CN105956681B - A kind of Drop and pull transport dynamic path planning method based on rolling time horizon optimization - Google Patents
A kind of Drop and pull transport dynamic path planning method based on rolling time horizon optimization Download PDFInfo
- Publication number
- CN105956681B CN105956681B CN201610241242.9A CN201610241242A CN105956681B CN 105956681 B CN105956681 B CN 105956681B CN 201610241242 A CN201610241242 A CN 201610241242A CN 105956681 B CN105956681 B CN 105956681B
- Authority
- CN
- China
- Prior art keywords
- node
- tractor
- time
- delta
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 40
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000005096 rolling process Methods 0.000 title claims abstract description 32
- 239000011159 matrix material Substances 0.000 claims abstract description 22
- 238000002922 simulated annealing Methods 0.000 claims abstract description 8
- 230000000694 effects Effects 0.000 abstract description 2
- 238000005070 sampling Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 230000003068 static effect Effects 0.000 description 4
- 230000007547 defect Effects 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
-
- 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
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Quality & Reliability (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of Drop and pull transport dynamic path planning method based on rolling time horizon optimization, it is characterized in that comprising the following steps:1st, fleet is divided into several time slices the activity duration using roller and the method for optimization;2nd, the interior distribution cost matrix of following several timeslices is obtained in each timing node;3rd, the Drop and pull transport active path planning model based on rolling time horizon optimization is established using distribution cost matrix;4th, the optimal case in the following several timeslices of Drop and pull transport active path planning model solution is utilized;5th, initial solution is generated at random, and solution is iterated using simulated annealing to initial solution, so as to obtain optimal solution, and performed.The present invention can carry out structured decision to Drop and pull transport active path planning problem, and quickly provide the scheme of active path planning, improve real-time, reasonability and the accuracy of path planning scheme, and reduce cost of transportation.
Description
Technical Field
The invention relates to a rolling type and optimization method, in particular to a drop-and-pull transport dynamic path planning method based on rolling time domain optimization, and belongs to the field of combination optimization, dynamic decision or dynamic planning.
Background
The drop and pull transportation means that the tractor unloads the trailer carried by the tractor at a cargo loading and unloading operation point and replaces the trailer with other trailers to continue operation in a logistics organization mode. The swing-hanging transportation is widely popularized and applied internationally, practice proves that the swing-hanging transportation is an efficient, green and advanced cargo transportation organization mode, and the swing-hanging transportation has important significance for reducing logistics cost, promoting modern logistics development and improving national overall economic operation quality.
The problem of planning the static drop-and-pull transport path can be described as follows: and (4) towing the trailer by using the tractor, visiting the drop-off and hang-up transportation user from the yard according to a plan, and returning to the yard after finishing all visiting requirements. The requirements of the drop and pull transport users can be divided into two types: requiring tractor access alone and tractor-trailer towing co-access. Meanwhile, the towing operation of the tractor can be completed only at a specific station. The problem is to minimize the total delivery cost while satisfying the relevant constraints by constructing a suitable vehicle travel route and a hang-off method.
At present, the related research on the problem of container fixed-hanging transportation at home and abroad is more, but the literature for specially researching the problem of swing-hanging transportation scheduling is relatively less; meanwhile, the research on the drop-and-pull transportation scheduling problem mainly takes path planning and transportation mode design in a static environment as a main part, relatively few researches are carried out in dynamic and uncertain environments, and the cooperativity among all parts of a distribution system in the scheduling process is not fully considered, so that the research result is difficult to effectively respond to the influence of the actual distribution environment change. The core human task of dynamic scheduling is to better realize the cooperation of the scheme and the implementation and obtain the best actual execution effect when the scheduling scheme faces a time-varying and complex actual implementation environment.
Disclosure of Invention
In order to reduce the deviation between the actual running process of a motorcade and a path planning scheme and overcome the defects of the conventional path planning method, the invention provides a drop and hang transportation dynamic path planning method based on rolling time domain optimization, so that the structured decision can be carried out on the drop and hang transportation dynamic path planning problem based on the rolling time domain optimization, and the dynamic path planning scheme can be rapidly provided, thereby improving the instantaneity, reasonability and accuracy of the path planning scheme and reducing the transportation cost.
