with the first charging station.
Removing the full charging station: the operator finds the charging stations in the solution path that are fully charged for the electric logistic vehicle and removes them randomly
Figure BDA0001642157210000131
And (4) respectively.
After removal of the charging station, an infeasible solution may result. Take the paths of fig. 5a and 5b as an example. FIG. 5a is a schematic diagram of the first charging station removal operation before, with the current feasible path solution being (D-C1-C2-S1-C3-S2-D). The percentages on the path represent the amount of power in arriving and leaving the node, while the numbers under the path show the arrival and departure times. When S1 is removed from the path, the electric logistics cart can still access C3 in the given sequence. However, since the battery level is insufficient, it takes longer for charging at S2, thereby delaying arrival at D. Fig. 5b is a schematic diagram after the first charge station removal operation, with arrival time at D being 370 longer than arrival time 350 before the charge station removal operation.
Taking fig. 6a and 6b as an example, it is explained how the situation that results in an infeasible solution occurs after the charging station is removed. Fig. 6a is a schematic diagram before the second charging station removing operation, and in the feasible path solution of fig. 6a, the electric logistic vehicle is charged at S1. Fig. 6b is a schematic diagram after the second charging station removing operation, and in fig. 6b, when S1 is removed, the battery power is returned to the distribution center in a state of negative level when leaving C3.
To do so, the charging station insertion algorithm is re-executed and the current path condition is restored, and let NS←NS+1, the insertion algorithm includes shortest distance charging station insertion algorithm, charging station insertion operator based on distance comparisonThe optimal charging station insertion operator comprises the following specific operator operations:
shortest distance charging station insertion operator: the algorithm determines the first customer point at which the vehicle arrives with a negative battery charge and inserts the charging station with the shortest added distance on the road segment between that customer point and the previous customer point. If such an insertion is not feasible, application to the previous road segment is attempted in the same way.
Charging station insertion operator based on distance comparison: the algorithm first determines the best charging station that caused the electric logistics vehicle to reach the road segment with negative customer electricity, and then compares the result with the situation that the corresponding best charging station is plugged into the previous road segment. The operation that minimizes the increase in the path distance will be performed. If both operators produce infeasible solutions, then the shortest distance charging station operator is used on the previous road segment.
Optimal charging station insertion operator: according to the method, firstly, road sections between a customer point visited when the electric quantity of the electric logistics vehicle is negative and a distribution center or a visited power exchange station are determined, then the road sections are traced back, and an optimal charging station insertion position is selected and executed. And after all the execution is finished, jumping to S024.
As shown in fig. 7a and 7b, the description is made before and after the charging station removing and charging station inserting operation. One possible path is depicted in fig. 7 a. If the charging station removal algorithm is used to delete S3, then S1 is inserted between C6 and D in order to maintain the feasibility of the time window and battery. The resulting path in fig. 7b is shorter than the initial path.
If not, selecting a customer point removal operator to remove the corresponding customer point, and enabling NRR←NRR+1，NC←NC+1, the customer point removal operator includes a customer point and front charging station removal operator and a customer point and rear charging station removal operator, and the two removal operators specifically operate as follows:
customer site and front charging station removal operators: the present invention will delete customer sites in the removal list while deleting their possible front-facing charging stations. The main purpose of this operator is to delete the previous swap station that is not necessary to access to save cost if the electric logistics vehicle does not access the deleted customer site any more, and the electric logistics vehicle does not affect the electric quantity of the electric logistics vehicle.
Removing operators of the customer points and the rear charging stations: the invention will remove customer sites on the list and potential charging stations behind them. Recharging may be required after leaving a customer site in order to be able to reach the next customer site in the path. However, if the leaving customer site is removed, charging is unnecessary and the corresponding charging station can be removed. After execution, the process jumps to S023.
S023, judging whether the solution obtained by using the destroy operator is feasible, if not, executing a charging station insert operator with the shortest distance, a customer point insert operator and a repair solution. Jump to S025.
S024, judging i and N at the momentRRWhether the remainder of the division is constantly equal to 0. If so, an attempt will be made to reduce the number of vehicles n used by iterationRRSelecting a random path deleting operator and a greedy path deleting operator as cycle times to delete corresponding client points, restoring the current path condition by using a client point inserting operator, and enabling N to be the order of NC←NC+1. The specific algorithm operates as follows:
the random path deletion operator randomly selects sigma paths and deletes all client points visited in those paths. σ depends on the number of paths in the current solution and is randomly determined between 10% and m of the total number of paths (10% < m < 100%). The greedy path deletion operator deletes the σ paths greedy. σ is determined in the same way as the random path deletion algorithm. The paths are ordered in order of the number of customer points served, from few to many, and σ paths are deleted from the first path in the sequence. The aim is to distribute the customers in the shorter path to other paths in the solution and reduce the number of active vehicles as much as possible. When σ is 2, the path change is represented as shown in fig. 4a and 4 b. FIG. 4a is a schematic diagram before a random path deletion operation, where the current feasible path solutions are (D-C5-S1-C2-C3-D), (D-C1-C5-D) and (D-C4-S2-C5-C6-D) before the random path deletion operation; fig. 4a is a schematic diagram after the random path deleting operation, and when σ is 2, the feasible path solutions after the random path deleting operation are (D-C5-S1-C2-C3-D) and (D-C4-S2-C5-C6-D).
