CN109739087A - Multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm - Google Patents
Multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm Download PDFInfo
- Publication number
- CN109739087A CN109739087A CN201811637000.7A CN201811637000A CN109739087A CN 109739087 A CN109739087 A CN 109739087A CN 201811637000 A CN201811637000 A CN 201811637000A CN 109739087 A CN109739087 A CN 109739087A
- Authority
- CN
- China
- Prior art keywords
- agent
- particle
- manufacturing process
- swarm algorithm
- objective
- 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
- 239000002245 particle Substances 0.000 title claims abstract description 141
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 114
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000005457 optimization Methods 0.000 title claims abstract description 43
- 230000008569 process Effects 0.000 claims abstract description 30
- 239000002994 raw material Substances 0.000 claims abstract description 21
- 239000002699 waste material Substances 0.000 claims abstract description 12
- 230000003993 interaction Effects 0.000 claims abstract description 11
- 238000004891 communication Methods 0.000 claims abstract description 7
- 238000013499 data model Methods 0.000 claims abstract description 7
- 230000008859 change Effects 0.000 claims abstract description 6
- 239000003795 chemical substances by application Substances 0.000 claims description 143
- 239000004568 cement Substances 0.000 claims description 35
- 239000000654 additive Substances 0.000 claims description 20
- 230000000996 additive effect Effects 0.000 claims description 18
- 238000007726 management method Methods 0.000 claims description 10
- 229910052500 inorganic mineral Inorganic materials 0.000 claims description 9
- 239000011707 mineral Substances 0.000 claims description 9
- 238000003860 storage Methods 0.000 claims description 9
- 239000000463 material Substances 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 5
- 238000013178 mathematical model Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 238000005553 drilling Methods 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 238000011112 process operation Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 12
- NUFBIAUZAMHTSP-UHFFFAOYSA-N 3-(n-morpholino)-2-hydroxypropanesulfonic acid Chemical compound OS(=O)(=O)CC(O)CN1CCOCC1 NUFBIAUZAMHTSP-UHFFFAOYSA-N 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 6
- 239000000203 mixture Substances 0.000 description 6
- 230000006399 behavior Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 244000144992 flock Species 0.000 description 2
- 238000009776 industrial production Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 241000251468 Actinopterygii Species 0.000 description 1
- 241000238631 Hexapoda Species 0.000 description 1
- 101150039208 KCNK3 gene Proteins 0.000 description 1
- 101150083764 KCNK9 gene Proteins 0.000 description 1
- 235000008331 Pinus X rigitaeda Nutrition 0.000 description 1
- 235000011613 Pinus brutia Nutrition 0.000 description 1
- 241000018646 Pinus brutia Species 0.000 description 1
- 240000003834 Triticum spelta Species 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000004927 clay Substances 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000009833 condensation Methods 0.000 description 1
- 230000005494 condensation Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000000227 grinding Methods 0.000 description 1
- 229910052602 gypsum Inorganic materials 0.000 description 1
- 239000010440 gypsum Substances 0.000 description 1
- 244000144980 herd Species 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000013433 optimization analysis Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 241000894007 species Species 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Landscapes
- Feedback Control In General (AREA)
Abstract
The invention discloses the multiple agent manufacturing process optimization methods based on multi-objective particle swarm algorithm, it include: the manufacturing process for process industry, construct the manufacturing process Optimized model based on multiple agent, including double-layer structure, wherein, superstructure is master control Agent and the pond Agent for being used to store algorithm and data information, understructure is raw material A gent, equipment Agent, manages Agent and waste material Agent, each intelligent Agent can carry out the interaction of information based on itself communication module and others Agent；Using Agent as the particle in multi-objective particle swarm algorithm, the ability of Agent Evolution of Population is assigned, corresponding data model is established for the manufacturing process Optimized model based on multiple agent, is solved using multi-objective particle swarm algorithm, obtains efficient solution set.Cost after change is integrally lower than cost spent by reality, so the Noninferior Solution Set by optimization can be used as the reference in actual production, low consumption, the high yield, the required cost of high-quality and reduction of process operation manufacture are improved with this.
Description
Technical field
This disclosure relates to process control technology field, more particularly to the multiple agent system based on multi-objective particle swarm algorithm
Make process optimization method.
Background technique
Process industry mainly includes the industries such as chemical industry, cement, papermaking, and important ground is occupied in the industrial production in China
Position, the manufacturing process of process industry is carried out continuously, cannot interrupt, and the sequence of manufacturing process is fixed and invariable,
Management characteristic is to ensure that continuous feeding and ensures that each production link must operate normally during operation.
Due between different the products either multiple batches of production of identical product there are time and resource-sharing, from
And it is easy to happen conflict, in order to solve problems, it is necessary to reasonably be distributed all kinds of resources in production process, with this
To guarantee that the production procedure of product is not disrupted.Therefore the system control and the optimization of manufacturing process of process industry are also more multiple
It is miscellaneous, this result in increasing we for raw material, the temperature etc. in manufacturing process influence production cost and product qualification rate because
The control difficulty of element.Meanwhile in order to meet the requirement of each target in manufacturing process, for example it should guarantee the quality of product again
To reduce the activity duration as far as possible to reach the requirement of minimum production cost.
