CN109255546A - Flexible job shop scheduling method based on more heuristic information Ant ColonySystems - Google Patents
Flexible job shop scheduling method based on more heuristic information Ant ColonySystems Download PDFInfo
- Publication number
- CN109255546A CN109255546A CN201811149243.6A CN201811149243A CN109255546A CN 109255546 A CN109255546 A CN 109255546A CN 201811149243 A CN201811149243 A CN 201811149243A CN 109255546 A CN109255546 A CN 109255546A
- Authority
- CN
- China
- Prior art keywords
- heuristic information
- ant
- node
- heuristic
- workpiece
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Operations Research (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- General Factory Administration (AREA)
Abstract
The invention belongs to intelligent algorithm and production management field, it is related to the flexible job shop scheduling method based on more heuristic information Ant ColonySystems, comprising: establish graph model of extracting;Initialization;Construct the disaggregation of FJSP;Pheromone update;Heuristic information archives update;Termination condition inspection.The present invention introduces six kinds of heuristic informations in Ant ColonySystem, make algorithm that can obtain ideal performance in different problem-instances, it avoids falling into local optimum, and the adaptive strategy of heuristic information and its directed force control parameter can be automatically selected according to open problems example and current search state using one kind, the heuristic information of entire ant colony is designed to restrain to the current optimal design that can generate high-quality solution, improve speed of convergence, algorithm optimization efficiency is improved, flexible job shop scheduling efficiency is improved.
Description
Technical field
The invention belongs to intelligent algorithm and production management field, it is related to the flexible job based on more heuristic information Ant ColonySystems
Job-Shop method.
Background technique
Flexible Job-shop Scheduling Problems (Flexible Job-shop Scheduling Problem, FJSP) are tradition
The extension of job-shop scheduling problem (Job-shop Scheduling Problem, JSP), each process can by it is multiple not
It is completed with machine, and the deadline on different machines is different.Specifically, the FJSP of a N × M is described as follows: having N
A workpiece to be processed and M platform available machines used, workpiece collection are combined into { p1,p2,...,pN, collection of machines is { m1,m2,...,mM,
Each workpiece includes one or multi-channel process, the n of workpiece i (i=1,2 ..., N)iProcedure is by set opi={ opi1,
opi2,...,opiniIndicate.The process processing sequence of workpiece be it is determining, every procedure can add on more different machines
Work, the target of problem are to find out a sequence of operation under constraint condition limitation for every machine in M platform machine and make N
The deadline of operation is minimum.FJSP reduces machine constraint, expands the search range of feasible solution, increases the complexity of problem
Degree is the difficult problem of nondeterministic polynomial (Non-deterministic Polynomial, NP).
The existing method for solving FJSP is roughly divided into two classes: layered approach and integration method.Layered approach is by FJSP points
It is solved for two problems of machine assignment and Operation Sequencing;Integration is that two question synthesis get up to solve.Layering side
Method can reduce the complexity of FJSP, but layered approach is not typically available the overall situation due to that can not comprehensively consider global information
Optimal solution.Integration method can theoretically guarantee to obtain globally optimal solution.Nevertheless, most of existing integration methods by
In FJSP high complexity and the problem of be faced with inefficiency.
Currently, some intelligent algorithms such as genetic algorithm (Genetic Algorithm, GA), Ant ColonySystem (Ant Colony
System, ACS) the methods of be widely used in FJSP problem due to the characteristics of it is simple and efficient.Compared to ACS, GA cannot fill
Divide and utilizes a large amount of heuristic information, and in different problem-instances, the emphasis of constraint condition is different, so utilizing GA
Solving FJSP has certain limitation.ACS method can make full use of a variety of heuristic informations in the construction process of solution, tool
There is a wider applicability, but that there are speed of convergences is slow, the shortcomings that easily falling into local optimum for ACS method.
Summary of the invention
In view of the above shortcomings of the prior art, the present invention proposes a kind of flexible job based on more heuristic information Ant ColonySystems
Job-Shop method, the present invention introduce six kinds of heuristic informations in Ant ColonySystem, make algorithm equal energy in different problem-instances
Ideal performance is obtained, avoids falling into local optimum, and can be according to open problems example and current search state using one kind
The adaptive strategy of heuristic information and its directed force control parameter is automatically selected, designs the heuristic information of entire ant colony
It is restrained to the current optimal design that can generate high-quality solution, improves speed of convergence, improve algorithm optimization efficiency, improve flexible job vehicle
Between dispatching efficiency.