The invention adopts the following technical scheme for solving the technical problems:
the invention relates to a drop and pull transport dynamic path planning method based on rolling time domain optimization, which is characterized by comprising the following steps: is applied to a target node U consisting of 1 yard A and R 1 V road network nodes U 2 In a distribution service area formed by position nodes S of P tractors, setting one vertex of an external square of the distribution service area as a coordinate origin o, and setting two adjacent sides in the horizontal and vertical directions connected with the coordinate origin o as an x axis and a y axis respectively so as to form a rectangular coordinate system xoy; in the coordinate system xoy:
define the location of yard A as (x) 0 ,y 0 ) Defining the position (x) of each of said P tractors from said yard A 0 ,y 0 ) Starting; recording the P tractors as K = { K = (1) ,K (2) ,…,K (p) ,…,K (P) },1≤p≤P,K (p) Representing the p-th tractor;
record R target nodes as Representing the r-th target node, the r-th target nodeIs denoted by (x) r ,y r ) (ii) a The r < th > target nodeIs recorded as e as the earliest access time r (ii) a The r < th > target nodeIs given as the latest access time of r ;e r And l r Time window [ e ] constituting a target node r ,l r ];
Marking V road network nodes as Representing a nth road network node; the v road network nodeIs denoted by (x) R+v ,y R+v );
Marking the position node of P tractors K as S = { S = { S } (R+V+1) ,S (R+V+2) ,…,S (R+V+p) ,…,S (R+V+P) },S (R +V+p) Represents the R + V + p nodes and represents the p tractor K (p) The location of the device; p-th tractor K (p) At the location node S (R +S+p) Is denoted by (x) R+V+p ,y R+V+p );
The 1 yard A, R target nodes U 1 V road network nodes U 2 And P nodes S of the positions of the tractors are recorded as a point set U = { A, U = (N, N) } 1 ,U 2 And S }, then the yard A represents the 0 th node; r target nodes U 1 Represents the 1 st node to the R < th > node; v road network nodes U 2 Represents the R +1 th to the R + S th nodes; the node S where the P tractors are located represents an R + V +1 th node to an R + V + P th node;
definition edge set E = &<i,j>, | i, j ∈ U, i ≠ j } represents a set of straight-line paths of any two nodes in the point set U,<i,j> represents a straight line path between any ith node and any jth node; recording a distribution cost matrix between any two nodes in the point set U as C, wherein C = { C = { (C) } ij |i,j∈U,i≠j};c ij Representing the distribution cost between the ith node and the jth node;
the drop-and-pull transport dynamic path planning method based on rolling time domain optimization is carried out according to the following steps:
step 1, calculating the working time [0, max l ] of all tractors K r ]Equally spaced and divided into N continuous time segments, which are recorded as { [0, delta t { ]],[Δt,2Δt],…,[(n-1)Δt,nΔt],…,[(N-1)Δt,maxl r ]}; wherein [ (n-1) Δ t, n Δ t]Represents the nth time segment; Δ t represents the divided interval;
step 2, letting n =1; planning an actual driving route of the vehicle in the nth to the nth + m time periods [ (n-1) delta t, (n + m) delta t ]; m is more than or equal to 1 and less than or equal to N; n is the total number of time segments;
step 3, enabling the nth to the nth + mth time periods [ (n-1) delta t, (n + m) delta t]Inner p-th tractor K (p) Node at initial positionInitializing initial position nodes of all tractors;
step 4, obtaining the n-th to the n + m-th time periods [ (n-1) delta t, (n + m) delta t]Distribution cost between the ith node and the jth nodeThereby obtaining a distribution cost matrix C n ;
Step 5, according to the p-th tractor K (p) The target node to be accessed falls into the time periods from the n th to the n + m th according to the sequence of the earliest access time of the target node to be accessed, wherein the time periods from (n-1) delta t to (n + m) delta t]All target nodes in the system form a p-th tractor K in sequence (p) Initial access sequence of
Step 6, taking out the p-th tractor K (p) Initial access sequence ofThe last target node in (2) is taken as the destination node and is recorded asThereby forming an updated access sequence
Step 7, according to the p-th tractor K (p) Node of home positionDistribution cost matrix C n Access sequenceDestination nodeEstablishing a drop and pull transport dynamic path planning model based on rolling time domain optimization;
step 8, randomly generating the nth initial solution X n ;
Step 9, utilizing simulated annealing algorithm to solve X for the nth initial solution n Carrying out iterative solution to obtain the nth optimal solution;
step 10, assigning n +1 to n;
step 11, updating the n-th to the n + m-th time periods [ (n-1) delta t, (n + m) delta t]Inner p-th tractor K (p) Node of the positionDistribution cost matrix C n ;
Step 12, in the (n-1) th time period [ (n-2) delta t, (n-1) delta t]Before finishing, acquiring the p-th tractor K (p) During the period from n to n + m [ (n-1) Δ t, (n + m) Δ t]Access sequence withinThe access sequenceThe method comprises the following steps: (n-1) th period of time [ (n-2) Δ t, ((n-2))n-1)Δt]Target node with incomplete access and p-th tractor K (p) The earliest access time of the target node to be accessed falls within the n-th to n + m-th time periods [ (n-1) Δ t, (n + m) Δ t]The target node of (2);
step 13, taking out the n-th to the n + m-th time periods [ (n-1) delta t, (n + m) delta t]Inner p-th tractor K (p) Access sequence ofAs a destination nodeThereby forming an updated access sequence
Step 14, according to the p-th tractor K (p) Node of home positionDistribution cost matrix C n Access sequenceDestination nodeInputting the dynamic path planning model of drop and pull transportation;
step 15, randomly generating the nth initial solution X n ,
Step 16, utilizing a simulated annealing algorithm to solve the nth initial solution X n Carrying out iterative solution to obtain the nth optimal solution;
step 17, judging whether N = N is established, and if so, indicating that all tractors K complete the operation; otherwise, return to step 10.