Such issues that in order to solve, is had been greatly developed based on didactic multiple-objection optimization technology, by a large amount of
Scientific investigations showed that, the technology is more more practical and effective than traditional method.It is more mainly to have: ACO (ant colony
Algorithm), NSGA (genetic algorithm), PAES (Pareto achieves evolution strategy algorithm), PSO (particle swarm algorithm).Wherein PSO algorithm
It is a kind of simulation social action, the evolution technology based on swarm intelligence, it has oneself distinctive search mechanisms, and convenient for real
Now, convergence capabilities are strong, so having obtained preferable development in manufacture field.
Based on above-mentioned analysis, how process industry to be carried out using multi-Agent modeling and utilizing multi-objective particle swarm algorithm
It is the main technical problems to be solved of the application institute to model optimization.
Summary of the invention
In order to solve the deficiencies in the prior art, embodiment of the disclosure is provided based on the more of multi-objective particle swarm algorithm
Intelligent body manufacturing process optimization method is improved on the basis of PSO algorithm, it is made to directly act on the optimization of multiple target
On.
To achieve the goals above, the application uses following technical scheme:
Multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm, comprising:
For the manufacturing process of process industry, the manufacturing process Optimized model based on multiple agent, including two layers of knot are constructed
Structure, wherein superstructure is master control Agent and the pond Agent for being used to store algorithm and data information, and understructure is raw material
Agent, equipment Agent, management Agent and waste material Agent, each intelligent Agent can based on itself communication module with
Other Agent carry out the interaction of information；
Using Agent as the particle in multi-objective particle swarm algorithm, the ability of Agent Evolution of Population is assigned, for being based on
The manufacturing process Optimized model of multiple agent establishes corresponding data model, is solved, is obtained using multi-objective particle swarm algorithm
Obtain efficient solution set.
As the further technical solution of the application, master control Agent is responsible for the coordination of the resource of entire manufacturing process, it
Positioned at the upper layer of modeling, while the dynamic of lower layer Agent is record, manufacturing process is timely carried out according to obtained information
Adjustment corresponding algorithm is taken out from the Agent of pond, then basis when optimizing analysis to a variety of different production decisions
The optimum results of algorithm carry out operation instruction；
The characteristic and various proportion information of unused material needed for raw material A gent stores production different product；
Equipment Agent is the machinery equipment in actual production, and record has temperature, pressure, the kiln speed and wind-force of machine, to protect
Demonstrate,prove the utilization rate of raw material and machine in production process；
Management Agent is used to the information that collecting device Agent and raw material A gent interaction comes, while it is fed back to
The master control Agent of layer；
Waste material Agent is then responsible for carrying out secondary use to productional surplus material, reduces the pollution to environment.
As the further technical solution of the application, one of them stage for cement manufacture is that " two mills one are burnt " matches
Than the foundation that mix stages carry out mathematical model:
Specific function is as follows:
Min H (x)=(f1(x),f2(x))T (3)
The purpose of formula (1) is exactly to make error h1(x) minimum as far as possible, improve cement performance, M according to this1It indicates to be added
Additive quantity, NkThe dosage of the additive of required investment, n when to meet standard contentkIt (x) is to reach required cement
The dosage of additive needed for performance indicator minimum zone, wherein the content of all mineral and the hardenability of current cement are resistance to
Long property etc. has the detector in specific manufacturing process to obtain；
Formula (2) is required totle drilling cost, and optimization is needed to obtain cost h2(x) minimum value, M2For the number of required additive
Amount, gtFor the additional amount of additive in t, ptFor the unit price of every kind of additive；
Formula (3) is the integration of two above target.
It as the further technical solution of the application, is solved, is carried out first just using multi-objective particle swarm algorithm
The size of population, initial position is arranged in beginningization, and initial velocity calculates the objective function of each particle, finds each particle
Current individual extreme value finds the current globally optimal solution of entire population, updates speed and the position of each particle, judges whether
Reach termination condition, if so, otherwise output optimal solution recalculates the objective function of each particle, until reaching termination item
Part.
It as the further technical solution of the application, is solved using multi-objective particle swarm algorithm, obtains effectively solution
The detailed process of collection are as follows:
Step 1: the population size Z of particle being set, and includes the position x of particle in random given populationiWith speed vi
And the number of iterations t, concurrently set the termination condition of algorithm；
Step 2: noninferior solution being added to external archival and is concentrated, is found out respectively with formula (1) and formula (2) each in population
The value of the fitness of particle；
Step 3: when the number of iterations is less than given the number of iterations t, repeating step 4-7, conversely, algorithm terminates；
Step 4: global extremum is chosen according to the fitness value of the particle in archiveEach grain is updated according to formula (3)
The position of son, while the adaptive value of contained particle is found out again；
Step 5: the process of step 2 is repeated, and removes in contrast more ineligible solution according to dominance relation,
Then the external archive of next iteration is formed；
Step 6: it is assumed that the optimal extreme value of the history for the particle chosen at present is good not as good as current location, then just using currently most
New particle works as best extreme value, conversely, then keeping current state constant；
Step 7: if the position of the random particle in population is than the best extreme value for the entire population previously chosen
If being better, then taking the processing mode such as step 6, the position extreme value previously chosen is replaced with this position
Step 8: after iteration, the particle that obtained external archival is concentrated then is the efficient solution set of entire algorithm.
