CN101266628A - Automatic real-time emulation and its paralleling method based on emulated knowledge library - Google Patents
Automatic real-time emulation and its paralleling method based on emulated knowledge library Download PDFInfo
- Publication number
- CN101266628A CN101266628A CNA2008100664810A CN200810066481A CN101266628A CN 101266628 A CN101266628 A CN 101266628A CN A2008100664810 A CNA2008100664810 A CN A2008100664810A CN 200810066481 A CN200810066481 A CN 200810066481A CN 101266628 A CN101266628 A CN 101266628A
- Authority
- CN
- China
- Prior art keywords
- simulation
- preparation
- incident
- emulation
- real
- 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
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an automatic real time simulation based on a simulation knowledge library and a method of parallelizing the automatic real time simulation. The method comprises: collecting and finishing preparation events, and classifying and joining the preparation events into a preparation event library; collecting and finishing simulation rules of current simulation technology, and classifying and joining the simulation rules of the current simulation technology into a simulation library; calling the simulation rules to process simulation for the preparation events, and classifying and joining result into a simulation result library, and establishing a mapping relationship for the preparation events and corresponding expected results; when a reality event happening, searching matched preparation event and finding out the preparation result according to the mapping relationship, outputting the preparation result as a real time simulation result of the reality event; when the result is not ideal, joining the event as the preparation event into the preparation event library for the time or after the event, and extending the preparation event and the expected result of the preparation event into the simulation knowledge library so that next time similar event realizes the automatic real time simulation. The invention satisfies a requirement of high real-time and has a high precision, and is not easy to make wrong decision or conclusion.
Description
Technical field
The present invention relates to a kind of emulation technology and method thereof, relate in particular to a kind of emulation and parallel method thereof based on simulation knowledge base.
Technical background
Emulation technology is a multi-disciplinary integrated technology, and it is instrument based on kybernetics, systematology, the principle of similitude and infotech with computing machine and specialized equipment, utilizes system model that system reality or imagination is carried out dynamic test.
Modeling and simulation is the human effective ways of handling practical problems, and it and human history exist simultaneously.People always use " spiritual model " to go to understand better reality, do plan, go to consider various possibilities, go to exchange thought with other people, go to work out the action plan of some idea, or go to confirm some irrealizable idea.
Even before several thousand, when people make boats and ships and plant equipment, also be to test with the model of little boats and ships or plant equipment earlier.Children's toy always be unable to do without the emulation of real world, and these toys are the model of tool under people, animal, object and the traffic normally.The model of said here boats and ships, plant equipment, toy for children, promptly so-called physics mould plough.Physical model be with by the material object of simulation object geometric similarity.Mathematical model is exactly that the mathematics of physical model is described.
Spirit model, physical model, mathematical model have been summarized all models in the emulation.The spirit model only is a thought process, and to a great extent and imprecision, it can only provide a rough qualitative conclusion.When doing simulation study, always to become physical model or mathematical model to spiritual model conversation, promptly usefulness material object such as the ship of imagining in the brain, truck or math equation are expressed.This promptly is the modeling process that will study.Utilize physical model to carry out emulation, be called physical simulation.The theoretical foundation of physical simulation is similarity theory, and its necessary condition is a geometric similarity.And, also to satisfy the condition that each relevant similarity criterion number equates for dynamic process.Utilize mathematical model to carry out emulation, be called mathematical simulation.Mathematical simulation is found the solution this mathematical model in fact exactly.If find the solution, just be called Computer Simulation with computing machine.
Present many application programs are all utilized knowledge, and what wherein have has also reached very high level, and still, these application programs may not be KBS Knowledge Based System, and they do not have knowledge base yet.General application program be based on the difference between the knowledge system: generally application program impliedly is coded in the knowledge of problem solving in the program, KBS Knowledge Based System is then expressed the problem solving knowledge of application with showing, and forms a relatively independent program entity.
Knowledge base has following several characteristics:
1, the knowledge in the knowledge base is configured the organizational form of being convenient to utilize, structure is arranged according to their application feature, background characteristics (background information when obtaining), use characteristic, attributive character etc.The knowledge sheet generally is modular.
2, the knowledge of knowledge base is stratified, and lowermost layer is a fact knowledge; The middle layer is to be used for controlling true knowledge (representing with rule, process etc. usually); Highest level is a strategy, and it is a controlling object with middle layer knowledge.Strategy also usually is considered to the rule of rule, and therefore, the basic structure of knowledge base is hierarchical structure, is determined by its knowledge self character.In knowledge base, all there is relation of interdependence between the knowledge sheet usually.Rule is typical, the most the most frequently used a kind of knowledge sheet.
3, a kind of knowledge that not only belongs to the special shape of a certain level (level in office in other words all exists) can be arranged in the knowledge base---confidence level (or claiming degree of belief, confidence measure etc.).To a certain problem, relevant facts, rule and strategy all can be marked with confidence level, like this, have just formed the augmentation knowledge base.In database, there is not uncertainty measure, because all belong to definite type in database processing.
4, also can there be a special part that is commonly referred to as the typical method storehouse in the knowledge base.If the solution route for some problem is to affirm with inevitable, just can directly be stored in it in typical method storehouse as the quite sure problem solution route of a part.The storage of this macroscopic view will constitute another part of knowledge base.When using this part, machine inference will be only limited to certain one deck body portion of selecting for use in the typical method storehouse.
In addition, knowledge base also can realize on distributed network.Like this, just need to build distributed knowledge base.The superiority of building distributed knowledge base has 3 points:
1, can under lower price, construct bigger knowledge base;
2, the problem-solving task of the knowledge base correspondence of different levels or different field is relatively more simple comparatively speaking, thereby can constitute the system of more efficient;
3, can be suitable for geographic distribution vast in territory.
The structure of knowledge base must make knowledge access and search effectively in the process that is used wherein, and the knowledge in the storehouse can be revised and edit easily, simultaneously, the consistance and the complete performance of knowledge in the storehouse is tested.
At present existing concurrent technique, walking abreast has two kinds of implications: the one, simultaneity refers to that two or more incidents take place at the same instant; The 2nd, concurrency refers to the interior at interval at one time generation of two or more incidents.Parallel computation is a kind of effective means that improves computer system computing velocity and processing power.Its basic thought is to work in coordination with the same problem of finding the solution with a plurality of processors, and the PROBLEM DECOMPOSITION that is about to be found the solution becomes several parts, each several part by one independently processor come parallel computation.Concurrent computational system both can be a supercomputer custom-designed, that contain a plurality of processors, also can be the cluster that the some platform independent computing machines that interconnect in some way constitute.
Parallel computation is based on a simple idea: N platform computing machine should be able to provide N times of computing power, finishes in the time of 1/N no matter the speed of current computer how, can be expected the problem of being found the solution.Obviously, this is an ideal situation, because the problem of being found the solution all can not be broken down into fully independently various piece under normal conditions, but need carry out data necessary exchange and synchronous.
However, parallel computation still can make whole performance of computer systems obtain substantial improvement, and improved degree depends on the degree of concurrence of desiring the problem of finding the solution self.
The advantage of parallel computation is to have the googol value to calculate and data-handling capacity, can be widely used in the key subjects that have profound influence in national economy, national defense construction and the development in science and technology, as petroleum prospecting, earthquake prediction and forecast, climatic simulation and weather forecast on a large scale, new weapon design, nuclear weapon systematic research simulation, aerospace flight vehicle, satellite image processing, celestial body and geoscience, real-time Film Animation system and virtual reality system or the like.
Because these systems are very complicated, and in large scale, directly carry out relatively difficulty of entity research, the expense of design and research is also very high, and risk is very big simultaneously.And emulation technology has characteristics such as low-risk and efficientibility, therefore set up and implement this type systematic before carry out the simulation study particular importance that just seems earlier.
The software of emulation is a lot, as ANSYS, and Femlab, Fluent.The existing serial version of some simulation software has parallel version again, as ANSYS.
Because the parallelization in these analogue systems is handled a lot, if adopt serialization emulation, will seriously hinder the simulated effect and the simulation time of system, influence the efficient of emulation, do not reach the requirement of real-time simulation, lost the purpose and the meaning of emulation.And the appearance along with low-cost parallel computer architecture and express network computing platform makes parallel artificial become possibility.The appearance of parallel artificial software is the precondition that parallel simulation system is realized, an outstanding parallel artificial software should have following three features:
1, can realize easily that but mathematical model arrives the conversion of operational simulation software model;
2, can the simultaneous adaptation distributed frame and the parallel computing platform of shared drive, and support multiple parallel artificial tasks allocation strategy and synchronization policy;
3, support the model development of visual programming and stratification;
Data sharing and message transmission are main two kinds of structures that present parallel artificial language is adopted, and wherein message delivery method also has the minority language to adopt the data sharing mode because its adaptability becomes most of language employings by force.