The drop-off and hang-up transportation dynamic path planning method based on rolling time domain optimization is characterized in that the model of the drop-off and hang-up transportation dynamic path planning based on rolling time domain optimization in the step 7 is as follows:
(1) An objective function:
(2) Constraint conditions
Equation (1) is an objective function, which represents the minimization of the total delivery cost of the tractor K; when in useWhen it is, the p-th tractor K is shown (p) Passing through a straight-line path (i, j) between the ith node and the jth node whenWhen, represents the p-th tractor K (p) Does not pass through a straight-line path (i, j) between the ith node and the jth node;
the expression (2) represents the p-th tractor K (p) Must be from the location nodeStarting; when in useWhen it is, the p-th tractor K is shown (p) Passing through location nodeAnd the jth nodeWhen in useWhen it is, the p-th tractor K is shown (p) Non-transit location nodeAnd the jth node
Equation (3) indicates that the p-th tractor must reach the destination nodeWhen in useWhen it is, the p-th tractor K is shown (p) Passing through the ith node and the destination nodeThe path betweenWhen the temperature is higher than the set temperatureWhen, represents the p-th tractor K (p) Does not pass through the ith node and the destination nodePath between
Equation (4) is a time window constraint and represents that the p-th tractor K is driven (p) At the location nodeThe Time of (1) is the current Time;showing the p-th tractor K (p) Arrival position nodeThe time of (d);
the formula (5) is as shownWhen the vehicle is in the right position, the p-th tractor K is obtained (p) Access time AT jth node, AT i (p) Showing the p-th tractor K (p) Time of arrival at the ith node;w i showing the p-th tractor K (p) Latency at the ith node; t is t ij Showing the p-th tractor K (p) The travel time from the ith node to the jth node;showing the p-th tractor K (p) Time of arrival at the jth node;
equation (6) represents the pth tractor K, as a time window constraint (p) To the r-th target nodeAt the earliest time of access e r And latest access time l r In the middle of;
equation (7) represents all the road network nodes U 2 And access sequenceThe target node in (2) satisfies the balance of access and exit;showing the p-th tractor K (p) A straight-line path (i, w) passing between the ith node and the w-th node&WhenWhen it is, the p-th tractor K is shown (p) Does not pass through a straight-line path between the ith node and the w-th node<i,w> when in useWhen, represents the p-th tractor K (p) A straight line path passing through the w-th node and the j-th node<w,j&WhenWhen, represents the p-th tractor K (p) The straight-line path does not pass through the w node and the j node;
equation (8) represents the access sequenceVisiting order constraint of the middle node; the describedShowing the p-th tractor K (p) Reach access sequenceTime of the z-th target node; the above-mentionedShowing the p-th tractor K (p) Reach access sequenceTime of the z +1 th target node;
equation (9) represents the access sequenceThe node in (1) is visited by one traction train only once; when in useWhen it is, the p-th tractor K is shown (p) A straight-line path passing between the ith node and the u-th node<i,u&WhenWhen it is, the p-th tractor K is shown (p) Does not pass through a straight-line path between the ith node and the u-th node<i,u> when in useWhen it is, the p-th tractor K is shown (p) A straight line path passing through the u-th node and the j-th node<u,j&WhenWhen it is, the p-th tractor K is shown (p) Does not pass through a straight-line path between the u-th node and the j-th node<u,j>;
The equation (10) indicates that the decision variable takes a value of "0" or "1".
Compared with the prior art, the invention has the beneficial effects that:
1. compared with a static drop and pull transport path planning method, the invention provides a rolling time domain optimization method applied to solve the problem of dynamic drop and pull transport vehicle scheduling, and breaks through the limitation of the original static path planning method. The drop and pull transport dynamic path planning method based on rolling time domain optimization overcomes the control defects in the path planning scheme execution process, so that the path planning scheme can be dynamically generated in real time, and the path planning scheme can better meet the actual execution environment.
2. The invention provides a rolling time domain optimization method framework. The rolling finite time domain optimization strategy is adopted in the rolling finite time domain optimization method, namely, the optimization strategy of the whole task process is not finished off line at one time, but is repeatedly carried out on line along with the time. The specific path optimization strategy is carried out in a future limited time segment (optimization time window) based on the current sampling moment, and the path optimization result is executed in the optimization time window based on the sampling time. When the next sampling moment comes, the optimization time window can move simultaneously, so that the dynamic distribution cost is introduced by the rolling time domain optimization, and the dynamic performance of the rolling time domain optimization is that dynamic cost matrixes at different sampling moments can also change at moments. In order to simplify the model, the invention provides that starting from the execution of the mission by the fleet, a fixed sampling moment is provided per interval Δ t.
3. The invention designs and establishes a dynamic vehicle path planning method model for drop and pull transportation, so that the model can be solved in a standardized way in each target optimization time window. The solution algorithm of the model is a simulated annealing algorithm which has the characteristics of strong robustness and universality, so that an optimized path planning scheme can be rapidly provided.
