CN109102203A - A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms - Google Patents
A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms Download PDFInfo
- Publication number
- CN109102203A CN109102203A CN201810994621.4A CN201810994621A CN109102203A CN 109102203 A CN109102203 A CN 109102203A CN 201810994621 A CN201810994621 A CN 201810994621A CN 109102203 A CN109102203 A CN 109102203A
- Authority
- CN
- China
- Prior art keywords
- string
- chromosome
- matrix
- operator
- task object
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/123—DNA computing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- 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
Abstract
The present invention discloses a kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms, step are as follows: 1, propose that one kind goes here and there the chromosome coding mode of chromosome more, gone here and there one more chromosome instead of traditional monosome representative genetic algorithm an individual.2, genetic operator.3, generation primary: the purpose of generation primary is to fill up ND more string dyeing volume matrixes, and ND is population scale.4, crossover operator: chromosome string cross method is used.5, mutation operator: mutation operation successively carries out in each chromosome string.6, fitness function.7, roulette selection method and elite retention strategy selection operator: are used.The present invention realizes the operation of a resource platform matching multiple tasks and pinned task number of targets;It can be applied to a traditional resource platform and match a task object, being different from common genetic algorithm is that rule in this can be kept not to be destroyed, and advantage is more flexible;The coding mode of 0-1, increases computational efficiency.
Description
Technical field
The present invention relates to a kind of Target Assignment optimization methods based on more string chromosomal inheritance algorithms, it can be in target point
Timing flexibly controls the resource quantity that each target is distributed, and belongs to intelligent algorithm module and scheduling of resource field.
Background technique
When the programme of complex task generates, Target Assignment is often related to.Target Assignment refers to for various resources
It is rationally and effectively adjusted and is measured and analyzed and used, normally behave as existing resource platform being matched to required execution
Task or target on.And with the development of technology, it is desirable that various Target Assignments require fast and automatically, accurately to carry out, institute
Be with the research for Target Assignment optimization method it is necessary, have good development prospect and practical application value.
Target assignment problem often has more optimizing index and a variety of methods of salary distribution, therefore often uses when solving the problems, such as this
Genetic algorithm optimizes.In the optimization process of traditional genetic algorithm, sequential coding, monosome coding staff are generally used
Formula assigns each position on chromosome as a resource platform, the position that is, when resource platform is matched to task object
Gene representative the resource platform is matched to the number of task object.3 are encoded to i.e. on gene loci 5, represents and No. 5 is provided
Source platform is matched to No. 3 task objects.Although this coding mode comparison is succinct, under this coding mode, a gene
Site can only store a coding, also mean that a resource platform can only be matched to a task object, various complicated
Task distribution when have significant limitation.For example, when certain large-scale resource platform is participated in the distribution, it can only under this coding mode
A task is carried out, this is undoubtedly unreasonable.Meanwhile that there is also computational efficiencies is low for such coding mode, is carrying out matrix fortune
The problem of additional addressing operation is needed when calculation.So a kind of new coding mode urgently proposes.
Summary of the invention
It is an object of the invention to the coding modes for traditional Target Assignment genetic algorithm to improve, and propose one
Target Assignment optimization method of the kind based on more string chromosomal inheritance algorithms.
A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms, specific steps are as follows:
S1, the chromosome coding mode for proposing one kind mostly string chromosome, chromosome of going here and there one replace traditional simple stain more
An individual for body representative genetic algorithm.More string chromosomes can regard that the matrix of m row n column, m represent task object as
Number, n represent resource platform number.(1≤i≤m) row i-th in matrix, the element of jth (1≤j≤n) column are fij, fijEqual to 1 or
0,1, which represents resource platform j, distributes to task object i, and 0 representative does not distribute.In this way, the chromosome of string more than one just has m × n 0-1
Encode the element of composition.
S2, genetic operator: chromosome coding is the matrix of a m × n, it is clear that cannot directly use genetic algorithm.
Next this more string chromosome is cut by column, forms n chromosome string, has m gene on every string, then each chromosome string generation
Task object information assigned by each resource platform of table.
S3, generation primary: the purpose of generation primary is to fill up ND more string dyeing volume matrixes, and ND is population scale.Each
Chromosome matrix decomposition shares n, each chromosome string inserts x by necessary requirement at random at the chromosome string in above-mentioned S2
A " 1 ", x is the task object number that the defined resource platform most multipotency is got, if each resource platform can only without particular/special requirement
If getting a task object, x takes " 1 "." 0 " cover is filled out in remaining position.So fill up ND matrix.
S4, crossover operator: to guarantee that task object number that required resource platform most multipotency is got is constant and target
Distribution condition, the crossover operator in the method for the present invention use chromosome string cross method.Routinely when the crossover operation of chromosome,
The chromosome string of interdigital anchor point is changed to by the gene of interdigital anchor point.That intersect every time in this way is exactly two fathers to be intersected
The task object of the same resource platform of body distributes information.