If the FJSP problem of N × M are as follows: have N number of workpiece to be processed and M platform available machines used, workpiece collection is combined into { p1,
p2,...,pN, collection of machines is { m1,m2,...,mM, each workpiece include one or multi-channel process, workpiece i (i=1,2 ...,
N n)iProcedure is by set opi={ opi1,opi2,...,opiniIndicate.Ant ColonySystem of the invention contains M sequence with one
The set S of columnanTo indicate a scheduling scheme, i.e. problem a solution.Wherein, each sequenceThose are had recorded by process
Distribute to the node of machine k, k=1,2 ..., M.
The basic step of flexible job shop scheduling method based on more heuristic information Ant ColonySystems are as follows:
S1 establishes graph model of extracting: extracting figure is also referred to as the structural map solved, is one oriented for describing the search space of solution
Connected graph.Starting point (v is respectively indicated in addition to twobegin) and terminal (vend) virtual nodes except, each node indicates one of work
Feasible distribution of the sequence to a machine.If indicating the number of all feasible distribution, T value with T are as follows:
More specifically, in addition to vbeginAnd vendOutside, each node vs(1≤s≤T) contains workpiece ps∈ [1, N], work
SequenceAnd some machine m of this process can be executeds, wherein all nodes (altogether T+2) are by having
It is connected to side, the direction on side meets the sequence constraint between process.
S2 initialization: the pheromones distribution in initialization extracting figure assigns every directed edge one initial information element value;
And a heuristic information archives I is initialized for every ant anan;
Heuristic information archives include the selection scheme y to heuristic informationanWith to two kinds of directed force control parameter values
q0 anAnd βan, i.e. Ian={ yan,q0 an,βan}。
The disaggregation of S3 construction FJSP: every ant all represents a disaggregation, a legal disaggregation SanBy M sequence group
At kth (k=1,2 ..., M) a sequence is denoted asIndicate all processes for distributing to machine k.
S4 Pheromone update: every ant one node of every selection, updating Policy Updates using local information element should mutually have
Pheromone concentration on side;After all ants all complete the construction of solution, it is all that Policy Updates are updated according to global information element
Pheromone concentration on directed edge.
S5 heuristic information archives update: carrying out the update of a heuristic information archives every G iteration.It inspires twice
Between formula news file regeneration interval, the duration of the scheduling scheme of every ant construction is by the heuristic information archives with the ant
It is saved in an external archive F jointly.Adaptive strategy selects the heuristic information archives of duration shortest λ % element in F
It is counted, is updated according to heuristic information archives of the statistic to every ant an.
S6 termination condition inspection: when the assessment number for the disaggregation that algorithm constructs ant reaches preset maximum time
When number, termination algorithm simultaneously exports result.
Further, in step S3, the disaggregation construction process of an ant is as follows:
S31 initialization: initialization is eachAnd a candidate solution point list L is constructed, wherein comprising all
First procedure of workpiece, i.e. opi1(i=1,2 ..., N);
S32 is each node v in Ls,Middle such a node v of searchpre, so that by vsAs vpreIt is subsequent
When node, will not constraint condition in violation problem, by find first vpreNode is known as vsIt is optimal before after node, this
When, ant is calculated by vsThe Probability p of disaggregation is addeds;
S33 selects a node v using pseudo-random process from Ls;
The node v that S34 will choose in S33sAnd then vpreIt is inserted into sequenceIn, i.e., by vsIt is placed in S32 and finds
VpreBelow, meanwhile, vsInsertion mean workpiece psProcess opsIt has been be assigned that, all be associated with work so removing in L
Sequence opsNode, if opsIt is not workpiece psThe last one process, then by all related workpiece psProcess ops+1Knot
Point is added in L;
If S35 L is empty set, disaggregation construction is finished;Otherwise S32 is returned.