4. The dynamic planning method for the drop-and-pull transport dynamic path based on the rolling time domain optimization is realized by dynamically generating the distribution cost matrix. On the basis of the fact that the road surface cost under the condition of smooth road surface is defined in advance, the dynamic path cost matrix meeting the current condition can be obtained by combining the traffic capacity and the prediction coefficient of the condition of the road section in a period of time in the future. The dynamic property of the model can be well reflected by applying the dynamic path cost matrix.
Drawings
FIG. 1 is a flow chart of a method of practicing the present invention;
fig. 2 is a schematic diagram of a drop and pull transport dynamic path planning based on rolling time domain optimization.
Detailed Description
In this embodiment, a drop and pull transportation dynamic path planning method based on rolling time domain optimization is applied to a yard comprising 1 yard a and R target nodes U 1 V road network nodes U 2 In a distribution service area formed by the position nodes S of the P tractors, one vertex of an external square of the distribution service area is set as a coordinate origin o, and two adjacent sides in the horizontal and vertical directions connected with the coordinate origin o are respectively set as an x axis and a y axis, so that a rectangular coordinate system xoy is formed; in the coordinate system xoy:
define the location of yard A as (x) 0 ,y 0 ) Defining the location (x) of P tractors each from yard A 0 ,y 0 ) Starting; let P tractors as K = { K = { (K) (1) ,K (2) ,…,K (p) ,…,K (P) },1≤p≤P,K (p) Representing the p-th tractor; the P tractors correspond to the position nodes where the P tractors are located one by one;
record R target nodes as Representing the r-th target node, the r-th target nodeIs denoted by (x) r ,y r ) (ii) a The r < th > target nodeIs recorded as e as the earliest access time r (ii) a The r-th target nodeIs recorded as the latest access time of l r ;e r And l r Time window [ e ] constituting a target node r ,l r ];
Marking V road network nodes as Representing a nth road network node; the v road network nodeIs denoted by (x) R+v ,y R+v );
Marking the position node of P tractors K as S = { S = { S } (R+V+1) ,S (R+V+2) ,…,S (R+V+p) ,…,S (R+V+P) },S (R +V+p) Represents the R + V + p nodes and represents the p tractor K (p) The location of the device; p-th tractor K (p) At the location node S (R +S+p) Is denoted by (x) R+V+p ,y R+V+p ) (ii) a p represents the number of tractors and also represents the number of nodes at which the tractors are located;
1 yard A, R target nodes U 1 V road network nodes U 2 And P nodes S of the positions of the tractors are recorded as a point set U = { A, U = (N, N) } 1 ,U 2 And S, the parking lot A represents the 0 th node; r target nodes U 1 Represents the 1 st node to the R < th > node; v road network nodes U 2 Represents the R +1 th node to the R + S th node; the node S where the P tractors are located represents an R + V +1 th node to an R + V + P th node; a total of R + V + P nodes in the network;
defining edge set E =:<i,j&| i, j belongs to U, i is not equal to j } represents a linear path set of any two nodes in the point set U, and < i, j > represents a linear path between any ith node and any jth node; the distribution cost matrix between any two nodes in the point set U is marked as C, and C = { ci = j |i,j∈U,i≠j};ci j Representing the distribution cost between the ith node and the jth node;
in this embodiment, as shown in fig. 1, the drop-and-pull transport dynamic path planning method based on rolling time domain optimization is performed according to the following steps:
step 1, calculating the working time [0, max l ] of all tractors K r ]Equally spaced, dividing into N successive time segments, with the tractor visiting the last target node as the end of the entire task, where max l r Represents the latest possible ending time of all tasks, and is marked as { [0, Δ t { ]],[Δt,2Δt],…,[(n-1)Δt,nΔt],…,[(N-1)Δt,max l r ]}; wherein [ (n-1) Δ t, n Δ t]Represents the nth time segment; Δ t represents the divided interval;
step 2, letting n =1; planning an actual driving route of the vehicle in the nth to the nth + m time periods [ (n-1) delta t, (n + m) delta t ]; m is more than or equal to 1 and less than or equal to N; n is the total number of time segments; the value of m represents a target node in m +1 time periods of each rolling plan;
step 3, enabling the nth to the nth + mth time periods [ (n-1) delta t, (n + m) delta t]Inner p-th tractor K (p) Node at initial positionInitializing initial position nodes of all tractors; the tractors are stopped at the parking lot at the beginning, and the initial position nodes of all the tractors are obtained in the step;
step 4, obtaining the n-th to the n + m-th time periods [ (n-1) delta t, (n + m) delta t]Distribution cost between ith node and jth node in the networkThereby obtaining a distribution cost matrix C n (ii) a Obtain a distribution cost matrix C n The method of (1): road surface cost C under the condition of reserving road surface smoothness in advance 0 On the basis, a dynamic path cost matrix capable of reflecting the traffic capacity and the condition of the road section can be obtained by combining the prediction coefficients of the traffic capacity and the condition of the road section in a period of time in the future;
