CN1224515A - Adaptive object-oriented optimization software system - Google Patents
Adaptive object-oriented optimization software system Download PDFInfo
- Publication number
- CN1224515A CN1224515A CN199898800513A CN98800513A CN1224515A CN 1224515 A CN1224515 A CN 1224515A CN 199898800513 A CN199898800513 A CN 199898800513A CN 98800513 A CN98800513 A CN 98800513A CN 1224515 A CN1224515 A CN 1224515A
- Authority
- CN
- China
- Prior art keywords
- software object
- software
- model
- target
- optimizer
- 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
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 206
- 238000000034 method Methods 0.000 claims abstract description 283
- 230000008569 process Effects 0.000 claims abstract description 204
- 238000004891 communication Methods 0.000 claims abstract description 98
- 238000005457 optimization Methods 0.000 claims abstract description 86
- 238000004886 process control Methods 0.000 claims abstract description 85
- 238000012549 training Methods 0.000 claims description 27
- 238000012360 testing method Methods 0.000 claims description 23
- 238000013528 artificial neural network Methods 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 13
- 230000000875 corresponding effect Effects 0.000 claims description 13
- 238000005070 sampling Methods 0.000 claims description 13
- 230000002596 correlated effect Effects 0.000 claims description 3
- 230000002068 genetic effect Effects 0.000 claims description 3
- 239000000463 material Substances 0.000 abstract description 9
- 238000013519 translation Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 32
- 230000006399 behavior Effects 0.000 description 20
- 238000003860 storage Methods 0.000 description 14
- 230000008859 change Effects 0.000 description 9
- 238000005065 mining Methods 0.000 description 9
- 238000012545 processing Methods 0.000 description 9
- 230000033228 biological regulation Effects 0.000 description 8
- 239000000284 extract Substances 0.000 description 8
- 230000007246 mechanism Effects 0.000 description 7
- 238000007726 management method Methods 0.000 description 6
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 230000003068 static effect Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 239000012141 concentrate Substances 0.000 description 3
- 238000011217 control strategy Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 230000002452 interceptive effect Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000011084 recovery Methods 0.000 description 3
- 239000007858 starting material Substances 0.000 description 3
- 238000012546 transfer Methods 0.000 description 3
- 241001269238 Data Species 0.000 description 2
- 230000009471 action Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005538 encapsulation Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000002650 habitual effect Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 239000013307 optical fiber Substances 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000013515 script Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 238000013502 data validation Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011423 initialization method Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000002045 lasting effect Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
- GOLXNESZZPUPJE-UHFFFAOYSA-N spiromesifen Chemical compound CC1=CC(C)=CC(C)=C1C(C(O1)=O)=C(OC(=O)CC(C)(C)C)C11CCCC1 GOLXNESZZPUPJE-UHFFFAOYSA-N 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
- G06N5/043—Distributed expert systems; Blackboards
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/048—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators using a predictor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH OR ROCK DRILLING; MINING
- E21B—EARTH OR ROCK DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
- E21B2200/00—Special features related to earth drilling for obtaining oil, gas or water
- E21B2200/22—Fuzzy logic, artificial intelligence, neural networks or the like
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Automation & Control Theory (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Stored Programmes (AREA)
- Feedback Control In General (AREA)
Abstract
The present invention relates to process control optimization systems which utilize an adaptive optimization software systems comprising goal seeking intelligent software objects; the goal seeking intelligent software objects further comprise internal software objects which include expert system objects, adaptive models objects, optimizer objects, predictor objects, sensor objects, and communication translation objects. The goal seeking intelligent software objects can be arranged in a hierarchical relationship whereby the goal seeking behavior of each intelligent software object can be modified by goal seeking intelligent software objects higher in the hierarchical structure. The goal seeking intelligent software objects can also be arranged in a relationship which representationally corresponds to the controlled process's flow of materials or data.
Description
The present invention relates generally to Process Control System, particularly, the present invention relates to utilize the optimizing process control system of adaptive optimization software systems.Specifically, the present invention relates to a kind of adaptive optimization software systems that comprise the intelligence software object with the classification relationship structural arrangement (after this being referred to as " ISO " sometimes), wherein the target of each ISO is searched state and can be revised by the higher level ISO in described ISO hierarchy.Again specifically, the present invention relates to comprise the ISO of the in house software object that contains expert system object, adaptive model object, optimizer object, fallout predictor object, detecting device object, communication/translation object.Another concrete and last man-machine interaction method that a bit relates to people and described adaptive optimization software systems of the present invention.
Process Control System is used to various devices with testing process state and adjustment process operating parameter, so that be used in the performance optimization of given set target.Current a lot of traditional Process Control System is used the static representations of controlled process, can not represent the variation of employed process control model in real time.
In traditional Adaptive Control Theory, the parameter of using static rule to select the adaptive controller structure and regulate described controller, thus make the output of described process progressively follow the output of a model reference.Described static rule does not allow the automatic and variation of procedure of adaptation state optimally of Process Control System.
Whether no matter be self-adaptation, a distinct disadvantage of prior art Process Control System is all to lack a user interface intuitively in system of initial configuration or with this system real-time, interactive the time.
Another distinct disadvantage of prior art Process Control System is, no matter whether be self-adaptation, described Process Control System all can not automatically perform control action, thereby can not provide one to constitute powerful unified system with described Process Control System and search mechanism (global goal-seeking mechanism) with the optimized omnidistance target that realization and management objectives and purpose adapt.
In addition, a lot of Process Control Systems of prior art only provide the hierarchy that reference mark, element and/or the system model of limited level are set up or controlled.
Therefore, whether no matter be self-adaptation, the Process Control System of a lot of prior aries all can not provide the parallel multistage optimization of scope from specific, oriented-component, narrow pack level to the wideest omnidistance level.
Traditional Process Control System is made up of the concrete element (being detecting device, controller) that works alone and lack rudimentary optimization.Some system has carried out full optimization system-level, but does not have the optimization of consideration in each component-level, and some other system then only optimizes in component-level.Search mechanism when the omnidistance target of neither one and constitute a powerful unified system when realizing management objectives with described parts, whole process just can not realize integrated and optimization rudimentary or component-level and whole process or system-level optimization.
A lot of systems that parallel multistage optimization to a certain degree is provided really depend on that only one or both realize the method for desired parallel multistage optimization, rather than depend on and comprise expert system, can use multiplicity and diversity on the adaptive model of one or more method for establishing model that comprise neural network and the method that other forecast model is set up technology.
In the system of limited use several different methods, there is not a kind of Process Control System to use interactive different adaptive approach to go dynamically to change its selected forecast model in real time and needn't stop just in controlled process or described Process Control System.
In addition, the operation that depends on the people of a lot of current habitual Process Control Systems determines in real time in whole Process Control System territory and implements to optimize the set-point.These Process Control Systems need people's intervention just can make process and system optimization, but, since human operators they the control inferential capability and experience aspect very big-difference is arranged, so this artificial factor will cause the variation in a big way on described Process Control System overall efficiency.
Expert system provides effective improvement to the conventional procedure control system that can not use expert system really.But the Process Control System of a lot of current techniques can not use that expert system is helped the process control algorithm operation, algorithm is selected or algorithm parameter estimates to carry out self-adaptation.
In addition, in case be mounted, do not use the Process Control System of the current techniques of expert system not have robotization, systematized means to optimize the algorithm of its expert system and described expert system adaptively.
Neural network be a kind of be used to guarantee described process model can free in accurately predicting be established the powerful modelling technology of model process performance.But neural network has known " memory " problem, and becomes " static state " and the different models that can not find to make a variation whereby, thus can not be in the whole time forecasting process performance more accurately.
In addition, the definition of neural network depends on that the user understands function entirely truely, in the time can not guaranteeing that the user can entirely truely understand, is subjected to the infringement that user's omniscient complete solution is relied on based on the neural network of system.
In addition, some neural network need be estimated the weights that adopted from the neural network that the required arbitrary data function of the constraint reconciliation method that will adopt is derived; As the input of described network, these weights all are unavailable, therefore, calculate these weights how in real time owing to can not acquire, so, limited the applicability of neural network to described Process Control System.
Therefore, thus an object of the present invention is to provide a kind of dynamic expression of using controlled process provides the optimizing process control system that is just changing in the using process controlling models in real time.
Another object of the present invention provide a kind of can be automatically and optimize the optimizing process control system of adaptive change process condition.
A further object of the present invention provides a kind of optimizing process control system with the intuitive user interface that is used for system of initial formation and the dialogue of real-time and described system.
A further object of the present invention provides a kind of optimizing process control system that omnidistance target is searched mechanism that has, described omnidistance target is searched mechanism and is constituted a powerful unified system with a Process Control System, with the realization optimization consistent with management objectives and purpose.
A further object of the present invention provides a kind of modelling with infinite stages in fact and control classification and the optimizing process control system of the component-level process control point of unlimited amount in fact.
A further object of the present invention provides a kind of optimizing process control system with scope parallel multistage optimization from specific, oriented-component, narrow pack level to the wideest, omnidistance level.
A further object of the present invention provides a kind of integrated and optimized optimizing process control system that can realize that senior or system-level optimization and rudimentary or component-level are optimized.
The whole bag of tricks that a further object of the present invention is to use the adaptive model that comprises expert system, can use one or more method for establishing model that comprise neural network and other forecast model to set up technology provides a kind of optimizing process control system with described parallel multistage optimization.
A further object of the present invention is to allow an optimizing process control system to use one or more interactive, different adaptive approachs to come Real-time and Dynamic to change its selected forecast model, and needn't stop just in controlled process or described Process Control System.
A further object of the present invention is a kind of optimizing process control system, and this system can not have to make process and system and management objectives as one man carry out optimization under the artificial situation of intervening continuously.
A further object of the present invention provides a kind of use expert system and assists algorithm operating, algorithm selection and algorithm parameter to estimate to carry out adaptive optimizing process control system.
A further object of the present invention provide a kind of can be automatically, systematization carries out its expert system of adaptive optimization and expert system optimization Algorithm Process Control System.
A further object of the present invention provides a kind of optimizing process control system of can self-adaptation using neural network to avoid the memory problems that caused by described neural network.
Last purpose of the present invention provides the optimizing process control system of a kind of user of not relying on to the entirely true understanding of function.
As be discussed in more detail below, the invention provides a kind of optimizing process control system, by comprise intelligence software object (after this being referred to as ISO), comprise described ISO the described adaptive optimization software systems of adaptive optimization software systems, initialization method and with the man-machine interaction method of described adaptive optimization software systems, this system can have enough means to realize above-mentioned these purposes.
Referring to accompanying drawing, here, the like in a series of accompanying drawings uses identical label below.
The block scheme of Fig. 1 shows according to present invention resides in an in house software object among the ISO;
The block scheme of Fig. 2 shows according to the present invention an ISO interconnecting to another ISO;
The block scheme of Fig. 3 shows according to the present invention the interconnecting of ISO that other ISO relatively is positioned at a hierarchy;
The block scheme of Fig. 4 shows the ISO that is connected to according to the present invention on the real world instrument;
Another block scheme of Fig. 5 shows the ISO that is connected to according to the present invention on the real world opertaing device;
The block scheme of Fig. 6 shows the ISO that is connected to the real world detecting device according to the present invention and calculated, the economic state variable based on fallout predictor and optimizer;
The block scheme of Fig. 7 shows the ISO that is interconnected with one another through long-range connection according to the present invention;
The block scheme of Fig. 8 shows the flow process between a plurality of ISO that constitute with the hierarchy form according to the present invention;
The block scheme of Fig. 9 shows the optimizer object and how mutual expert system is;
The block scheme of Figure 10 shows the adaptive model object and how mutual the fallout predictor object is;
The block scheme of Figure 11 shows the optimizer object and how mutual adaptive model is;
It is how mutual with the translater object of communicating by letter that the block scheme of Figure 12 shows the optimizer object;
The block scheme of Figure 13 shows the optimizer object and how mutual the detecting device object is;
The block scheme of Figure 14 shows the expert system object and how mutual the adaptive model object is;
The block scheme of Figure 15 shows the expert system object and how mutual the fallout predictor object is;
It is how mutual with the translater object of communicating by letter that the block scheme of Figure 16 shows the expert system object;
The block scheme of Figure 17 shows the expert system object and how mutual the detecting device object is;
It is how mutual with the translater object of communicating by letter that the block scheme of Figure 18 shows the adaptive model object;
The block scheme of Figure 19 shows expert's adaptive model object and how mutual the detecting device object is;
It is how mutual with the translater object of communicating by letter that the block scheme of Figure 20 shows the fallout predictor object;
The block scheme of Figure 21 shows the scholarly forecast device and how mutual the detecting device object is;
The block scheme of Figure 22 shows the detecting device object and how mutual signal procedure translater object is;
Figure 23 shows the present invention and represents the how graphical user interface of an ISO of initialization of user;
Figure 24 shows the present invention and represents how the user is connected to a graphical user interface on the real world equipment with a detecting device object;
Figure 25 shows the present invention and represents how the user makes an ISO connect the graphical user interface that goes up opening relationships with the 2nd ISO by flow process;
Figure 26 shows the present invention and represents how the user is classified to the graphical user interface of initialization the one ISO, the 2nd ISO to the three ISO; With
Figure 27 show the present invention represent the user how with the interface of an ISO graphical user interface with the described ISO behavior of real time altering.