It, can simultaneously for the more new formula of particle rapidity in the step 4 as the further technical solution of the application
It is divided into three parts:
First is that " inertia portion ", the formula description are as follows: w*v to keeping in the one of displacement state is representedi, wherein w generation
Table weight, viFor the speed of current particle；
Second is that " individual understanding part ", is derived from oneself individual experience, it will be appreciated that best with oneself for particle current location
The distance between position.Formula description are as follows: c1r1(hi(t)-xi(t))；
Third is that " overall situation understanding " part, the experience of other outstanding particles in group, it will be appreciated that be particle i present bit
Set the distance between group desired positions, formula description are as follows: c2v2(ki(t)-xi(t))；
The sum of formula is described using three above to be updated to the speed of particle；
xi(t+1)=vi(t+1)+xi(t) (4)
Wherein w is inertia weight, uses changeable weight set regular, it is specified that weight limit is 0.9, minimal weight 0.4,
During the test can value between the selection 0.4-0.9 of dynamic random calculated, c1c2For non-negative constant, it is same I
During the test given value be 2, r1r2For the random number between 0-1.
Embodiment of the disclosure also discloses the optimization of the multiple agent manufacturing process based on multi-objective particle swarm algorithm system
System, comprising:
Manufacturing process Optimized model construction unit based on multiple agent constructs base for the manufacturing process of process industry
In the manufacturing process Optimized model of multiple agent, including double-layer structure, wherein superstructure is master control Agent and is used to store
The pond Agent of algorithm and data information, understructure be raw material A gent, equipment Agent, management Agent and waste material Agent,
Each intelligent Agent can carry out the interaction of information based on itself communication module and others Agent；
Multi-objective particle swarm algorithm solves unit, using Agent as the particle in multi-objective particle swarm algorithm, assigns
The ability of Agent Evolution of Population establishes corresponding data model for the manufacturing process Optimized model based on multiple agent, utilizes
Multi-objective particle swarm algorithm is solved, and efficient solution set is obtained.
As the further technical solution of the application, the multi-objective particle swarm algorithm is solved in unit, utilizes more mesh
Mark particle swarm algorithm is solved, and is initialized first, and the size of population, initial position is arranged, and initial velocity calculates each
The objective function of a particle finds the current individual extreme value of each particle, finds the current globally optimal solution of entire population, more
The speed of new each particle and position, judge whether to reach termination condition, if so, otherwise output optimal solution recalculates each
The objective function of a particle, until reaching termination condition.
Embodiment of the disclosure also discloses a kind of computer readable storage medium, wherein it is stored with a plurality of instruction, institute
Instruction is stated to be suitable for being loaded by the processor of terminal device and being executed the multiple agent system based on multi-objective particle swarm algorithm
Make process optimization method.
Embodiment of the disclosure also discloses a kind of terminal device, including processor and computer readable storage medium,
Processor is for realizing each instruction；Computer readable storage medium is suitable for for storing a plurality of instruction, described instruction by processor
It loads and executes the multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm.
Compared with prior art, the beneficial effect of the disclosure is:
The multi-objective particle swarm algorithm that embodiment of the disclosure is related to is changed on the basis of PSO algorithm
Into, directly act on it in optimization of multiple target, the cost after optimization is integrally lower than cost spent by reality, so by
The Noninferior Solution Set of optimization can be used as the reference in actual production, and the low consumption, high yield, high-quality of process operation manufacture is improved with this
And cost needed for reducing.
Detailed description of the invention
The accompanying drawings constituting a part of this application is used to provide further understanding of the present application, and the application's shows
Meaning property embodiment and its explanation are not constituted an undue limitation on the present application for explaining the application.
Fig. 1 is that the manufacturing process Optimized model structure of multiple agent constructed by disclosure one or more examples of implementation is shown
It is intended to；
Fig. 2 is PSO Algorithm flow diagram of the disclosure one or more examples of implementation based on multiple agent；
Fig. 3 is that disclosure one or more examples of implementation optimize verifying schematic diagram to the particle of MOPSO algorithm；
Fig. 4 is the Pareto disaggregation curve graph of disclosure one or more examples of implementation optimum results；
Fig. 5 is Cost comparisons' schematic diagram after the practical Expenses Cost of disclosure one or more examples of implementation and optimization.
Specific embodiment
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the application.Unless another
It indicates, all technical and scientific terms used in this application have logical with the application person of an ordinary skill in the technical field
The identical meanings understood.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
Term explains part:
Particle swarm algorithm (PSO): PSO is a kind of random optimization technology based on population, by Eberhart and Kennedy in
Nineteen ninety-five proposes.Particle swarm algorithm imitates the collective behaviour of insect, herd, flock of birds and shoal of fish etc., these groups are according to a kind of conjunction
The mode search of food of work, each member in group are constantly changed by the experience and the experience of other members for learning own
Become its search pattern.
Agent system: the network structure that Agent system (MAS) is made of the Agent of several loose couplings.
These entities are either physically loose in logic but their behavior is autonomous, pass through the friendship of information each other
Mutually, negotiate to complete the control to complex task, and the solution to multi-objective problem jointly.
Process industry: also referred to as process industrial refers to the production process by physical change and chemical change progress.And it is every
A link is uninterruptedly, is carried out continuously.Such as the industries such as chemical industry, oil refining, papermaking, cement.
In a kind of typical embodiment of the application, the multiple agent manufacture based on multi-objective particle swarm algorithm is provided
Process optimization method, comprising: for the manufacturing process of process industry, the manufacturing process Optimized model based on multiple agent is constructed,
Including double-layer structure, wherein superstructure is master control Agent and the pond Agent for being used to store algorithm and data information, lower layer's knot
Structure is raw material A gent, equipment Agent, manages Agent and waste material Agent, each intelligent Agent can be based on the logical of itself
Believe that module and others Agent carry out the interaction of information；
Using Agent as the particle in multi-objective particle swarm algorithm, the ability of Agent Evolution of Population is assigned, for being based on
The manufacturing process Optimized model of multiple agent establishes corresponding data model, is solved, is obtained using multi-objective particle swarm algorithm
Obtain efficient solution set.