step 5, according to the p-th tractor K (p) The target node to be accessed falls into the time periods from the n th to the n + m th according to the sequence of the earliest access time of the target node to be accessed, wherein the time periods from (n-1) delta t to (n + m) delta t]All target nodes in the system form a p-th tractor K in sequence (p) Initial access sequence of (2)
Step 6, taking out the p-th tractor K (p) Initial access sequence of (2)The last target node in (2) is taken as the destination node and is recorded asThereby forming an updated access sequence Is thatOn the basis of, get rid ofIs obtained by the last target node in (1);
step 7, according to the p-th tractor K (p) Node of home positionDistribution cost matrix C n Access sequenceDestination nodeEstablishing a drop and pull transport dynamic path planning model based on rolling time domain optimization; 4 pieces of parameter information are required to be obtained in advance when a dynamic path planning model is used for path planning;
the model of the drop and pull transport dynamic path planning based on the rolling time domain optimization in the step 7 is as follows:
(1) An objective function:
(2) Constraint conditions
Equation (1) is an objective function, representing the minimization of the total delivery cost of the tractor K; when in useWhen, represents the p-th tractor K (p) Passing through a straight-line path (i, j) between the ith node and the jth nodeWhen it is, the p-th tractor K is shown (p) Does not pass through a straight-line path (i, j) between the ith node and the jth node;
the expression (2) represents the p-th tractor K (p) Must be from the location nodeStarting; when in useWhen it is, the p-th tractor K is shown (p) Passing through location nodeAnd the jth nodeWhen in useWhen it is, the p-th tractor K is shown (p) Non-transit location nodeAnd the jth node
Equation (3) indicates that the p-th tractor must reach the destination nodeWhen in useWhen it is, the p-th tractor K is shown (p) Through the ith node and the destination nodePath betweenWhen in useWhen it is, the p-th tractor K is shown (p) Does not pass through the ith node and the destination nodePath between
Equation (4) is a time window constraint and represents that the p-th tractor K is driven (p) At the location nodeThe Time of (2) is the current Time (Time);showing the p-th tractor K (p) Arrival position nodeThe time of (d);
the formula (5) is as shownWhen the vehicle is in the right position, the p-th tractor K is obtained (p) Access time on jth node, AT i (p) Representation of the p-th tractor K (p) Time of arrival at the ith node; w is a i Showing the p-th tractor K (p) Latency at the ith node; t is t ij Showing the p-th tractor K (p) The travel time from the ith node to the jth node;showing the p-th tractor K (p) Time of arrival at the jth node;
equation (6) represents the pth tractor K as a time window constraint (p) To the r-th target nodeAt the earliest access e r And the latest access time l r To (c) to (d);
equation (7) represents all the road network nodes U 2 And access sequenceThe target node in (1) meets the access balance;showing the p-th tractor K (p) Through the ith nodeAnd the w-th node, whenWhen, represents the p-th tractor K (p) Does not pass through a straight line path (i, w) between the ith node and the w-th node; when the temperature is higher than the set temperatureWhen it is, the p-th tractor K is shown (p) Passing through a straight line path between the w-th node and the j-th node<, w, j, whenWhen, represents the p-th tractor K (p) Does not pass through a straight line path between the w-th node and the j-th node;
equation (8) represents the access sequenceVisiting order constraint of the middle node;showing the p-th tractor K (p) Reach access sequenceTime of the z-th target node;showing the p-th tractor K (p) Reach access sequenceTime of the z +1 th target node;representing access sequencesThe z-th target node of (1);
equation (9) represents the access sequenceThe node in (1) is visited by one traction vehicle only once; when in useWhen it is, the p-th tractor K is shown (p) Passing through the straight-line path (i, u) between the ith node and the u th node whenWhen, represents the p-th tractor K (p) Does not pass through the straight-line path (i, u) between the ith node and the u th node; when the temperature is higher than the set temperatureWhen it is, the p-th tractor K is shown (p) Passing through a straight-line path (u, j) between the u-th node and the j-th nodeWhen, represents the p-th tractor K (p) Does not pass through the straight-line path (u, j) between the u-th node and the j-th node;
the equation (10) indicates that the decision variable takes a value of "0" or "1".
Fig. 2 is a diagram of dynamically planning target nodes in a plurality of future time segments at a certain time node according to a model of drop-and-pull transport dynamic path planning based on rolling time domain optimization. The diagram is a schematic diagram of a path planning scheme generated by aiming at 2 tractors, corresponding position nodes where the 2 tractors are located, 7 target nodes and road network nodes.