S5, mutation operator: in the methods of the invention, mutation operation successively carries out in each chromosome string.First by dye
Colour solid string mutation probability randomly selects a chromosome string, then presses the specific gene position of the genetic mutation probability selection chromosome string
After point, once judged: if gene is " 0 " on this site, being changed into " 1 ", later by remaining in this chromosome string
" 1 " equiprobability in position, which chooses one (being not required to select if only one " 1 "), becomes " 0 ";Similarly, if gene is on this site
" 1 " is then changed into " 0 ", and " 0 " the equiprobability selection one of remaining position in this string chromosome is become 1.
S6, fitness function: in the method, can be by resource platform information and task object information to majority of case
It is organized into the information matrix of m × n, information matrix and more string dyeing volume matrixes are carried out point-to-point multiplication and obtain a fitness square
Battle array, then directly summed or weighted sum obtain this it is more string chromosomes fitness.
S7, selection operator: roulette selection method and elite retention strategy are used.It can be achieved with chromosomes of going here and there in this way to compile more
The operation of the next resource platform matching multiple tasks target of code mode and pinned task number of targets.
A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms of the invention, advantage and effect are
1, the resource platform matching multiple tasks that common genetic algorithm cannot achieve and pinned task target are realized
Several operations.
2, this method can also be applied to a traditional resource platform and match a task object, is different from common heredity and calculates
Method is that rule in this can be kept not to be destroyed, and advantage is more flexible.
3, the coding mode of 0-1, increases computational efficiency.
Detailed description of the invention
Fig. 1 is the basic flow chart based on more string chromosomal inheritance algorithms.
Fig. 2 a, Fig. 2 b are more string chromosome coding citings.
Fig. 3 is that same bit string intersects schematic diagram.
Fig. 4 is dibit variation schematic diagram.
Specific embodiment
The invention will now be described in further detail with reference to the accompanying drawings.
A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms of the present invention, basic procedure such as Fig. 1 institute
Show, core methed is the coding and calculating operation of more string chromosomes.This process is pressed below, is made in conjunction with specific embodiments specifically
It is bright.
Embodiment is as follows: existing ten resource platforms distribute to five task objects, wherein No.1 platform can match two
A task object.
S1, generation primary
Method user determines crossover probability P according to the size that required outstanding parent retains firstxSize, this implementation
0.97 is taken by often customary in example;String mutation probability P is determined according to the size of required filial generation result search rangeiAnd genetic mutation
Probability PjSize, take 0.01 and 0.2 respectively as usual in this example.Px、PiAnd PjThe determination of probability, the value and convergence taken
Speed is directly proportional.Simultaneously according to the needs for calculating the time, determine that population scale ND, the present embodiment are determined as 30;Determine that maximum changes
Generation number G, the present embodiment 100.
There are ten resource platforms, five task objects in the present embodiment, i.e. m is 5, n 10, then more string chromosomes are 1
× 10 matrix.More string chromosomes are decomposed into 10 chromosome strings by column, whether each chromosome string represents the resource platform
Select one information in this five task objects.More string chromosome matrix coder examples of two male parents to be intersected are as schemed
Shown in 2a, Fig. 2 b.Population scale is 30, i.e., string dyeing volume matrix more than 30.This goes here and there dyeing volume matrix as parent, step more 30
The operation of rapid S2, S3 carry out this parent.
S2 intersects with bit string
30 in the population primary that S1 is generated are intersected by adjacent principle as parent.Adjacent male parent to be intersected will
10 judgements are carried out, judge to carry out in the chromosome string of its corresponding position every time, the 1st string and more strings of such as more string chromosomes 1
1st string of chromosome 2.Use crossover probability PxJudge whether to be intersected, concrete operation method is to generate one at random
Number x (0≤x≤1), if x < Px, then crossover operation is carried out, if x > Px, then it is not processed.If crossover operation is carried out, in this time
Judge that, by two string chromosome interchange positions of this position, the same bit strings of as more string chromosomal inheritance algorithms are intersected, such as Fig. 3.