Further, in step S4, after every ant completes solution construction process, part letter will be executed according to the solution constructed
Breath element updates.In a preferred embodiment, pheromones are placed on the arc between the node for being associated with same machines.Just
When the beginning, the pheromones on all arcs are uniformly set as τ0=(mk0)–1, wherein mk0It is to be produced by the rule of most fast operator precedence
The deadline of raw solution.During solving construction, when an ant is by vsInsertionIn vpreWhen later, a kind of part letter
Breath element updates rule and is used to update arc (vpre, vs) on pheromones τpre→s.More specifically, pheromones in the following manner into
Row local information element updates:
τpre→s=(1- ε) τpre→s+ετ0
Wherein ε ∈ (0,1) is a parameter.After all ants complete solution construction process, following global information elements are executed
It updates:
Wherein ρ ∈ (0,1) is a parameter, mkbestIt is optimal solution S so farbestDeadline.
Further, in step S4, the influence of the solution construction process of ant while be inspired formula information and pheromones, for
Each of extracting figure node vs, define the heuristic information of six seed types.Wherein, first three types are static heuristic letters
Breath, vsHeuristic information value search process at the beginning when calculated and remained unchanged from this.In addition the inspiration of three types
Formula information, vsHeuristic information value must be calculated according to the part solution of ant present construction.
Based on six kinds of heuristic informations, yanIt is a 0/1 sextuple vector, node vsHeuristic information value calculate such as
Under:
To make algorithm that can obtain ideal performance in different problem-instances, present invention introduces an adaptive strategies, should
Strategy adjusts the heuristic information archives I of ant according to passing search experiencean, including yanWith the guidance of heuristic information
Power control parameter q0(ant an is in q0On value be denoted as) and β (value of the ant an on β is denoted as βan)。
The invention has the following advantages:
1, the characteristics of being directed to Flexible Job-shop Scheduling Problems, it is contemplated that six kinds of heuristic informations, devising one kind can be with
The ACS algorithm for adapting to different problem-instances avoids falling into local optimum.
2, the adaptive design strategy for introducing a kind of heuristic information, enables ACS algorithm to be searched according to passing
Rope experience is adaptive selected heuristic information and adjusts its directed force control parameter to improve the optimization efficiency of algorithm
It efficiently solves Flexible Job-shop Scheduling Problems and provides a new approach.
Detailed description of the invention
Fig. 1 is the flow chart of flexible job shop scheduling method of the present invention.
Specific embodiment
Below by specific embodiment, the present invention is described in further detail, but embodiments of the present invention are not
It is limited to this.
The FJSP problem of an existing N × M: having N number of workpiece to be processed and M platform available machines used, and workpiece collection is combined into { p1,
p2,...,pN, collection of machines is { m1,m2,...,mM, each workpiece include one or multi-channel process, workpiece i (i=1,2 ...,
N n)iProcedure is by gatheringIt indicates.Based on the building method of extracting figure in summary of the invention,
One set S containing M sequence of Ant ColonySystem of the inventionanTo indicate a scheduling scheme, i.e. problem a solution.Its
In, each sequenceHave recorded those nodes that process is distributed to machine k, k=1,2 ..., M.The disaggregation of ant an constructs
Process comprises the steps of:
Step (1): it usesEach node sequence that initialization solution is concentrated.A time is selected from extracting figure
Choosing solution point list L, wherein including the first procedure of all workpiece, i.e. opi1(i=1,2 ..., N).
Step (2): for each node v in Ls,Middle such a node v of searchpre, so that by vsAs vpre's
When successor node, will not constraint condition in violation problem, in the present embodiment, constraint condition are as follows: the node free time can call, will
First v foundpreNode is known as vsIt is optimal before after node, at this moment, calculate ant for vsThe Probability p of disaggregation is addeds, ps's
Calculation method are as follows:
ps=τpre→s(ηs)β
Wherein: τpre→sIt indicates in directed edge (vpre, vs) on information cellulose content, ηsIt is vsThe value of heuristic information, β are
Control the parameter that heuristic information influences.
Step (3): a node is selected from L with pseudorandom ratio rules.Specifically, pseudorandom ratio rules are pressed
As under type is realized:
Wherein: r ∈ [0,1] is a standardized random number, q0∈ (0,1) is a control parameter of algorithm, and L ' is
The randomly selected a subset from L, L's ' is dimensioned to
Step (4): the v that will be selected in step (3)sInsertionAfter node before it optimal found in middle step (2)
Below.By vsInsertionShow workpiece psProcess opsIt has been be scheduled that, therefore deleted from L and be associated with opsAll nodes.