Step 8, randomly generating the nth initial solution X n ;
Step 9, utilizing simulated annealing algorithm to solve X for the nth initial solution n Carrying out iterative solution to obtain the nth optimal solution;
step 10, assigning n +1 to n;
step 11, updating the n-th to the n + m-th time periods [ (n-1) deltat,(n+m)Δt]Inner p-th tractor K (p) Node of the positionDistribution cost matrix C n ;And C n The obtaining method is the same as the steps 3 and 4;
step 12, in order to plan the subsequent time slices better, an advance period Δ T may be set, that is, before each time slice ends, the subsequent time slices are planned in advance by a time length Δ T, where Δ T =0 in order to simplify the model; in the (n-1) th time period [ (n-2) Delta t, (n-1) Delta t]Before finishing, acquiring the p-th tractor K (p) During the period from n to n + m [ (n-1) Δ t, (n + m) Δ t]Access sequence withinAccess sequenceThe method comprises the following steps: the (n-1) th time period [ (n-2) Δ t, (n-1) Δ t]Target node with incomplete access and p-th tractor K (p) The earliest access time of the target node to be accessed falls within the n-th to n + m-th time periods [ (n-1) Δ t, (n + m) Δ t]The target node of (1); here, theThe obtaining mode of the step (5) is different from that of the step (5);
step 13, taking out the n-th to the n + m-th time periods [ (n-1) delta t, (n + m) delta t]Inner p-th tractor K (p) Access sequence ofAs a destination nodeThereby forming an updated access sequence
Step 14, according to the p-th tractor K (p) Node at initial positionDistribution cost matrix C n Access sequenceDestination nodeInputting the data into a drop and pull transport dynamic path planning model; 4 pieces of parameter information are required to be obtained in advance when a dynamic path planning model is used for path planning;
step 15, randomly generating the nth initial solution X n ,
Step 16, utilizing a simulated annealing algorithm to solve the Nth initial solution X n Carrying out iterative solution to obtain the nth optimal solution;
step 17, judging whether N = N is established, and if so, indicating that all tractors K finish the operation; otherwise, return to step 10.
Claims (1)
1. A drop and pull transport dynamic path planning method based on rolling time domain optimization is characterized by comprising the following steps: is applied to a target node U consisting of 1 yard A and R 1 V road network nodes U 2 In a distribution service area formed by position nodes S of P tractors, setting one vertex of an external square of the distribution service area as a coordinate origin o, and setting two adjacent sides in the horizontal and vertical directions connected with the coordinate origin o as an x axis and a y axis respectively so as to form a rectangular coordinate system xoy; in the coordinate system xoy:
define the location of yard A as (x) 0 ,y 0 ) Defining the location (x) of each of said P tractors from said yard A 0 ,y 0 ) Starting; recording the P tractors as K = { K = { (K) } (1) ,K (2) ,…,K (p) ,…,K (P) },1≤p≤P,K (p) Represents the p-th tractor;
record R target nodes as1≤r≤R,Representing the r-th target node, the r-th target nodeIs denoted by (x) r ,y r ) (ii) a The r < th > target nodeIs recorded as e r (ii) a The r < th > target nodeIs recorded as the latest access time of l r ;e r And l r Time window [ e ] constituting a target node r ,l r ];
Marking V road network nodes as1≤v≤V,Representing a nth road network node; the v-th road network nodeIs denoted by (x) R+v ,y R+v );
P tractors K are location node as S = { S = (R+V+1) ,S (R+V+2) ,…,S (R+V+p) ,…,S (R+V+P) },S (R+V+p) Represents the R + V + p nodes and represents the p tractor K (p) The location of the location; p-th tractor K (p) At the location node S (R+S+p) Is denoted by (x) R+V+p ,y R+V+p );
The 1 yard A, R target nodes U 1 V road network nodes U 2 And P nodes S of the positions of the tractors are recorded as a point set U = { A, U = (N, N) } 1 ,U 2 And S, the yard A represents the 0 th node; r target nodes U 1 Represents the 1 st node to the R < th > node; v road network nodes U 2 Represents the R +1 th to the R + S th nodes; the node S where the P tractors are located represents an R + V +1 th node to an R + V + P th node;
defining edge set E =:<i,j>, | i, j ∈ U, i ≠ j } represents a set of straight-line paths of any two nodes in the point set U,<i,j> represents a straight line path between any ith node and any jth node; recording a distribution cost matrix between any two nodes in the point set U as C, wherein C = { C = { (C) } ij |i,j∈U,i≠j};c ij Representing the distribution cost between the ith node and the jth node;
the drop-and-pull transport dynamic path planning method based on rolling time domain optimization is carried out according to the following steps:
step 1, calculating the operation time [0, maxl ] of all tractors K r ]Equally divided into N continuous time segments, which are marked as { [0, delta t { [0 ]],[Δt,2Δt],…,[(n-1)Δt,nΔt],…,[(N-1)Δt,maxl r ]}; wherein [ (n-1) Δ t, n Δ t]Representing the nth time slice; Δ t represents the divided interval;
step 2, letting n =1; planning an actual driving route of the vehicle in the time periods from the nth to the n + mth [ (n-1) delta t, (n + m) delta t ]; m is more than or equal to 1 and less than or equal to N; n is the total number of time segments;
step 3, enabling the nth to the nth + mth time periods [ (n-1) delta t, (n + m) delta t]Inner p-th tractor K (p) Node at initial positionInitializing initial position nodes of all tractors;
step 4, obtaining the n-th to the n + m-thA period of time [ (n-1) Δ t, (n + m) Δ t]Distribution cost between the ith node and the jth nodeThereby obtaining a distribution cost matrix C n ;