S3, dibit variation
The method that mutation operator in the method for the present invention is made a variation simultaneously using dibit.Concrete operations are the method for the present invention step
A string mutation probability P is given in rapid S1iWith a genetic mutation probability Pj.A string of progress every to 10 chromosome strings first
String mutation probability PiJudgement, concrete operation method be generate a random number x (0≤x≤1), if x < Pi, then make a variation
Operation, if x > Pi, then it is not processed;If desired it makes a variation, then each to 5 gene locis of the chromosome string again
Gene carries out genetic mutation probability P againjJudgement.Concrete operation method be each generate random number x (0≤x≤
1), if x < Pj, then mutation operation is carried out, if x > Pj, then it is not processed.If the position gene needs to make a variation, there is the following two kinds
Situation: if the position gene is " 1 ", this " 1 " being become " 0 ", then again that " 0 " in remaining position of string chromosome etc. is general
Rate, which chooses one, becomes " 1 ";If the position gene is " 0 ", this " 0 " is become " 1 ", then again by remaining in the string chromosome
" 1 " equiprobability in position, which chooses one, becomes " 0 " (directly becoming " 0 " if only one " 1 ").Resource is ensured that in this way
The constant dibit variation of the destination number that platform matches.Such as Fig. 4, by No. 3 genes on No. 3 chromosome strings of determine the probability
" 0 " variation is " 1 ", wherein only one " 1 " will become " 0 ".
By the processed more string dyeing volume matrixes of step S2, S3, generates 30 new more strings and dye volume matrix, as son
Generation.By parent and filial generation, the dyeing volume matrix of string more than totally 60 merges, as merging population.
S4, selection
Fitness function calculating is carried out first.An individual in the methods of the invention is more strings dyeing of a m × n
Volume matrix F is one 10 × 5 matrix in the present embodiment.The information of resource platform and task object is organized into respectively first
Information matrix, meaning are when this resource platform is matched to the task object, and obtained income degree size is denoted as resource platform
Information matrix R and task object information matrix M, they are 10 × 5 matrix;Due to going here and there chromosome matrix F as 0-1 square more
Battle array, therefore a coefficient matrix can be regarded as, therefore will go here and there more and dye matrix F, resource platform information matrix R, task object information matrix M work
Fitness matrix V, such as following formula can be obtained in dot product:
V=FRM
In formula, " " represents the dot product of matrix, or is denoted as " .* ".
Fitness matrix V, then sum, the as fitness fitness of more string chromosomes, such as following formula:
Fitness=∑i∑jVij
The roulette strategy and elite retention strategy obtained after fitness function in routinely genetic algorithm carries out selection behaviour
Make, concrete operations are as follows:
Roulette strategy: carrying out the selection of individual using fitness ratio selection mode, in the method, per each and every one
The select probability of body is directly proportional to its fitness.If population scale is ND, i-th (i=1,2 ..., N) individual in population
Fitness is fi, then i-th individual select probability pi, shown in following formula:
Elite retention strategy: the maximum individual of fitness in population will be merged and remained into population of new generation.
S5, the number of iterations judgement
The judgement for carrying out current iteration number, when the number of iterations reaches maximum number of iterations or nearest ten generations adaptive optimal control degree
Variance less than 0.5 i.e. enter step S6, otherwise return to step S2.
S6, end loop
End loop exports outstanding population.Wherein the highest individual of fitness is optimal objective allocation plan.
Claims (1)
1. a kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms, it is characterised in that: the specific step of this method
Suddenly are as follows:
S1, the chromosome coding mode for proposing one kind mostly string chromosome, chromosome of going here and there one replace traditional monosome generation more
An individual for table genetic algorithm;More string chromosomes regard that the matrix of m row n column, m represent task object number as, and n represents money
Source platform number;(1≤i≤m) row i-th in matrix, the element of jth (1≤j≤n) column are fij, fijResource is represented equal to 1 or 0,1
Platform j distributes to task object i, and 0 representative does not distribute;In this way, the chromosome of string more than one is just encoded the member formed by m × n 0-1
Element;
S2, genetic operator: more string chromosomes are cut by column, n chromosome string is formed, has m gene on every string, then each
Chromosome string represents task object information assigned by each resource platform;
S3, generation primary: the purpose of generation primary is to fill up ND more string dyeing volume matrixes, and ND is population scale;Each dyeing
Volume matrix resolves into the chromosome string in above-mentioned steps S2, shares n, each chromosome string inserts x by necessary requirement at random
A " 1 ", x is the task object number that the defined resource platform most multipotency is got, if each resource platform can only without particular/special requirement
If getting a task object, x takes " 1 ";" 0 " cover is filled out in remaining position;So fill up ND matrix;
S4, crossover operator: the crossover operator uses chromosome string cross method;That is the routinely crossover operation of chromosome