If opsIt is not operation psLast procedure (i.e. ops<ns), workpiece p will be associated withsProcess ops+1Node join arrive
L, otherwise, it is not necessary to new node be added.
Step (5): if L non-empty, return step (2) dispatches remaining process.Otherwise, all process steps have been scheduled,
One complete solution construction finishes.
After every ant completes solution construction process, local Pheromone update will be executed according to the solution constructed.The present embodiment
In, pheromones are placed on the arc between the node for being associated with same machines.When initial, the pheromones on all arcs are unified
It is set as τ0=(mk0)–1, wherein mk0It is the deadline of the solution generated by the rule of most fast operator precedence.It was constructed in solution
Cheng Zhong, when an ant is by vsInsertionIn vpreWhen later, a kind of local information element updates rule and is used to update arc
(vpre, vs) on pheromones τpre→s.More specifically, pheromones carry out local information element update in the following manner:
τpre→s=(1- ε) τpre→s+ετ0
Wherein ε ∈ (0,1) is a parameter.After all ants complete solution construction process, following global information elements are executed
It updates:
Wherein ρ ∈ (0,1) is a parameter, mkbestIt is optimal solution S so farbestDeadline.
Each of extracting figure is tied in the influence of the solution construction process of ant while be inspired formula information and pheromones
Point vs, the heuristic information of six seed types is defined in the method for the present invention.Wherein, first three types are static heuristic information, vs's
Heuristic information value search process at the beginning when calculated and remained unchanged from this.In addition the heuristic information of three types,
vsHeuristic information value must be calculated according to the part solution of ant present construction.Tool about above-mentioned six kinds of heuristic informations
Body is defined as follows:
※ static state heuristic information
(1) heuristic information based on the process deadline: a procedure is distributed to use by heuristic information foundation
The minimum time completes the thought definition of its machine.Assuming that in machine mkUpper completion process opijThe time it takes isThen
Node vsHeuristic information value based on the process deadline calculates are as follows:
Wherein:Representing all can execute workpiece psIn process opsMachine set.
(2) heuristic information based on workpiece residue process: the heuristic information has according to preferential selection compared with multiresidue
The thought of the workpiece of process defines, node vsHeuristic information value based on workpiece residue process calculates are as follows:
Wherein, nsIt is the quantity of each tasks leave process.
(3) heuristic information based on workpiece available machines used number: the heuristic information is according to preferential selection available machines used
The thought of less process defines, node vsHeuristic information value based on workpiece available machines used number calculates are as follows:
Wherein,Representing all can execute workpiece psIn process opsMachine set, mijFor i-th of task
J-th of process be currently available that machine quantity.
※ dynamic heuristic information
(4) based on the heuristic information of time started: the heuristic information in order to preferentially select have the more early time started
Process and define, give a node vs, the process op that is associatedsAt the beginning of staTsIt determines are as follows:
Wherein:Indicate machine msThe available time,Indicate same workpiece psIn upper procedure ops-1
Deadline.Use time started staTs, node vsHeuristic information value based on the time started calculates are as follows:
(5) heuristic information based on the process deadline: the heuristic information is relatively early completed to preferentially select to have
The process of time, node vsHeuristic information value based on the process deadline calculates are as follows:
Wherein: exeTsEqual to staTsWithSum, staTsFor with node vsAssociated process opsAt the beginning of,
For in machine msUpper completion process opsThe time it takes.
(6) heuristic information based on machine work on hand amount: the heuristic information, which is based on process to distribute to, to be had most
The workload of the thought of the machine of small workload, machine can be measured with its pot life, therefore, node vsBased on machine
The heuristic information value of work on hand amount calculates are as follows:
Wherein:Indicate machine msThe available time.
Notice three kinds of dynamic heuristic information to vsIt is significant only to work as opsWhen becoming executable, opsWhether can hold
Row whether there is v according to the part solution judgement constructed at presents∈L.Every ant an is according to itself heuristic information archives Ian
In heuristic information define yanTo calculate the heuristic information value an of node.Based on above-mentioned six kinds of heuristic informations, yanIt is
One 0/1 sextuple vector, node vsHeuristic information value calculate it is as follows:
To make algorithm that can obtain ideal performance in different problem-instances, present invention introduces an adaptive strategies, should
Strategy adjusts the heuristic information archives I of ant according to passing search experiencean, including yanWith the guidance of heuristic information
Power control parameter q0(ant an is in q0On value be denoted as) and β (value of the ant an on β is denoted as βan), it is specific to adjust
Adjusting method is as follows:
When algorithm initialization, the heuristic information archives of every ant are randomly generated.Specifically, yanEach elementIt will be randomly generated between { 0,1 }, q0 anAnd βanValue will be in the predefined codomain [q of corresponding parametermin,qmax] and [βmin,
βmax] in generate.