Step 5, according to the p-th tractor K (p) The target node to be accessed falls into the time periods from (n-1) delta t to (n + m) delta t according to the sequence of the earliest access time of the target node to be accessed]All target nodes in the system form a p-th tractor K in sequence (p) Initial access sequence of (2)
Step 6, taking out the p-th tractor K (p) Initial access sequence ofAs the destination node, is recorded asThereby forming an updated access sequence
Step 7, according to the p-th tractor K (p) Node of home positionDistribution cost matrix C n Access sequenceDestination nodeEstablishing a drop and pull transport dynamic path planning model based on rolling time domain optimization;
the drop and pull transport dynamic path planning model based on rolling time domain optimization is as follows:
(1) An objective function:
(2) Constraint conditions
Equation (1) is an objective function, which represents the minimization of the total delivery cost of the tractor K; when the temperature is higher than the set temperatureWhen it is, the p-th tractor K is shown (p) A straight line path passing through the ith node and the jth node<i,j>, whenWhen it is, the p-th tractor K is shown (p) Does not pass through a straight-line path between the ith node and the jth node<i,j>;
The formula (2) represents the p-th tractor K (p) Must be from the location nodeStarting; when the temperature is higher than the set temperatureWhen it is, the p-th tractor K is shown (p) Passing through location nodeAnd the jth nodeWhen in useWhen, represents the p-th tractor K (p) Non-transit location nodeAnd the jth node
Equation (3) indicates that the p-th tractor must reach the destination nodeWhen in useWhen, represents the p-th tractor K (p) Through the ith node and the destination nodePath betweenWhen in useWhen it is, the p-th tractor K is shown (p) Does not pass through the ith node and the destination nodePath between
Equation (4) is a time window constraint, which means that the pth tractor K is driven (p) At the position nodeThe Time of (2) is the current Time (Time);showing the p-th tractor K (p) Arrival position nodeThe time of (d);
the formula (5) is as shownWhen the vehicle is standing, the p-th tractor K is obtained (p) At the time of access on the jth node,representation of the p-th tractor K (p) Time of arrival at the ith node; w is a i Showing the p-th tractor K (p) Latency at the ith node; t is t ij Showing the p-th tractor K (p) The travel time from the ith node to the jth node;representation of the p-th tractor K (p) Time of arrival at the jth node;
equation (6) represents the pth tractor K as a time window constraint (p) To the r-th target nodeAt the earliest access e r And the latest access time l r In the middle of;
equation (7) represents all the road network nodes U 2 And access sequenceThe target node in (2) satisfies the balance of access and exit;showing the p-th tractor K (p) A straight line path passing between the ith node and the w-th node<, i, w whenWhen it is, the p-th tractor K is shown (p) Does not pass through a straight-line path between the ith node and the w-th node<, i, w >; when in useWhen it is, the p-th tractor K is shown (p) A straight line path passing through the w-th node and the j-th node<w,j&WhenWhen, represents the p-th tractor K (p) The straight-line path does not pass through the w node and the j node;
equation (8) represents the access sequenceVisiting order constraint of the middle node; the describedShowing the p-th tractor K (p) Reach access sequenceTime of the z-th target node; the above-mentionedShowing the p-th tractor K (p) Reach access sequenceTime of the z +1 th target node;
equation (9) represents the access sequenceThe node in (1) is visited by one traction train only once; when in useWhen it is, the p-th tractor K is shown (p) A straight line path passing through the ith node and the u-th node<i,u&WhenWhen it is, the p-th tractor K is shown (p) Does not pass through a straight-line path between the ith node and the u-th node<i,u> when in useWhen it is, the p-th tractor K is shown (p) A straight line path passing through the u-th node and the j-th node<u,j&WhenWhen it is, the p-th tractor K is shown (p) Does not pass through a straight-line path between the u-th node and the j-th node<u,j>;
Equation (10) represents that the decision variable takes a value of "0" or "1";
step 8, randomly generating the nth initial solution X n ;
Step 9, utilizing simulated annealing algorithm to solve X for the nth initial solution n Carrying out iterative solution to obtain the nth optimal solution;
step 10, assigning n +1 to n;
step 11, updating the n-th to the n + m-th time periods [ (n-1) delta t, (n + m) delta t]Inner p-th tractor K (p) Node of the positionDistribution cost matrix C n ;
Step 12, in the (n-1) th time period [ (n-2) delta t, (n-1) delta t]Before finishing, acquiring the p-th tractor K (p) In the period from the n-th time to the n + m-th time [ (n-1) delta t, (n + m) delta t]Access sequence withinThe access sequenceThe method comprises the following steps: (n-1) th time period [ (n-2) Δ t, (n-1) Δt]Target node with incomplete access and p-th tractor K (p) The earliest access time of the target node to be accessed falls within the n-th to n + m-th periods of time (n-1) Δ t, (n + m) Δ t]The target node of (1);
step 13, taking out the n-th to the n + m-th time periods [ (n-1) delta t, (n + m) delta t]Inner p-th tractor K (p) Access sequence ofAs a destination nodeThereby forming an updated access sequence