Auto-adaptive process control system uses the model based on computing machine of controlled process, is being used for representing that the model of described process exists under parameter or the structural uncertainties situation, helps that process of control.According to one group of target and purpose, auto-adaptive process control system changes their model to adapt to the active procedure condition and to make process performance optimization.In traditional Adaptive Control Theory, the parameter of using adaptation rule to select a suitable controller architecture and regulate this controller is so that the output of model reference is progressively followed in the output of described process.
Referring to Fig. 1, adaptive optimization software systems of the present invention comprise intelligence software object or ISO10, these software systems control is provided and/or optimize to use in various useful functions and can connecting in every way or be grouped together.ISO10 comprises the in house software object; In most preferred embodiment, the present invention use one known as object based programming, usually use such as SmallTalk
TMOr the software program method carried out of the computerese of C++ removes to carry out the in house software object of described ISO10, thus, sets up a self-adaptation, OO Optimization Software system.Should be appreciated that in applicant's plan, these software objects of in house software object and other software object can both but must not be restricted to " object-oriented " software object.Adaptive optimization software systems of the present invention are by having the control function that one or more ISO10 that are configured to collaborative expression controlled process carry out this system; By between the ISO10 in house software object and be configured and the co-operating that is used as between the ISO10 of a system can also realize optimizing.
ISO10 of the present invention can keep and described process, concrete element and/or the relevant logout of abstract element represented by these ISO10 configurablely.Be configured in below with the detecting device object of describing in detail 25 for each ISO10.Detecting device object 25 is used as the data management system of controlled process state, and these states comprise the state of described process control variable.Use these detecting device objects 25, ISO10 expert system object 12, fallout predictor object 18, adaptive model object 20 and optimizer object 22 to work together finding, to calculate, to explain and to deduce the new state of described control variable, thereby reach desired process status or implementation procedure target.
Except real world process, concrete element and/or abstract element, designed adaptive optimization software systems of the present invention, according to its initial optimization target, its current optimization aim with by the target of system user regulation, monitor its performance and revise its initial configuration adaptively to improve performance.
Referring to Fig. 1,, usually show the block scheme of ISO as an example with label 10 here.ISO10 in house software object allow each ISO10 to represent almost to be imaginabale anything.At first describe the general property of routine ISO10, describe the example institutional framework of specific ISP10 then.
ISO10 is the essential structure element of adaptive optimization software systems of the present invention.Each ISO can represent and set up the model of a plurality of physics things, abstract things or generalities things and comprise a plurality of in house software objects that will be discussed in more detail below at first, and these objects can be enabled, forbid or not constitute.Described ISO10 in house software object comprises: the expert system object 12 that can utilize the one or more rule-based knowledge bases 13 that comprise distinct logic rules 14 and/or fuzzy logic ordination 16; Can use multistage, parallel, different method for establishing model to produce the adaptive model object 20 of adaptive model, each described adaptive model mutually " competition " in real time in ISO10 is to predict the real-time process result on the basis of current, past and forecasting process parameter; Fallout predictor object 18 is used for selecting from the adaptive model object of vying each other 20 that adaptive model of can optimum prediction measured real-time process result; Optimizer object 22, be used to determine by ISO10 use at the calculating of given process status, optimization or the parameters optimization of element; Communication translater object 26, it can handle the communication between inside and outside the ISO10; With detecting device object 25, it can partly be used as intelligent data repository device and search library.
Though in ISO10, there is the in house software object functionality, enable these functions then by user's decision, or selectively, by expert system object 12 decisions of ISO itself through it.Self-adaptation of the present invention to be described in detail below, OO Optimization Software user interface, the initial initial sets that ISO10 in house software object is provided to the user; Then, the user is according to this initial sets of possibility in house software object, dispose specific I SO or ISO10 group, representing concrete element (for example automobile) or abstract element (for example calculating of per gallon kilometer), promptly to set up model with ISO is relevant with process, thus the abstraction process or the actual life of expression such as shop equipment, process, thought or system.Mechanical hook-up, electronic installation, controlled process, abstract calculating or can Be Controlled or optimization almost anything can be represented with ISO10.
With regard to the Process Control System of current techniques, ISO10 uses expert system object 12 based on computing machine, is a kind of specialized control program, and its uses inference engine and programming rule to come effect near an expert operator in calculating and carry out the one group of step that comprises setting optimization setting point.In most preferred embodiment, only there is an expert system object 12.Expert system object 12 can utilize one or more rule-based knowledge bases 13.Expert system object 12 provides an intelligent script environment (intelligentscripting environment), be used for influencing the behavior of ISO10 in adaptive optimization software systems of the present invention, and the modelling, prediction, optimization and the control that are used to realize and encourage ISO10.In addition, expert system object 12 provides a script environment, is used to make detecting device object 25 intelligence to be kept at the data that ISO10 meets with or produces in the whole time.Expert system object 12 can carry out " memory " by the relevant data of storage and they operation itself.These data can be by other in house software object or other ISO10 visit among the ISO10 identical with described expert system object 12, but cannot be directly by described user access.
No matter the those of ordinary skill in this professional skill field is appreciated that an expert system rule blurs or distinct, all reflected the element of expert's knowledge, text books relation, process model and area shops equipment knowledge.As shown in Figure 1, expert system object 12 rule-based knowledge bases 13 can use distinct rule 14, fuzzy rule 16 one or both of to go the knowledge of regulation itself and its state variable, regulation is considered the knowledge of ISO10 in house software object interaction in ISO10, regulation considers with the mutual knowledge of other ISO10 and regulation itself and it how to become the meta-knoeledge (meta-knowledge) of " living " in described computing environment and adaptive optimization software systems.Expert system object 12 can also provide a kind of ability, can (for example, its state select in whole set of) past, current and predicted value, and influence the behavior of ISO10 according to the selection of being carried out in given it self.The rule-based knowledge base 13 of expert system object 12 comprises linguistics, mathematics and/or the semiology rule of describing with distinctness or ambiguous term, these all are and (for example can comprise business rules by user's configuration, if per unit cost＞4, so, use relatively inexpensive materials) and habitual more control law is (for example, if state=overflow so, is connected valve).In addition, ambiguous term comprises fuzzy grammar and such as the usage of the Fuzzy logic design of fuzzy logic subordinate function and fuzzy set.
The key that in the adaptive optimization software systems, adopts process model be to guarantee described process model can be in the whole time accurately predicting by the performance of modelling process.The key feature of each ISO10 is to comprise functional and do not limit the method that is used to realize or provide that function.Adaptive model object 20 can use parallel, the adaptive model method for building up of some to go to provide desirable dirigibility, and these methods comprise empirical model, phenomenological model, first principle model, system identification model, neural network, linear regression and other method for establishing model.The ability of the one or more different method for establishing model of adaptive model object 20 parallel uses is advantages of the present invention.Because adaptive model object 20 can use different method for establishing model, so practical methods or method parameter can change in the whole time adaptively and automatically, like this, described method can adapt to the task that it is designed more accurate and effectively.
In most preferred embodiment, adaptive model object 20 uses the optimization method of neural network as them.The neural network of adaptive model object 20 " study " judges by the ability of described neural network, regulates the weights and/or the conjunctive tissue structure of neural network with data rule of thumb, produces " training " model whereby.20 uses of adaptive model object are trained with the corresponding training process of method of described adaptive model object 20 or are learnt; Different adaptive model methods have different training process.But in general, adaptive model object 20 uses each input in the training data relevant with described adaptive model object 20 gathered, come emulation, calculating or predict described training dataset close in that corresponding output.Adaptive model object 20 uses training process to upgrade described model parameter, so that the difference minimum between the output that is produced by adaptive model object 20 and the output of being write down.After carrying out suitable training, adaptive model object 20 uses so validity of the described adaptive model of generation of battery of tests data validation.The output of each in 20 emulation of adaptive model object, calculating or the prediction experiment data acquisition, and use the output that produces by described adaptive model and in described test figure set difference or the error between the output of this item, determine the ability of adaptive model, so that accurately emulation or predict described process.
Each ISO10 has: the input state variable, and they provide the data that are used for the described model of initialization to described adaptive model; The output state variable, they are data that described adaptive model produces; The state of a control variable, they are data relevant with target and purpose.Each ISO10 comprises the ability of predicting its to-be by the adaptive model object 20 that concerns between one or more fallout predictor objects 18 and the understanding input relevant with fallout predictor object 18 or state of a control variable and the output state variable, and prerequisite is the statement that provides when precondition and required following condition.
Use its distinct methods, adaptive model object 20 produces various different, " competition " adaptive models of may mode finishing assignment of mission with best attempted; The fallout predictor object 18 relevant with described adaptive model object 20 selects to be best suited for for given sampling increment indivedual adaptive models of real data then.Sampling increment (sampling delta) can be as the time in most preferred embodiment, but also can be such as the quality that changes, volume, color, flow, sound or dollar etc.Fallout predictor object 18 and adaptive model object 20 have how " study " set up model for their represented processes ability, fortune require they to detecting device object 25 and other ISO10 and 25 inquiries of their detecting device object by fallout predictor object 18 and the required data of adaptive model object 20 employed training methods.Fallout predictor object 20 can carry out " memory " by the relevant data of storage and they self operation.For as other in house software object or other ISO10 among the same ISO10 of fallout predictor object 20, these data can be accessed, but do not allow the user directly to visit.
Though each adaptive model is competed each other, by instructing other ISO10 promptly to provide the model 24 of being trained to other ISO10, ISO10 collaborative work in ISO10 and the hierarchy level identical or inequality, thereby allow a given ISO10 to use the model 24 of the quilt training of another ISO10, and collaborative work is to realize the target and the function of self-adaptation, object-oriented software system.
The optimizing process control system must play the optimal control effect in the mode consistent with management objectives.In the present invention, optimizer object 22 uses fallout predictor object 18 selected optimal self-adaptive models, determines to be used for realizing the optimal value of desired ISO10 to-be condition, for example controls set point.The selected optimal self-adaptive model of fallout predictor object is to be the best model of the process foundation relevant with this optimizer object 22 under the condition of the current state that provides ISO10 output.Optimizer object 22 can use the optimization method of some; In most preferred embodiment, optimizer object 22 uses generic (genetic) algorithm that their adaptive ability is provided.
ISO10 can also comprise the detecting device object 25 of any amount, with the input and output state of collection, storage and managing I SO10.Detecting device object 25 is used as data-carrier store and retrieval impact damper (datastorage and retrieval moderator), and it can provide the current data and the historical data of ISO10 of reflection or real world equipment current state for ISO10 and its in house software object; More abstract says, current data can be the computational data such as the per unit most current cost, and whether providing of historical data can be by user or ISO configuration itself.Each detecting device object 25 can comprise a plurality of built-in state variables, and they comprise that expression ISO10 software represents that the state variable (being the real world detector value) in the external world of self, ISO10 self determine the state variable of (being computing) and the state variable that other ISO10 provides.The economic scene variable is the significant capability of each ISO10.The economic scene variable is a new state variable based on the cost of the calculated value of for example " per unit " or other state variable.Detecting device object 25 also keeps the history and the statistical knowledge of the own state variable of ISO10, and these knowledge comprise the value of ISO10 and the built-in expression and the knowledge of cost.
