CN109190900A - A kind of method that distribution Constraint Anchored Optimization solves AGV scheduling system task distribution - Google Patents
A kind of method that distribution Constraint Anchored Optimization solves AGV scheduling system task distribution Download PDFInfo
- Publication number
- CN109190900A CN109190900A CN201810867224.0A CN201810867224A CN109190900A CN 109190900 A CN109190900 A CN 109190900A CN 201810867224 A CN201810867224 A CN 201810867224A CN 109190900 A CN109190900 A CN 109190900A
- Authority
- CN
- China
- Prior art keywords
- agent
- message
- value
- constraint
- cop
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Educational Administration (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to dispatching method technical fields, it is related to a kind of method of AGV scheduling system task distribution, a kind of method for solving AGV scheduling system task distribution more particularly to distributed Constraint Anchored Optimization, the present invention does not need image set Chinese style method for solving and the knowledge of respective agent and subproblem is focused on a core agent like that, but cooperate with other agent, the global preferred plan of problem is obtained by local decision-making.This mode effectively increases the efficiency solved the problems, such as, while agent not being required externally to expose all individual informations, enhances privacy.Application distribution formula method for solving, local solution procedure are often confined to the expense of some agent breath transmitting, reduce system loading.Centralized method for solving does not usually have fault-tolerance to system structure, and an agent, which goes wrong, will lead to solve failure.And in a distributed system, single agent, which goes wrong, does not interfere with the solution of entire agent system, has good robustness.
Description
Technical field
The invention belongs to dispatching method technical fields, are related to a kind of method of AGV scheduling system task distribution, and in particular to
A kind of method that distribution Constraint Anchored Optimization solves AGV scheduling system task distribution.
Background technique
Distributed constrained optimization problem (Distributed Constraint Optimization Problem, DCOP)
It is to be developed based on constraint satisfaction problemx (Constraint Satisfied Problem, CSP), is research multiple agent
The important frame of Agent.Currently, intellectual Agent in academia's not unified definition, typically refers to have autonomous
Property and dynamic role, can be carried out between each other communication entity or virtual system.There are multiple Agent in DCOP, and they it
Between have the constraint relationship, the target of DCOP is exactly the constraint coordinated between Agent, to reach an objective function global optimum
Purpose.Meanwhile a centralized constrained optimization problem is formed inside each Agent, Agent has the decision of self-contr ol
Ability.
AGV (Automated Guided Vehicle), which refers to, magnetically or optically waits homing guidances device equipped with electricity, by counting
The control of calculation machine, characterized by wheel type mobile, self-powered or power switching device, and can be automatic along defined guide path
The means of transport of traveling generally has the multiple functions such as security protection, transfer.Popular says, AGV is exactly one and is used to transport
Mobile robot, it is a porter, and cargo is transported at B from A.Most of research of AGV is also included in moving machine
In device people field.
AGV scheduling system is to ensure that task is efficiently run, i.e., by known and unknown task according to it is certain it is regular simultaneously
In the case where meeting system task time and the constraint such as limited of various resources, AGV is reasonably distributed to.Times of AGV system at present
Business scheduling is mainly by two types: offline task schedule and online task schedule.
Summary of the invention
According to the above-mentioned deficiencies of the prior art, the present invention provides a kind of distributed Constraint Anchored Optimization solution AGV scheduling system
The method of system task distribution, is based on offline task schedule, problem is abstracted as distributed constrained optimization problem (DCOP), is asked
Solution.