The client insertion operators are divided into a shortest time insertion operator and a region insertion operator. In the shortest time insertion operator, the insertion cost is calculated as the difference between the total path time before and after the insertion of the customer point. For each customer point, the algorithm determines the best insertion position in all paths based on this insertion cost. The customer point with the least added path time is selected and inserted. This process is repeated for the remaining customer points until all customer points are inserted. The purpose of this operator is to increase the number of customer sites visited by the electronic logistics vehicles by combining time windows or distances of compatible customers.
In the region interpolation operator, when selecting a customer point, the operator uses the above time-shortest interpolation operator. However, rather than looking at all paths in the solution, only paths within a randomly selected region are considered. The region is determined in the same way as the region removal operator.
In order to determine the battery state of charge and the amount of charge of the charging station accessed during the implementation of the customer point insertion operator, the invention makes the following assumptions: the electric logistics vehicle leaves from the distribution center with full battery capacity, and if the electric logistics vehicle is charged at least once on the driving path, enough electric quantity returns to the distribution center. Therefore, if the electric logistics vehicle is charged only once in the travel path: (i) if a customer site is inserted between the distribution center and the charging station, the insertion only affects the battery state of charge when the charging station is reached; (ii) if the customer site is interposed between the charging station and the distribution center, the amount of charge is increased so that the electric logistics vehicle has enough electric power to return to the distribution center.
And S025, determining the acceptance or rejection criterion of the path solution by using simulated annealing.
The acceptance or rejection criteria for simulated annealing to determine the path solution are implemented as follows:
if the number of electric logistics vehicles in the new solution is less than that in the current best solutionThe new solution is accepted if the number of logistics vehicles or the number of logistics vehicles is the same but the total distance of the new solution is shorter. On the other hand, if more electric logistics car numbers are needed, the new solution is rejected. When the number of electric logistics vehicles is equal but the distance is long, the new solution is accepted with a certain probability. The probability is calculated as follows:
Figure BDA0001642157210000161
wherein f (R) represents the total distance to R, RNAnd RCRespectively representing the new solution and the current best solution, T representing the current temperature, and the initial setting of T being TiAnd at each iteration, using the formula T-T x lambda to reduce, lambda representing a cooling rate parameter with a value range of 0 < lambda <1, and using an initial temperature control parameter gamma to determine TiSo that a solution that is gamma% worse than the initial solution is accepted with a probability of 0.5.
And enabling i ← i +1 after accepting or rejecting the solution by using a simulated annealing algorithm.
S026, judging i and N at this timeCWhether the remainder of the division is constantly equal to 0. And if so, updating the adaptive weights of the customer point removal operator and the customer point insertion operator. And jumping to S028.
S027, judging i and N at this timeSWhether the remainder of the division is constantly equal to 0. And if so, updating the adaptive weights of the charging station removing operator and the charging station inserting operator.
And S028, judging whether a termination condition is met, if so, ending, otherwise, skipping to S022.
Based on the same inventive concept, the embodiment of the invention also provides an electric logistics vehicle dispatching system with a time window, and the implementation of the system can be realized by referring to the process of the method, and repeated parts are not described in detail.
Fig. 8 is a schematic structural diagram of an electric logistics vehicle dispatching system with a time window according to a second embodiment of the present invention, including a model building module 100, a constraint condition setting module 200, an optimizing module 300, and a dispatching module 400; the model establishing module 100 is configured to obtain distribution parameters of the electric logistics vehicle, and establish a mixed integer planning model according to the distribution parameters; the constraint condition setting module 200 is configured to obtain a planning constraint parameter of the electric logistics vehicle, and determine a planning constraint condition of the mixed integer planning model according to the planning constraint parameter; the planning constraint condition comprises a planning demand constraint condition and a charging constraint condition; the optimization module 300 is configured to perform optimization calculation on the mixed integer programming model by using a self-adaptive neighborhood search algorithm and a simulated annealing algorithm and combining a programming demand constraint condition and a charging constraint condition to obtain distribution path information; the scheduling module 400 is configured to complete optimization of scheduling of the electric logistics vehicles according to the delivery path information.
The invention provides an electric logistics vehicle dispatching system with a time window, which comprises a model establishing module 100, a constraint condition setting module 200, an optimizing module 300 and a dispatching module 400; the distribution route that can rationally arrange electronic logistics car accomplishes the dispatch, improves the efficiency with the help of electronic logistics car goods delivery service, thereby practices thrift the electric energy and further reduces the pollution to the environment.