When, the chromosome string of interdigital anchor point is changed to by the gene of interdigital anchor point;What is intersected every time in this way is exactly two wait hand over
The task object for pitching the same resource platform of male parent distributes information;
S5, mutation operator: mutation operation successively carries out in each chromosome string;It is selected at random by chromosome string mutation probability first
Take a chromosome string, then by the specific gene site of the genetic mutation probability selection chromosome string after, once judged:
If gene is " 0 " on this site, it is changed into " 1 ", later chooses " 1 " equiprobability in this chromosome string in remaining position
One becomes " 0 ", is not required to select if only one " 1 ";Similarly, it if gene is " 1 " on this site, is changed into " 0 ", and will
" 0 " equiprobability of remaining position, which chooses one, in this string chromosome becomes 1;
S6, fitness function: by resource platform information and task object finish message at the information matrix of m × n, by information matrix
Point-to-point multiplication are carried out with more string dyeing volume matrixes and obtain a fitness matrix, then are directly summed or weighted sum obtains
To the fitness of more string chromosomes;
S7, selection operator: using roulette selection method and elite retention strategy, that is, realizes one under more string chromosome coding modes
The operation of a resource platform matching multiple tasks target and pinned task number of targets.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810994621.4A CN109102203A (en) | 2018-08-28 | 2018-08-28 | A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810994621.4A CN109102203A (en) | 2018-08-28 | 2018-08-28 | A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109102203A true CN109102203A (en) | 2018-12-28 |
Family
ID=64864011
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810994621.4A Pending CN109102203A (en) | 2018-08-28 | 2018-08-28 | A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109102203A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111461591A (en) * | 2020-03-12 | 2020-07-28 | 哈尔滨工业大学(威海) | Crowdsourcing task allocation method based on genetic algorithm |
CN112990515A (en) * | 2019-12-02 | 2021-06-18 | 中船重工信息科技有限公司 | Workshop resource scheduling method based on heuristic optimization algorithm |
CN113487236A (en) * | 2021-07-30 | 2021-10-08 | 大连海事大学 | Airplane scheduling method based on genetic algorithm |
-
2018
- 2018-08-28 CN CN201810994621.4A patent/CN109102203A/en active Pending
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112990515A (en) * | 2019-12-02 | 2021-06-18 | 中船重工信息科技有限公司 | Workshop resource scheduling method based on heuristic optimization algorithm |
CN111461591A (en) * | 2020-03-12 | 2020-07-28 | 哈尔滨工业大学(威海) | Crowdsourcing task allocation method based on genetic algorithm |
CN113487236A (en) * | 2021-07-30 | 2021-10-08 | 大连海事大学 | Airplane scheduling method based on genetic algorithm |
CN113487236B (en) * | 2021-07-30 | 2023-09-15 | 大连海事大学 | Airplane scheduling method based on genetic algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Ghosh et al. | Evolutionary algorithms for multi-criteria optimization: A survey | |
Wardlaw et al. | Evaluation of genetic algorithms for optimal reservoir system operation | |
Jalali et al. | Multi-colony ant algorithm for continuous multi-reservoir operation optimization problem | |
Chang et al. | Real-coded genetic algorithm for rule-based flood control reservoir management | |
CN109886589A (en) | A method of low-carbon Job-Shop is solved based on whale optimization algorithm is improved | |
CN104035816B (en) | Cloud computing task scheduling method based on improved NSGA-II | |
CN106484512B (en) | The dispatching method of computing unit | |
CN109102203A (en) | A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms | |
CN101118611A (en) | Business process model resource configuring optimizing method based on inheritance algorithm | |
CN105929690B (en) | A kind of Flexible Workshop Robust Scheduling method based on decomposition multi-objective Evolutionary Algorithm | |
CN111325356A (en) | Neural network search distributed training system and training method based on evolutionary computation | |
CN106250650A (en) | The resource allocation and optimization method of model in high flux emulation | |
CN104616062B (en) | A kind of Nonlinear System Identification planned based on multi-objective Genetic | |
CN103324954A (en) | Image classification method based on tree structure and system using same | |
CN110471762A (en) | A kind of cloud resource distribution method and system based on multiple-objection optimization | |
CN111047272A (en) | Project scheduling method and device for multi-language collaborative development | |
CN113341889B (en) | Distributed blocking flow workshop scheduling method and system with assembly stage and energy consumption | |
CN106611378A (en) | Priority encoding-based hybrid genetic algorithm for solving job-shop scheduling problem | |
CN110909787A (en) | Method and system for multi-objective batch scheduling optimization based on clustering evolutionary algorithm | |
CN1450493A (en) | Nerve network system for realizing genetic algorithm | |
CN112348323A (en) | Multi-target energy supply and operation flexible scheduling method | |
CN104392317A (en) | Project scheduling method based on genetic culture gene algorithm | |
CN114217580B (en) | Functional fiber production scheduling method based on improved differential evolution algorithm | |
Ehtesham Rasi | Optimization of the multi-objective flexible job shop scheduling model by applying NSGAII and NRGA algorithms | |
CN108153254B (en) | A kind of part based on glowworm swarm algorithm is clustered to process route optimization method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181228 |
|
RJ01 | Rejection of invention patent application after publication |