The method that a kind of distributed Constraint Anchored Optimization of the present invention solves AGV scheduling system task distribution, it is special
Sign be the following steps are included:
(1) discrete constraint optimization problem COP is defined by triple<X, D, R>composition, and target is for variable XiFind one
Make the maximized instance X of total utility functional value*, wherein
X={ X1,...,XnBe model variable set;
D={ d1,...,dnBe the corresponding limited codomain of variable set;
R={ r1,...,rmIt is one group of utility function;Such function is that each of the variate-value in the range of function can
It can combine and assign an effectiveness (reward), and negative means cost.Hard constraint (limiting certain particular values) is utility function
Special case, feasible tuple is assigned a value of 0 by it, and it is infinite that infeasible tuple assignment is negative;
Wherein, riValue be corresponding particular instance X value of utility;
(2) distribution Constraint Anchored Optimization DCOP is defined by triple < A, COP, Ria> composition, wherein
A={ A1,...,AkBe one group of Agent set;
COP={ COP1,...COPkBe one group of non-intersecting, centralized COP set, each COPiReferred to as Agent
AiLocal subproblem, and by Agent AiPossess and controls;
Ria={ r1,...rnBe one group of interactional utility function set;They are according to from several different sheets
Ground subproblem COPiVariable-definition get;As COP, hard constraint is simulated by utility function, it is by 0 assignment
To feasible tuple, and it will bear and infinite become infeasible tuple.
(3) according to DCOP model it can be concluded that the constraint relationship figure between multi-Agent, for the constraint relationship figure, obtain with
Corresponding pseudo- tree, be based on Depth Priority Algorithm DFS for the special case of puppet tree and generate DFS tree;
(4) bucket is carried out on DFS tree based on dynamic programming algorithm DPOP to disappear member.The main advantage of DPOP is its only needs
The message of linear quantity, to introduce exponential less network overhead than searching algorithm when applying in distribution setting.It
Complexity be the size of UTIL message, the induction width that it is sorted by the DFS selected exponentially bounded.Therefore,
DPOP is the optimal selection for solving the problems, such as DCOP, because these problems have lower induction width.DPOP is a kind of complete
Algorithm has the important advantage for the message for only generating linear quantity.This is critically important in distribution setting, because sending a large amount of small
Message (such as searching algorithm) usually requires a large amount of communication overhead.
Wherein, preferred embodiment is as follows:
In step (2), triple DCOP is further simplified into five-tuple five-tuple < A, X, D, F, Ria>, wherein
A={ A1,...,AkBe one group of Agent set;
X={ X1,...,XnBe model variable set;
D={ d1,...,dnBe the corresponding limited codomain of model variable set;
F=﹛ f1... .., Fm﹜ is the set of constraint function cost;
Ria={ r1,...rnBe one group of interactional utility function set.
In step (4), bucket is carried out on DFS tree to disappear first to include UTIL message propagation stage and VALUE message propagation stage.
Complete DPOP algorithm includes 3 stages, and the first stage substantially includes to generate this single order of DFS tree using DFS algorithm in step (3)
Section, UTIL message propagation stage and VALUE message propagation stage are respectively as second stage and phase III.
UTIL message propagation stage specifically: this is a process from bottom to top, since lowest level, only passes through tree side
Upwardly propagate, in this process, Agent to its parents send UTIL message, these message summarize this Agent and its
Entire influence of the subtree to rest part, whereinRefer to from Agent XiIt is sent to XjUTIL message a multidimensional
Matrix.
VALUE message propagation stage specifically: this is a top-down process, is completed by UTIL message propagation stage
When started by root node, each node disappears according to the calculated result of UTIL message propagation stage and from the VALUE that his father's item receives
It ceases and determines its optimum value, then, this value is sent to its filial generation by VALUE message by it, until all variables have value
When, algorithm terminates.Obviously, DPOP can generate the message of linear quantity.Its complexity is the size of UTIL message, this be by
The time of DFS sequence width used and space index determine.
The knowledge and son of respective agent are asked like that the present invention has the advantages that (1) does not need image set Chinese style method for solving
Topic focuses on a core agent, but cooperates with other agent, the best side of the overall situation for obtaining problem by local decision-making
Case.This mode effectively increases the efficiency solved the problems, such as, while agent not being required externally to expose all individual informations, increases
Strong privacy.(2) in centralized method for solving, certain agent are far apart from core agent, need a large amount of information into
Row interaction.And in distributed method for solving, local solution procedure is often confined to the expense of some agent breath transmitting, reduces
System loading.(3) centralized method for solving does not usually have fault-tolerance to system structure, and an agent goes wrong will
Cause to solve and fail.And in a distributed system, single agent, which goes wrong, does not interfere with the solution of entire agent system,
With good robustness.