In order to improve the operation efficiency, the constraint condition setting module 200 includes a filtering unit 210;
the screening unit 210 is configured to perform correlation screening on the planning constraint parameters and the distribution parameters in the mixed integer planning model, and establish planning constraint conditions according to the planning constraint parameters after the correlation screening.
Further, the constraint setting module 200 is configured to: the charging constraint condition comprises a charging state constraint condition and a charging time window constraint condition;
the mathematical formula for the state of charge constraint is expressed as:
Figure BDA0001642157210000171
wherein m represents the mth charging station; bmRepresenting the electric quantity state of the battery of the electric logistics vehicle when the electric logistics vehicle reaches the charging station m; z represents the capacity of the battery of the electric logistics vehicle, BmA decision variable representing the battery charging state of the battery of the electric logistics vehicle when the electric logistics vehicle leaves the charging station m; f representsA set of charging stations, wherein F' represents a virtual network point set generated according to F and allows each power exchanging station in the set F to be accessed for multiple times; in the set subscripts 0 and I +1, 0 represents the distribution center as a starting point, I +1 represents the distribution center as an end point, and each path starts from 0 and ends at I + 1; f'0F 'is a union of a distribution center and a virtual network point set'0＝F′Y{0}；
The mathematical formula for the charging time window constraint is expressed as:
Figure BDA0001642157210000172
where θ represents the start time of customer service, θmRepresents the start time, theta, of customer service at the mth charging stationnIndicating a start time of customer service at the nth charging station; p is a radical ofmnThe decision variables are binary decision variables, when the electric logistics vehicle passes through a road section from a client point m to n, the value is 1, otherwise, the value is 0; t is tmnRepresenting the time for the electric logistics vehicle to travel between the customer point m and the customer point n; r ismRepresents the service time at customer site m; l0Represents an upper limit of a time window when the distribution center point is used as a starting point; q represents a charge rate of the battery; d ═ 1.. times, I } represents a set of customer points, D0Representing the union of the starting point and the set of customer points, i.e. D, with the distribution centre as the starting point0DY {0 }; d 'represents a set of F' and D; d'I+1Indicates the end point, the union of the F 'and D sets, that is, D'I+1＝D′Y{I+1}。
Fig. 9 is a schematic structural diagram of the optimization module 300; comprises a construction unit 310, a definition unit 320 and a path optimization unit 330; the constructing unit 310 is configured to construct an initial path solution of the mixed integer programming model according to a preset path selection rule; the defining unit 320 is configured to use the adaptive neighborhood search algorithm as a basic frame, and process the mixed integer programming model by combining a programming requirement constraint condition and a charging constraint condition, so as to define an initial iteration value of each preset operator; the path optimization unit 330 is configured to sequentially execute initial iteration values corresponding to the preset operators, and optimize an initial path solution to obtain a first path solution; and carrying out probability random optimization on the first path solution by using a simulated annealing algorithm, and obtaining distribution path information according to a probability random optimization result and a preset receiving or rejecting criterion.
Fig. 10 is a schematic structural diagram of the path optimizing unit 330; comprises a judgment subunit 331, a loop optimization subunit 332 and an acceptance subunit 333; the judging subunit 331 is configured to obtain a second path solution according to the probabilistic random optimization result, and evaluate and judge whether the second path solution satisfies a termination condition; the loop optimization subunit 332 is configured to, if the second path solution does not satisfy the termination condition, take the evaluation determination result as a new termination condition, and change the adaptive weight of each operator by using an adaptive neighborhood search algorithm according to the evaluation determination result to adjust the initial iteration value; sequentially executing the adjusted initial iteration values corresponding to the preset operators, and optimizing the second path solution to obtain a third path solution; carrying out probability random optimization on the third path solution by using a simulated annealing algorithm to obtain a fourth path solution; continuously and circularly evaluating and judging until the probability is randomly optimized to obtain the current path solution which meets the last termination condition; the receiving subunit 333 is configured to use the second path solution as the distribution path information if the second path solution satisfies the termination condition.
Further, the model building module 100 is arranged to: the mixed integer programming model is:
Figure BDA0001642157210000181
wherein m represents the mth charging station; n represents an nth charging station; x is the number ofmnRepresents the distance from the mth customer point to the nth customer point; p is a radical ofmnThe decision variables are binary decision variables, when the electric logistics vehicle passes through a road section from a customer point m to a customer point n, the decision variables are 1, otherwise, the decision variables are 0; d'0Represents a union of a set of origin points, F 'and D, that is, D'0＝D′Y{0}。
Although the present invention has been described with reference to the preferred embodiments, it is not intended to limit the present invention, and those skilled in the art can make variations and modifications of the present invention without departing from the spirit and scope of the present invention by using the methods and technical contents disclosed above.