Algorithm carries out the adjustment of a heuristic information archives, the twice adjustment of heuristic information archives every G iteration
The heuristic information archives of interim, ant remain unchanged, and the duration of the scheduling scheme of every ant construction will be with this
The heuristic information archives of ant are saved in an external archive F jointly.Assuming that the scale of ant colony is m, after G iteration in archives
MG element will be preserved.The heuristic information archives of duration shortest λ % element are united in adaptive strategy selection F
Meter, and adjusted as follows according to heuristic information archives of the statistic to every ant an:
βan=C (θβ,αβ)
Wherein, r is the standardization random number between one (0,1), and B indicates the subset of duration shortest λ % element in F, C
(θ, α) indicates the density function of Cauchy's distribution,WithExpression parameter q0The average and standard deviation of value, θ in BβAnd αβ
The then average and standard deviation of expression parameter β value in B, every time after adjustment, emptying external archive F is to adjust next time
It prepares;The general value 5~10 of G, the general value 10~25 of λ.
From the above process as can be seen that adaptive strategy proposed by the present invention drives ant from generating more excellent solution in history
Learn in heuristic information archives, sets the heuristic information of entire ant colony by adjusting the heuristic information archives of each ant
It counts and is restrained to the current optimal design that can generate high-quality solution, to promote the raising of algorithm optimization efficiency.
Experimental result on FJSP standard testing library is as shown in table 1, flexible job shop scheduling side disclosed by the invention
It is most short or close to shortest scheduling scheme that method can efficiently find the duration.
1 FJSP standard testing library experimental result of table
Conspicuousness experimental result shows that HDA-ACO is significantly better than ACOTH-FJSP
Conspicuousness experimental result shows that HDA-ACO is significantly better than ACOCTH-FJSP
The above embodiment is a preferred embodiment of the present invention, but embodiments of the present invention are not by above-described embodiment
Limitation, other any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the present invention,
It should be equivalent substitute mode, be included within the scope of the present invention.
Claims (10)
1. based on the flexible job shop scheduling method of more heuristic information Ant ColonySystems, if the FJSP problem of N × M are as follows: have it is N number of to
The workpiece and M platform available machines used, workpiece collection of processing are combined into { p1,p2,...,pN, collection of machines is { m1,m2,...,mM, each
Workpiece includes one or multi-channel process, the n of workpiece iiProcedure is by gatheringIt indicates, ant colony system
One set S containing M sequence of systemanTo indicate a scheduling scheme, i.e. problem a solution;Wherein, each sequenceHave recorded those nodes that process is distributed to machine k, i=1,2 ..., N, k=1,2 ..., M, which is characterized in that flexible
Job-shop scheduling method the following steps are included:
S1 establishes graph model of extracting;
S2 initialization: the pheromones distribution in initialization extracting figure assigns every directed edge one initial information element value;And
A heuristic information archives I is initialized for every ant anan;
The disaggregation of S3 construction FJSP: every ant all represents a disaggregation, a legal disaggregation SanIt is made of M sequence, kth
A sequence is denoted asIndicate all processes for distributing to machine k;
S4 Pheromone update: every ant one node of every selection updates the corresponding directed edge of Policy Updates using local information element
On pheromone concentration;After all ants all complete the construction of solution, it is all oriented that Policy Updates are updated according to global information element
Pheromone concentration on side;
S5 heuristic information archives update: carrying out the update of a heuristic information archives every G iteration, twice heuristic letter
Between ceasing archives regeneration interval, the duration of the scheduling scheme of every ant construction will be common with the heuristic information archives of the ant
It is saved in an external archive F;Adaptive strategy selects the heuristic information archives of duration shortest λ % element in F to carry out
Statistics, is updated according to heuristic information archives of the statistic to every ant an;
S6 termination condition inspection: when the assessment number for the disaggregation that algorithm constructs ant reaches preset maximum times
When, termination algorithm simultaneously exports result.