Detailed description of the invention
Fig. 1 is the pseudo- tree graph of AGV1 in embodiment 2;
Fig. 2 is the pseudo- tree graph of AGV2 in embodiment 2;
Fig. 3 is the pseudo- tree graph of AGV3 in embodiment 2;
Fig. 4 is the pseudo- tree graph of AGV4 in embodiment 2.
Specific embodiment
The present invention will be further described with reference to embodiments.
Embodiment 1:
It is a kind of distribution Constraint Anchored Optimization solve AGV scheduling system task distribution method, it is characterised in that including with
Lower step:
(1) discrete constraint optimization problem COP is defined by triple<X, D, R>composition, and target is for variable XiFind one
Make the maximized instance X of total utility functional value*, wherein
X={ X1,...,XnBe model variable set;
D={ d1,...,dnBe the corresponding limited codomain of variable set;
R={ r1,...,rmIt is one group of utility function;Such function is that each of the variate-value in the range of function can
It can combine and assign an effectiveness (reward), and negative means cost.Hard constraint (limiting certain particular values) is utility function
Special case, feasible tuple is assigned a value of 0 by it, and it is infinite that infeasible tuple assignment is negative;
Wherein, riValue be corresponding particular instance X value of utility;
(2) distribution Constraint Anchored Optimization DCOP is defined by triple < A, COP, Ria> composition, wherein
A={ A1,...,AkBe one group of Agent set;
COP={ COP1,...COPkBe one group of non-intersecting, centralized COP set, each COPiReferred to as Agent
AiLocal subproblem, and by Agent AiPossess and controls;
Ria={ r1,...rnBe one group of interactional utility function set;They are according to from several different sheets
Ground subproblem COPiVariable-definition get;As COP, hard constraint is simulated by utility function, it is by 0 assignment
To feasible tuple, and it will bear and infinite become infeasible tuple;
Further, triple DCOP is further simplified into five-tuple five-tuple < A, X, D, F, Ria>, wherein
A={ A1,...,AkBe one group of Agent set;
X={ X1,...,XnBe model variable set;
D={ d1,...,dnBe the corresponding limited codomain of model variable set;
F=﹛ f1,.....,fm﹜ is the set of constraint function cost;
Ria={ r1,...rnBe one group of interactional utility function set.
(3) according to DCOP model it can be concluded that the constraint relationship figure between multi-Agent, for the constraint relationship figure, obtain with
Corresponding pseudo- tree, be based on Depth Priority Algorithm DFS for the special case of puppet tree and generate DFS tree;
(4) bucket is carried out on DFS tree based on dynamic programming algorithm DPOP to disappear member, including UTIL message propagation stage with
VALUE message propagation stage.Complete DPOP algorithm includes 3 stages, and the first stage substantially includes that DFS is used in step (3)
Algorithm generates this stage of DFS tree, and UTIL message propagation stage and VALUE message propagation stage are respectively as second stage and the
Three stages.The main advantage of DPOP is that it only needs the message of linear quantity, so that application when ratio is searched in distribution setting
Rope algorithm introduces exponential less network overhead.Its complexity is the size of UTIL message, it is sorted by the DFS selected
Induction width exponentially bounded.Therefore, DPOP is the optimal selection for solving the problems, such as DCOP because these problems have compared with
Low induction width.DPOP is a kind of complete algorithm, has the important advantage for the message for only generating linear quantity.This is being distributed
It is critically important in formula setting, because sending a large amount of small message (such as searching algorithm) usually requires a large amount of communication overhead.
UTIL message propagation stage specifically: this is a process from bottom to top, since lowest level, only passes through tree side
Upwardly propagate, in this process, Agent to its parents send UTIL message, these message summarize this Agent and its
Entire influence of the subtree to rest part, whereinRefer to from Agent XiIt is sent to XjUTIL message a multidimensional
Matrix.
VALUE message propagation stage specifically: this is a top-down process, is completed by UTIL message propagation stage
When started by root node, each node disappears according to the calculated result of UTIL message propagation stage and from the VALUE that his father's item receives
It ceases and determines its optimum value, then, this value is sent to its filial generation by VALUE message by it, until all variables have value
When, algorithm terminates.Obviously, DPOP can generate the message of linear quantity.Its complexity is the size of UTIL message, this be by
The time of DFS sequence width used and space index determine.
Embodiment 2:
The task distribution of the present embodiment, which refers to, is assigned to 4 AGV for known Transport Vehicle task, and total delay is punished
Cost minimization.
Therefore, the DCOP model of task distribution is as follows:
Agent:4 AGV, A={ A1,A2,A3,A4};
Variable are as follows:Whether i-th of task is assigned to AGVj;
Codomain are as follows:
Objective function:Total weighting drags phase punishment cost minimum;
Constraint condition:
Constraint (1) indicates each task energy and can only distribute to an AGV completion;When constraining the general assignment of (2) each AGV
Between no more than existing electricity can be done time of task;Constrain the tardiness time no more than one that (3) indicate that task is completed
There are a maximum value in definite value, the i.e. waiting time of car owner, and in this problem, we define tmaxwaitingtime≤150s。”
FRODO is the Java Open Framework of distributed Combinatorial Optimization, initially by the Lausanne Institute of Technology of Zwiterland
(EPFL) Artificial Intelligence Laboratory (LIA) develops.The present embodiment has selected FRODO version 2 .x, and the version is by Adrian Petcu
It is redesigned completely on initial FRODO platform base and implements exploitation again.
In this experiment, there are task 20, AGV4, Constraint is converted into the maximum time of execution task herein
It is as shown in table 1:
Table 1: Constraint
AGV serial number | 1 | 2 | 3 | 4 |
Total Mission Time (s) | 1000 | 1500 | 900 | 1300 |
Executing the time is to have the path planning calculating of task to get.According to these given datas, as shown in table 2, by 20
A task distributes to 4 AGV, meanwhile, meet the constraint condition in model.
Table 2: task primary data
According to above example data, it can be seen that there are 4 Agent to be initialized according to constraint condition, 5 obtained time
Initial solution is as shown in table 3.
Table 3:DCOP initial solution
Serial number | AGV1 | AGV 2 | AGV 3 | AGV 4 | Target value |
1 | 9-8-1-6-17 | 3-19-10-14-2-18 | 4-12-11-15 | 7-20-13-16-5 | 54125 |
2 | 5-13-15-19-3 | 18-10-17-20-6 | 11-12-19-2 | 14-16-1-4-9-8-7 | 46288 |
3 | 2-8-16-7-1 | 12-19-20-17-5-11 | 13-4-8-14 | 18-3-6-9-10 | 50791 |
4 | 9-11-3-6 | 2-8-19-20-15-16 | 13-1-7-4 | 12-17-14-18-5-10 | 47264 |
5 | 15-3-16-2 | 18-17-1-10-7-4 | 12-5-8-13 | 20-14-9-19-11-6 | 50465 |
And obtaining corresponding pseudo- tree according to constraint is shown in FIG. 1 to FIG. 4:
It is calculated that the results are shown in Table 4 using algorithm DPOP.
Table 4:DCOP result
AGV1 | AGV 2 | AGV 3 | AGV 4 | Target value |
20-18-19-10 | 16-17-5-6-2 | 15-14-7-11-9 | 8-4-13-3-1-12 | 134 |
Claims (5)
1. a kind of method that distribution Constraint Anchored Optimization solves AGV scheduling system task distribution, it is characterised in that including following
Step:
(1) discrete constraint optimization problem COP is defined by triple<X, D, R>composition, and target is for variable XiFinding one makes always to imitate
With the maximized instance X of functional value*, wherein
X={ X1,...,XnBe model variable set;
D={ d1,...,dnBe the corresponding limited codomain of variable set;
R={ r1,...,rmIt is one group of utility function;
Wherein, riValue be corresponding particular instance X value of utility;
(2) distribution Constraint Anchored Optimization DCOP is defined by triple < A, COP, Ria> composition, wherein
A={ A1,...,AkBe one group of Agent set;
COP={ COP1,...COPkBe one group of non-intersecting, centralized COP set, each COPiReferred to as Agent Ai's
Local subproblem, and by Agent AiPossess and controls;
Ria={ r1,...rnBe one group of interactional utility function set;
(3) according to DCOP model it can be concluded that the constraint relationship figure between multi-Agent, for the constraint relationship figure, it is right therewith to obtain
The pseudo- tree answered generates DFS tree based on Depth Priority Algorithm DFS for the special case of puppet tree;
(4) bucket is carried out on DFS tree based on dynamic programming algorithm DPOP to disappear member.