Claims (6)

1. A method for dispatching an electric logistics vehicle with a time window is characterized by comprising the following steps;
acquiring distribution parameters of the electric logistics vehicle, and establishing a mixed integer planning model according to the distribution parameters;
acquiring planning constraint parameters of the electric logistics vehicle, and determining planning constraint conditions of the mixed integer planning model according to the planning constraint parameters; the planning constraint condition comprises a planning demand constraint condition and a charging constraint condition;
performing optimization calculation on the mixed integer programming model by using a self-adaptive neighborhood search algorithm and a simulated annealing algorithm and combining the programming demand constraint condition and the charging constraint condition to obtain distribution path information, and the method comprises the following steps;
constructing an initial path solution of the mixed integer programming model according to a preset path selection rule;
taking an adaptive neighborhood search algorithm as a basic frame, and combining the planning demand constraint condition and the charging constraint condition to process the mixed integer planning model so as to define the initial iteration value of each preset operator; sequentially executing initial iteration values corresponding to all preset operators, and optimizing the initial path solution to obtain a first path solution;
carrying out probability random optimization on the first path solution by using a simulated annealing algorithm, and obtaining distribution path information according to a probability random optimization result and a preset receiving or rejecting criterion, wherein the method comprises the following steps;
obtaining a second path solution according to the probability random optimization result, and evaluating and judging whether the second path solution meets the termination condition;
if the second path solution does not meet the termination condition, taking an evaluation judgment result as a new termination condition, and changing the adaptive weight of each operator by using an adaptive neighborhood search algorithm according to the evaluation judgment result so as to adjust the initial iteration value; sequentially executing the adjusted initial iteration values corresponding to the preset operators, and optimizing the second path solution to obtain a third path solution; carrying out probability random optimization on the third path solution by using a simulated annealing algorithm to obtain a fourth path solution; continuously and circularly evaluating and judging until the probability is randomly optimized to obtain the current path solution which meets the last termination condition;
if the second path solution meets the termination condition, taking the second path solution as distribution path information;
completing the optimization of the dispatching of the electric logistics vehicles according to the distribution path information;
the method utilizing the self-adaptive neighborhood search algorithm and the simulated annealing algorithm further comprises the following steps of:
defining an initial iteration value i ← 1 and an iteration value N of an initial charging station removal operatorSROid No. 1, iteration value N of the client point removal operatorRROid ← 1, sum N of numbers of operator iterations for client point insertion and removalC← 1, sum N of iterative values of charge station insertion and removal operatorsS←1；
Determining i and N at the momentSRIf the remainder of the division is constantly equal to 0, if yes, executing a charging station removal operator and removing the corresponding charging station, and enabling N to beSR←NSR+1，NS←NS+1, the removal operators include an invalid charging station removal operator and a full charging station removal operator.
2. The method for dispatching electric logistics vehicles with time windows as claimed in claim 1, wherein the determining the planning constraint conditions of the mixed integer planning model according to the planning constraint parameters comprises the following steps;
and performing correlation screening on the planning constraint parameters and the distribution parameters in the mixed integer planning model, and establishing planning constraint conditions according to the planning constraint parameters after the correlation screening.
3. The method for dispatching electric logistics vehicles with time windows as set forth in claim 2, wherein the charging constraints comprise charging state constraints and charging time window constraints;
the mathematical formula for the state of charge constraint is represented as: bm≤Bm≤Z
Figure FDA0002934091490000021