2. flexible job shop scheduling method according to claim 1, which is characterized in that heuristic information archives IanIncluding
To the selection scheme y of heuristic informationanWith to two kinds of directed force control parameter value q0 anAnd βan, i.e. Ian={ yan,q0 an,
βan}。
3. flexible job shop scheduling method according to claim 2, which is characterized in that for each of extracting figure
Node vs, define the heuristic information of six seed types.
4. flexible job shop scheduling method according to claim 3, which is characterized in that node vsSix seed types open
Hairdo information are as follows:
(1) heuristic information based on the process deadline: the heuristic information is distributed to according to by a procedure using minimum
Time completes the thought definition of its machine, it is assumed that in machine mkUpper completion process opijThe time it takes isThen node
vsHeuristic information value based on the process deadline calculates are as follows:
Wherein:Representing all can execute workpiece psIn process opsMachine set;
(2) heuristic information based on workpiece residue process: the heuristic information has according to preferential selection compared with multiresidue process
Workpiece thought definition, node vsHeuristic information value based on workpiece residue process calculates are as follows:
Wherein, nsIt is the quantity of each tasks leave process;
(3) heuristic information based on workpiece available machines used number: the heuristic information is less according to preferential selection available machines used
Process thought definition, node vsHeuristic information value based on workpiece available machines used number calculates are as follows:
Wherein,Representing all can execute workpiece psIn process opsMachine set, mijIt is the of i-th of task
J process is currently available that machine quantity;
(4) based on the heuristic information of time started: the heuristic information is in order to preferentially select the work with the more early time started
Sequence and define, give a node vs, the process op that is associatedsAt the beginning of staTsIt determines are as follows:
Wherein:Indicate machine msThe available time,Indicate same workpiece psIn upper procedure ops-1It is complete
At the time, time started staT is useds, node vsHeuristic information value based on the time started calculates are as follows:
(5) heuristic information based on the process deadline: the heuristic information in order to preferentially select have the more early deadline
Process, node vsHeuristic information value based on the process deadline calculates are as follows:
Wherein: exeTsEqual to staTsWithSum, staTsFor with node vsAssociated process opsAt the beginning of,For
Machine msUpper completion process opsThe time it takes;
(6) heuristic information based on machine work on hand amount: the heuristic information, which is based on process to distribute to, has most unskilled labourer
The workload of the thought for the machine that work is measured, machine can be measured with its pot life, therefore, node vsIt is existing based on machine
The heuristic information value of workload calculates are as follows:
Wherein:Indicate machine msThe available time.
5. flexible job shop scheduling method according to claim 4, which is characterized in that yanBe one sextuple 0/1 to
Amount, node vsHeuristic information value calculate it is as follows:
6. the flexible job shop scheduling method according to any one of claim 2-5, which is characterized in that adaptive strategy
Specific method of adjustment is as follows:
When algorithm initialization, the heuristic information archives of every ant, y is randomly generatedanEach elementIt will be between { 0,1 }
It is randomly generated, q0 anAnd βanValue will be in the predefined codomain [q of corresponding parametermin,qmax] and [βmin,βmax] in generate;
Algorithm carries out the adjustment of a heuristic information archives every G iteration, twice the adjustment interval of heuristic information archives
The heuristic information archives of period, ant remain unchanged, and the duration of the scheduling scheme of every ant construction will be with the ant
Heuristic information archives be saved in an external archive F jointly;Assuming that the scale of ant colony is m, will be protected in archives after G iteration
There is mG element, the heuristic information archives of duration shortest λ % element are counted in adaptive strategy selection F, and
It is adjusted as follows according to heuristic information archives of the statistic to every ant an:
βan=C (θβ,αβ)
Wherein, r is the standardization random number between one (0,1), the subset of duration shortest λ % element in B expression F, C (θ,
α) indicate the density function of Cauchy's distribution,WithExpression parameter q0The average and standard deviation of value, θ in BβAnd αβThen table
Show the average and standard deviation of parameter beta value in B, every time after adjustment, emptying external archive F is that standard is done in adjustment next time
It is standby.