In specific embodiment, as shown in Figure 1, the concept in pond is introduced in the multi-Agent Systems Model of building,
Procedure information in optimization algorithm and production process required for being used to store.Each lower layer Agent and master control Agent are being produced
With algorithm and corresponding information can be obtained when optimizing to data from the Agent of pond, can again after the utilization of resources is complete
It puts back in pond, so that other Agent are called.
Manufacturing process Optimized model it can be seen from Fig. 1 based on multiple agent mainly includes two layers, and upper layer is master control
The Agent and pond Agent for being used to store algorithm and data information.Lower layer be raw material A gent, equipment Agent, management Agent with
And waste material Agent, each intelligent Agent can carry out the interaction of information based on itself communication module and others Agent.Its
In, master control Agent is responsible for the coordination of the resource of entire manufacturing process, it is located at the upper layer of modeling, while it will record
The dynamic of lower layer Agent timely carries out the adjustment of manufacturing process according to obtained information, to a variety of different production decisions into
When row optimization analysis, corresponding algorithm can be taken out from the Agent of pond, then carry out production according to the optimum results of algorithm and refer to
It leads.Characteristic and various proportion information of unused material needed for raw material A gent stores production different product etc..Equipment
Agent is the machinery equipment in actual production, and the inside record has temperature, pressure, kiln speed and wind-force of machine etc., to guarantee to produce
The utilization rate of raw material and machine in the process.Management Agent is used to the information that collecting device Agent and raw material A gent interaction comes,
It is fed back to simultaneously the master control Agent on upper layer.Waste material Agent is then responsible for carrying out secondary use to productional surplus material, reduces
Pollution to environment.
As shown in Fig. 2, the particle swarm algorithm based on multiple agent, Agent is one and is capable of the abstract of sensing external environment
Entity, and can be communicated with other Agent, the common solution for completing problem, at the same it also have intelligent thinking and
Intelligent behavior, Agent system is then by the collaborative network of several Agent loose couplings formed.And particle swarm algorithm is then
It is the mankind to the abstract and simulation during flock of birds predation, is a kind of random search iterative algorithm.It is all essentially pine
Dissipate coupling " colony intelligence " structure.Therefore, in this application, using Agent as the particle in multi-objective particle swarm algorithm, assign
The ability of Agent Evolution of Population.Multi-objective problem involved in manufacturing process is handled with this.
For the validity of verification algorithm, Binh and Korn function is had chosen:
In calculating process, the number of iterations is 5000 times, number of particles 30, and maximum number of run is 1000 times, is accelerated
Factor c1=c2=2 takes log2 operation to result, then performs image display to preferably show convergence effect.As a result such as
Shown in Fig. 3.
From figure 3, it can be seen that the effect of acquirement is better than when being optimized with Agent as the particle of MOPSO algorithm
Common MOPSO.
This embodiment example can be explained in detail the technical solution of invention by taking manufacture of cement process as an example, make to it
The analysis that process carries out substantially is made to describe.The people for understanding cement production process mentions " two mills in the production Shi Douhui for mentioning cement
One burns ", they are that raw material prepare (mill), clinker burning (one burns), cement grinding (two mills), in cement factory, cement
Production mainly have following several stages: the preparation of raw material, raw material are ground, clinker is fired, cement is ground, are stored and shipped.Its
Middle cement grind be cement manufacture last procedure, while being also the most process of power consumption, before carrying out this step,
It clinker can be carried out proportion with gypsum and additive mixes, thus required for the mineral and finished product that guarantee in clinker meet us
Performance indicator.This experiment is to be directed to water under the premise of having used multi-Agent technology to carry out modeling to manufacturing process
Clay one of them stage made i.e. said ratio mix stages carry out the foundation and optimization of mathematical model.
Specific function is as follows:
Min H (x)=(f1(x),f2(x))T (3)
The purpose of formula (1) is exactly to make error h1(x) minimum as far as possible, improve cement performance according to this.M1It indicates to be added
Additive quantity, NkThe dosage of the additive of required investment, n when to meet standard contentkIt (x) is to reach required cement
The dosage of additive needed for performance indicator minimum zone.Wherein the content of all mineral and the hardenability of current cement are resistance to
Long property etc. has the detector in specific manufacturing process to obtain.
Formula (2) is required totle drilling cost, and optimization is needed to obtain cost h2(x) minimum value, M2For the number of required additive
Amount, gtFor the additional amount of additive in t, ptFor the unit price of every kind of additive.
Formula (3) is the integration of two above target.
The specific implementation process of Case-based Reasoning
Task1: being arranged the population size Z of particle, and includes the position x of particle in random given populationiWith speed vi
And the number of iterations t, concurrently set the termination condition of algorithm.
Task2: being added to external archival for noninferior solution and concentrate, and is found out respectively with formula (1) and formula (2) each in population
The value of the fitness of particle is the fitness value that each particle is calculated separately according to the objective function of setting, and is achieved；
Task3: when the number of iterations is less than given the number of iterations t, repeat step 4-7, conversely, algorithm terminates.
Task4: individual extreme value is chosen according to the fitness value of the particle in achieving in Task2And global extremum
The adaptive value for updating the position of each particle according to formula (4), while finding out contained particle again has updated the position of particle
It sets with after speed, is calculated according to updated particle at this time；
xi(t+1)=vi(t+1)+xi(t) (4)
Three parts can be divided into simultaneously for the more new formula of particle rapidity:
First is that " inertia portion ", the formula description are as follows: w*v to keeping in the one of displacement state is representedi, wherein w generation
Table weight, viFor the speed of current particle；
Second is that " individual understanding part ", is derived from oneself individual experience, it will be appreciated that best with oneself for particle current location
The distance between position.Formula description are as follows: c1r1(hi(t)-xi(t))；
Third is that " overall situation understanding " part, the experience of other outstanding particles in group, it will be appreciated that be particle i present bit
Set the distance between group desired positions, formula description are as follows: c2v2(ki(t)-xi(t))。
The sum of formula is described using three above to be updated to the speed of particle.
Wherein w is inertia weight, this is tested us and changeable weight use to set regular, it is specified that weight limit is 0.9, most
Small weight is 0.4, during the test can value between the selection 0.4-0.9 of dynamic random calculated.c1c2It is non-negative
Constant, it is same we given value is 2, r during the test1r2For the random number between 0-1.
Task5: repeating the process of Task2, and remove in contrast more ineligible solution according to dominance relation,
Then the external archive of next iteration is formed.
Task6: it is assumed that the optimal extreme value of the history for the particle chosen at present is good not as good as current location, then just using currently most
New particle works as best extreme value, conversely, then keeping current state constant.
Task7: if the position of the random particle in population is than the best extreme value for the entire population previously chosen
If being better, then taking the processing mode as Task6, the position extreme value previously chosen is replaced with this position
Task8: after iteration, the particle that obtained external archival is concentrated then is the efficient solution set of entire algorithm.
After establishing the objective function that need to optimize, it is optimized using MOPSO algorithm, emulation experiment we adopt
With a kind of concrete model (A of certain cement production plants1) cement creation data carry out experimental verification.
In A1In the manufacturing process of model cement, content table needed for required various mineral and oxide is as shown in table 1:
Table 1: the content and range of cement Minerals, oxide
The schedule of rates of various additives is shown in table 2:
Table 2: the unit price of various additives
| Mineral species | Monovalent (yuan/ton) |
| Early strength agent | 25000 |
| Antifreezing agent | 5000 |
| Dry solution | 1800 |
| Slow setting (accelerator) | 33150 |
When optimizing to data, the number of iterations selected is 200, and the scale of population size and external archival collection is
100.Experimental result is illustrated in fig. 4 shown below, for the obtained forward position pareto of partial data intercepted:
It can be seen that the cost when composition error drops to 0.03 from 0.42, in manufacturing process from 1.13 from the result of Fig. 4
Ten thousand yuan increase 4.2 ten thousand yuan, illustrate the cost inverse relation ratio in the error and the manufacturing for reducing mineral content as far as possible
It is more apparent.And in actual manufacturing process, we often more focus on the degree up to standard of cement, and are carrying out large size
Whether cement can fully meet requirement of the people for its intensity, pressure-proof and snap-resistent, aquation condensation etc. when fortification, so generally existing
The extreme solution in noninferior solution, that is, minimum composition error can be selected in actual production to carry out the manufacturing of cement.
It gives in 50 groups of manufacturing costs after MOPSO algorithm optimization and practical manufacture of cement and is consumed in figure 5 above
Cost, can be relatively clear conclude that optimization after cost be integrally lower than cost spent by reality, so by excellent
The Noninferior Solution Set of change can be used as the reference in actual production, and low consumption, the high yield, high-quality and drop of cement manufacture are improved with this
Cost needed for low.
The model constructed in the application using the manufacturing process of certain concrete model cement, and it is based on MOPSO algorithm optimization
Two objective functions of minimum composition error and required minimum process cost of various mineral contents and specified value in cement, verifying
It is feasible and effective that multi-objective particle swarm algorithm, which is applied on process industry,.
The present invention be by taking cement manufactures as an example, for the complicated bad control of its manufacturing process, caused serious waste of resources,
The problem of increased costs, is optimized using manufacturing process of the multi-objective particle to process industry, is found out with this
The governing factor of one group of preferable cement production process reaches the rationalization utilization of resource, reduces the purpose of production cost.
It specifically, is that MOPSO algorithm is applied to the manufacture of process industry using Optimizing Flow industrial production as target
On Cheng Youhua, on the basis of the manufacturing process analysis to process industry, it is with the specific manufacturing process of certain flow industry enterprise
Example, establishes with the mathematical model of total cost and the minimum optimization aim of composition error, and give the specific of algorithm
Realization process.Experiment simulation has been carried out using existing control system model and MOPSO algorithm, the results showed that, MOPSO algorithm pair
In the optimization of manufacturing process be feasible.
Embodiment of the disclosure also discloses the optimization of the multiple agent manufacturing process based on multi-objective particle swarm algorithm system
System, comprising:
Manufacturing process Optimized model construction unit based on multiple agent constructs base for the manufacturing process of process industry
In the manufacturing process Optimized model of multiple agent, including double-layer structure, wherein superstructure is master control Agent and is used to store
The pond Agent of algorithm and data information, understructure be raw material A gent, equipment Agent, management Agent and waste material Agent,
Each intelligent Agent can carry out the interaction of information based on itself communication module and others Agent；
Multi-objective particle swarm algorithm solves unit, using Agent as the particle in multi-objective particle swarm algorithm, assigns
The ability of Agent Evolution of Population establishes corresponding data model for the manufacturing process Optimized model based on multiple agent, utilizes
Multi-objective particle swarm algorithm is solved, and efficient solution set is obtained.
The multi-objective particle swarm algorithm solves in unit, is solved using multi-objective particle swarm algorithm, is carried out first
Initialization, is arranged the size of population, initial position, initial velocity calculates the objective function of each particle, finds each particle
Current individual extreme value, find the current globally optimal solution of entire population, update speed and the position of each particle, judgement is
It is no to reach termination condition, if so, otherwise output optimal solution recalculates the objective function of each particle, until reaching termination
Condition.
Embodiment of the disclosure also discloses a kind of computer readable storage medium, wherein it is stored with a plurality of instruction, institute
Instruction is stated to be suitable for being loaded by the processor of terminal device and being executed the multiple agent system based on multi-objective particle swarm algorithm
Make process optimization method.
Embodiment of the disclosure also discloses a kind of terminal device, including processor and computer readable storage medium,
Processor is for realizing each instruction；Computer readable storage medium is suitable for for storing a plurality of instruction, described instruction by processor
It loads and executes the multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm.
The foregoing is merely preferred embodiment of the present application, are not intended to limit this application, for the skill of this field
For art personnel, various changes and changes are possible in this application.Within the spirit and principles of this application, made any to repair
Change, equivalent replacement, improvement etc., should be included within the scope of protection of this application.
Claims (10)
1. the multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm, characterized in that include:
For the manufacturing process of process industry, the manufacturing process Optimized model based on multiple agent, including double-layer structure are constructed,
In, superstructure is master control Agent and is used to store the pond Agent of algorithm and data information, understructure be raw material A gent,
Equipment Agent, management Agent and waste material Agent, each intelligent Agent can be based on itself communication modules and other
The interaction of Agent progress information；
Using Agent as the particle in multi-objective particle swarm algorithm, the ability of Agent Evolution of Population is assigned, for based on more intelligence
The manufacturing process Optimized model of energy body establishes corresponding data model, is solved, is had using multi-objective particle swarm algorithm
Imitate disaggregation.
2. the multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm as described in claim 1, feature
It is that master control Agent is responsible for the coordination of the resource of entire manufacturing process, it is located at the upper layer of modeling, while record lower layer
The dynamic of Agent timely carries out the adjustment of manufacturing process according to obtained information, excellent to a variety of different production decisions progress
When changing analysis, corresponding algorithm is taken out from the Agent of pond, then carries out operation instruction according to the optimum results of algorithm；
The characteristic and various proportion information of unused material needed for raw material A gent stores production different product；
Equipment Agent is the machinery equipment in actual production, and record has temperature, pressure, the kiln speed and wind-force of machine, to guarantee to give birth to
The utilization rate of raw material and machine during production；
Management Agent is used to the information that collecting device Agent and raw material A gent interaction comes, while it is fed back to upper layer
Master control Agent；
Waste material Agent is then responsible for carrying out secondary use to productional surplus material, reduces the pollution to environment.
3. the multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm as described in claim 1, feature
It is that one of them stage for cement manufacture is the foundation that " two mills one are burnt " proportion mix stages carry out mathematical model:
Specific function is as follows:
MinH (x)=(f1(x),f2(x))T (3)
The purpose of formula (1) is exactly to make error h1(x) minimum as far as possible, improve cement performance, M according to this1Indicate that is be added adds
Add the quantity of agent, NkThe dosage of the additive of required investment, n when to meet standard contentkIt (x) is to reach required cement performance
The dosage of additive needed for index minimum zone, wherein the hardenability durability of the content of all mineral and current cement
Deng thering is the detector in specific manufacturing process to obtain；
Formula (2) is required totle drilling cost, and optimization is needed to obtain cost h2(x) minimum value, M2For the quantity of required additive, gt
For the additional amount of additive in t, ptFor the unit price of every kind of additive；
Formula (3) is the integration of two above target.
4. the multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm as described in claim 1, feature
It is to be solved using multi-objective particle swarm algorithm, is initialized first, the size of population is set, initial position, initially
Speed calculates the objective function of each particle, finds the current individual extreme value of each particle, finds the current complete of entire population
Office's optimal solution, updates speed and the position of each particle, judges whether to reach termination condition, if so, output optimal solution, no
Then, the objective function of each particle is recalculated, until reaching termination condition.
5. the multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm as claimed in claim 3, feature
It is to be solved using multi-objective particle swarm algorithm, obtains the detailed process of efficient solution set are as follows:
Step 1: the population size Z of particle being set, and includes the position x of particle in random given populationiWith speed viAnd
The number of iterations t concurrently sets the termination condition of algorithm；
Step 2: noninferior solution being added to external archival and is concentrated, finds out each particle in population respectively with formula (1) and formula (2)
Fitness value；
Step 3: when the number of iterations is less than given the number of iterations t, repeating step 4-7, conversely, algorithm terminates；
Step 4: global extremum X is chosen according to the fitness value of the particle in archivei gbEach particle is updated according to formula (3)
Position, while the adaptive value of contained particle is found out again；
Step 5: repeating the process of step 2, and remove in contrast more ineligible solution according to dominance relation, then
Form the external archive of next iteration；
Step 6: it is assumed that the optimal extreme value of the history for the particle chosen at present is good not as good as current location, then just with current newest
Particle works as best extreme value, conversely, then keeping current state constant；
Step 7: if the position of the random particle in population is got well than the best extreme value for the entire population previously chosen
If, then taking the processing mode such as step 6, the position extreme value X previously chosen is replaced with this positioni gb；
Step 8: after iteration, the particle that obtained external archival is concentrated then is the efficient solution set of entire algorithm.