Wherein m represents the mth charging station; bmRepresenting the electric quantity state of the battery of the electric logistics vehicle when the electric logistics vehicle reaches the charging station m; z represents the capacity of the battery of the electric logistics vehicle, BmA decision variable representing the battery charging state of the battery of the electric logistics vehicle when the electric logistics vehicle leaves the charging station m; f represents a set of charging stations, F' represents a virtual network point set generated according to F, and each power conversion station in the set F is allowed to be accessed for multiple times; in the set subscripts 0 and I +1, 0 represents the distribution center as a starting point, I +1 represents the distribution center as an end point, and each path starts from 0 and ends at I + 1; f'0F 'is a union of a distribution center and a virtual network point set'0＝F′∪{0}；
The mathematical formula of the charging time window constraint is expressed as:
Figure FDA0002934091490000022
where θ represents the start time of customer service, θmRepresents the start time, theta, of customer service at the mth charging stationnIndicating a start time of customer service at the nth charging station; p is a radical ofmnAs binary decision variables, as electrokinetic logisticsThe value of the vehicle passing through the road section from the client point m to the client point n is 1, otherwise, the value is 0; t is tmnRepresenting the time for the electric logistics vehicle to travel between the customer point m and the customer point n; r ismRepresents the service time at customer site m; l0Represents an upper limit of a time window when the distribution center point is used as a starting point; q represents a charge rate of the battery; d ═ 1.. times, I } represents a set of customer points, D0Representing the union of the starting point and the set of customer points, i.e. D, with the distribution centre as the starting point0Tuo {0 }; d 'represents a set of F' and D; d'I+1Indicates the end point, the union of the F 'and D sets, that is, D'I+1＝D′∪{I+1}。
4. The method for dispatching electric logistics vehicles with time windows as claimed in claim 3, wherein the mixed integer programming model is:
Figure FDA0002934091490000023
wherein m represents the mth charging station; n represents an nth charging station; x is the number ofmnRepresents the distance from the mth customer point to the nth customer point; p is a radical ofmnThe decision variables are binary decision variables, when the electric logistics vehicle passes through a road section from a customer point m to a customer point n, the decision variables are 1, otherwise, the decision variables are 0; d'0Represents a union of node sets of origin, F 'and D, i.e., D'0＝D′∪{0}。
5. The electric logistics vehicle dispatching system with the time window is characterized by comprising a model establishing module, a constraint condition setting module, an optimizing module and a dispatching module;
the model establishing module is used for acquiring distribution parameters of the electric logistics vehicle and establishing a mixed integer planning model according to the distribution parameters;
the constraint condition setting module is used for acquiring planning constraint parameters of the electric logistics vehicle and determining planning constraint conditions of the mixed integer planning model according to the planning constraint parameters; the planning constraint condition comprises a planning demand constraint condition and a charging constraint condition;
the optimization module is used for carrying out optimization calculation on the mixed integer programming model by utilizing a self-adaptive neighborhood search algorithm and a simulated annealing algorithm and combining the programming demand constraint condition and the charging constraint condition to obtain distribution path information;
the optimization module comprises a construction unit, a definition unit and a path optimization unit;
the construction unit is used for constructing an initial path solution of the mixed integer programming model according to a preset path selection rule;
the defining unit is used for processing the mixed integer programming model by taking the adaptive neighborhood search algorithm as a basic frame and combining the programming demand constraint condition and the charging constraint condition so as to define the initial iteration value of each preset operator; sequentially executing initial iteration values corresponding to all preset operators, and optimizing the initial path solution to obtain a first path solution;
the path optimization unit is used for carrying out probability random optimization on the first path solution by using a simulated annealing algorithm and obtaining distribution path information according to a probability random optimization result and a preset receiving or rejecting criterion;
the path optimization unit comprises a judgment subunit, a circulation optimization subunit and an acceptance subunit;