Step 14, according to the p-th tractor K (p) Node of home positionDistribution cost matrix C n Access sequenceDestination nodeInputting the dynamic path planning model of drop and pull transportation;
step 15, randomly generating the nth initial solution X n ,
Step 16, utilizing a simulated annealing algorithm to solve the nth initial solution X n Carrying out iterative solution to obtain the nth optimal solution;
step 17, judging whether N = N is established, and if so, indicating that all tractors K complete the operation; otherwise, return to step 10.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610241242.9A CN105956681B (en) | 2016-04-15 | 2016-04-15 | A kind of Drop and pull transport dynamic path planning method based on rolling time horizon optimization |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610241242.9A CN105956681B (en) | 2016-04-15 | 2016-04-15 | A kind of Drop and pull transport dynamic path planning method based on rolling time horizon optimization |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105956681A CN105956681A (en) | 2016-09-21 |
CN105956681B true CN105956681B (en) | 2018-02-06 |
Family
ID=56917724
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610241242.9A Active CN105956681B (en) | 2016-04-15 | 2016-04-15 | A kind of Drop and pull transport dynamic path planning method based on rolling time horizon optimization |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105956681B (en) |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107392358B (en) * | 2017-06-30 | 2021-07-23 | 上海汽车集团股份有限公司 | Vehicle distribution method and device |
CN109919348A (en) * | 2017-12-12 | 2019-06-21 | 顺丰科技有限公司 | A kind of method for optimizing route, device, equipment, storage medium |
CN109961162B (en) * | 2017-12-22 | 2023-04-07 | 株式会社日立制作所 | Path planning method and path planning device |
CN109858752A (en) * | 2018-12-27 | 2019-06-07 | 安庆师范大学 | Dynamic based on roll stablized loop takes out the method and device of dispatching |
CN109948990A (en) * | 2019-03-29 | 2019-06-28 | 合肥工业大学智能制造技术研究院 | Method for optimizing route, system and storage medium under cloud logistics environment |
CN110147971B (en) * | 2019-04-08 | 2023-02-03 | 合肥工业大学 | Method, system and storage medium for planning a vehicle path |
CN111459108B (en) * | 2020-04-08 | 2021-07-06 | 北京理工大学 | Task allocation and conflict-free path planning method for pull-type multi-AGV system |
CN111709570B (en) * | 2020-06-11 | 2022-09-27 | 西南交通大学 | Optimization method for network drop-and-drop transport scheduling |
CN112862414B (en) * | 2021-04-12 | 2023-12-05 | 中南大学 | Collaborative distribution path optimization method based on cluster travel business problem |
CN113077106B (en) * | 2021-04-16 | 2024-06-18 | 北京京东振世信息技术有限公司 | Article transportation method and device based on time window |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103295117A (en) * | 2013-05-20 | 2013-09-11 | 东南大学 | Method and system for matching tractor and trailer |
-
2016
- 2016-04-15 CN CN201610241242.9A patent/CN105956681B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103295117A (en) * | 2013-05-20 | 2013-09-11 | 东南大学 | Method and system for matching tractor and trailer |
Also Published As
Publication number | Publication date |
---|---|
CN105956681A (en) | 2016-09-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105956681B (en) | A kind of Drop and pull transport dynamic path planning method based on rolling time horizon optimization | |
CN109635917B (en) | Multi-agent cooperation decision and training method | |
CN109117993B (en) | Processing method for optimizing vehicle path | |
CN106020189B (en) | Vacant lot heterogeneous robot system paths planning method based on neighborhood constraint | |
DE102019113389B4 (en) | SYSTEM AND METHOD FOR PREDICTING ENTITY BEHAVIOR | |
CN112074845A (en) | Deep reinforcement learning for optimizing car pooling strategies | |
CN113276915B (en) | Subway departure scheduling method and system | |
CN110599236A (en) | Short-time parking demand prediction method based on GRU model | |
US10869066B2 (en) | Media content delivery system | |
CN112071062B (en) | Driving time estimation method based on graph convolution network and graph attention network | |
Qin et al. | Reinforcement learning for ridesharing: A survey | |
CN112068515A (en) | Full-automatic parking lot scheduling method based on deep reinforcement learning | |
CN108596385A (en) | The queuing strategy and device of vehicle, readable medium, logistics system | |
CN113672846A (en) | Network appointment scheduling method and device, electronic equipment and storage medium | |
Siddique et al. | Puzzle-based parking | |
CN110889530A (en) | Destination prediction method based on recurrent neural network and server | |
CN106274994A (en) | A kind of train arrival time Forecasting Methodology and system completely | |
CN113536692B (en) | Intelligent dispatching method and system for high-speed rail train under uncertain environment | |
CN114676911A (en) | Method and device for determining driving route of transport vehicle | |
CN114626755A (en) | Steel logistics intelligent scheduling method and system based on Markov decision | |
CN117252318B (en) | Intelligent networking automobile group machine collaborative carpooling scheduling method and system | |
DE102021114724A1 (en) | IMPROVED VEHICLE OPERATION | |
CN115170006B (en) | Dispatching method, device, equipment and storage medium | |
Ma | A cross entropy multiagent learning algorithm for solving vehicle routing problems with time windows | |
CN114141029B (en) | Ramp control method based on offline reinforcement learning and macroscopic model |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |