CN110046851A - Unmanned vehicle logistics method for allocating tasks based on Multi-Paxos - Google Patents
Unmanned vehicle logistics method for allocating tasks based on Multi-Paxos Download PDFInfo
- Publication number
- CN110046851A CN110046851A CN201910268757.1A CN201910268757A CN110046851A CN 110046851 A CN110046851 A CN 110046851A CN 201910268757 A CN201910268757 A CN 201910268757A CN 110046851 A CN110046851 A CN 110046851A
- Authority
- CN
- China
- Prior art keywords
- task
- mission
- num
- trolley
- proposal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/083—Shipping
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Abstract
The invention discloses the unmanned vehicle logistics method for allocating tasks based on Multi-Paxos, specifically includes the following steps: step 1: two level logistics physical models of building and three phased mission state models form the concrete scene with city logistics transportation；Step 2: unmanned vehicle campaigns for task using Multi-paxos election algorithm；Step 3: unmanned vehicle determines the task packet sequence obtained based on TSP traveling salesman problem with Branch-and-Bound Algorithm.The present invention carries out task distribution using Multi-Paxos election algorithm, the determination of branch and bound method progress task point sequence, in control decentralization logistics transportation system scene, it is more applicable for the current same city logistic pattern such as taking out, it realizes more efficient dynamic unmanned vehicle task distribution, creates shorter average task completion time and more steady scheduling system.
Description
Technical field
The invention belongs to unmanned logistics and task allocation technique field, and in particular to a kind of nothing based on Multi-Paxos
People's vehicle logistics method for allocating tasks.
Background technique
As the scale of construction of e-commerce industry constantly expands and the continuous improvement of cost of labor, what dispatching was operated as logistics
Basic link, since its logistics form is increasingly various, data update is more frequent, and dispensing task is increasingly heavy, more has " in delivery
The raising of the service request of door ", wisdom logistics are increasingly prevailing, and Chinese logistics have turned on the unmanned epoch.
Full-automation intelligent scheduling technology in warehouse has developed quite mature at home, however unmanned intelligence dispenses on the spot
Technology is also in the budding stage.Unmanned vehicle dispatching in city has just been tried in Jingdone district early in 2016, however so far still without popularizing, it can
See that distribution technology development still has huge obstruction on the spot.
Method development based on the distribution of centralization scheduler task is more early, also relative maturity, has numerous based on the calculation of such as ant colony
The centralization dispatching algorithm of method, particle swarm algorithm, cluster etc. has a central control machine device people, be responsible for collecting environmental information,
And divide numerous tasks, it distributes to different robots, junior and goes to execute.However this centralization scheduler task distribution
Method is only applicable in static scene, and central robot, once breaking down, whole system will be in paralyzed state.
Summary of the invention
The purpose of invention is to provide a kind of suitable for dynamic scene and the high nothing based on Multi-Paxos of system robustness
People's Che Tongcheng logistics method for allocating tasks.
To realize the above-mentioned technical purpose, present invention employs following technical solutions: the unmanned vehicle object based on Multi-Paxos
Method for allocating tasks is flowed, specifically includes the following steps:
Step 1: two level logistics physical models of building and three phased mission state models are formed with city logistics transportation
Concrete scene；
Step 2: unmanned vehicle campaigns for task using Multi-paxos election algorithm；
Step 3: unmanned vehicle determines the task packet sequence obtained based on TSP traveling salesman problem with Branch-and-Bound Algorithm.
Further, the unmanned vehicle logistics method for allocating tasks above-mentioned based on Multi-Paxos, in which: in step 1
In, specific step is as follows for the building of logistics physical model and information initializing:
Divide multiple units according to actual needs based on geographic area, in each unit one cluster of setting and Duo Tai without
People's trolley；Cluster is mainly responsible for the energy supply and maintenance of trolley, the transfer across unit assignment and temporary, the letter of said units
Breath is collected, summarizes and submit；Unmanned trolley is only responsible for the election contest of task, and task point sequence is determining and executes；
Step (1.1): cluster information initializing Concentrator (id, i, n, Nmax), wherein
Concentrator.id is cluster geographical location information, and Concentrator.i is untreated across the unit assignment number that enters the station,
Concentrator.n is the number of tasks (including Concentrator.i) that cluster is temporarily put, and Concentrator.Nmax is collection
Middle station peak load amount of storage；Early warning is sent to monitoring platform when Concentrator.n is close to Concentrator.Nmax
Information；
Step (1.2): it is responsible for collection and the unit order store (i.e. task library) of the region order by each cluster
Generation change and management；
Wherein MissionBase.num is task library number to task library MissionBase (num, t, n, list []),
MissionBase.t is the limited timeline of task library, and MissionBase.n is number of tasks in task library,
MissionBase.list [] is in task library according to task issuing time (across the unit assignment packet come from the operating of other units
Equally abide by the rule) sort task sequence；Task library is substituted according to the period, as MissionBase.n=0,
MissionBase.t is moved back with time interval MissionBase.TInterval, until MissionBase.t be greater than etc.
When current；Only issuing before MissionBase.t for task just can enter task library and be assigned, namely only previous
The task of period is all assigned and executes, and the task of the latter period just can enter assigned ready state；Appoint entering the station
The larger period is flowed in business, and MissionBase.t may be earlier than current time；In the task flow that the enters the station smaller period, MissionBase.t
Current time may be later than, then it can real-time update into library task；
Step (1.3): unmanned trolley task initialization AGV (num, type, id, f, situation, n) is wherein
AGV.num is trolley number, and AGV.type is trolley type, and AGV.id is trolley geographical location information, and AGV.f is trolley residue
Fuel quantity, AGV.situation are trolley current states, and AGV.n is the task packet number that trolley executes；As AGV.type=1,
Unmanned vehicle is GENERAL TYPE, normal to participate in task distribution；As AGV.type=2 and AGV.type=3, unmanned vehicle is respectively rapidly
Type and super large load type, only transport indicates the task packet that urgent part and great good identify respectively, since two kinds of task packets are opposite
It is minimum in the ratio of general task packet, and the feature that its timeliness and volume mass respectively is larger, two kinds of vehicles are all made of one
The rule of secondary task packet；When trolley, which does not compete task packet, to be in idle condition, AGV.situation=0；When trolley exists
When execution task packet, AGV.situation=1；When trolley is competing task packet, AGV.situation=2；Work as trolley
In maintenance can not working condition when, AGV.situation=3.Trolley is only approached in the task consumption summation campaigned for
AGV.f or in the case where having campaigned for task packet in task library, can just go execution task；
Step (1.4): different units span unit assignment then uses one or more large-scale unmanned vehicle polling units collection
The mode at middle station is transported.
Further, the unmanned vehicle logistics method for allocating tasks above-mentioned based on Multi-Paxos, in which: in step 1
In, specific step is as follows for the building of three phased mission state models and mission bit stream initialization:
Step (2.1): mission bit stream initialization Mission (num, type, t, src, dst, src ', dst ', f,
Situation, step, chosen), wherein Mission.num is mission number, and Mission.type is task type,
Mission.t is the task initiation time, and Mission.src and Mission.dst are respectively the true source address and purpose of task
Address, Mission.src ' and Mission.dst ' be respectively task in source unit when or purpose unit in when unit it is endogenous
Address and destination address are then that source unit concentrates station address and purpose unit to concentrate station address when transporting between unit,
Mission.f estimates consumption trolley energy when being transport in TU task unit, Mission.stiuation is shape locating for task packet
State, Mission.step are the stage belonging to task packet transports, and Mission.chosen is to be assigned unmanned trolley number；Task packet
Corresponding unmanned vehicle type is divided into three types, when Mission.type=1, is GENERAL TYPE, need to participate in the distribution into task library；
Mission.type=2 and Mission.type=3 is respectively urgent type and ultra-large type, is not involved in distribution, is directly entered each
From sequence, the unmanned vehicle of idle affiliated vehicle is distributed to according to Mission.t；
Step (2.2): the building of Three-stage Model: for the task across unit, as Mission.step=1, task
Packet is in source unit, Mission.src '=Mission.src, Mission.dst '=source unit cluster；When
When Mission.step=2, task packet transports between being in unit；As Mission.step=3, task packet is in purpose unit
It is interior, Mission.src '=purpose unit cluster, Mission.dst '=Mission.dst；For non-across unit assignment,
Mission.step=0 is constant always；Mission.step, Mission.src ', Mission.dst ' respectively by source unit and
Purpose unit cluster is modified；Mission.situation=1 is that task packet is not acquired in source address,
Mission.situation=2 is task packet just in travel position, and Mission.situation=3 is that task packet arrived mesh
Address.
Further, the unmanned vehicle logistics method for allocating tasks above-mentioned based on Multi-Paxos, in which: in step 2
In, unmanned vehicle campaigns for task using Multi-paxos, and specific step is as follows:
Step (3.1): after a unmanned vehicle has executed selected all tasks, AGV.f lower than minimum angle value then into
Enter maintenance state, otherwise enters idle state；Task library is updated in an idle state, if task library is sky, continues to monitor, it is no
Then enter election contest task bag-like state；According to Mission.t be priority to Mission.f < AGV.f exceed early the issuing of the task into
Row selection；
Step (3.2): time shaft vertical analysis is carried out to the election contest process of some task in task library, i.e., individually
Paxos process, specifically includes:
Step (3.2.1): it is intended to select the unmanned trolley Proposer of the task, (is more than in the unit to majority's trolley
Other unmanned trolleies of half) and cluster transmission proposal Proposal request Prepare (Mission.num,
Proposer.num, Proposal.n) wherein Mission.num is the number of task to, Proposer.num is the Proposer
Number, Proposal.n is number about task Proposal；
Step (3.2.2): being connected to the Acceptor of Prepare request, if what the Proposal.n was received before being less than it
The number that other Prepare about the task are requested, then directly abandon the secondary request and ignore；If more than, then promise to undertake by
It will not receive and propose to number any request smaller than the Proposal.n, and return to promise response.If not receiving proposal before
It then directly receives, and returns to promise response；
It promises to undertake response Promised (Mission.num, Preci.num, Preci.f), wherein Mission.num is task
Number, the trolley number that Preci.num is the Proposal of a upper Accept, Preci.f is a upper Accept
The trolley remaining fuel amounts of Proposal；If not receiving any Accept before trolley, Promised content is sky；
Step (3.2.3): after the Promised that Proposer has obtained in the unit more than the trolley of half is responded, then to
Majority's trolley and cluster transmission please adopt request, and (this majority's trolley set can be small with the majority in Step1
Vehicle is unequal)；
Please adopt request Accept (Mission.num, Proposer.num, Proposal.n, winner.num,
Winner.f), wherein Mission.num is the number of task, and Proposer.num is the number of the Proposer,
Proposal.n is the number about task Proposal, and
Precik.f, promise response of the Preci.num from multiple returns in Step2；
If not obtaining responding more than the Promised of half trolley, the secondary Proposal failure, trolley is initiated again
Proposal, and Proposal.n is added 1 on the basis of original, but Proposal.n is no more than NPmax (maximum ballot
Number), i.e., the number that trolley initiates Proposal is limited；
Step (3.2.4): Acceptor is received after please adopting request, if Proposal.n holding of not violating that do-it-yourself crosses
Promise is then modified and saves the mission bit stream:
And return to have adopted and respond Accepted (Mission.num, Proposal.n),
As the Proposal.n==NPmax that please adopt request of receiving, then chosen.num is the final task
Assigned trolley；
Adopt if Acceptor occur and not having yet in the timeoff time after adopting certain Proposal
Proposal, then chosen.num at this time is the assigned trolley of the final task；
Step (3.2.5): have been determined that the trolley of final task distribution sends confirm to cluster
(Mission.num, chosen.num) determines information, and wherein Mission.num is the number of task, and chosen.num is most
The assigned trolley of the task eventually；The confirmation information content that cluster receives the task that number is Mission.num for the first time
It modifies and updates task library；
Step (3.3): by the process distributed using simplicity paxos election algorithm individual task of step (3.2) description
As an instance, then each task is independently of each other, not interfere with each other, can synchronize progress in task library
Instance allows the same moment to have multiple tasks being campaigned for, unmanned small workshop and the communication between cluster
It is identified by mission number Mission.num as instance.num, prevents the conflict of different task instance.
Further, the unmanned vehicle logistics method for allocating tasks above-mentioned based on Multi-Paxos, in which: in step 2
In, the election contest task operating of unmanned trolley, which is built upon, to carry out on the basis of local task library, therefore, to assure that the sheet of each trolley
Task library real-time synchronization in election contest in ground updates, and guarantee accurately learns that situation is deleted in the addition of task in task library in real time, is based on
Specific step is as follows for the task library synchronized update of Multi-paxos:
Step (4.1): the special instance* stack that the update of task library is realized separately as one using paxos is (with step
Instance in rapid 3 is different) specification is put, it is the consistency operation of the increase delete operation to task, each time one
The operating process of cause property is an instance, the specific implementation process is as follows:
Step (4.1.1): the main cluster of the instance (can temporarily use when failure and act on behalf of trolley) as
Proposer, unmanned trolley are Acceptor.Cluster to all trolleies issue Prepare (instance.num,
Proposal.n) propose request, wherein instance.num is the number that instance* is different from instance in step 3,
Proposal.n is the number of the secondary proposal；Instance.num on the basis of previous round instance.num plus 1, only before
One wheel instance completion just can be carried out next round instance；
Step (4.1.2): receiving the unmanned trolley of Prepare, if Proposal.n is less than before in the instance
Some Proposal.n received is then abandoned and is ignored；If it is greater than or equal to then receiving and return to Promised
(instance.num, Proposal.n) promises to undertake response；
Step (4.1.3): cluster has obtained after responding more than the Promised for the trolley in the unit being more than half, then
To majority's trolley transmission please adopt request Accept (instance.num, Proposal.n, add (Mission (...))/
Delete (Mission.num) ... }), wherein add (Mission (...)) indicates to increase the operation of some task, delete
(Mission.num) indicate that deleting some has distributed the operation of task；Multiple addition delete operation requests can once be sent；
When the Promised received, which is responded, is no more than half, Prepare is initiated again and proposes request, Proposal.n exists
On the basis of original plus one, Proposal.n increases no maximum；But excessive then issue to administrator of Proposal.n is alarmed, and is concentrated
It stands out of touch with more unmanned trolleies；
Step (4.1.4): unmanned trolley receives after please adopting request, if Proposal.n holding of not violating that do-it-yourself crosses
Promise is then updated Accept request content to task library, and return adopted respond Accepted (instance.num,
Proposal.n)；
Step (4.2): for the loss for preventing the part communication information, study mechanism is introduced, i.e., unmanned trolley is each in election contest
Learning tasks library before task, the specific steps are as follows:
Step (4.2.1): unmanned trolley initiates study request Learn to cluster or the unmanned trolley of surrounding
(latest.num, deficiency.num, deficiency1.num... }), wherein latest.num is the trolley
Maximum instance.num in instance* stack, deficiency.num are that should lack in continuous instance.num
Instance.num, and may have multiple；
Step (4.2.2): the cluster of study request learn is received, is found bigger than latest.num
instance.num；And the deficiency.num of missing, return response message Response (instance.num,
Operation (...) }, { instance.num, operation (...) } ...), wherein instance.num and operation
That is the sequence of operation that the instance is determined collectively constitutes one to reply.
Further, the unmanned vehicle logistics method for allocating tasks above-mentioned based on Multi-Paxos, in which: in step 3
In, with the determining task packet sequence obtained of Branch-and-Bound Algorithm, specific step is as follows based on TSP traveling salesman problem for unmanned vehicle:
Step (5.1): next functionDetermination:
Step (5.1.1): the next function of root: building cost matrix R, rijFor the i-th place to jth place path away from
From the r as i=jij=∞；Then reduction is carried out, then
L is the approximate number of cost matrix, tiFor the approximate number (0≤i < n) of the i-th row, kjFor the approximate number (0≤j < n) of jth row；
Step (5.1.2): it non-leaf state next time function: is calculated according to the cost matrix A of its parent node；I.e.
(1) it enablesGenerate A'；
(2) reduction is implemented to A' again and obtains B, the matrix approximate number of reduction is L at this timeB, then
Step (5.1.3): the next function of leaf node: page status node S'sFor the road traversed along the paths
Electrical path length；
Step (5.2): previous function u (X)=∞；
Step (5.3): the step of LC branch and bound method, is as follows:
Step (5.3.1): generating root node, and trolley is at pause at this time, from pause place, then by all numbers
Place generates child nodes, sequentially enters priority queue, next functional value is priority；
Step (5.3.2): priority is highest from priority queue falls out as E- node, and once generates its child
Node sequentially enters priority queue；
Step (5.3.3): repeating step (5.3.2) operation, until state space tree reaches at (n+1)th layer of leaf node,
A paths have been generated at this time；
Step (5.3.4): whether priority queue root node priority is higher than the priority of the leaf node at this time for judgementStep (5.3.2), step (5.3.3) are repeated if being higher than, and are generated new sequence and are obtained the priority of its leaf nodeIfThen ans pointer is directed toward the leaf node；
Step (5.3.5): repetitive operation step (5.3.4) is lower than leaf node until priority queue root node priority
PriorityThen ans pointer is directed toward the leaf node, and minimum cost sequence is sequence of the root node to the leaf node
Column；
Step (5.4): child's node selection range: selected place is formed on certain state node to root node path
Set be M, then child's node set from two aspect:
It is final to determine task execution sequence.
Through the implementation of the above technical solution, the beneficial effects of the present invention are:
(1) it is suitable for dynamic scene, can be realized the reduction with the cluster traffic and traffic load, and avoid center
The problems such as " single point failure " problem under change model, telecommunication quality is bad, most importantly there is lost contact in single trolley
Or in the case where machine, access system after total system is operated and easily repaired is not influenced；
(2) the unmanned wisdom logistics framework in same city an of decentralization is created；
(3) present invention uses Multi-paxos election algorithm, carries out more unmanned vehicle election contest formula task distribution, improves system
Robustness avoids " single point failure " problem, reduces traffic and communication load, meets while meeting flexibility in real time
Property；
(4) study mechanism is introduced, to greatly allow the failure for working as machine and any communication information of any trolley
Afterwards, Information recovering works, and improves serious forgiveness；
(5) present invention carries out task distribution using Multi-Paxos election algorithm, and branch and bound method carries out task point sequence
Determination be more applicable for the current same city logistics mould such as taking out in control decentralization logistics transportation system scene
Formula realizes more efficient dynamic unmanned vehicle task distribution, creates shorter average task completion time and more steady tune
Degree system.
Detailed description of the invention
Fig. 1 is the flow diagram that Multi-Paxos elects task allocation algorithms in the present invention.
Fig. 2 is the flow diagram of task library more new algorithm in the present invention.
Fig. 3 is the flow diagram that branch and bound method determines task execution sequence in the present invention.
Specific implementation method
Unmanned vehicle logistics method for allocating tasks based on Multi-Paxos, specifically includes the following steps:
Step 1: two level logistics physical models of building and three phased mission state models are formed with city logistics transportation
Concrete scene；
Wherein, specific step is as follows for the building of logistics physical model and information initializing:
Divide multiple units according to actual needs based on geographic area, in each unit one cluster of setting and Duo Tai without
People's trolley；Cluster is mainly responsible for the energy supply and maintenance of trolley, the transfer across unit assignment and temporary, the letter of said units
Breath is collected, summarizes and submit；Unmanned trolley is only responsible for the election contest of task, and task point sequence is determining and executes；
Step (1.1): cluster information initializing Concentrator (id, i, n, Nmax), wherein
Concentrator.id is cluster geographical location information, and Concentrator.i is untreated across the unit assignment number that enters the station,
Concentrator.n is the number of tasks (including Concentrator.i) that cluster is temporarily put, and Concentrator.Nmax is collection
Middle station peak load amount of storage；Early warning is sent to monitoring platform when Concentrator.n is close to Concentrator.Nmax
Information；
Step (1.2): it is responsible for collection and the unit order store (i.e. task library) of the region order by each cluster
Generation change and management；
Wherein MissionBase.num is task library number to task library MissionBase (num, t, n, list []),
MissionBase.t is the limited timeline of task library, and MissionBase.n is number of tasks in task library,
MissionBase.list [] is in task library according to task issuing time (across the unit assignment packet come from the operating of other units
Equally abide by the rule) sort task sequence；Task library is substituted according to the period, as MissionBase.n=0,
MissionBase.t is moved back with time interval MissionBase.TInterval, until MissionBase.t be greater than etc.
When current；Only issuing before MissionBase.t for task just can enter task library and be assigned, namely only previous
The task of period is all assigned and executes, and the task of the latter period just can enter assigned ready state；Appoint entering the station
The larger period is flowed in business, and MissionBase.t may be earlier than current time；In the task flow that the enters the station smaller period, MissionBase.t
Current time may be later than, then it can real-time update into library task；
Step (1.3): unmanned trolley task initialization AGV (num, type, id, f, situation, n) is wherein
AGV.num is trolley number, and AGV.type is trolley type, and AGV.id is trolley geographical location information, and AGV.f is trolley residue
Fuel quantity, AGV.situation are trolley current states, and AGV.n is the task packet number that trolley executes；As AGV.type=1,
Unmanned vehicle is GENERAL TYPE, normal to participate in task distribution；As AGV.type=2 and AGV.type=3, unmanned vehicle is respectively rapidly
Type and super large load type, only transport indicates the task packet that urgent part and great good identify respectively, since two kinds of task packets are opposite
It is minimum in the ratio of general task packet, and the feature that its timeliness and volume mass respectively is larger, two kinds of vehicles are all made of one
The rule of secondary task packet；When trolley, which does not compete task packet, to be in idle condition, AGV.situation=0；When trolley exists
When execution task packet, AGV.situation=1；When trolley is competing task packet, AGV.situation=2；Work as trolley
In maintenance can not working condition when, AGV.situation=3.Trolley is only approached in the task consumption summation campaigned for
AGV.f or in the case where having campaigned for task packet in task library, can just go execution task；
Step (1.4): different units span unit assignment then uses one or more large-scale unmanned vehicle polling units collection
The mode at middle station is transported.If region is larger, poll duration is too long, in order to shorten poll time, can take multiple lists
Tuple answer print, each is built a Ge Pian cluster on the basis of being based on unit cluster, in across piece task packet
It transports defeated.It can also continue to multilayer nest according to actual conditions, realize the transport of the larger task packet of span, realization principle is similar, under
Text only realizes the distribution of task packet in description piece；
Wherein, specific step is as follows for the building of three phased mission state models and mission bit stream initialization:
Step (2.1): mission bit stream initialization Mission (num, type, t, src, dst, src ', dst ', f,
Situation, step, chosen), wherein Mission.num is mission number, and Mission.type is task type,
Mission.t is the task initiation time, and Mission.src and Mission.dst are respectively the true source address and purpose of task
Address, Mission.src ' and Mission.dst ' be respectively task in source unit when or purpose unit in when unit it is endogenous
Address and destination address are then that source unit concentrates station address and purpose unit to concentrate station address when transporting between unit,
Mission.f estimates consumption trolley energy when being transport in TU task unit, Mission.stiuation is shape locating for task packet
State, Mission.step are the stage belonging to task packet transports, and Mission.chosen is to be assigned unmanned trolley number；Task packet
Corresponding unmanned vehicle type is divided into three types, when Mission.type=1, is GENERAL TYPE, need to participate in the distribution into task library；
Mission.type=2 and Mission.type=3 is respectively urgent type and ultra-large type, is not involved in distribution, is directly entered each
From sequence, the unmanned vehicle of idle affiliated vehicle is distributed to according to Mission.t；
Step (2.2): the building of Three-stage Model: for the task across unit, as Mission.step=1, task
Packet is in source unit, Mission.src '=Mission.src, Mission.dst '=source unit cluster；When
When Mission.step=2, task packet transports between being in unit；As Mission.step=3, task packet is in purpose unit
It is interior, Mission.src '=purpose unit cluster, Mission.dst '=Mission.dst；For non-across unit assignment,
Mission.step=0 is constant always；Mission.step, Mission.src ', Mission.dst ' respectively by source unit and
Purpose unit cluster is modified；Mission.situation=1 is that task packet is not acquired in source address,
Mission.situation=2 is task packet just in travel position, and Mission.situation=3 is that task packet arrived mesh
Address；
Step 2: unmanned vehicle campaigns for task using Multi-paxos election algorithm, and each car is according to itself shape
Each task estimates consumption indicators in condition and task library, and election contest is lower than upper limit value task, and is tried to be the first with ballot and energy storage residence
Mostly priority principle, as shown in Figure 1, the specific steps are as follows:
Step (3.1): after a unmanned vehicle has executed selected all tasks, AGV.f lower than minimum angle value then into
Enter maintenance state, otherwise enters idle state；Task library is updated in an idle state, if task library is sky, continues to monitor, it is no
Then enter election contest task bag-like state；According to Mission.t be priority to Mission.f < AGV.f exceed early the issuing of the task into
Row selection；
Step (3.2): time shaft vertical analysis is carried out to the election contest process of some task in task library, i.e., individually
Paxos process, specifically includes:
Step (3.2.1): it is intended to select the unmanned trolley Proposer of the task, (is more than in the unit to majority's trolley
Other unmanned trolleies of half) and cluster transmission proposal Proposal request Prepare (Mission.num,
Proposer.num, Proposal.n) wherein Mission.num is the number of task to, Proposer.num is the Proposer
Number, Proposal.n is number about task Proposal；
Step (3.2.2): being connected to the Acceptor of Prepare request, if what the Proposal.n was received before being less than it
The number that other Prepare about the task are requested, then directly abandon the secondary request and ignore；If more than, then promise to undertake by
It will not receive and propose to number any request smaller than the Proposal.n, and return to promise response.If not receiving proposal before
It then directly receives, and returns to promise response；
Promise response Promised (Mission.num, Preci.num, Preci.f }), wherein Mission.num is to appoint
The number of business, the trolley number that Preci.num is the Proposal of upper one promise, Preci.f are upper one promise
The trolley remaining fuel amounts of Proposal；If not receiving any promise before trolley, Preci.num, Preci.f }
It omits；
Step (3.2.3): after the Promised that Proposer has obtained in the unit more than the trolley of half is responded, then to
Majority's trolley and cluster transmission please adopt request (this majority's trolley set can with it is more in step (3.2.1)
Number sends trolley unequal)；
Please adopt request Accept (Mission.num, Proposer.num, Proposal.n, winner.num,
Winner.f), wherein Mission.num is the number of task, and Proposer.num is the number of the Proposer,
Proposal.n is the number about task Proposal, and
Precik.f, promise response of the Preci.num from multiple returns in Step2；
If not obtaining responding more than the Promised of half trolley, the secondary Proposal failure, trolley is initiated again
Proposal, and Proposal.n is added 1 on the basis of original, but Proposal.n is no more than NPmax (maximum ballot
Number), i.e., the number that trolley initiates Proposal is limited；
Step (3.2.4): Acceptor is received after please adopting request, if Proposal.n holding of not violating that do-it-yourself crosses
Promise is then modified and saves the mission bit stream:
And return to have adopted and respond Accepted (Mission.num, Proposal.n),
As the Proposal.n==NPmax that please adopt request of receiving, then chosen.num is the final task
Assigned trolley；
Adopt if Acceptor occur and not having yet in the timeoff time after adopting certain Proposal
Proposal, then chosen.num at this time is the assigned trolley of the final task；
Step (3.2.5): have been determined that the trolley of final task distribution sends confirm to cluster
(Mission.num, chosen.num) determines information, and wherein Mission.num is the number of task, and chosen.num is most
The assigned trolley of the task eventually；The confirmation information content that cluster receives the task that number is Mission.num for the first time
It modifies and updates task library；
Step (3.3): by the process distributed using simplicity paxos election algorithm individual task of step (3.2) description
As an instance, then each task is independently of each other, not interfere with each other, can synchronize progress in task library
Instance allows the same moment to have multiple tasks being campaigned for, unmanned small workshop and the communication between cluster
It is identified by mission number Mission.num as instance.num, prevents the conflict of different task instance；
Wherein, the election contest task operating of unmanned trolley, which is built upon, carries out on the basis of local task library, therefore, to assure that
The local task library of each trolley real-time synchronization in election contest updates, and guarantees the addition feelings for accurately learning task in task library in real time
Condition, as shown in Fig. 2, specific step is as follows for the task library synchronized update based on Multi-paxos:
Step (4.1): the special instance* stack that the update of task library is realized separately as one using paxos is (with step
Instance in rapid 3 is different) specification is put, it is the consistency operation of the increase delete operation to task, each time one
The operating process of cause property is an instance, the specific implementation process is as follows:
Step (4.1.1): the main cluster of the instance (can temporarily use when failure and act on behalf of trolley) as
Proposer, unmanned trolley are Acceptor.Cluster to all trolleies issue Prepare (instance.num,
Proposal.n) propose request, wherein instance.num is the number that instance* is different from instance in step 3,
Proposal.n is the number of the secondary proposal；Instance.num on the basis of previous round instance.num plus 1, only before
One wheel instance completion just can be carried out next round instance；
Step (4.1.2): receiving the unmanned trolley of Prepare, if Proposal.n is less than before in the instance
Some Proposal.n received is then abandoned and is ignored；If it is greater than or equal to then receiving and return to Promised
(instance.num, Proposal.n) promises to undertake response；
Step (4.1.3): cluster has obtained after responding more than the Promised for the trolley in the unit being more than half, then
Request Accept (instance.num, Proposal.n, add (Mission (...))) please be adopt to the transmission of majority's trolley,
Wherein add (Mission (...)) indicates to increase the operation of some task；Multiple addition operation requests can once be sent；
When the Promised received, which is responded, is no more than half, Prepare is initiated again and proposes request, Proposal.n exists
On the basis of original plus one, Proposal.n increases no maximum；But excessive then issue to administrator of Proposal.n is alarmed, and is concentrated
It stands out of touch with more unmanned trolleies；
Step (4.1.4): unmanned trolley receives after please adopting request, if Proposal.n holding of not violating that do-it-yourself crosses
Promise is then updated Accept request content to task library, and return adopted respond Accepted (instance.num,
Proposal.n)；
Step (4.2): for the loss for preventing the part communication information, study mechanism is introduced, i.e., unmanned trolley is each in election contest
Learning tasks library before task, the specific steps are as follows:
Step (4.2.1): unmanned trolley initiates study request Learn to cluster or the unmanned trolley of surrounding
(latest.num, deficiency.num, deficiency1.num... }), wherein latest.num is the trolley
Maximum instance.num in instance* stack, deficiency.num are that should lack in continuous instance.num
Instance.num, and may have multiple；
Step (4.2.2): the cluster of study request learn is received, is found bigger than latest.num
instance.num；And the deficiency.num of missing, return response message Response (instance.num,
Operation (...) }, { instance.num, operation (...) } ...), wherein instance.num and operation
That is the sequence of operation that the instance is determined collectively constitutes one to reply；
Step 3: unmanned vehicle determines the task packet sequence obtained based on TSP traveling salesman problem with Branch-and-Bound Algorithm, is based on
The source address of each task and destination address mark task point in practical map, and middle road distance is side right value, meter according to the map
Calculating meets source address task point in the preceding posterior task point sequence of destination address task point；As shown in figure 3, specific steps are such as
Under:
Trolley campaign for after task, the determination of task execution sequence is the deformation based on TSP problem；Each task has one
A key-value pair (source place, destination), and each place is numbered；The determination of task execution sequence is i.e. to gained institute
The source place and point of destination for having task mix sequence, and the place for coming front first reaches, and obtain an optimal arrival sequence；It wants
Asking the corresponding source place of each task in sequence must traverse before destination, while unlike TSP problem
It is not necessarily to return starting point after all the points；
Since according to onboard capabilities limitation to can be obtained number of tasks limited for each trolley, for each trolley in task
Although the determination problem of task sequence after the completion of election contest is the solution based on this NP difficulty problem of TSP travelling salesman, using institute
Actual demand can be met by obtaining the accurate branch and bound method of result；
Step (5.1): next functionDetermination:
Step (5.1.1): the next function of root: building cost matrix R, rijFor the i-th place to jth place path away from
From the r as i=jij=∞；Then reduction is carried out, then
L is the approximate number of cost matrix, tiFor the approximate number (0≤i < n) of the i-th row, kjFor the approximate number (0≤j < n) of jth row；
Step (5.1.2): it non-leaf state next time function: is calculated according to the cost matrix A of its parent node；I.e.
(3) it enablesGenerate A'；
(4) reduction is implemented to A' again and obtains B, the matrix approximate number of reduction is L at this timeB, then
Step (5.1.3): the next function of leaf node: page status node S'sFor the road traversed along the paths
Electrical path length；
Step (5.2): previous function u (X)=∞；
Step (5.3): the step of LC branch and bound method, is as follows:
Step (5.3.1): generating root node, and trolley is at pause at this time, from pause place, then by all numbers
Place generates child nodes, sequentially enters priority queue, next functional value is priority；
Step (5.3.2): priority is highest from priority queue falls out as E- node, and once generates its child
Node sequentially enters priority queue；
Step (5.3.3): repeating step (5.3.2) operation, until state space tree reaches at (n+1)th layer of leaf node,
A paths have been generated at this time；
Step (5.3.4): whether priority queue root node priority is higher than the priority of the leaf node at this time for judgementStep (5.3.2), step (5.3.3) are repeated if being higher than, and are generated new sequence and are obtained the priority of its leaf nodeIfThen ans pointer is directed toward the leaf node；
Step (5.3.5): repetitive operation step (5.3.4) is lower than leaf node until priority queue root node priority
PriorityThen ans pointer is directed toward the leaf node, and minimum cost sequence is sequence of the root node to the leaf node
Column；
Step (5.4): child's node selection range: selected place is formed on certain state node to root node path
Set be M, then child's node set from two aspect:
It is final to determine task execution sequence.
The invention has the advantages that
(1) it is suitable for dynamic scene, can be realized the reduction with the cluster traffic and traffic load, and avoid center
The problems such as " single point failure " problem under change model, telecommunication quality is bad, most importantly there is lost contact in single trolley
Or in the case where machine, access system after total system is operated and easily repaired is not influenced；
(2) the unmanned wisdom logistics framework in same city an of decentralization is created；
(3) present invention uses Multi-paxos election algorithm, carries out more unmanned vehicle election contest formula task distribution, improves system
Robustness avoids " single point failure " problem, reduces traffic and communication load, meets again while meeting flexibility in real time
Property；
(4) study mechanism is introduced, to greatly allow the failure for working as machine and any communication information of any trolley
Afterwards, Information recovering works, and improves serious forgiveness；
(5) present invention carries out task distribution using Multi-Paxos election algorithm, and branch and bound method carries out task point sequence
Determination be more applicable for the current same city logistics mould such as taking out in control decentralization logistics transportation system scene
Formula realizes more efficient dynamic unmanned vehicle task distribution, creates shorter average task completion time and more steady tune
Degree system.
Claims (6)
1. the unmanned vehicle logistics method for allocating tasks based on Multi-Paxos, it is characterised in that: specifically includes the following steps:
Step 1: two level logistics physical models of building and three phased mission state models are formed with the specific of city logistics transportation
Scene；
Step 2: unmanned vehicle campaigns for task using Multi-paxos election algorithm；
Step 3: unmanned vehicle determines the task packet sequence obtained based on TSP traveling salesman problem with Branch-and-Bound Algorithm.
2. the unmanned vehicle logistics method for allocating tasks according to claim 1 based on Multi-Paxos, it is characterised in that:
In step 1, specific step is as follows for the building of logistics physical model and information initializing:
Multiple units are divided according to actual needs based on geographic area, nobody is small by one cluster of setting and Duo Tai in each unit
Vehicle；Cluster is mainly responsible for the energy supply and maintenance of trolley, the transfer across unit assignment and temporary, the information receipts of said units
Collect, summarize and submits；Unmanned trolley is only responsible for the election contest of task, and task point sequence is determining and executes；
Step (1.1): cluster information initializing Concentrator (id, i, n, Nmax), wherein Concentrator.id is
Cluster geographical location information, Concentrator.i are untreated across the unit assignment numbers that enters the station, and Concentrator.n is
The number of tasks (including Concentrator.i) that cluster is temporarily put, Concentrator.Nmax are the storage of cluster peak load
Amount；Warning information is sent to monitoring platform when Concentrator.n is close to Concentrator.Nmax；
Step (1.2): it is responsible for the collection of the region order and the life of the unit order store (i.e. task library) by each cluster
At variation and management；
Wherein MissionBase.num is task library number to task library MissionBase (num, t, n, list []),
MissionBase.t is the limited timeline of task library, and MissionBase.n is number of tasks in task library,
MissionBase.list [] is in task library according to task issuing time (across the unit assignment packet come from the operating of other units
Equally abide by the rule) sort task sequence；Task library is substituted according to the period, as MissionBase.n=0,
MissionBase.t is moved back with time interval MissionBase.TInterval, until MissionBase.t be greater than etc.
When current；Only issuing before MissionBase.t for task just can enter task library and be assigned, namely only previous
The task of period is all assigned and executes, and the task of the latter period just can enter assigned ready state；Appoint entering the station
The larger period is flowed in business, and MissionBase.t may be earlier than current time；In the task flow that the enters the station smaller period, MissionBase.t
Current time may be later than, then it can real-time update into library task；
Step (1.3): wherein AGV.num is unmanned trolley task initialization AGV (num, type, id, f, situation, n)
Trolley number, AGV.type is trolley type, and AGV.id is trolley geographical location information, and AGV.f is trolley remaining fuel amounts,
AGV.situation is trolley current state, and AGV.n is the task packet number that trolley executes；As AGV.type=1, unmanned vehicle
It is normal to participate in task distribution for GENERAL TYPE；As AGV.type=2 and AGV.type=3, unmanned vehicle is respectively rapidly type and super
Big load type, only transport indicates the task packet that urgent part and great good identify respectively, since two kinds of task packets are relative to general
The ratio of task packet is minimum, and its timeliness and the larger feature of volume mass respectively, two kinds of vehicles be all made of one time one
The rule of business packet；When trolley, which does not compete task packet, to be in idle condition, AGV.situation=0；When trolley is appointed in execution
When business packet, AGV.situation=1；When trolley is competing task packet, AGV.situation=2；When trolley is in dimension
Shield can not working condition when, AGV.situation=3.Trolley only campaign for task consumption summation approach AGV.f or
In the case that person has campaigned for the task packet in task library, execution task can be just gone；
Step (1.4): different units span unit assignment then uses one or more large-scale unmanned vehicle polling units cluster
Mode transported.
3. the unmanned vehicle logistics method for allocating tasks according to claim 1 based on Multi-Paxos, it is characterised in that:
In step 1, specific step is as follows for the building of three phased mission state models and mission bit stream initialization:
Step (2.1): mission bit stream initialization Mission (num, type, t, src, dst, src ', dst ', f,
Situation, step, chosen), wherein Mission.num is mission number, and Mission.type is task type,
Mission.t is the task initiation time, and Mission.src and Mission.dst are respectively the true source address and purpose of task
Address, Mission.src ' and Mission.dst ' be respectively task in source unit when or purpose unit in when unit it is endogenous
Address and destination address are then that source unit concentrates station address and purpose unit to concentrate station address when transporting between unit,
Mission.f estimates consumption trolley energy when being transport in TU task unit, Mission.stiuation is shape locating for task packet
State, Mission.step are the stage belonging to task packet transports, and Mission.chosen is to be assigned unmanned trolley number；Task packet
Corresponding unmanned vehicle type is divided into three types, when Mission.type=1, is GENERAL TYPE, need to participate in the distribution into task library；
Mission.type=2 and Mission.type=3 is respectively urgent type and ultra-large type, is not involved in distribution, is directly entered each
From sequence, the unmanned vehicle of idle affiliated vehicle is distributed to according to Mission.t；
Step (2.2): the building of Three-stage Model: for the task across unit, as Mission.step=1, at task packet
In source unit, Mission.src '=Mission.src, Mission.dst '=source unit cluster；When
When Mission.step=2, task packet transports between being in unit；As Mission.step=3, task packet is in purpose unit
It is interior, Mission.src '=purpose unit cluster, Mission.dst '=Mission.dst；For non-across unit assignment,
Mission.step=0 is constant always；Mission.step, Mission.src ', Mission.dst ' respectively by source unit and
Purpose unit cluster is modified；Mission.situation=1 is that task packet is not acquired in source address,
Mission.situation=2 is task packet just in travel position, and Mission.situation=3 is that task packet arrived mesh
Address.
4. the unmanned vehicle logistics method for allocating tasks according to claim 1 based on Multi-Paxos, it is characterised in that:
In step 2, unmanned vehicle campaigns for task using Multi-paxos, and specific step is as follows:
Step (3.1): after a unmanned vehicle has executed selected all tasks, AGV.f then enters dimension lower than minimum angle value
Otherwise shield state enters idle state；Update task library in an idle state, if task library be sky, continue to monitor, otherwise into
Enter to campaign for task bag-like state；It is that priority selects early issuing for the task that exceedes of Mission.f < AGV.f according to Mission.t
It selects；
Step (3.2): time shaft vertical analysis, i.e., single paxos are carried out to the election contest process of some task in task library
Process specifically includes:
Step (3.2.1): being intended to select the unmanned trolley Proposer of the task, (is more than half in the unit to majority's trolley
Other unmanned trolleies) and cluster send propose Proposal request Prepare (Mission.num, Proposer.num,
Proposal.n) wherein Mission.num is the number of task to, and Proposer.num is the number of the Proposer,
Proposal.n is the number about task Proposal；
Step (3.2.2): be connected to Prepare request Acceptor, if the Proposal.n be less than its before receive about
The number of other Prepare request of the task, then directly abandon the secondary request and ignore；If more than then promising to undertake will not
It receives and proposes to number any request smaller than the Proposal.n, and return to promise response.It is straight if do not receive to propose before
It receives, and returns to promise response；
Promise response Promised (Mission.num, Preci.num, Preci.f }), wherein Mission.num is task
Number, the trolley number that Preci.num is the Proposal of upper one promise, Preci.f is the Proposal of upper one promise
Trolley remaining fuel amounts；If not receiving any promise before trolley, and, Preci.num, Preci.f } and it omits；
Step (3.2.3): after the Promised that Proposer has obtained in the unit more than the trolley of half is responded, then to majority
Sending trolley and cluster transmission that please adopt request, (this majority's trolley set can be with the majority in step (3.2.1)
Trolley set is unequal)；
Please adopt request Accept (Mission.num, Proposer.num, Proposal.n, winner.num,
Winner.f), wherein Mission.num is the number of task, and Proposer.num is the number of the Proposer,
Proposal.n is the number about task Proposal, and
Precik.f, promise response of the Preci.num from multiple returns in step (3.2.2)；
If not obtaining responding more than the Promised of half trolley, the secondary Proposal failure, trolley is initiated again
Proposal, and Proposal.n is added 1 on the basis of original, but Proposal.n is no more than NPmax (maximum ballot
Number), i.e., the number that trolley initiates Proposal is limited；
Step (3.2.4): Acceptor is received after please adopting request, if Proposal.n does not violate the promise that do-it-yourself is crossed,
It then modifies and saves the mission bit stream:
And return to have adopted and respond Accepted (Mission.num, Proposal.n),
As the Proposal.n==NPmax that please adopt request of receiving, then chosen.num is that the final task is divided
The trolley matched；
If Acceptor occur the Proposal adopted yet after adopting certain Proposal in the timeoff time,
Then chosen.num at this time is the assigned trolley of the final task；
Step (3.2.5): have been determined the trolley of final task distribution to cluster send confirm (Mission.num,
Chosen.num information) is determined, wherein Mission.num is the number of task, and chosen.num is that the final task is divided
The trolley matched；To the task that number is Mission.num, the confirmation information content that receives is modified update for the first time for cluster
Task library；
Step (3.3): using the process that individual task is distributed using simplicity paxos election algorithm of step (3.2) description as
One instance, then each task is independently of each other, not interfere with each other, can synchronize progress in task library
Instance allows the same moment to have multiple tasks being campaigned for, unmanned small workshop and the communication between cluster
It is identified by mission number Mission.num as instance.num, prevents the conflict of different task instance.
5. the unmanned vehicle logistics method for allocating tasks according to claim 4 based on Multi-Paxos, it is characterised in that:
In step 2, the election contest task operating of unmanned trolley, which is built upon, to be carried out on the basis of local task library, therefore, to assure that each
The local task library of trolley real-time synchronization in election contest updates, and guarantees the addition situation for accurately learning task in task library in real time
(only doing addition operation, delete operation is completed in the deletion of task at the end of corresponding instance), is based on Multi-paxos
Task library synchronized update specific step is as follows:
Step (4.1): the special instance* stack that the update of task library is realized separately as one using paxos is (with step 3
In instance it is different), be to task addition consistency operation, consistency operation process is one each time
Instance, the specific implementation process is as follows:
Step (4.1.1): the instance mainly by cluster (can temporarily use when failure and act on behalf of trolley) as
Proposer, unmanned trolley are Acceptor.Cluster to all trolleies issue Prepare (instance.num,
Proposal.n) propose request, wherein instance.num is the number that instance* is different from instance in step 3,
Proposal.n is the number of the secondary proposal；Instance.num on the basis of previous round instance.num plus 1, only before
One wheel instance completion just can be carried out next round instance；
Step (4.1.2): receiving the unmanned trolley of Prepare, if Proposal.n is received in the instance before being less than
Some Proposal.n, then abandon ignore；If it is greater than or equal to, then receive and return Promised (instance.num,
Proposal.n response) is promised to undertake；
Step (4.1.3): cluster has obtained after responding more than the Promised of half trolley in the unit, then small to majority
Vehicle transmission please adopt request Accept (instance.num, Proposal.n, add (Mission (...))), wherein add
(Mission (...)) indicates to increase the operation of some task；Multiple addition delete operation requests can once be sent；
When the Promised received, which is responded, is no more than half, Prepare is initiated again and proposes request, Proposal.n is original
On the basis of plus one, Proposal.n increase no maximum；But Proposal.n it is excessive then to administrator issue alarm, cluster with
More unmanned trolleies are out of touch；
Step (4.1.4): unmanned trolley receives after please adopting request, if Proposal.n does not violate the promise that do-it-yourself is crossed,
Then Accept request content is updated task library, and return adopted respond Accepted (instance.num,
Proposal.n)；
Step (4.2): for the loss for preventing the part communication information, study mechanism is introduced, i.e., unmanned trolley is campaigning for each task
Preceding learning tasks library, the specific steps are as follows:
Step (4.2.1): unmanned trolley initiates study request Learn to cluster or the unmanned trolley of surrounding
(latest.num, deficiency.num, deficiency1.num... }), wherein latest.num is the trolley
Maximum instance.num in instance* stack, deficiency.num are that should lack in continuous instance.num
Instance.num, and may have multiple；
Step (4.2.2): the cluster of study request learn is received, the instance.num bigger than latest.num is found；With
And the deficiency.num of missing, return response message Response ({ instance.num, operation (...) },
{ instance.num, operation (...) } ...), wherein the instance.num and i.e. instance of operation is determined
The sequence of operation collectively constitute one to reply.
6. the unmanned vehicle logistics method for allocating tasks according to claim 1 based on Multi-Paxos, it is characterised in that:
In step 3, unmanned vehicle determines the specific step of the task packet sequence obtained based on TSP traveling salesman problem with Branch-and-Bound Algorithm
It is rapid as follows:
Step (5.1): next functionDetermination:
Step (5.1.1): the next function of root: building cost matrix R, rijIt is the i-th place to the path distance in jth place, works as i
R when=jij=∞；Then reduction is carried out, then
L is the approximate number of cost matrix, tiFor the approximate number (0≤i < n) of the i-th row, kjFor the approximate number (0≤j < n) of jth row；
Step (5.1.2): it non-leaf state next time function: is calculated according to the cost matrix A of its parent node；I.e.
(1) it enablesGenerate A'；
(2) reduction is implemented to A' again and obtains B, the matrix approximate number of reduction is L at this timeB, then
Step (5.1.3): the next function of leaf node: page status node S'sFor the path length traversed along the paths
Degree；
Step (5.2): previous function u (X)=∞；
Step (5.3): the step of LC branch and bound method, is as follows:
Step (5.3.1): generating root node, and trolley is at pause at this time, from pause place, then by all number places
Child nodes are generated, sequentially enter priority queue, next functional value is priority；
Step (5.3.2): priority is highest from priority queue falls out as E- node, and once generates its child knot
Point, sequentially enters priority queue；
Step (5.3.3): step (5.3.2) operation is repeated, until state space tree reaches at (n+1)th layer of leaf node, at this time
A paths are generated；
Step (5.3.4): whether priority queue root node priority is higher than the priority of the leaf node at this time for judgement
Step (5.3.2), step (5.3.3) are repeated if being higher than, and are generated new sequence and are obtained the priority of its leaf node
IfThen ans pointer is directed toward the leaf node；
Step (5.3.5): repetitive operation step (5.3.4) is excellent lower than leaf node until priority queue root node priority
First weighThen ans pointer is directed toward the leaf node, and minimum cost sequence is sequence of the root node to the leaf node；
Step (5.4): child's node selection range: collect composed by selected place on certain state node to root node path
Conjunction is M, then child's node set is from two aspects:
It is final to determine task execution sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910268757.1A CN110046851B (en) | 2019-04-03 | 2019-04-03 | Unmanned vehicle logistics task allocation method based on Multi-Paxos |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910268757.1A CN110046851B (en) | 2019-04-03 | 2019-04-03 | Unmanned vehicle logistics task allocation method based on Multi-Paxos |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110046851A true CN110046851A (en) | 2019-07-23 |
| CN110046851B CN110046851B (en) | 2022-08-23 |
Family
ID=67276180
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910268757.1A Active CN110046851B (en) | 2019-04-03 | 2019-04-03 | Unmanned vehicle logistics task allocation method based on Multi-Paxos |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110046851B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111582804A (en) * | 2020-05-28 | 2020-08-25 | 中国人民解放军军事科学院国防科技创新研究院 | Task allocation method of unmanned transportation device |
| CN115440043A (en) * | 2021-09-26 | 2022-12-06 | 国信军创(岳阳)六九零六科技有限公司 | Real-time road condition information road monitoring management system based on RFID vehicle identification |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107727099A (en) * | 2017-09-29 | 2018-02-23 | 山东大学 | The more AGV scheduling of material transportation and paths planning method in a kind of factory |
| CN107845013A (en) * | 2017-09-25 | 2018-03-27 | 北京科东电力控制系统有限责任公司 | A kind of electric power transaction platform distributed application services calling processing method |
-
2019
- 2019-04-03 CN CN201910268757.1A patent/CN110046851B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107845013A (en) * | 2017-09-25 | 2018-03-27 | 北京科东电力控制系统有限责任公司 | A kind of electric power transaction platform distributed application services calling processing method |
| CN107727099A (en) * | 2017-09-29 | 2018-02-23 | 山东大学 | The more AGV scheduling of material transportation and paths planning method in a kind of factory |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111582804A (en) * | 2020-05-28 | 2020-08-25 | 中国人民解放军军事科学院国防科技创新研究院 | Task allocation method of unmanned transportation device |
| CN115440043A (en) * | 2021-09-26 | 2022-12-06 | 国信军创(岳阳)六九零六科技有限公司 | Real-time road condition information road monitoring management system based on RFID vehicle identification |
| CN115440043B (en) * | 2021-09-26 | 2023-07-25 | 国信军创(岳阳)六九零六科技有限公司 | Real-time road condition information road monitoring management system based on RFID vehicle identification |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110046851B (en) | 2022-08-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| de Weerdt et al. | Intention-aware routing of electric vehicles | |
| CN109117993B (en) | Processing method for optimizing vehicle path | |
| Guo et al. | A deep reinforcement learning approach to ride-sharing vehicle dispatching in autonomous mobility-on-demand systems | |
| CN111862579A (en) | Taxi scheduling method and system based on deep reinforcement learning | |
| Sun et al. | Dynamic path planning algorithms with load balancing based on data prediction for smart transportation systems | |
| CN113811915A (en) | Unified order serving and fleet management for online shared travel platform | |
| CN107203190A (en) | A kind of inertial navigation AGV dispatching methods and system based on pahtfinder hard | |
| CN112163720A (en) | Multi-agent unmanned electric vehicle battery replacement scheduling method based on Internet of vehicles | |
| CN104598994A (en) | Related logistics transportation optimized dispatching method with time-varying time window | |
| CN102136104A (en) | Load balance and Lin-Kernighan (LK) algorithm based vehicle route planning method | |
| CN103337167A (en) | Method for preventing traffic jams through determining final driving route by urban road scheduled allocation | |
| CN110110903A (en) | A kind of distribution vehicle paths planning method based on neural evolution | |
| US11774947B2 (en) | Industrial internet of things for material transportation control, control methods and media thereof | |
| CN110046851A (en) | Unmanned vehicle logistics method for allocating tasks based on Multi-Paxos | |
| Dai et al. | A balanced assignment mechanism for online taxi recommendation | |
| CN109920250A (en) | Dynamic prediction urban road intelligent traffic administration system method | |
| Wang et al. | Assessing the potential of the strategic formation of urban platoons for shared automated vehicle fleets | |
| Wang et al. | Cross-regional customized bus route planning considering staggered commuting during the COVID-19 | |
| Wang et al. | Towards accessible shared autonomous electric mobility with dynamic deadlines | |
| CN113140108B (en) | Cloud traffic situation prediction method in internet-connected intelligent traffic system | |
| CN109859475A (en) | A kind of intersection signal control method based on DBSCAN Density Clustering, apparatus and system | |
| Namoun et al. | A multi-agent system for modelling urban transport infrastructure using intelligent traffic forecasts | |
| CN115062868B (en) | Pre-polymerization type vehicle distribution path planning method and device | |
| Maipradit et al. | An adaptive traffic signal control scheme based on back-pressure with global information | |
| CN103336799B (en) | Railway traffic transfer plan querying method and device based on adaptive quantum ant colony |
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 |