As impact damper, detecting device object 25 can filtering, calculating is with statistics ground deal with data and storage data and provide with ISO10 and use relevant storage data.ISO10 maintains a detecting device object 25 respectively to each state variable of the state variable that comprises each instrument relevant with ISO10 or actuator.Detecting device object 25 can comprise: " input " detecting device, and it will be managed as the input data and to it by the ISO10 received data; " output " detecting device, its with ISO10 in or data of output relevant and it is managed; Use communication translater object 26, the data in the detecting device object 25 can with the real world equipment opening relationships such as instrument or actuator, thereby between the ISO10 and the described world, set up contact.Detecting device object 25 can for example store and cushion the single value data of " 1 " or for example variable codomain or collective data, comprise the color data of red, blue and green value set etc.Detecting device object 25 also is the main in house software object of ISO10, and it allows to have the user interface of ISO10 data.
In addition, detecting device object 25 is stored in adaptive model prediction, adaptive mode shape parameter and/or the adaptive model state of the ISO10 that uses in the ISO10.ISO10 also keeps output or the result from all fallout predictor objects 18 and optimizer object 22 as the ISO10 state.In the process of finishing the target that is included in the described rule description among the expert system object 12, described expert system object 12 can accumulate and use the data of detecting device object 25.Though the data of detecting device object 25 are integer, floating number, character string, symbol, logic symbol or their any arrangement normally, detecting device object 25 can be stored a table of entire I SO10, any one ISO10 in house software object or ISO10 or their in house software object.
As shown in Figure 1, each ISO10 also comprises communication translater object 26.Described communication translater object 26 can be between ISO10 and one need be by the given real world process of real time monitoring or control or elements, between ISO10 and abstract data source (for example calculated or by the data source of emulation) and/or between an ISO10 and other ISO synchronously or communication asynchronously.Therefore, communication translater object 26 can be according to the request of adaptive optimization software systems, carry out synchronously or asynchronous communication through serial line interface, parallel interface, lan interfaces, Wide Area Network interface and other hardware interface, these hardware interfaces are including, but not limited to radio frequency (RF) technology, modulator-demodular unit, satellite, electric wire, optical fiber and telemetry.
The communication of communication translater object 26 can be coding or noncoding according to the ability of receiving entity.Communication between ISO10 is encoded for according to the rules the agreement of described ISO10.Communication between ISO10 and non-ISO by that non-ISO (for example can meet, the vision of quantity, alphanumeric, text, character string, pictures and sounds or non-visual representation) needed any one agreement, these agreements comprise serial data, parallel data, LAN (Local Area Network) and wan protocol.When ISO10 ran into a new input source, communication translater object 26 also provided a new mechanism, by means of this mechanism, a plurality of new means of communication can be added on the ISO10 successively.Consider the data that they are operated oneself by storage, communication translater object 26 can " be remembered ".For as other in house software object or other ISO10 among the same ISO10 of described communication translater object 26, these data can be accessed, but cannot directly be visited by described user.
Though through communication translater 26 and other ISO10 and real world devices communicating, ISO10 in house software element is understood the attribute and the behavior of ISO10 in house software object to ISO10, and can directly communicate by letter with other ISO10 in house software element usually.For example, if request msg is understood and how to be known to the optimizer object that is disposed in an ISO10 22, then this optimizer object 22 can send data to the optimizer object 22 of another ISO10, and the intervention of the translater 26 that do not need to communicate by letter.
Each ISO10 is mutual with its environment and other ISO10 in every way.ISO10 has the ability of generation itself or copy itself, perhaps can be copied, condition be provide comprise relate to ISO10 self with it how with the meta-knoeledge of the ISO self of the mutual relevant rule of computer environment and other ISO10.As long as the support of computing environment ability is parallel and/or distributed process, in fact ISO10 just can survive under distributed situation and work.Therefore, ISO10 can be established also any old place in " being present in " one or more computer systems, and, if described computer system can provide a communicator that can be subjected to outside or internal control, described ISO10 can with also be that to be present in other ISO10 of any old place in that computer system or these computer systems mutual.In this manner, ISO10 can be connected to or be embedded within the physics things with the classification encapsulation that needs controlled process.For example, ISO10 can be embedded into (physics setting) in a physical equipment, be provided for described ISO10 state whereby and be positioned at the status detection input of other ISO10 outside the described physical equipment that it is embedded into, thereby convert the physical entity of the described ISO10 of comprising to the intelligent physical entity.The ISO10 that is arranged in the setting of described physical equipment physics has special application the robot field.
In addition, use the ISO10 of detecting device object 25 data, fallout predictor object 18 and adaptive model object 20, can emulation and/or the following process performance of prediction and determine that the ISO10 of following process performance sets up the validity of model.Therefore, described adaptive optimization software systems can be used by simulation model; Under described simulation model, can be used to replace the data of real world detecting device and actuator by emulation and/or calculated detecting device and actuator data.For the emulation of reality or abstract system is by ISO10 being estimated or the model of inquiry reality or abstract things or system, or by estimating and/or carrying out alternately with reality or the relevant rule of abstract things.
The user at first can be configurable ground or in real time with the ISO10 interface, to revise or to replace alternatively described rule, target and optimization criterion.In addition, expert system object 12, optimizer object 22, fallout predictor object 18 and adaptive model object 22 be intercommunication and mutual mutually adaptively each other, automatically real time altering each other behavior and do not need artificial intervention, these behaviors comprise sets up and deletes other in house software project; For example, optimizer object 22 can be revised the rule-based knowledge base 13 of expert system object 12, and expert system object 12 can be revised the optimization aim of the optimizer object 22 that need be searched.
Because the adaptive characteristic of described ISO10 is so adaptive optimization software systems of the present invention have the ability of given problem being set up new solution.The expert system object 12 that acts under knowledge, target and process that comprises in expert system object 12 rule-based knowledge bases 13 and the specific ISO10 state value can be set up, revise or or even breaking-up fallout predictor object 18, adaptive model object 20, optimizer object 22, communication translater object 26 and detecting device object 25.The ability of other ISO10 internal object rule of ISO10 in house software object modification, method and optimization criterion, the model and the realization that allow adaptive optimization software systems that are made of described ISO10 to change it are real-time dynamicly optimized, and needn't stop step or the described Process Control System that controlled process is carried out.
ISO10 comes " study " by continuing map and rebuilding the I/O relation.In rule, model and the equation that uses by described ISO10 any one can with ISO10 can the whole time adapt to automatically and change in response to model prediction and the practical detector of generation the prediction after read between the form of built-in measurement of error represent.Therefore, no matter version how, its model of modification that ISO10 can both be continuous, with better with these model maps to by its each detecting device object 25 continuous recording multidimensional fact (multidimensional reality).
In addition, adaptive model object 20 allows optimizer object 22 to be identified for " top condition " of a given sampling increment.If the model of adaptive model object 20 comprises be used to describe by the physical Design and the function of the represented entity of ISO10 and influence the design parameter of performance, then can actually use the optimization method of ISO10 to go to create to be used to the better mode of the ISO10 that reaches its basic task.
Each ISO10 is presented the definition of it oneself " appropriateness (fitness) " or performance objective or purpose.Described appropriateness, promptly localize performance objective can based on economy, based on arithmetic function, based on rule or set arbitrarily by the user during ISO10 in configuration.During carrying out initial configuration by the user or by one of software scripts template, these targets or purpose are prescribed and activate, can also be by the user, use the expert system object 12 or optimizer object 22 real time alterings of self rule-based knowledge base 13.
In adaptive optimization software systems of the present invention, ISO10 can also influence other ISO10.Except the model of training it, provide the model of being trained, regular collection, fuzzy logic subordinate function or by having thought employed target of ISO10 and the purpose that other effective with correct ISO10 optimizer object 22 is communicated by letter to transmitting ISO10, ISO10 can train or instruct other ISO10 by ISO10 to other.ISO10 can also become receiving entity expression to understand to instruct by expression (for example, the mathematical model) map that will be encoded.An example like this is the output that the multi-dimensional surface map of ISO10 input is become to comprise the language in distinct expression or fuzzy being illustrated in.Can also convert mathematical notation to and described result is depicted as bidimensional, three-dimensional or four-dimensional shape from the language representation.
Similarly, by transmitting appropriateness or performance objective or purpose to other ISO10, ISO10 can influence those ISO10 in the described adaptive optimization software systems.
The present invention includes a graphical user interface intuitively, be used to provide a formatting process, with regulation ISO10 interconnect and will be by the objective function of ISO10 inter-process, appropriateness criterion and target.In order to form adaptive optimization software systems, the method that the user uses the user interface of described adaptive optimization software systems to dispose, stipulate and the ISO10 in house software object selecting to be enabled and initial parameters, target and the in house software object that will be enabled use.One group parameter, target and the method initial according to ISO10, the user selects, disposes and enable one group of internal object and stipulate a group of parameter, target and method that they are initial, to be created in that ISO10 and need to be monitored and the process of optimal control or the physical relationship between other ISO10.Described user continues this process, till described user thinks that to needs controlled process is set up the satisfied model of other people fully.
Specifically, arrive shown in Figure 8 as Fig. 2, the user can need the mode of the order of information communication between the Be Controlled element to dispose two or more ISO10 or organize ISO10 more with imitation, promptly, they are to set up " flow process " ISO10 that model or expression ground connect corresponding to the relation that needs Be Controlled material or information flow.The connection imbody of flow process ISO10 be reflected in the characteristic of the sequence flow of information between the element (tangible or invisible) or material, and be used to communicate and keep control them with abstract equipment and/or real world process.
Use graphical user interface of the present invention, the user can also be configurablely with two or more ISO10 or organize ISO10 more and connect into the classification set of relationship, make these ISO10 become " classification " ISO10, stipulate priority and scale relationships between ISO10 or the ISO10 group then." classification " ISO10 is by the higher rank value or the higher-priority of the set of for example higher optimization aim of configuration, logically encapsulates one group of other ISO10, and makes more senior abstract the specializing in the described adaptive optimization software systems.Though " classification " ISO10 that is prescribed and other ISO10 in their classifications communicate,, the ISO10 of classification also can be configured to group of other ISO10, the ISO10 in their classifications not or real world devices communicating or not communicate by letter.
Therefore, ISO10 can be used as " flow process " ISO10, " classification " ISO10 or have flow process ISO10 and " classification flow process " ISO10 of classification ISO10 binding characteristic is connected.The ability of ISO10 among logic encapsulation (being classification relationship) another ISO10 allows the in fact unlimited classification in the ISO10 hierarchy.Because each ISO10 can have the in fact detecting device object 25 of unlimited amount, and the ISO10 of described in fact unlimited amount can the invention provides the in fact component-level process control set-point of unlimited amount when flow process connects each other.Therefore, after a user had been configured with initialization adaptive optimization software systems, " flow process " ISO10 related to process, the equipment and abstract that for example motor or per unit cost etc. need Be Controlled and/or optimization.In addition, in the ordinary course of things, the user is with an adaptive optimization software systems configuration and be initialized as many groups of " flow process " ISO10 that form hierarchy, need the higher and more senior abstract of Be Controlled and optimizing process so that form, and next represent total system by adaptive optimization software systems of the present invention control, for example, motor ISO10, driving shaft ISO10 and wheel ISO10 are organized into the hierarchy that expression drives train ISO10, and driving train ISO10 and other ISO10 or ISO10 organize together with expression and control a train.
In most preferred embodiment, each ISO10 concentrates level and the following level to ISO10 described in the hierarchy to be optimized; ISO10 on the lowest hierarchical structure level concentrates and carries out the highly optimization of localization.The optimization that progressively enlarge, that scope is wideer of higher levels in each hierarchy, searching in the process senior, more general objectives up to described adaptive optimization software systems and to arrive the summit promptly till the highest ISO10 in hierarchy (" pyramid "), thereby carrying out the optimization of " whole process " or system scope.In this manner, utilize the adaptive optimization software systems of described ISO10 to provide the part (element) and the omnidistance or comprehensive target that constitute a powerful unified optimizing process control system with Process Control System to search behavior, can realize simultaneously and can comprise whole process, the comprehensive consistent optimization of user's define objective of business goal.By user's initial configuration or by the hierarchy characteristic of the present invention that ISO10 or user's real-time adaptive are revised, provide desired optimization concurrently from all levels of described hierarchy that are optimized to optimization the wideest, omnidistance level special, oriented-component, narrow boundling.It can also make the user to be different from the hierarchy that head-to-foot mode disposes ISO10 or adaptive optimization software systems flexibly and at random.
A main target of described adaptive optimization software systems is, automatically use ISO10 specific optimisation objective function and concentrate on target omnidistance, that manage the system scope of definition more, to be identified for doing the optimizing process set-point of as a whole described system.Therefore, comprise the self-adaptation of the ISO10 that is configured to flow process ISO10, classification ISO10 and classification flow process ISO10 relation, OO Optimization Software system, the optimization aim of setting up at real-time process, estimate the total system performance continuously, automatically adapt to and revise rule and the method for each ISO10, so that reach user's define objective that can comprise whole process, comprehensive business goal.In this manner, the present invention and the management object as one man described process of Automatic Optimal and system and do not need artificial lasting intervention.Because adaptive model object 20 can use multiple parallel competitive model method for building up, comprise that the adaptive optimization software systems of ISO10 can realize parallel multistage optimization.
For example, an enterprise can have two operation factories.Described user can interrelate an ISO10 and this enterprise on the whole, as first or top hierarchy " pyramid " ISO10; An ISO10 and each independent factory are interrelated, as the next one in the described hierarchy lower or second level.Even each among the ISO10 of the second level can have their optimal control target, these optimal control targets can not realize the optimization of system scope.Five-star ISO10 is that enterprise-level ISO10 can influence each second level ISO10, makes the optimal control target maximization of enterprise-level on the whole, rather than makes each second level ISO10 keep their optimal control target separately.
In addition, a self-adaptation object-oriented Optimization Software system can be used to the process of emulation real world, and can train the terminal user in by the operation of emulation real world process or work.
In Fig. 2 to 22, ISO10 is marked as the back with having a letter or letter and number, for example ISO10A or ISO10B1 to represent the example of an ISO10.Similarly, the in house software object that is marked with letter or letter and number in the back also means it is an example of described object, and for example, real world instrument 30A1 is an example of real world instrument 30.
Referring to Fig. 2, ISO10A to ISO10H be used as expression information each other sequence flow flow process ISO10 and connect.The flow process of specifying information from an ISO10 to another ISO10 such as the physical item information of all parts in this way of expression or material represented in the connection of (between for example described ISO10A and the ISO10C) between the described ISO10, or such as the flow process of abstracted information from an ISO10 to another ISO10 that is the restriction of instruction or cost.
Referring to Fig. 3, ISO10 can be configured to one " container " (being that classification is relevant) of other ISO10, then, sets up the hierarchy of an ISO10 even ISO10 group.In Fig. 3, " flow process " ISO10A1 is encapsulated in by classification to ISO10A4 the ISO10 that is included among another ISO10 is described in the ISO10A, like this, sets up a hierarchy between many group ISO10.In this case, the in house software object relevant with described " container " ISO10A and other have been set up relation with ISO10 that comprises among the described container ISO10A or ISO10 group (for example ISOA1 is to ISOA4).Container ISO10A in house software object can be used to transmit target, purpose, constraint or order (instruction or direction), and to the ISO10 that is comprised, promptly ISO10A1 is to ISO10A4.
Referring to Fig. 4, ISO10A to 10F communicate each other and, through being arranged in the translater object 26 of communicating by letter of ISO10A1 and 10A2, to communicate by letter with real world instrument 30A2 with real world instrument 30A1, their state is dynamically caught by the detecting device object in ISO10A1 and ISO10A2.Shown in ISO10B, the real world state variable can be the static variable of real world instrument 30B1 such as the expression color, or such as the dynamic variable of the real world instrument 30BA of expression pressure.
Referring to Fig. 5, ISO10A communicates by letter with real world opertaing device 32 to ISO10F, with the notion of expression control action.For example, through its communication translater object 26 and use its detecting device object 25, expert system object 12, rule-based knowledge base 13, fallout predictor object 18, adaptive model object 20 and optimizer object 22, ISO10A transmits a plurality of state values and gives ISO10C.ISO10A1 gives real world opertaing device 32A1 through a control command that transmits governing speed in the ISO10A1 communication translater object 26.
Referring to Fig. 6, enumerated the example of adaptive optimization software systems here.ISO10A communicates by letter with economic scene variable 34A1 with real world instrument 30A1, real world instrument 30A2; ISO10B communicates by letter with real world instrument 30B2 with real world instrument 30B1; ISO10C communicates by letter with economic scene variable 34C1; ISO10D communicates by letter with real world instrument 30D1 and has a predicted state variable 36D1; ISO10E communicates by letter with real world instrument 30E1 and has optimizer state variable 38E1; Communicate by letter with real world instrument 30F1 with ISO10F and have economic scene variable 34F1.
Referring to Fig. 7, a plurality of ISO10 communicate with one another (being used to communicate with one another) through remote connector 40.Described remote connector 40 forms in many ways, including, but not limited to radio-frequency technique, modulator-demodular unit, satellite, electric wire, optical fiber, remote measurement, LAN (Local Area Network) and wide area network.
Referring to Fig. 8, a plurality of ISO10 interconnect with the hierarchy form.ISO10 is classified to comprise ISO2A, ISO2B and ISO2C.ISO2A is classified to comprise ISO10A; ISO2B is classified to comprise ISO10B; Be classified to comprise ISO3A and ISO3B with ISO2C.In addition, ISO3A is classified to comprise ISO10C, ISO10D and ISO10E; ISO3B is classified to comprise ISO10F, ISO10G and ISO10H.When ISO10 was arranged in such a way, they can be represented by be included in the abstract concept such as company, factory or factory area that ISO10 constituted in the ISO group by classification.
Mutual, mutual with other ISO10 and real world equipment of ISO10 special internal software object and they and other in house software object are described below.
Fig. 9 shows optimizer object 22 and how mutual with ISO10 expert system object 12 is.Expert system object 12 can utilize one or more rule-based knowledge bases 13, and the behavior of the target, purpose and the business logic that are provided for realizing being suitable for ISO10 and the language rule of control strategy are preserved in these storehouses.Use is included in distinct logic rules 14 and/or the fuzzy logic ordination 16 in the rule-based knowledge base 13, and expert system object 12 can pass through the regulation and the configuration of the target/purpose function 50 of change optimizer object 22, revises optimizer object 22.For example, expert system object 12 can utilize the rule-based knowledge base 13 that comprises distinct rule 14 and/or fuzzy rule 16 to dispose, and revises the target/purpose function 50 of optimizer object 22 on the basis that obtains the original supply material relevant with the ISO10 process.
Figure 10 shows the mutual of adaptive model object 20 and fallout predictor object 18.Fallout predictor object 18 will need Be Controlled and adaptive model and those adaptive model data of the described process represented by ISO10 offer ISO10.The actual prediction and the described response of each current adaptive model of described process are compared, and in fallout predictor object 18 its all adaptive models of identification which calculated to a nicety most; Then, that adaptive model is identified as the best model 60 of fallout predictor object 18.
Figure 11 shows optimizer object 22 and how mutual adaptive model object 20 is.In adaptive optimization software systems, optimizer object 22 disposes according to the target that is suitable for ISO10/purpose function 50.Optimizer object 22 disposes according to process constraint 52, and process constraint 52 is restrictions of the controlled process that optimizer object 22 can not be violated when work is with the target/purpose function 50 that realizes the ISO10 regulation.It is how mutual with fallout predictor object 18 that Figure 11 also shows optimizer object 22.Optimizer object 22 uses the best model 60 of fallout predictor object 18 and fallout predictor object 18, determine the best condition that realizes the target/purpose function 50 of optimizer object 22, represent the future performance of process with prediction ISO10 on the basis of the described condition that offers fallout predictor object 18 by optimizer object 22.Optimizer object 22 uses adaptive model object 20 to come the future performance of emulation, calculating or forecasting process can not run counter to process constraint 52 with the condition of finding best its target/purpose function 50 of realization.Optimizer object 22 knows which adaptive model object 20 will be made the best model 60 that spends discovery fallout predictor object 18 by inquiry fallout predictor object 18.When optimizer object 22 is chosen as the test condition of possible solution of the target/purpose function 50 that realizes optimizer object 22, optimizing process will carry out repeatedly, send these test conditions to adaptive model object 20 then.Adaptive model object 20 uses the condition emulation of optimizer object 22 to need controlled process, and the result of adaptive model object 20 returned to optimizer object 22, then, optimizer object 22 determines whether these results have realized optimizer object 22 desired target/purpose functions 50 and the process constraint 52 of not running counter to optimizer object 22.If optimizer object 22 keeps described result; If not, then optimizer object 22 is attempted one group of new test condition.
Figure 12 shows optimizer object 22 and how to transmit and receive data to ISO10B or from ISO10A to real world equipment 100 from ISO10A alternately with the translater 26 of communicating by letter; Shown in Figure 12 to 22, real world equipment 100 can comprise real world opertaing device 32 and real world instrument 30.ISO10A optimizer object 22 provides ISO benchmark 52 for an ISO10A communication translater object 26, and the described ISO10A of these ISO benchmark 52 identifications wishes the ISO10B that communicates by letter with it; The 26 storage requests of communication translater object are extracted needed scheduling of data (schedule) and ISO benchmark 52 from ISO10B.In addition, communication translater object 26 uses the agreement that is applicable to ISO10B 66 known to the ISO10A, sets up the request from the ISO10B data.At a suitable sampling increment place, communication translater object 26 is surveyed ISO10B, extracts data and send these data to optimizer object 22 from ISO10B.When by optimizer object 22 when communication translater 26 sends data request command and makes optimizer object 22 request from the data of ISO10B, can from ISO10B, extract data through communication translater 26, perhaps when optimizer object 22 used transfer data command to send described data to communication translater object 26, described data can be communicated by letter with ISO10B.
As shown in figure 12, the optimizer object 22 of ISO10A can by provide ISO10A communication translater 26 requests with the real world benchmark 64 that is used to discern real world equipment 100 from or transmit data to real world equipment 100; The 26 storage requests of communication translater object are extracted required scheduling and the real world benchmark 64 of data from real world equipment 100, or use the agreement that is applicable to real world equipment 100 66 that ISO10A understood, and come to carry out data communication with real world equipment 100.At a suitable sampling increment place, communication translater object 26 is surveyed real world equipment 100, extracts data and send these data to optimizer object 22 from real world equipment 100.When by optimizer object 22 when communication translater 26 sends a data request command and makes optimizer object 22 request from the data of real world equipment 100, can from real world equipment 100, extract data through communication translater 26, perhaps reportedly lose one's life order when sending data to communication translater object 26 when optimizer object 22 uses a number, described data can be communicated by letter with real world equipment 100.
In another case, when needs, data be read or be write to optimizer object 22 can asynchronous request on demand.
It is how mutual with detecting device object 25 that Figure 13 shows optimizer object 22.Optimizer object 22 uses the current data in data-carrier store 74 of detecting device objects 25 or arbitrary combination of history and/or statistics in historical and statistical processor 72, come the target/purpose function 50 of Estimation Optimization device object 22, and the optimization of optimizer object 22 is carried out in use constraint 52.
It is how mutual with adaptive model object 20 that Figure 14 shows expert system object 12.Expert system object 12 can utilize one or more rule-based knowledge bases 13 and can use the rule that is included in the described rule-based knowledge base 13, the performance and the configuration of adaptively modifying adaptive model object 20.For example, in some configuration or arbitrarily on the basis of condition, adaptive model object 20 may need to stop the sampling to new training data 58 or test data 56.Rule in the rule-based knowledge base 13 of expert system object 12 should be configured to such an extent that can realize this effect.In addition, expert system object 12 rule-based knowledge bases 13 can be included in the rule that all stops training process on the basis of model parameter or training error value.
It is how mutual with fallout predictor object 18 that Figure 15 shows expert system object 12.Expert system object 12 can use the rule that is included in the rule-based knowledge base 13, the performance or the configuration of adaptively modifying fallout predictor object 18.For example, when described best model 60 can not accurately be described the controlled process of needs, expert system object 12 can comprise the rule that makes fallout predictor object 18 produce a new adaptive model.
It is how mutual with the translater object 26 of communicating by letter that Figure 16 shows expert system object 12.Expert system object 12 is provided for discerning the ISO benchmark 62 of ISO10B or is used to discern the real world benchmark 64 that ISO10A wishes the real world equipment 100 that communicates with to communication translater object 26.Required scheduling of data and suitable benchmark are extracted in the translater object 26 storage requests of communicating by letter from ISO10B or real world equipment 100.Communication translater object 26 uses the agreement 66 that is suitable for ISO10B or real world equipment 100, sets up the request of extracting data from ISO10B or real world equipment 100.Expert system object 12 is asked the data from ISO10B or real world equipment 100 by sending data request command to communication translater object 26.At a suitable sampling increment place, communication translater object 26 is surveyed ISO10B or real world equipment 100 in due course, extracts data and sends described data to expert system object 12.For data being sent to ISO10B or real world equipment 100, the data that expert system object 12 uses data transfer command will need to communicate by letter send communication translater object 26 to.Data be read and be write to expert system object 12 can asynchronous when needed request on demand.
It is how mutual with detecting device object 25 that Figure 17 shows expert system object 12.Expert system object 12 can use the acting rules that are included in the rule-based knowledge base 13 in the current data in data-carrier store 74 or the history in historical and statistical processor 72 and/or the combination in any of statistics of detecting device object 25.
It is how mutual with the translater object 26 of communicating by letter that Figure 18 shows adaptive model object 20.Adaptive model object 20 is provided for discerning ISO10B or ISO10A expectation the ISO benchmark 62 or the real world benchmark 64 of the real world equipment 100 of communication with it to communication translater object 26.The translater object 26 storage requests of communicating by letter are extracted required scheduling of data and suitable benchmark from ISO10B or real world equipment 100, use is applicable to the agreement 66 of ISO10B or real world equipment 100, sets up the request of extracting data from ISO10B or real world equipment 100.At a suitable sampling increment place, communication translater object 26 is surveyed ISO10B or real world equipment 100, extracts data and described data is sent to the adaptive model object 20 of the request of sending.In order to transmit data to ISO10B or real world equipment 100, adaptive model object 20 uses data to transmit order and sends data to communication translater object 26; Adaptive model object 20 is by sending the request msg order to ISO10B or real world equipment 100 request msgs.When needs, can utilize adaptive model object 20 to carry out the asynchronous request of reading on demand or writing.
It is how mutual with detecting device object 25 that Figure 19 shows adaptive model object 20.Adaptive model object 20 makes ISO10 have description, sets up the ability of model and prediction ISO10 and its detecting device object 25 represented processes.Adaptive model object 20 is established as its adaptive model of training and learns how to control its required data of correlated process.Adaptive model object 20 is stored the data relevant with training in adaptive model object 20 training datas 58; In addition, adaptive model object 20 is set up the required data of correct degree of described adaptive model object 20 study of test, and with these data storage in the test data 56 of adaptive model object 20.In order to set up training data 58 and/or test data 56, adaptive model object 20 is consulted one or more detecting device objects 25 that comprise adaptive model object 20 input tables, when the relevant process of emulation or prediction and adaptive model object 20, described input provides initialization adaptive model object 20 required data.Adaptive model object 20 is also consulted one or more detecting device objects 25 that comprise adaptive model object 20 output tables, and described output provides the adaptive model object 20 described data.The inlet of training data 58 or test data 56 comprises from the input with the sample interval increment of corresponding output.
It is how mutual with the translater object 26 of communicating by letter that Figure 20 shows fallout predictor object 18.Fallout predictor object 18 is provided for discerning the ISO10A expectation ISO benchmark 62 of the ISO10B of communication with it to communication translater object 26.Required scheduling and the ISO benchmark 62 of data extracted in the 26 storage requests of communication translater object from ISO10B, and the agreement that is suitable for ISO10B 66 of using ISO10A when needed and being understood is set up the request of extracting data from ISO10B.At suitable sampling increment place, communication translater object 26 is surveyed ISO10B, extracts the fallout predictor object 18 that data also send these data to the request of sending.In order to transmit data to ISO10B, fallout predictor object 18 uses data to transmit order and sends those data to communication translater object 26; Fallout predictor object 18 is by sending the request msg order to the ISO10B request msg.When needs, can utilize fallout predictor object 18 to read or write the asynchronous request of data on demand.
Figure 21 shows fallout predictor object 18 and how mutual detecting device object 25 is.Fallout predictor object 18 provides ISO10 to represent the data of adaptive model He these adaptive models of process to ISO10.Fallout predictor object 18 can relate to one or more adaptive model objects 20, and identification is used for the best model 60 of that sampling increment, promptly discerns in all adaptive model objects 20 relevant with fallout predictor 18 which by current accurately predicting.Fallout predictor object 18 can be set up and be used to store and select best model and write down current best model efficient and a detecting device object 25 of fallout predictor object 18 efficient relevant datas.Difference between current best model efficient and fallout predictor object 18 efficient is: current best model efficient only relates to the single performance of that model, and the efficient of fallout predictor object 18 can not only be considered a model, but must reflect and relate to the history of described fallout predictor object 18 performances.Therefore, the efficient of fallout predictor object 18 combines the history of selected each model as fallout predictor object 18 best models 60.
It is how mutual with the translater object 26 of communicating by letter that Figure 22 shows detecting device object 25.Detecting device object 25 provides real world benchmark 64 to communication translater object 26, promptly discerns the needed characteristic of data in the real world equipment 100.If desired, required scheduling and the described benchmark of data extracted in the 26 storage requests of communication translater object from real world equipment 100; Communication translater object 26 uses the known agreement 66 relevant with real world equipment 100 of communication translater object 26, and foundation is to the request from the data of real world equipment 100.At a suitable increment place, communication translater object 26 is surveyed real world equipment 100, extracts data and sends these data to detecting device object 25.When needs, the asynchronous request of data can be read or write to detecting device object 25 on demand.In order to transmit data to real world equipment 100, the data that detecting device object 25 uses transfer data command will need to communicate by letter send communication translater object 26 to.Detecting device object 25 is by sending the request msg order, and request is from the data of real world equipment 100.No matter data are being requested still to be transmitted, and detecting device object 25 all adopts suitable calculating/filtering 70 methods that its result is stored in its data-carrier store 74.
The function of interconnection ISO10 provides control represented to them, that set up the real world entities of model, optimization and control.The present invention provides a graphical user interface intuitively to the user, thereby also realizes the interconnection of ISO10 with it alternately with the optimizing process control system that initialization is made of the present invention.By initialization ISO10 or one group of ISO10 and the mode, order and the priority that are based upon information flow between the ISO10, the ability that the invention provides a uniqueness is handled described diagram interface and the controlled described process of expression needs diagrammatically to allow the user.The example that Figure 23 to 27 provides a user how to be related through graphical user interface of the present invention and the present invention.
Figure 23 shows the graphical user interface for user's configuration and the described ISO10 of initialization of the present invention.Usually, the user selects one or more menu options 82 so that generic ISO10 is added to the arrangement plan of being drawn from menu 80; Each ISO10 can have by adaptive optimization software systems or the unique identifier that provided by described user.Use is from the menu option 82 of menu 80, the user just can edit the configuration of described ISO10A, and this comprises the configuration of rule-based knowledge base 13, fallout predictor object 18, adaptive model object 20 and the optimizer object 22 of ISO10A detecting device object 25, expert system object 12.
Though the user can dispose an ISO10 when the configuration starter system, Figure 24 shows an example, and how the expression user uses the behavior of graphical user interface real time altering ISO10 of the present invention.In Figure 24, be initialised before described graphical user interface is used to be chosen in and dispose and discernible ISO10, also be used to dispose the rule-based knowledge base 13 that is used for expert system object 12.Figure 24 shows the fuzzy rule 16 of utilization at preceding configuration change ISO:
" if ISO detecting device-1 is low, then ISO detecting device-A height "
" if ISO detecting device-1 is low, and then ISO detecting device-A is medium "
Under similar mode, the user can use interface shown in Figure 24 to change the configuration and the behavior of detecting device object 25, fallout predictor object 18, adaptive model object 20 and optimizer object 22.
In case the user has disposed two or more ISO10, the user just can dispose flow process and/or the classification relationship between the described ISO10.Graphical user interface of the present invention when Figure 25 shows the ISO10 of flow process connection contact that passes through to another ISO10.After having set up two ISO10, be ISO10A and ISO10B through graphical user interface shown in Figure 23, the user is choice menus option 82 from allow the user to draw to be connected with the flow process of configuration from ISO10A to ISO10B 84 menu 80.Then, the user uses the indication device such as mouse that cursor is placed on the ISO10A, and uses the system of selection of described indication device, for example knocks mouse button and select ISO10A.After having selected ISO10A, the user uses described indication device that cursor is placed on the ISO10B, and breaks away from the system of selection of described indication device, thereby the line 84 that ends in the arrow point is added in the drawing of describing the flow process relation.Described arrow point is pointed out for example flow direction from ISO10A to ISO10B.
Figure 26 shows and is used for classification and sets up the user interface that concerns between ISO10A, ISO10B and the ISO10C.In this example, the user at first sets up and disposes as ISO10A and ISO10B in conjunction with Figure 23 discussed.In this example, ISO10A and ISO10B also are connected through flow process described in conjunction with Figure 25 and are relative to each other.The user uses the indication device such as mouse to place ISO10A to go up selection ISO10A and ISO10B cursor then, and the selective power of utilization indication device for example mouse button is selected ISO10A, and the selection relevant with ISO10B with repetition handled.Then the user from menu 80 choice menus option 82 to set up and to show ISO10C; ISO10C is higher one-level in described hierarchy.By they logically being packaged into next highest in their hierarchies, ISO10C will the two sets up classification relationship with ISO10A and ISO10B.
After configuration ISO10, the user can be added to detecting device object 25 on the ISO10, and arbitrary detecting device object 25 is relevant with real world equipment.
Figure 27 shows the present invention and is being used to set up the graphical user interface that concerns between ISO10 detecting device object 25 and the real world devices.In this example, the user selects a real world equipment in the drop-down combo box that is marked as " data server " 86.On the basis of giving configuration protocol that is used for that equipment, configuration field 88 appears at the below part of demonstration.For the agreement of selection like this, the user stipulates the attribute of all some real world equipment as shown in figure 27 in configuration field 88: " name ", " attribute " and " FCM ".Can in the input frame that is marked as " renewal rate " 90, dispose the renewal rate that is used to arrange automatic data acquisition.When described user had correctly disposed these options, he selected " acceptance " button 92.
Showing and describing under the situation of most preferred embodiment, can make various modifications and replacement and can not break away from the spirit and scope of the present invention.Therefore, should be appreciated that explanation of the present invention is exemplary rather than restrictive.
As an example of commercial Application of the present invention, the enterprise that moves the mining processing plant of the Process Control System that needs the mining processing provides a chance of understanding purposes of the present invention in depth.Example hereto, the mining processing plant of enterprise comprises two treatment schemees: process of lapping and floatation process.
The business goal relevant with operation mining processing plant comprises to be realized maximum productivity simultaneously, reclaims valuable ore to greatest extent in rock is supplied with, the product of production E.B.B. most possibly and preservation original product material.Example supposes that described plant configuration is at first through floatation process the starting material of supplying with to be handled then through process of lapping hereto.In addition, the business goal of enterprise has alternative.For example, increase throughput rate and may make recovery reduce, reclaim the decline that may cause product quality and increase.
In addition, preserve starting material and produce contradiction with maximization throughput rate, the highest recovery of realization and production E.B.B. product.Therefore, be very complicated in variation and the economic relation between the raw materials consumption, and be difficult to realize simultaneously production, recovery, product quality growth property.
Utilize self-adaptation object-oriented software system to dispose the ISO that expression has the mining treatment plant of grinding and floatation process.ISO is placed on the diagram drawing interface, is used to represent the classification relationship between material and information flow and the described ISO.Being added to details at different levels in the described drawing just depends in controlled process and the controlled target that is used to optimize described process.
For the current example that is being considered, in described drawing, set up or draw the ISO of the described mining of expression processing plant and the ISO of expression process of lapping and floatation process.These process of lapping and floatation process ISO should be positioned at the described mining of expression processing plant and comprise within the ISO of mining processing plant of the grinding and the such fact of floatation process.
Also should set up and be used for each treatment loop relevant and the ISO of instrument with grinding and floatation process.The all ISO of set represent through the material of described factory and the flow process of information with described hierarchy then.
The real world equipment of expression instrument and with described mining processing plant, treatment loop and device-dependent data, get in touch through described detecting device object and suitable ISO foundation.Communication translater object provides the data that need transmit between real world equipment and described adaptive optimization software systems.In addition, the detecting device object that is used to represent to stipulate the data of each ISO target also is configured to such an extent that be convenient to optimize described process.
Be the expert system object of ISO configuration use fuzzy logic or distinct logical type regular expression, promptly the control strategy rule comes the execution scope to be suitable for the control of described ISO.For example, expert system object control strategy rule can be configured to such an extent that be used for described grinding and floatation process, discerns the cycle of operation during with described processing of box lunch or instrument overload.In addition, these rules can be configured to can not the accurately predicting process performance at described adaptive model the described process of cycle inner control.
For ISO configuration fallout predictor object and adaptive model object, so that the process of representing for described ISO is set up model.Adaptive model is predicted as the ability that ISO provides the prediction future performance, and this ability comprises the data of the performance objective that the described ISO of definition is represented.
Be ISO configuration optimization device object and optimization aim, will cause the condition of described process optimum performance so that search.Selected condition can directly be sent to described process or is stored in the described ISO detecting device object.
Statement
International office is received the modification of mailing on August 25th, 1998, the 1st, 2 and the 10th original claims of corresponding replacement of wherein new claims, and claims 3-9 item originally, the 11-21 item does not become.
Claims
Modification according to the 19th of treaty
1. a target is searched intelligence software object (goal seeking intelligent software object), it reside among the computer usable medium and can be from this medium accesses, can represent and influence a Process Control System, and the process that can control by described Process Control System by a plurality of experience state representation and influence
It is characterized in that
Search the experience state of intelligence software object for each described target, described target is searched the intelligence software object and is adapted to described controlled process by setting up, select, learn, train and remembering one or more forecasting software models
Whereby
For each the described experience state that has more than the described process of a described forecasting software model, described target is searched the one or more described forecasting software model of intelligence software Object Selection training in advance, comes to set up model for described controlled process.
2. a target is searched the intelligence software object, it reside among the computer usable medium and can be from this medium accesses, can represent and influence a Process Control System, and the process that can control by described Process Control System by a plurality of experience state representation and influence, described intelligence software object comprises a plurality of in house software objects
It is characterized in that
At least one described in house software to as if be different from the in house software object of other described a plurality of in house software objects, described a plurality of in house software objects intercom mutually,
Each described in house software object represents individual target and searches behavior, described in house software object represent alternately mutually comprehensive target search behavior and
The described individual target that one or more described in house software objects can be revised one or more described other in house software objects is searched behavior.
3. a target is searched the intelligence software object, comprising:
A plurality of in house software objects, described a plurality of in house software objects also comprise
At least one expert system software object,
At least one adaptive model software object,
At least one fallout predictor software object and
At least one optimizer software object,
Wherein
Described expert system software object is communicated by letter with described optimizer software object and the behavior that can revise described optimizer software object,
Described expert system software object is communicated by letter with described fallout predictor software object and the behavior that can revise described fallout predictor software object,
Described expert system software object is communicated by letter with described adaptive model software object and the behavior that can revise described adaptive model software object,
Described optimizer software object and described expert system software object communication and the behavior that can revise described expert system software object,
Described optimizer software object communicate by letter with described fallout predictor software object and can revise described fallout predictor software object behavior and
Described optimizer software object is communicated by letter with described adaptive model software object and the behavior that can revise described adaptive model software object,
Whereby
Described a plurality of in house software object manifests comprehensive target and searches behavior.
4. target according to claim 3 is searched the intelligence software object, it is characterized in that, described expert system software object also comprises a rule-based knowledge base, and this rule-based knowledge base comprises one or more rules of selecting from the set of distinct logic rules and fuzzy logic ordination.
5. target according to claim 3 is searched the intelligence software object, it is characterized in that, described optimizer software object also comprises a plurality of optimization methods, and wherein, described optimization method comprises one or more generic algorithms (genetic algorithm).
6. target according to claim 3 is searched the intelligence software object, it is characterized in that, also comprises a plurality of method for establishing model, and wherein, described method for establishing model comprises one or more neural networks.
7. target according to claim 3 is searched the intelligence software object, it is characterized in that described communication translater software object also comprises from comprising that internal agreement, described target search the communication protocol of selecting the set of the needed agreement of intelligence software object external unit, serial data communication agreement, parallel data communication agreement, LAN data communication protocol and wide area network data communication protocol.
8. a target is searched the intelligence software object, comprising:
At least one the expert system software object that comprises at least one rule-based knowledge base;
At least one adaptive model software object is used for producing one or more forecast models according to one or more input values with according to one or more method for establishing model;
At least one fallout predictor software object is used for using fallout predictor to select criterion and optimization method to select best one of described forecast model;
At least one optimizer software object is used to utilize one or more subject object and one or more optimization method to set up the optimization output data value that is used for described optimizer software object;
At least one communication translater software object is used to use one or more data communication protocols to carry out data communication; With
At least one detector software object, this detector software object can receive data from described expert system software object, described optimizer software object, described adaptive model software object, described fallout predictor software object, described communication translater software object and other detector software object; The described data of process are to store; Store described data; With on the basis of described expert system software object, described optimizer software object, described adaptive model software object, described fallout predictor software object, described communicate by letter translater software object and other detector software object requests, the data of being stored are offered described expert system software object, described optimizer software object, described adaptive model software object, described fallout predictor software object and the described translater software object of communicating by letter;
Whereby
Described expert system software object is communicated by letter with described optimizer software object and the described subject object that can revise described optimizer software object,
Described expert system software object is communicated by letter with described optimizer software object and the described optimization method that can revise described optimizer software object,
Described expert system software object is communicated by letter with described fallout predictor software object and the described fallout predictor that can revise described fallout predictor software object is selected criterion,
Described expert system software object is communicated by letter with described adaptive model software object and the described method for establishing model that can revise described adaptive model software object,
Described optimizer software object and described expert system software object communication and the described rule-based knowledge base that can revise described expert system software object,
Described optimizer software object communicate by letter with described fallout predictor software object and can revise described fallout predictor software object described selection criterion and
Described optimizer software object is communicated by letter with described adaptive model software object and the described method for establishing model that can revise described adaptive model software object.
9. target according to claim 8 is searched the intelligence software object, it is characterized in that, also comprise a plurality of described detector software objects, each described detecting device object can provide described stored data auxiliaryly on the basis of another described detector software object requests.
10. adaptive optimization software systems (adaptive optimization software system) comprising:
A plurality of targets are searched the intelligence software object, they reside among the computer usable medium and can access from this medium, can represent and influence a Process Control System, and the process that can be controlled by described Process Control System by a plurality of experience state representation and influence
Wherein
Search each experience state of intelligence software object for described target, described target is searched the intelligence software object and is adapted to described controlled process by setting up, select, learn, train and remembering one or more forecasting software models
Whereby
For each the described experience state that has more than the described process of a described forecasting software model, described target is searched a plurality of forecasting software models of intelligence software Object Selection in preceding training, come to set up model for described controlled process,
Described a plurality of described target is searched the intelligence software object and is searched the intelligence software object with other described target and communicate by letter relatively, with on one or more expression levels for described controlled process set up with representing model and
Described target search in the intelligence software object each can search in the intelligence software object external device communication of any with described target.
11. adaptive optimization software systems according to claim 10, it is characterized in that, also comprise a graphical user interface, wherein, described graphical user interface is used to according to the set of communicating by letter of a relation of being made up of the classification relationship of user's specified quantity and flow process relation, stipulate the described communication that concerns, whereby
Described classification relationship stipulate a plurality of described targets search priority between the intelligence software object and scale relationships and
Described flow process relation expression ground is corresponding to searching the data of communicating by letter between the intelligence software object in described a plurality of described targets.
12. adaptive optimization software systems according to claim 11 is characterized in that, concern that in highest ranked the described target of level searches the described priority of intelligence software object and scope based on economy.
13. adaptive optimization software systems according to claim 11 is characterized in that, concern that in lowest hierarchical the described target of level searches the described priority of intelligence software object and scope based on non-economy.
14. an optimizing process control system (process control optimization system) comprising:
Computer system, described computer system also comprises
One or more display device that comprise one or more viewing areas and
A plurality of terminal user's input medias;
Controlled process, it also comprises
External unit, described external unit also comprises
One or more controllable devices
Wherein
Described controllable device provides a plurality of interfaces, and described a plurality of interfaces are at described computing machine
Provide means of communication between system and the described controllable device; With
Be used to provide the instrument of data; With
The relevant optimization aim of one or more and described Process Control System; With
Described adaptive optimization software systems comprise that also a plurality of targets that can influence described controlled process search the intelligence software object
Wherein
Described target is searched the intelligence software object and is also comprised
One or more rule-based knowledge bases,
One or more method for establishing model,
One or more fallout predictors select criterions and
One or more optimization objects targets;
Search each experience state of intelligence software object for described target, described target is searched intelligence
Can software object by setting up, select, learn, train and to remember one or more predictions soft
The part model adapts to described controlled process; With
One or more targets are searched intelligence software object and one or more described external device communication
Whereby
To the described experience shape of each existence more than the described controlled process of a described forecast model
Attitude, described target are searched the forecasting software model of a plurality of training in advance of intelligence software Object Selection
So that set up model for described controlled process
Wherein
One or more described rule-based knowledge bases comprise a plurality of execution and described controlled mistakes of being used for
The rule of the described optimization aim of Cheng Xiangguan,
One or more described method for establishing model are used to produce the model of described controlled process,
One or more described fallout predictors are selected criterions, be used for describing with described controlled process mutually
The described optimization aim of closing and
One or more described optimization aim that are used for described controlled process comprise described process
The optimization of control system;
Described a plurality of described target is searched the intelligence software object and is searched intelligence with other described target relatively
The communication of energy software object; With
Search the intelligence software Object table with a described target of each of communicating by letter in the described external unit
Show corresponding to that described external unit.
15. optimizing process control system according to claim 14 is characterized in that, the described instrument that data are provided provides the data of actual real world equipment.
16. optimizing process control system according to claim 14 is characterized in that, the described instrument that data are provided provides the data of emulation real world equipment.
17. optimizing process control system according to claim 14, it is characterized in that, also comprise a graphical user interface, wherein, described graphical user interface is used for communicating by letter according to a relation of being made up of classification relationship and flow process relation and gathers stipulate arbitrary quantity described and concern communication
Whereby
Described classification relationship stipulates that described target searches priority and the scale relationships between the intelligence software object,
Described flow process relation expression ground corresponding to will by described target search the control of intelligence software object described process and
Described flow process relation represents that also ground is corresponding to search the data of communicating by letter between the intelligence software object in described target.
18. adaptive optimization method of optimizing the controlled process system, it is characterized in that, described optimizing process control system provides a plurality of targets to search the intelligence software object, and described target is searched the intelligence software object and also comprised the detector software object that is used to provide current data, historical data and statistics; Be used to provide the expert system software object of one or more dependency rule knowledge bases; Be used to provide the adaptive model software object of one or more method for establishing model; Be used to provide one or more fallout predictors to select the fallout predictor software object of criterion; Be used to provide the optimizer software object of one or more targets and purpose function and process constraint; And the communication translater software object that is used to provide the relevant data communication protocol of one or more and given sampling increment, described method comprises following parallel step;
Execution is by a process of described optimization controlled process system control;
Determine to optimize output data value in the described optimizer software object, this value can realize described target and purpose function best and not violate described process constraint;
Check the most suitable forecast model described in the described expert system software object, this model can be realized described target and purpose function and not violate described process constraint; With
Determine self-adaptation intervention suitable in the described expert system software object.
19. the adaptive optimization method of optimizing process control system according to claim 18 is characterized in that, described optimizing process control system is used to emulation to the useful real world process of training terminal user.
20. the adaptive optimization method of optimizing process control system according to claim 18 is characterized in that, the step of described definite optimization output data value also comprises the steps;
The described optimizer software object that has one or more input values and have described current data, historical data and statistics is provided,
Determine a plurality of trial values in the described optimizer software object, described trial value can best realize described target and purpose function and not violate described process constraint,
Estimate the described trial value in the described optimizer software object,
Described estimating step also comprises the steps
Determine the best test forecast model in the described fallout predictor software object, described best test is pre-
Survey model and be matched with described output data value most;
Use described trial value and the emulation of described best test forecast model, calculating and prediction and described
Be Controlled process described in the described adaptive model software object that the optimizer software object is correlated with
Future performance;
Determine whether best test forecast model can be realized described in the described optimizer software object
Described target and purpose function and don't violate the constraint of described process;
If described best test forecast model can be realized described target and purpose function and don't disobey
Described trial value is then preserved in anti-described process constraint in described optimizer software object; With
If described best test forecast model can not both realize described target and purpose function and don't
Violate described process constraint, then in described optimizer software object, attempt one group of new trial value.
21. the adaptive optimization method of optimizing process control system according to claim 18 is characterized in that, describedly determines that the step of suitable self-adaptation intervention also comprises the steps
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised and the relevant performance objective of described optimizer software object;
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised the configuration of described optimizer software object;
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised and the relevant performance objective of described adaptive model software object;
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised the configuration of described adaptive model software object;
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised and the relevant performance objective of described fallout predictor software object; With
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised the configuration of described fallout predictor software object.
Claims (21)
1. a target is searched intelligence software object (goal seeking intelligent software object), and it is applied to a Process Control System, can influence the process of being controlled by described Process Control System by a plurality of experience states,
It is characterized in that
Search the experience state of intelligence software object for each described target, described target is searched the intelligence software object and is adapted to described controlled process by setting up, select, learn, train and remembering one or more forecasting software models
Whereby
For each the described experience state that has more than the described process of a described forecasting software model, described target is searched the one or more described forecasting software model of intelligence software Object Selection training in advance, comes to set up model for described controlled process.
2. a target is searched the intelligence software object, and it comprises a plurality of in house software objects
It is characterized in that
Described a plurality of in house software object intercoms mutually,
Each described in house software object represents individual target and searches behavior,
Described in house software object represent alternately mutually comprehensive target search behavior and
The described individual target that one or more described in house software objects can be revised one or more described other in house software objects is searched behavior.
3. a target is searched the intelligence software object, comprising:
A plurality of in house software objects, described a plurality of in house software objects also comprise
At least one expert system software object,
At least one adaptive model software object,
At least one fallout predictor software object and
At least one optimizer software object,
Wherein
Described expert system software object is communicated by letter with described optimizer software object and the behavior that can revise described optimizer software object,
Described expert system software object is communicated by letter with described fallout predictor software object and the behavior that can revise described fallout predictor software object,
Described expert system software object is communicated by letter with described adaptive model software object and the behavior that can revise described adaptive model software object,
Described optimizer software object and described expert system software object communication and the behavior that can revise described expert system software object,
Described optimizer software object communicate by letter with described fallout predictor software object and can revise described fallout predictor software object behavior and
Described optimizer software object is communicated by letter with described adaptive model software object and the behavior that can revise described adaptive model software object,
Whereby
Described a plurality of in house software object manifests comprehensive target and searches behavior.
4. target according to claim 3 is searched the intelligence software object, it is characterized in that, described expert system software object also comprises a rule-based knowledge base, and this rule-based knowledge base comprises one or more rules of selecting from the set of distinct logic rules and fuzzy logic ordination.
5. target according to claim 3 is searched the intelligence software object, it is characterized in that, described optimizer software object also comprises a plurality of optimization methods, and wherein, described optimization method comprises one or more generic algorithms (genetic algorithm).
6. target according to claim 3 is searched the intelligence software object, it is characterized in that, also comprises a plurality of method for establishing model, and wherein, described method for establishing model comprises one or more neural networks.
7. target according to claim 3 is searched the intelligence software object, it is characterized in that described communication translater software object also comprises from comprising that internal agreement, described target search the communication protocol of selecting the set of the needed agreement of intelligence software object external unit, serial data communication agreement, parallel data communication agreement, LAN data communication protocol and wide area network data communication protocol.
8. a target is searched the intelligence software object, comprising:
At least one the expert system software object that comprises at least one rule-based knowledge base;
At least one adaptive model software object is used for producing one or more forecast models according to one or more input values with according to one or more method for establishing model;
At least one fallout predictor software object is used for using fallout predictor to select criterion and optimization method to select best one of described forecast model;
At least one optimizer software object is used to utilize one or more subject object and one or more optimization method to set up the optimization output data value that is used for described optimizer software object;
At least one communication translater software object is used to use one or more data communication protocols to carry out data communication; With
At least one detector software object, this detector software object can receive data from described expert system software object, described optimizer software object, described adaptive model software object, described fallout predictor software object, described communication translater software object and other detector software object; The described data of process are to store; Store described data; With on the basis of described expert system software object, described optimizer software object, described adaptive model software object, described fallout predictor software object, described communicate by letter translater software object and other detector software object requests, the data of being stored are offered described expert system software object, described optimizer software object, described adaptive model software object, described fallout predictor software object and the described translater software object of communicating by letter;
Whereby
Described expert system software object is communicated by letter with described optimizer software object and the described subject object that can revise described optimizer software object,
Described expert system software object is communicated by letter with described optimizer software object and the described optimization method that can revise described optimizer software object,
Described expert system software object is communicated by letter with described fallout predictor software object and the described fallout predictor that can revise described fallout predictor software object is selected criterion,
Described expert system software object is communicated by letter with described adaptive model software object and the described method for establishing model that can revise described adaptive model software object,
Described optimizer software object and described expert system software object communication and the described rule-based knowledge base that can revise described expert system software object,
Described optimizer software object communicate by letter with described fallout predictor software object and can revise described fallout predictor software object described selection criterion and
Described optimizer software object is communicated by letter with described adaptive model software object and the described method for establishing model that can revise described adaptive model software object.
9. target according to claim 8 is searched the intelligence software object, it is characterized in that, also comprise a plurality of described detector software objects, each described detecting device object can provide described stored data auxiliaryly on the basis of another described detector software object requests.
10. adaptive optimization software systems (adaptive optimization software system) comprising:
A plurality of targets are searched the intelligence software object, and they are applied to a Process Control System, can influence the process of being controlled by described Process Control System by a plurality of experience states
Wherein
Search each experience state of intelligence software object for described target, described target is searched the intelligence software object and is adapted to described controlled process by setting up, select, learn, train and remembering one or more forecasting software models
Whereby
For each the described experience state that has more than the described process of a described forecasting software model, described target is searched a plurality of forecasting software models of intelligence software Object Selection in preceding training, come to set up model for described controlled process,
Described a plurality of described target search that intelligence software object and other described target search that the intelligence software object is communicated by letter relatively and
Described target search in the intelligence software object each can search in the intelligence software object external device communication of any with described target.
11. adaptive optimization software systems according to claim 10, it is characterized in that, also comprise a graphical user interface, wherein, described graphical user interface is used to according to the set of communicating by letter of a relation of being made up of the classification relationship of user's specified quantity and flow process relation, stipulate the described communication that concerns, whereby
Described classification relationship stipulate a plurality of described targets search priority between the intelligence software object and scale relationships and
Described flow process relation expression ground is corresponding to searching the data of communicating by letter between the intelligence software object in described a plurality of described targets.
12. adaptive optimization software systems according to claim 11 is characterized in that, concern that in highest ranked the described target of level searches the described priority of intelligence software object and scope based on economy.
13. adaptive optimization software systems according to claim 11 is characterized in that, concern that in lowest hierarchical the described target of level searches the described priority of intelligence software object and scope based on non-economy.
14. an optimizing process control system (process control optimization system) comprising:
Computer system, described computer system also comprises
One or more display device that comprise one or more viewing areas and
A plurality of terminal user's input medias;
Controlled process, it also comprises
External unit, described external unit also comprises
One or more controllable devices
Wherein
Described controllable device provides a plurality of interfaces, and described a plurality of interfaces are at described computing machine
Provide means of communication between system and the described controllable device; With
Be used to provide the instrument of data; With
The relevant optimization aim of one or more and described Process Control System; With
Described adaptive optimization software systems comprise that also a plurality of targets that can influence described controlled process search the intelligence software object
Wherein
Described target is searched the intelligence software object and is also comprised
One or more rule-based knowledge bases,
One or more method for establishing model,
One or more fallout predictors select criterions and
One or more optimization objects targets;
Search each experience state of intelligence software object for described target, described target is searched intelligence
Can software object by setting up, select, learn, train and to remember one or more predictions soft
The part model adapts to described controlled process; With
One or more targets are searched intelligence software object and one or more described external device communication
Whereby
To the described experience shape of each existence more than the described controlled process of a described forecast model
Attitude, described target are searched the forecasting software model of a plurality of training in advance of intelligence software Object Selection
So that set up model for described controlled process
Wherein
One or more described rule-based knowledge bases comprise a plurality of execution and described controlled mistakes of being used for
The rule of the described optimization aim of Cheng Xiangguan,
One or more described method for establishing model are used to produce the model of described controlled process,
One or more described fallout predictors are selected criterions, be used for describing with described controlled process mutually
The described optimization aim of closing and
One or more described optimization aim that are used for described controlled process comprise described process
The optimization of control system;
Described a plurality of described target is searched the intelligence software object and is searched intelligence with other described target relatively
The communication of energy software object; With
Search the intelligence software Object table with a described target of each of communicating by letter in the described external unit
Show corresponding to that described external unit.
15. optimizing process control system according to claim 14 is characterized in that, the described instrument that data are provided provides the data of actual real world equipment.
16. optimizing process control system according to claim 14 is characterized in that, the described instrument that data are provided provides the data of emulation real world equipment.
17. optimizing process control system according to claim 14, it is characterized in that, also comprise a graphical user interface, wherein, described graphical user interface is used for communicating by letter according to a relation of being made up of classification relationship and flow process relation and gathers stipulate arbitrary quantity described and concern communication
Whereby
Described classification relationship stipulates that described target searches priority and the scale relationships between the intelligence software object,
Described flow process relation expression ground corresponding to will by described target search the control of intelligence software object described process and
Described flow process relation represents that also ground is corresponding to search the data of communicating by letter between the intelligence software object in described target.