the judging subunit is configured to obtain a second path solution according to the probabilistic random optimization result, and evaluate and judge whether the second path solution satisfies a termination condition;
the loop optimization subunit is configured to, if the second path solution does not satisfy the termination condition, take an evaluation judgment result as a new termination condition, and change the adaptive weight of each operator by using an adaptive neighborhood search algorithm according to the evaluation judgment result to adjust the initial iteration value; sequentially executing the adjusted initial iteration values corresponding to the preset operators, and optimizing the second path solution to obtain a third path solution; carrying out probability random optimization on the third path solution by using a simulated annealing algorithm to obtain a fourth path solution; continuously and circularly evaluating and judging until the probability is randomly optimized to obtain the current path solution which meets the last termination condition;
the receiving subunit is configured to, if the second path solution satisfies a termination condition, use the second path solution as distribution path information;
the dispatching module is used for finishing the optimization of the dispatching of the electric logistics vehicles according to the distribution path information;
the method utilizing the self-adaptive neighborhood search algorithm and the simulated annealing algorithm further comprises the following steps of:
defining an initial iteration value i ← 1 and an iteration value N of an initial charging station removal operatorSROid No. 1, iteration value N of the client point removal operatorRROid ← 1, sum N of numbers of operator iterations for client point insertion and removalC← 1, sum N of iterative values of charge station insertion and removal operatorsS←1；
Determining i and N at the momentSRIf the remainder of the division is constantly equal to 0, if yes, executing a charging station removal operator and removing the corresponding charging station, and enabling N to beSR←NSR+1，NS←NS+1, the removal operators include an invalid charging station removal operator and a full charging station removal operator.
6. The time-windowed electronic logistics vehicular dispatch system of claim 5, wherein the constraint condition setting module is configured to: the charging constraint condition comprises a charging state constraint condition and a charging time window constraint condition;
the mathematical formula for the state of charge constraint is represented as: bm≤Bm≤Z
Figure FDA0002934091490000041
Wherein m represents the mth charging station; bmRepresenting the electric quantity state of the battery of the electric logistics vehicle when the electric logistics vehicle reaches the charging station m; z represents the capacity of the battery of the electric logistics vehicle, BmThe decision variable represents that the electric logistics vehicle is powered when the electric logistics vehicle leaves the charging station mA battery state of charge of the battery; f represents a set of charging stations, F' represents a virtual network point set generated according to F, and each power conversion station in the set F is allowed to be accessed for multiple times; in the set subscripts 0 and I +1, 0 represents the distribution center as a starting point, I +1 represents the distribution center as an end point, and each path starts from 0 and ends at I + 1; f'0F 'is a union of a distribution center and a virtual network point set'0＝F′∪{0}；
The mathematical formula of the charging time window constraint is expressed as:
Figure FDA0002934091490000042
where θ represents the start time of customer service, θmRepresents the start time, theta, of customer service at the mth charging stationnIndicating a start time of customer service at the nth charging station; p is a radical ofmnThe decision variables are binary decision variables, when the electric logistics vehicle passes through a road section from a client point m to n, the value is 1, otherwise, the value is 0; t is tmnRepresenting the time for the electric logistics vehicle to travel between the customer point m and the customer point n; r ismRepresents the service time at customer site m; l0Represents an upper limit of a time window when the distribution center point is used as a starting point; q represents a charge rate of the battery; d ═ 1.. times, I } represents a set of customer points, D0Representing the union of the starting point and the set of customer points, i.e. D, with the distribution centre as the starting point0Tuo {0 }; d 'represents a set of F' and D; d'I+1Indicates the end point, the union of the F 'and D sets, that is, D'I+1＝D′∪{I+1}。
CN201810385664.2A 2018-04-26 2018-04-26 Electric logistics vehicle scheduling method and system with time window Active CN108764777B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810385664.2A CN108764777B (en) 2018-04-26 2018-04-26 Electric logistics vehicle scheduling method and system with time window