7. flexible job shop scheduling method according to claim 6, which is characterized in that G value 5~10, λ value 10~
25。
8. flexible job shop scheduling method according to any one of claims 1-5, which is characterized in that analysed in step S1
It takes figure to be used to describe the search space of solution, is a directed connected graph, respectively indicate starting point v in addition to twobeginWith terminal vend's
Except virtual nodes, each node indicates the feasible distribution of one of process a to machine.
9. flexible job shop scheduling method according to claim 8, which is characterized in that the solution of an ant in step S3
It is as follows to collect construction process:
S31 initialization: initialization is eachAnd a candidate solution point list L is constructed, wherein including all workpiece
The first procedure, i.e. opi1(i=1,2 ..., N);
S32 is each node v in Ls,Middle such a node v of searchpre, so that by vsAs vpreSuccessor node
When, will not constraint condition in violation problem, by find first vpreNode is known as vsIt is optimal before after node, at this moment, meter
Ant is calculated by vsThe Probability p of disaggregation is addeds;
S33 selects a node v using pseudo-random process from Ls;
The node v that S34 will choose in S33sAnd then vpreIt is inserted into sequenceIn, i.e., by vsIt is placed on and finds in S32
vpreBelow, remove in L and all be associated with process opsNode, if opsIt is not workpiece psThe last one process, then
By all related workpiece psProcess ops+1Node join into L;
If S35 L is empty set, disaggregation construction is finished;Otherwise S32 is returned.
10. flexible job shop scheduling method according to claim 9, which is characterized in that during solving construction, when one
Ant is by vsInsertionIn vpreWhen later, arc (vpre, vs) on pheromones τpre→sLocal letter is carried out in the following manner
Breath element updates:
τpre→s=(1- ε) τpre→s+ετ0
Wherein ε ∈ (0,1) is a parameter;
After all ants complete solution construction process, executes following global information elements and updates:
Wherein ρ ∈ (0,1) is a parameter, mkbestIt is optimal solution S so farbestDeadline.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811149243.6A CN109255546A (en) | 2018-09-29 | 2018-09-29 | Flexible job shop scheduling method based on more heuristic information Ant ColonySystems |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811149243.6A CN109255546A (en) | 2018-09-29 | 2018-09-29 | Flexible job shop scheduling method based on more heuristic information Ant ColonySystems |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109255546A true CN109255546A (en) | 2019-01-22 |
Family
ID=65044752
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811149243.6A Pending CN109255546A (en) | 2018-09-29 | 2018-09-29 | Flexible job shop scheduling method based on more heuristic information Ant ColonySystems |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109255546A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109872091A (en) * | 2019-04-18 | 2019-06-11 | 安徽大学 | A kind of Job Scheduling method and device based on ant group algorithm |
CN109963262A (en) * | 2019-01-28 | 2019-07-02 | 华南理工大学 | Wireless sensor method for optimizing scheduling in a kind of wireless sensor network |
CN110378579A (en) * | 2019-07-03 | 2019-10-25 | 东南大学 | Color batch processes are pressed for coating workshop |
CN110942251A (en) * | 2019-11-27 | 2020-03-31 | 安徽大学 | Batch scheduling method based on joint ant colony algorithm |
CN111210116A (en) * | 2019-12-23 | 2020-05-29 | 华南理工大学 | Cross transportation receiving door and transportation door distribution method based on double ant colony algorithm |
CN112327621A (en) * | 2020-11-02 | 2021-02-05 | 金航数码科技有限责任公司 | Flexible production line self-adaptive control system and method based on ant colony algorithm |
CN118153927A (en) * | 2024-05-11 | 2024-06-07 | 浙江大学 | Workshop scheduling method, device and equipment based on ant colony algorithm |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103246938A (en) * | 2013-04-25 | 2013-08-14 | 中山大学 | Self-adaptive ant colony optimization based flexible workshop dispatching technology |
CN103246811A (en) * | 2013-04-25 | 2013-08-14 | 中山大学 | Medical system admission scheduling method based on ant colony optimization |
CN104914835A (en) * | 2015-05-22 | 2015-09-16 | 齐鲁工业大学 | Flexible job-shop scheduling multi-objective method |
-
2018