2. the method that a kind of distributed Constraint Anchored Optimization according to claim 1 solves AGV scheduling system task distribution,
It is characterized by: triple DCOP is further simplified into five-tuple five-tuple < A, X, D, F, R in step (2)ia>, wherein
A={ A1,...,AkBe one group of Agent set;
X={ X1,...,XnBe model variable set;
D={ d1,...,dnBe the corresponding limited codomain of model variable set;
F=﹛ f1,…..,fm﹜ is the set of constraint function cost;
Ria={ r1,...rnBe one group of interactional utility function set.
3. the method that a kind of distributed Constraint Anchored Optimization according to claim 1 solves AGV scheduling system task distribution,
It is characterized by: carrying out bucket on DFS tree to disappear first including that UTIL message propagation stage and VALUE message propagate rank in step (4)
Section.
4. the method that a kind of distributed Constraint Anchored Optimization according to claim 3 solves AGV scheduling system task distribution,
It is characterized in that UTIL message propagation stage specifically: this is a process from bottom to top, since lowest level, only passes through tree
While upwardly propagate, in this process, Agent sends UTIL message to its parents, these message summarize this Agent and
Its influence of entire subtree to rest part, whereinRefer to from Agent XiIt is sent to XjMore than one of UTIL message
Tie up matrix.
5. the method that a kind of distributed Constraint Anchored Optimization according to claim 3 solves AGV scheduling system task distribution,
It is characterized in that VALUE message propagation stage specifically: this is a top-down process, complete by UTIL message propagation stage
At when started by root node, each node is according to the calculated result of UTIL message propagation stage and the VALUE received from his father's item
Message determines its optimum value, and then, this value is sent to its filial generation by VALUE message by it, until all variables take
When value, algorithm terminates.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810867224.0A CN109190900A (en) | 2018-08-02 | 2018-08-02 | A kind of method that distribution Constraint Anchored Optimization solves AGV scheduling system task distribution |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810867224.0A CN109190900A (en) | 2018-08-02 | 2018-08-02 | A kind of method that distribution Constraint Anchored Optimization solves AGV scheduling system task distribution |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109190900A true CN109190900A (en) | 2019-01-11 |
Family
ID=64920349
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810867224.0A Pending CN109190900A (en) | 2018-08-02 | 2018-08-02 | A kind of method that distribution Constraint Anchored Optimization solves AGV scheduling system task distribution |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109190900A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110378663A (en) * | 2019-04-19 | 2019-10-25 | 西北工业大学 | A kind of complicated movement crowdsourcing method for allocating tasks based on Greedy strategy |
CN110533301A (en) * | 2019-08-09 | 2019-12-03 | 大连理工大学 | A kind of population dispatching method based on dynamic constrained matrix |
CN113408823A (en) * | 2021-07-13 | 2021-09-17 | 重庆理工大学 | Ant colony heredity-based distributed constraint optimization problem solving method and application thereof |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103164745A (en) * | 2011-12-13 | 2013-06-19 | 中国人民解放军第二炮兵工程学院 | Maintenance supply chain integration mechanism based on ant colony algorithm and multi-agent technology |
CN103995750A (en) * | 2014-06-04 | 2014-08-20 | 重庆大学 | Asymmetric distributed constrained optimization method for multi-Agent system |
US20150284010A1 (en) * | 2013-09-16 | 2015-10-08 | Disney Enterprises, Inc. | Shared control of semi-autonomous vehicles including collision avoidance in multi-agent scenarios |
CN106684913A (en) * | 2016-12-29 | 2017-05-17 | 中国电力科学研究院 | Energy storage power station tracking generation plan control system and method based on multiple agents |
-
2018
- 2018-08-02 CN CN201810867224.0A patent/CN109190900A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103164745A (en) * | 2011-12-13 | 2013-06-19 | 中国人民解放军第二炮兵工程学院 | Maintenance supply chain integration mechanism based on ant colony algorithm and multi-agent technology |