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810385664.2A CN108764777B (en) 2018-04-26 2018-04-26 Electric logistics vehicle scheduling method and system with time window

Publications (2)

Publication Number Publication Date
CN108764777A CN108764777A (en) 2018-11-06
CN108764777B true CN108764777B (en) 2021-03-30

Family

ID=64011878

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810385664.2A Active CN108764777B (en) 2018-04-26 2018-04-26 Electric logistics vehicle scheduling method and system with time window

Country Status (1)

Country Link
CN (1) CN108764777B (en)

Families Citing this family (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109559078B (en) * 2018-11-23 2021-10-08 南方科技大学 Vehicle scheduling method, device, equipment and storage medium
CN109583650B (en) * 2018-11-30 2021-03-30 浙江工商大学 Electric vehicle battery replacement station site selection and logistics distribution joint scheduling method
CN109919365B (en) * 2019-02-19 2020-12-01 清华大学 Electric vehicle path planning method and system based on double-strategy search
CN110309946A (en) * 2019-05-09 2019-10-08 上汽安吉物流股份有限公司 Logistics route method and device for planning, computer-readable medium and logistics system
CN111950950A (en) * 2019-05-17 2020-11-17 北京京东尚科信息技术有限公司 Order distribution path planning method and device, computer medium and electronic equipment
CN112529487A (en) * 2019-09-19 2021-03-19 北京京东振世信息技术有限公司 Vehicle scheduling method, device and storage medium
CN110942184B (en) * 2019-11-14 2022-08-19 南方科技大学 Self-adaptive addressing route-finding planning method, device, equipment and storage medium
CN111216571A (en) * 2019-11-21 2020-06-02 长沙理工大学 Battery-replacement type electric automobile navigation method participating in real-time logistics distribution
CN110956325B (en) * 2019-11-29 2022-07-22 浙江工业大学 Electric vehicle path planning method with time window
CN111178730A (en) * 2019-12-24 2020-05-19 中国航空工业集团公司西安飞机设计研究所 Method and device for planning supply of oiling machine
CN111144647B (en) * 2019-12-25 2021-06-08 华院计算技术(上海)股份有限公司 General vehicle path planning method and system based on large-scale neighborhood search algorithm
CN111160654B (en) * 2019-12-31 2022-06-24 哈工汇智(深圳)科技有限公司 Transportation path optimization method for reducing total cost based on fuzzy C-means-simulated annealing algorithm
JP7259774B2 (en) * 2020-01-15 2023-04-18 Ｊｆｅスチール株式会社 Delivery plan creation method and delivery plan creation device
CN111311158B (en) * 2020-03-04 2023-08-18 西华大学 Electric logistics vehicle path planning method under limited charging facility condition
CN111325409B (en) * 2020-03-09 2022-11-22 西南交通大学 Method and system for site selection of battery replacement station and route planning of hybrid fleet
CN111432449B (en) * 2020-03-26 2023-02-10 沈阳理工大学 Industrial wireless chargeable sensor network charging scheduling method based on new particle swarm
CN111598343A (en) * 2020-05-18 2020-08-28 武汉轻工大学 Distribution path optimization method, device and readable storage medium
CN111985676A (en) * 2020-06-28 2020-11-24 济南浪潮高新科技投资发展有限公司 Method and equipment for planning transportation line of electric truck
CN112488358B (en) * 2020-10-31 2023-04-07 海南电网有限责任公司 Electric vehicle charging path planning method and storage medium
CN112733272A (en) * 2021-01-13 2021-04-30 南昌航空大学 Method for solving vehicle path problem with soft time window
CN113592148B (en) * 2021-07-01 2024-03-15 合肥工业大学 Optimization method and system for improving delivery route of vehicle and unmanned aerial vehicle
CN117162845B (en) * 2023-11-01 2023-12-29 南通国轩新能源科技有限公司 Movable energy storage charging method and device