- 2018-09-29 CN CN201811149243.6A patent/CN109255546A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103246938A (en) * | 2013-04-25 | 2013-08-14 | 中山大学 | Self-adaptive ant colony optimization based flexible workshop dispatching technology |
CN103246811A (en) * | 2013-04-25 | 2013-08-14 | 中山大学 | Medical system admission scheduling method based on ant colony optimization |
CN104914835A (en) * | 2015-05-22 | 2015-09-16 | 齐鲁工业大学 | Flexible job-shop scheduling multi-objective method |
Non-Patent Citations (1)
Title |
---|
王万良,赵澄,熊婧,徐新黎: "基于改进蚁群算法的柔性作业车间调度问题的求解方法", 《系统仿真学报》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109963262A (en) * | 2019-01-28 | 2019-07-02 | 华南理工大学 | Wireless sensor method for optimizing scheduling in a kind of wireless sensor network |
CN109872091A (en) * | 2019-04-18 | 2019-06-11 | 安徽大学 | A kind of Job Scheduling method and device based on ant group algorithm |
CN110378579A (en) * | 2019-07-03 | 2019-10-25 | 东南大学 | Color batch processes are pressed for coating workshop |
CN110378579B (en) * | 2019-07-03 | 2023-04-07 | 东南大学 | Color-by-color batching method for automobile painting workshop |
CN110942251A (en) * | 2019-11-27 | 2020-03-31 | 安徽大学 | Batch scheduling method based on joint ant colony algorithm |
CN110942251B (en) * | 2019-11-27 | 2022-09-30 | 安徽大学 | Batch scheduling method based on joint ant colony algorithm |
CN111210116A (en) * | 2019-12-23 | 2020-05-29 | 华南理工大学 | Cross transportation receiving door and transportation door distribution method based on double ant colony algorithm |
CN112327621A (en) * | 2020-11-02 | 2021-02-05 | 金航数码科技有限责任公司 | Flexible production line self-adaptive control system and method based on ant colony algorithm |
CN112327621B (en) * | 2020-11-02 | 2022-07-08 | 金航数码科技有限责任公司 | Flexible production line self-adaptive control system and method based on ant colony algorithm |
CN118153927A (en) * | 2024-05-11 | 2024-06-07 | 浙江大学 | Workshop scheduling method, device and equipment based on ant colony algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109255546A (en) | Flexible job shop scheduling method based on more heuristic information Ant ColonySystems | |
Blum et al. | An ant colony optimization algorithm for shop scheduling problems | |
CN104580284B (en) | Traffic assignments device and method for distributing business | |
CN107544251B (en) | Single machine scheduling method for minimizing total stall period based on distributed robust model | |
CN107317699B (en) | Dynamic ant colony rapid optimization method of cloud manufacturing service combination | |
CN107168267A (en) | Based on the production scheduling method and system for improving population and heuristic strategies | |
CN110708259A (en) | Information-agnostic Coflow scheduling system capable of automatically adjusting queue threshold and scheduling method thereof | |
CN102495804A (en) | Automatic software testing method | |
CN115310794A (en) | Man-machine cooperation assembly line balancing method and device | |
CN112101674B (en) | Resource allocation matching method, device, equipment and medium based on group intelligent algorithm | |
CN112667260A (en) | OTA remote upgrading system and method based on intelligent brain | |
CN116050540A (en) | Self-adaptive federal edge learning method based on joint bi-dimensional user scheduling | |
Ma et al. | A deep reinforcement learning approach for dynamic contents caching in HetNets | |
CN114371925A (en) | Edge cloud cooperative multi-task scheduling method and system for guaranteeing edge cloud load ratio | |
CN114154685A (en) | Electric energy data scheduling method in smart power grid | |
Renting et al. | Automated configuration of negotiation strategies | |
CN109873406A (en) | A kind of electric system weakness route discrimination method | |
CN117762591A (en) | Task control method, task control device, computer device, and storage medium | |
CN112528524A (en) | Balanced and optimized scheduling method for mixed-flow assembly line | |
CN116089083A (en) | Multi-target data center resource scheduling method | |
CN103455525B (en) | The method and apparatus of popularization account number state is determined based on the search popularization behavior of user | |
CN114492849B (en) | Model updating method and device based on federal learning | |
CN102857939B (en) | The collocation method and device of cell parameter | |
CN106992962A (en) | Method of the virtual machine based on multi-tenant in cloud data center security configuration | |
Ahmadizar et al. | A novel ant colony algorithm for the single-machine total weighted tardiness problem with sequence dependent setup times |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190122 |