US20150284010A1 (en) * | 2013-09-16 | 2015-10-08 | Disney Enterprises, Inc. | Shared control of semi-autonomous vehicles including collision avoidance in multi-agent scenarios |
CN103995750A (en) * | 2014-06-04 | 2014-08-20 | 重庆大学 | Asymmetric distributed constrained optimization method for multi-Agent system |
CN106684913A (en) * | 2016-12-29 | 2017-05-17 | 中国电力科学研究院 | Energy storage power station tracking generation plan control system and method based on multiple agents |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110378663A (en) * | 2019-04-19 | 2019-10-25 | 西北工业大学 | A kind of complicated movement crowdsourcing method for allocating tasks based on Greedy strategy |
CN110533301A (en) * | 2019-08-09 | 2019-12-03 | 大连理工大学 | A kind of population dispatching method based on dynamic constrained matrix |
CN110533301B (en) * | 2019-08-09 | 2022-08-23 | 大连理工大学 | Particle swarm scheduling method based on dynamic constraint matrix |
CN113408823A (en) * | 2021-07-13 | 2021-09-17 | 重庆理工大学 | Ant colony heredity-based distributed constraint optimization problem solving method and application thereof |
CN113408823B (en) * | 2021-07-13 | 2022-12-13 | 重庆理工大学 | Emergency rescue method for urban emergency |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Shyalika et al. | Reinforcement learning in dynamic task scheduling: A review | |
CN109375601A (en) | Assembly line planning method and equipment based on the optimization of data-driven modeling and simulation | |
CN104408589A (en) | AGV optimization scheduling method based on mixed particle swarm optimization | |
CN107045667B (en) | Measurement and control data transmission integrated station network resource scheduling method | |
CN109409773B (en) | Dynamic planning method for earth observation resources based on contract network mechanism | |
CN108256671A (en) | A kind of more resources of multitask based on learning-oriented genetic algorithm roll distribution method | |
CN112215350B (en) | Method and device for controlling agent based on reinforcement learning | |
Khmeleva et al. | Fuzzy-logic controlled genetic algorithm for the rail-freight crew-scheduling problem | |
CN109190900A (en) | A kind of method that distribution Constraint Anchored Optimization solves AGV scheduling system task distribution | |
WO2019127945A1 (en) | Structured neural network-based imaging task schedulability prediction method | |
Zhang et al. | An improved particle swarm optimization algorithm for integrated scheduling model in AGV-served manufacturing systems | |
CN109767128A (en) | imaging satellite autonomous task planning method based on machine learning | |
CN101617328A (en) | The automatic configuration method and the device that in problem solving environment, are used for meta-heuristic algorithms | |
CN103246969B (en) | A kind of implementation method of logistics deployment and device | |
CN113760553B (en) | Mixed part cluster task scheduling method based on Monte Carlo tree search | |
CN112801540A (en) | Intelligent cooperative architecture design based on unmanned cluster | |
Yu et al. | An extended flexible job shop scheduling model for flight deck scheduling with priority, parallel operations, and sequence flexibility | |
CN109039428A (en) | Repeater satellite single access antenna based on conflict resolution dispatches stochastic search methods | |
CN116643877A (en) | Computing power resource scheduling method, training method and system of computing power resource scheduling model | |
CN109709985A (en) | A kind of unmanned plane task optimization method, apparatus and system | |
CN102566423A (en) | Balancing control method for single-sided assembly line | |
CN116739323A (en) | Intelligent evaluation method and system for emergency resource scheduling | |
Peng et al. | A constraint programming method for advanced planning and scheduling system with multilevel structured products | |
CN106096159B (en) | A kind of implementation method of distributed system behavior simulation analysis system under cloud platform | |
CN117472552A (en) | Service scene intelligent arrangement and dynamic scheduling method, device, equipment and medium |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190111 |