6. the multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm as claimed in claim 5, feature
It is that in the step 4, three parts can be divided into simultaneously for the more new formula of particle rapidity:
First is that " inertia portion ", the formula description are as follows: w*v to keeping in the one of displacement state is representedi, wherein w is represented
Weight, viFor the speed of current particle；
Second is that " individual understanding part ", is derived from oneself individual experience, it will be appreciated that for particle current location and oneself desired positions
The distance between.Formula description are as follows: c1r1(hi(t)-xi(t))；
Third is that " overall situation understanding " part, the experience of other outstanding particles in group, it will be appreciated that for the current location particle i with
The distance between group's desired positions, formula description are as follows: c2v2(ki(t)-xi(t))；
The sum of formula is described using three above to be updated to the speed of particle；
xi(t+1)=vi(t+1)+xi(t) (4)
Wherein w is inertia weight, uses changeable weight set regular, it is specified that weight limit is 0.9, minimal weight 0.4 is being tried
During testing can value between the selection 0.4-0.9 of dynamic random calculated, c1c2For non-negative constant, it is same we
Given value is 2, r during test1r2For the random number between 0-1.
7. the multiple agent manufacturing process optimization system based on multi-objective particle swarm algorithm, characterized in that include:
Manufacturing process Optimized model construction unit based on multiple agent, for the manufacturing process of process industry, building is based on more
The manufacturing process Optimized model of intelligent body, including double-layer structure, wherein superstructure is master control Agent and is used to store algorithm
With the pond Agent of data information, understructure is raw material A gent, equipment Agent, management Agent and waste material Agent, each
A intelligent Agent can carry out the interaction of information based on itself communication module and others Agent；
Multi-objective particle swarm algorithm solves unit, using Agent as the particle in multi-objective particle swarm algorithm, assigns Agent kind
The ability that group evolves, establishes corresponding data model for the manufacturing process Optimized model based on multiple agent, utilizes multiple target
Particle swarm algorithm is solved, and efficient solution set is obtained.
8. the multiple agent manufacturing process optimization system based on multi-objective particle swarm algorithm as claimed in claim 7, feature
It is that the multi-objective particle swarm algorithm solves in unit, is solved using multi-objective particle swarm algorithm, is carried out first initial
Change, the size of population, initial position are set, and initial velocity calculates the objective function of each particle, finds working as each particle
Preceding individual extreme value, finds the current globally optimal solution of entire population, updates speed and the position of each particle, judge whether to reach
To termination condition, if so, otherwise output optimal solution recalculates the objective function of each particle, until reaching termination item
Part.
9. a kind of computer readable storage medium, wherein being stored with a plurality of instruction, described instruction is suitable for the processing by terminal device
Device load and perform claim require any multiple agent manufacturing process optimization side based on multi-objective particle swarm algorithm 1-7
Method.
10. a kind of terminal device, including processor and computer readable storage medium, processor is for realizing each instruction；It calculates
Machine readable storage medium storing program for executing is for storing a plurality of instruction, and described instruction is suitable for by processor load and perform claim requires 1-7 any
The multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811637000.7A CN109739087A (en) | 2018-12-29 | 2018-12-29 | Multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811637000.7A CN109739087A (en) | 2018-12-29 | 2018-12-29 | Multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109739087A true CN109739087A (en) | 2019-05-10 |
Family
ID=66362413
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811637000.7A Pending CN109739087A (en) | 2018-12-29 | 2018-12-29 | Multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109739087A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110427006A (en) * | 2019-08-22 | 2019-11-08 | 齐鲁工业大学 | A kind of multi-agent cooperative control system and method for process industry |
| CN111077896A (en) * | 2019-12-27 | 2020-04-28 | 南京理工大学 | Liquid-filled flexible spacecraft parameter optimization method based on improved layering algorithm |
| CN111580564A (en) * | 2020-04-30 | 2020-08-25 | 湖南科技大学 | Parallel multi-target search cooperative operation method for three-dimensional group unmanned aerial vehicle |
| CN112379589A (en) * | 2020-10-13 | 2021-02-19 | 重庆大学 | Worm wheel machining shape controllable process |
| CN114970343A (en) * | 2022-05-24 | 2022-08-30 | 燕山大学 | Multi-objective optimization method for cement grinding process based on search space change |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101706002A (en) * | 2009-11-23 | 2010-05-12 | 湖南泵阀制造有限公司 | Bidirectional hard sealing structure of rotary ball valve |
| KR20120076528A (en) * | 2010-11-29 | 2012-07-09 | 엘에스산전 주식회사 | Method for selecting an available transfer capability |
| CN104914835A (en) * | 2015-05-22 | 2015-09-16 | 齐鲁工业大学 | Flexible job-shop scheduling multi-objective method |