18. adaptive optimization method of optimizing the controlled process system, it is characterized in that, described optimizing process control system provides a plurality of targets to search the intelligence software object, and described target is searched the intelligence software object and also comprised the detector software object that is used to provide current data, historical data and statistics; Be used to provide the expert system software object of one or more dependency rule knowledge bases; Be used to provide the adaptive model software object of one or more method for establishing model; Be used to provide one or more fallout predictors to select the fallout predictor software object of criterion; Be used to provide the optimizer software object of one or more targets and purpose function and process constraint; And the communication translater software object that is used to provide the relevant data communication protocol of one or more and given sampling increment, described method comprises following parallel step;
Execution is by a process of described optimization controlled process system control;
Determine to optimize output data value in the described optimizer software object, this value can realize described target and purpose function best and not violate described process constraint;
Check the most suitable forecast model described in the described expert system software object, this model can be realized described target and purpose function and not violate described process constraint; With
Determine self-adaptation intervention suitable in the described expert system software object.
19. the adaptive optimization method of optimizing process control system according to claim 18 is characterized in that, described optimizing process control system is used to emulation to the useful real world process of training terminal user.
20. the adaptive optimization method of optimizing process control system according to claim 18 is characterized in that, the step of described definite optimization output data value also comprises the steps;
The described optimizer software object that has one or more input values and have described current data, historical data and statistics is provided,
Determine a plurality of trial values in the described optimizer software object, described trial value can best realize described target and purpose function and not violate described process constraint,
Estimate the described trial value in the described optimizer software object,
Described estimating step also comprises the steps
Determine the best test forecast model in the described fallout predictor software object, described best test is pre-
Survey model and be matched with described output data value most;
Use described trial value and the emulation of described best test forecast model, calculating and prediction and described
Be Controlled process described in the described adaptive model software object that the optimizer software object is correlated with
Future performance;
Determine whether best test forecast model can be realized described in the described optimizer software object
Described target and purpose function and don't violate the constraint of described process;
If described best test forecast model can be realized described target and purpose function and don't disobey
Described trial value is then preserved in anti-described process constraint in described optimizer software object; With
If described best test forecast model can not both realize described target and purpose function and don't
Violate described process constraint, then in described optimizer software object, attempt one group of new trial value.
21. the adaptive optimization method of optimizing process control system according to claim 18 is characterized in that, describedly determines that the step of suitable self-adaptation intervention also comprises the steps
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised and the relevant performance objective of described optimizer software object;
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised the configuration of described optimizer software object;
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised and the relevant performance objective of described adaptive model software object;
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised the configuration of described adaptive model software object;
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised and the relevant performance objective of described fallout predictor software object; With
Utilize the one or more described rule-based knowledge base in the described expert system software object, self-adaptation is revised the configuration of described fallout predictor software object.
Applications Claiming Priority (5)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US3735597P | 1997-02-21 | 1997-02-21 | |
| US60/037,355 | 1997-02-21 | ||
| US6128597P | 1997-10-07 | 1997-10-07 | |
| US60/061,285 | 1997-10-07 | ||
| PCT/US1998/003356 WO1998037465A1 (en) | 1997-02-21 | 1998-02-20 | Adaptive objet-oriented optimization software system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1224515A true CN1224515A (en) | 1999-07-28 |
Family
ID=26714062
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN199898800513A Pending CN1224515A (en) | 1997-02-21 | 1998-02-20 | Adaptive object-oriented optimization software system |
Country Status (5)
| Country | Link |
|---|---|
| EP (1) | EP0898737A1 (en) |
| CN (1) | CN1224515A (en) |
| AU (1) | AU727326B2 (en) |
| CA (1) | CA2252512A1 (en) |
| WO (1) | WO1998037465A1 (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100409227C (en) * | 2002-10-16 | 2008-08-06 | 韩国电子通信研究院 | Method and system for transforming adaptively visual contents according to user's symptom characteristics of low vision impairment and user's presentation preferences |
| CN100454314C (en) * | 2003-12-24 | 2009-01-21 | 雅马哈发动机株式会社 | Multiobjective optimization apparatus, multiobjective optimization method and multiobjective optimization program |
| CN100543673C (en) * | 2007-07-26 | 2009-09-23 | 南京大学 | A kind of reflexion type architecture self-evolvement method based on body |
| CN102279563A (en) * | 2010-06-13 | 2011-12-14 | 天津理工大学 | Method for establishing parameterized kinematics model library for nonstandard digital equipment operator |
| CN103955368B (en) * | 2014-04-28 | 2017-08-04 | 中国人民解放军理工大学 | A kind of adaptive support system of software obfuscation and development approach |
| CN107075921A (en) * | 2014-08-28 | 2017-08-18 | 那伯斯工业公司 | The method and system verified for pipe fitting |
| CN108292437A (en) * | 2015-11-20 | 2018-07-17 | 索尼公司 | Image processing apparatus and method |
| CN110989540A (en) * | 2012-07-04 | 2020-04-10 | 诺尔斯海德公司 | Method for optimizing product characteristics and production costs of an industrial process |
| WO2022146416A1 (en) * | 2020-12-28 | 2022-07-07 | Landmark Graphics Corporation | Drilling data correction with machine learning and rules-based predictions |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1033637A3 (en) * | 1999-03-02 | 2001-05-02 | Yamaha Hatsudoki Kabushiki Kaisha | Method and apparatus for optimizing overall characteristic of device, using heuristic method |
| US6266619B1 (en) | 1999-07-20 | 2001-07-24 | Halliburton Energy Services, Inc. | System and method for real time reservoir management |
| US6853921B2 (en) | 1999-07-20 | 2005-02-08 | Halliburton Energy Services, Inc. | System and method for real time reservoir management |
| JP2001084237A (en) | 1999-09-14 | 2001-03-30 | Yamaha Motor Co Ltd | Real-time optimization method for controller provided with fuzzy inference system |
| US6434084B1 (en) | 1999-11-22 | 2002-08-13 | Halliburton Energy Services, Inc. | Adaptive acoustic channel equalizer & tuning method |
| JP2001152898A (en) | 1999-12-01 | 2001-06-05 | Sanshin Ind Co Ltd | Cruising characteristic control device |
| US6681387B1 (en) | 1999-12-01 | 2004-01-20 | Board Of Trustees Of The University Of Illinois | Method and apparatus for instruction execution hot spot detection and monitoring in a data processing unit |
| US6980940B1 (en) | 2000-02-22 | 2005-12-27 | Schlumberger Technology Corp. | Intergrated reservoir optimization |
| US6933856B2 (en) | 2001-08-02 | 2005-08-23 | Halliburton Energy Services, Inc. | Adaptive acoustic transmitter controller apparatus and method |
| US7765175B2 (en) | 2003-09-18 | 2010-07-27 | Optimum Power Technology, L.P. | Optimization expert system |
| US9032531B1 (en) | 2012-06-28 | 2015-05-12 | Middlegate, Inc. | Identification breach detection |
-
1998
- 1998-02-20 CN CN199898800513A patent/CN1224515A/en active Pending
- 1998-02-20 CA CA002252512A patent/CA2252512A1/en not_active Abandoned
- 1998-02-20 EP EP98908632A patent/EP0898737A1/en not_active Withdrawn
- 1998-02-20 WO PCT/US1998/003356 patent/WO1998037465A1/en not_active Application Discontinuation
- 1998-02-20 AU AU66619/98A patent/AU727326B2/en not_active Ceased
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100409227C (en) * | 2002-10-16 | 2008-08-06 | 韩国电子通信研究院 | Method and system for transforming adaptively visual contents according to user's symptom characteristics of low vision impairment and user's presentation preferences |
| CN100454314C (en) * | 2003-12-24 | 2009-01-21 | 雅马哈发动机株式会社 | Multiobjective optimization apparatus, multiobjective optimization method and multiobjective optimization program |
| CN100543673C (en) * | 2007-07-26 | 2009-09-23 | 南京大学 | A kind of reflexion type architecture self-evolvement method based on body |
| CN102279563A (en) * | 2010-06-13 | 2011-12-14 | 天津理工大学 | Method for establishing parameterized kinematics model library for nonstandard digital equipment operator |
| CN110989540A (en) * | 2012-07-04 | 2020-04-10 | 诺尔斯海德公司 | Method for optimizing product characteristics and production costs of an industrial process |
| CN103955368B (en) * | 2014-04-28 | 2017-08-04 | 中国人民解放军理工大学 | A kind of adaptive support system of software obfuscation and development approach |
| CN107075921A (en) * | 2014-08-28 | 2017-08-18 | 那伯斯工业公司 | The method and system verified for pipe fitting |
| US10107051B2 (en) | 2014-08-28 | 2018-10-23 | Nabors Industries, Inc. | Methods and systems for tubular validation |
| CN107075921B (en) * | 2014-08-28 | 2018-11-27 | 那伯斯工业公司 | Method and system for pipe fitting verifying |
| CN108292437A (en) * | 2015-11-20 | 2018-07-17 | 索尼公司 | Image processing apparatus and method |
| CN108292437B (en) * | 2015-11-20 | 2022-04-26 | 索尼公司 | Image processing apparatus and method |
| WO2022146416A1 (en) * | 2020-12-28 | 2022-07-07 | Landmark Graphics Corporation | Drilling data correction with machine learning and rules-based predictions |
| GB2612275A (en) * | 2020-12-28 | 2023-04-26 | Landmark Graphics Corp | Drilling data correction with machine learning and rules-based predictions |
Also Published As
| Publication number | Publication date |
|---|---|
| AU6661998A (en) | 1998-09-09 |
| EP0898737A1 (en) | 1999-03-03 |
| CA2252512A1 (en) | 1998-08-27 |
| WO1998037465A1 (en) | 1998-08-27 |
| AU727326B2 (en) | 2000-12-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1224515A (en) | Adaptive object-oriented optimization software system | |
| Nadimi-Shahraki et al. | A systematic review of the whale optimization algorithm: theoretical foundation, improvements, and hybridizations | |
| Slowik et al. | Nature inspired methods and their industry applications—Swarm intelligence algorithms | |
| Pourpanah et al. | A review of artificial fish swarm algorithms: Recent advances and applications | |
| Zhang et al. | Game theory based real‐time shop floor scheduling strategy and method for cloud manufacturing | |
| Parpinelli et al. | A review of techniques for online control of parameters in swarm intelligence and evolutionary computation algorithms | |
| Türkyılmaz et al. | A research survey: heuristic approaches for solving multi objective flexible job shop problems | |
| Agbaje et al. | Automatic data clustering using hybrid firefly particle swarm optimization algorithm | |
| Ciornei et al. | Hybrid ant colony-genetic algorithm (GAAPI) for global continuous optimization | |
| Felsberger et al. | A Review of Decision Support Systems for Manufacturing Systems. | |
| CN101064652B (en) | Service-to-device mapping for smart items using a genetic algorithm | |
| US9466023B1 (en) | Data mining technique with federated evolutionary coordination | |
| Ogundoyin et al. | Optimization techniques and applications in fog computing: An exhaustive survey | |
| US11663492B2 (en) | Alife machine learning system and method | |
| Rezaei et al. | GuASPSO: a new approach to hold a better exploration–exploitation balance in PSO algorithm | |
| KR20210067605A (en) | A method for controlling commercial laundry machine and system for the same using artificial intelligence | |
| Sumathi et al. | Computational intelligence paradigms for optimization problems using MATLAB®/SIMULINK® | |
| Kushwah et al. | A modified whale optimization algorithm to overcome delayed convergence in artificial neural networks | |
| Mishra et al. | Swarm intelligence optimization: editorial survey | |
| Xu et al. | Task allocation for unmanned aerial vehicles in mobile crowdsensing | |
| Zhou et al. | A chaotic parallel artificial fish swarm algorithm for water quality monitoring sensor networks 3D coverage optimization | |
| Singh et al. | A neighborhood search based cat swarm optimization algorithm for clustering problems | |
| Sundarakumar et al. | A heuristic approach to improve the data processing in big data using enhanced Salp Swarm algorithm (ESSA) and MK-means algorithm | |
| Liu et al. | Review of the grey wolf optimization algorithm: variants and applications | |
| Zhang et al. | Multi-species evolutionary algorithm for wireless visual sensor networks coverage optimization with changeable field of views |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C01 | Deemed withdrawal of patent application (patent law 1993) | ||
| WD01 | Invention patent application deemed withdrawn after publication |