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104331743A (en) * 2014-10-11 2015-02-04 清华大学 Electric vehicle travel planning method based on multi-target optimization
CN106251009A (en) * 2016-07-27 2016-12-21 清华大学 A kind of optimized algorithm of the Vehicle Routing Problems solving time window
CN106503836A (en) * 2016-10-09 2017-03-15 电子科技大学 A kind of pure electric automobile logistics distribution Optimization Scheduling of multiple-objection optimization
CN107194513A (en) * 2017-05-26 2017-09-22 中南大学 A kind of optimization method for solving full channel logistics distribution

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104331743A (en) * 2014-10-11 2015-02-04 清华大学 Electric vehicle travel planning method based on multi-target optimization
CN106251009A (en) * 2016-07-27 2016-12-21 清华大学 A kind of optimized algorithm of the Vehicle Routing Problems solving time window
CN106503836A (en) * 2016-10-09 2017-03-15 电子科技大学 A kind of pure electric automobile logistics distribution Optimization Scheduling of multiple-objection optimization
CN107194513A (en) * 2017-05-26 2017-09-22 中南大学 A kind of optimization method for solving full channel logistics distribution

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
基于时间窗约束的纯电动汽车路径规划模型分析;刘颖鑫;《物流科技》;20171231;第83-87页 *
多车场带时间窗车辆路径问题的变邻域搜索算法;王征 等;《中国管理科学》;20110430;第99-109页 *

Also Published As

Publication number Publication date
CN108764777A (en) 2018-11-06

Similar Documents

Publication Publication Date Title
CN108764777B (en) Electric logistics vehicle scheduling method and system with time window
Guo et al. The battery charging station location problem: Impact of users’ range anxiety and distance convenience
CN109034481B (en) Constraint programming-based vehicle path problem modeling and optimizing method with time window
James Two-stage request scheduling for autonomous vehicle logistic system
Li-ying et al. Multiple charging station location-routing problem with time window of electric vehicle.
Cheung et al. Dynamic routing model and solution methods for fleet management with mobile technologies
CN107766994A (en) A kind of shared bicycle dispatching method and scheduling system
CN111144568A (en) Multi-target urban logistics distribution path planning method
CN112801347B (en) Multi-target city two-stage distribution planning method based on mobile transfer station and crowdsourcing
Ma et al. Rebalancing stochastic demands for bike-sharing networks with multi-scenario characteristics
CN110909952B (en) City two-stage distribution and scheduling method with mobile distribution station
CN110097218B (en) Unmanned commodity distribution method and system in time-varying environment
Grzybowska et al. Decision support system for real-time urban freight management
CN111860957B (en) Multi-vehicle-type vehicle path planning method considering secondary distribution and balancing
CN113848970B (en) Multi-target cooperative path planning method for vehicle-unmanned aerial vehicle
Zhang et al. A data-driven dynamic repositioning model in bicycle-sharing systems
CN111445094B (en) Express vehicle path optimization method and system based on time requirement
Wallar et al. Optimizing multi-class fleet compositions for shared mobility-as-a-service
CN114444843A (en) Agricultural product green logistics distribution vehicle scheduling method and system based on large-scale variable neighborhood search strategy
CN114358233A (en) Multi-AGV path planning problem optimization method and system based on double-hybrid particle swarm
Murakami Formulation and algorithms for route planning problem of plug-in hybrid electric vehicles
Oda et al. Distributed fleet control with maximum entropy deep reinforcement learning
Xiao et al. A diversity-enhanced memetic algorithm for solving electric vehicle routing problems with time windows and mixed backhauls
Liu et al. An adaptive large neighborhood search method for rebalancing free-floating electric vehicle sharing systems
CN114841582A (en) Truck and unmanned aerial vehicle cooperative distribution method

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