| CN107479523A (en) * | 2017-09-28 | 2017-12-15 | 齐鲁工业大学 | Multiple agent based on QPSO manufactures process optimization method and apparatus |
| CN107730029A (en) * | 2017-09-22 | 2018-02-23 | 齐鲁工业大学 | Manufacturing process optimization method and apparatus based on quantum-behaved particle swarm optimization |
-
2018
- 2018-12-29 CN CN201811637000.7A patent/CN109739087A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101706002A (en) * | 2009-11-23 | 2010-05-12 | 湖南泵阀制造有限公司 | Bidirectional hard sealing structure of rotary ball valve |
| KR20120076528A (en) * | 2010-11-29 | 2012-07-09 | 엘에스산전 주식회사 | Method for selecting an available transfer capability |
| CN104914835A (en) * | 2015-05-22 | 2015-09-16 | 齐鲁工业大学 | Flexible job-shop scheduling multi-objective method |
| CN107730029A (en) * | 2017-09-22 | 2018-02-23 | 齐鲁工业大学 | Manufacturing process optimization method and apparatus based on quantum-behaved particle swarm optimization |
| CN107479523A (en) * | 2017-09-28 | 2017-12-15 | 齐鲁工业大学 | Multiple agent based on QPSO manufactures process optimization method and apparatus |
Non-Patent Citations (4)
| Title |
|---|
| DONGWANG LI,等: "Research on Manufacturing ProcessControl Based on Multi-Agent-System", 《IEEE》 * |
| 张鑫礼: "多目标粒子群算法原理及其应用研究", 《中国优秀硕士学位论文全文数据库 信息科技辑 》 * |
| 李艳丽: "基于多目标优化的粒子群算法研究及其应用", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 陈潇凯: "《车辆多学科设计优化方法》", 30 June 2018 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110427006A (en) * | 2019-08-22 | 2019-11-08 | 齐鲁工业大学 | A kind of multi-agent cooperative control system and method for process industry |
| CN111077896A (en) * | 2019-12-27 | 2020-04-28 | 南京理工大学 | Liquid-filled flexible spacecraft parameter optimization method based on improved layering algorithm |
| CN111077896B (en) * | 2019-12-27 | 2022-07-19 | 南京理工大学 | Liquid-filled flexible spacecraft parameter optimization method based on improved layering algorithm |
| CN111580564A (en) * | 2020-04-30 | 2020-08-25 | 湖南科技大学 | Parallel multi-target search cooperative operation method for three-dimensional group unmanned aerial vehicle |
| CN111580564B (en) * | 2020-04-30 | 2022-06-10 | 湖南科技大学 | Parallel multi-target search cooperative operation method for three-dimensional group unmanned aerial vehicle |
| CN112379589A (en) * | 2020-10-13 | 2021-02-19 | 重庆大学 | Worm wheel machining shape controllable process |
| CN112379589B (en) * | 2020-10-13 | 2022-02-18 | 重庆大学 | Worm wheel machining shape controllable process |
| CN114970343A (en) * | 2022-05-24 | 2022-08-30 | 燕山大学 | Multi-objective optimization method for cement grinding process based on search space change |
| CN114970343B (en) * | 2022-05-24 | 2024-07-02 | 燕山大学 | Multi-objective optimization method for cement grinding process based on search space variation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109739087A (en) | Multiple agent manufacturing process optimization method based on multi-objective particle swarm algorithm | |
| CN103279793B (en) | A kind of unmanned vehicle formation method for allocating tasks determined under environment | |
| Kaur et al. | A survey on the applications of bee colony optimization techniques | |
| CN104077630B (en) | Workshop layout method for complex job of simulating human cell evolution | |
| CN110298491B (en) | Traffic network node selection method based on slime mold group intelligence | |
| Yang et al. | The cascade reservoirs multi-objective ecological operation optimization considering different ecological flow demand | |
| CN102448123A (en) | Task allocation algorithm in wireless sensor network based on node property | |
| CN109685386A (en) | The complex network model construction method of multi-objective Job Shop based on data information | |
| Emami | Seasons optimization algorithm | |
| CN105976122A (en) | Multi-target resource allocation system | |
| CN109121105A (en) | Operator's competition slice intensified learning method based on Markov Game | |
| CN114565239A (en) | Comprehensive low-carbon energy scheduling method and system for industrial park | |
| Xu et al. | Competitive search algorithm: a new method for stochastic optimization | |
| CN111461284A (en) | Data discretization method, device, equipment and medium | |
| Abedinia et al. | A new stochastic search algorithm bundled honeybee mating for solving optimization problems | |
| Cheng et al. | Multi-strategy adaptive cuckoo search algorithm for numerical optimization | |
| Bekker | Applying the cross-entropy method in multi-objective optimisation of dynamic stochastic systems | |
| CN104239975B (en) | Based on the ship piloting scheduling method for improving discrete particle cluster algorithm | |
| Azizivahed et al. | Multi-area economic emission dispatch considering load uncertainty | |
| Liu et al. | Research on agricultural machine scheduling in hilly areas based on Improved Non-dominated Sorting Genetic Algorithm-III | |
| Guo et al. | Research on task scheduling of emergency repair team in wartime based on improved genetic algorithm | |
| Xuan et al. | An improved quantum particle swarm optimization and its application | |
| Zhang et al. | Weighted-based oriented cuckoo search | |
| Ji et al. | A variable neighborhood search algorithm for human resource selection and optimization problem in the home appliance manufacturing industry | |
| CN111260500A (en) | Hadoop-based distributed differential evolution scheduling method for small hydropower station |
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 | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: 250353 University Road, Changqing District, Ji'nan, Shandong Province, No. 3501 Applicant after: Qilu University of Technology (Shandong Academy of Sciences) Address before: 250353 University Road, Changqing District, Ji'nan, Shandong Province, No. 3501 Applicant before: Qilu University of Technology Country or region before: China |