Parallel artificial language commonly used now mainly contains following three classes:
1, parallel storehouse of exploitation and api interface at first will be developed the parallel artificial module library, call the module in the parallel storehouse then in the serial emulational language of standard.This method need not be learnt new language, and the user is easy to grasp, still, owing to do not have specific compiler, bug check ability not to have high-level emulational language strong, and so the parallel artificial modular design must be tried one's best simply.More representational this speech like sound has GTW, UPS.Compose etc.
2, adding the parallel processing function in the serial emulational language, mainly is in order to strengthen the ability of serial emulational language.Compare with the parallel artificial module library, this method has had compiler, for the programming personnel provides relatively more friendly interface and working environment easily.And in this language, can use multiple optimization method easily, reduce backoff interval etc. as control of the granularity of Apostle and Maisie.More representational this speech like sound has Sim++, Apostle.Maisie, Parsec etc.
3, add copying in the parallel language, SCE for example, it grows up on parallel language Ada.
Accident generally needs in time to handle, and this just needs emergent treatment system to handle.Need in the emergency command real-time and effective emulation is carried out in the scene of the accident, otherwise be difficult to control serious loss and accident.Science needs problem in science is carried out real-time and effective emulation in calculating, otherwise can influence the progress of scientific research.Need in the industry product is carried out real-time and effective emulation, otherwise can influence speed of production and quality.
When real incident takes place, analogue system will be carried out emulation to incident.When any system is carried out emulation, at first to try to achieve its mathematical model.Ask the method for mathematical model that following two kinds of diverse methods are arranged usually in the emulation technology at present, i.e. black box method and " white box " method.
So-called black box method is that a system is added different input (disturbance) signals, observes its output.According to the input of being write down, output signal, with input and the output relation that one or several mathematic(al) representation is expressed this system, this method is not gone the mechanism and the function of descriptive system inside.For the system of thermal power plant, be " curve of ascending to heaven test " method and " System Discrimination " method through black box method commonly used.
Ask the mathematical model of a system with " white box " method, need know many details of system itself, such as this system by several sections form, how to connect between them, how they influence each other etc.This method is not paid attention to the observation to system's behavior in the past, only pays attention to the description of system architecture and process.The mechanism of system has been had after the detailed understanding, just may obtain describing the mathematical model of this system.For the system of thermal power plant, used " white box " method is according to the principle modeling of " energy conservation ", " mass conservation " and " momentum conservation ".
The process of asking system mathematic model is called a modeling, obtain just can carrying out simulation process behind the model of system.Because the mathematic(al) representation of resulting descriptive system is generally forms such as the differential equation, partial differential equation, algebraic equation and difference equation, and the algorithmic language great majority of digital machine can not directly be found the solution the differential equation and partial differential equation, before to these model solutions, must convert them to can describe form with ALGOL.This transfer process is referred to as the secondary modeling.The secondary modeling uses a computer and finds the solution.
Fast development along with digital machine, some advanced algorithm language can directly be found the solution the differential equation, even partial differential equation (for example: the MATLAB language), also have the language that grows up for emulation specially, as emulational languages such as ESL-A, CSSL ' S, SIMULINKV31, CAE2000.When using these language that system is carried out Digital Simulation, the secondary modeling process does not just need.
The simulation calculation amount is very big, and is very consuming time, in order to accelerate simulation speed, adopts parallel computation and simplified model to accelerate simulation speed usually.
Prior art all is just to carry out emulation when demand is arranged, emulation, particularly large-scale emulation, very huge of calculated amount, even utilize high-performance computer also to be difficult in the time of expectation, finish, for example simulation of building blast, the urgent simulation of revealing of pollutant, generally all be difficult to when incident takes place, finish, if do not finish when incident takes place, the meaning of then carrying out emulation will reduce greatly, can only be used for the reason that the ex-post analysis simulated accident takes place.And these emergencies, be difficult to do prediction scheme with emulation, because the place that accident takes place is unpredictable in advance, such as: whoever can not expect that the Pentagon can explode in 911 incidents, and who can not notify in advance and adopts what method to explode, can not predict more aircraft will from which angle which highly clashes into, so this class prediction scheme can't be carried out.So existing emulation technology does not satisfy the demanding demands of applications of this class real-time.
Use existing emulation technology simultaneously, scientist need wait for the result that just can obtain scientific simulations for a long time, has delayed the process of scientific research greatly.The project planner need wait for the result that just can obtain industrial emulation for a long time, has delayed the process of research and development of products greatly.
Though emulation technology can be calculated by integrating parallel at present, but the division of parallel granularity and the concurrency that parallel acceleration depends on application, and be subject to the expense of communicating by letter between different concurrent processes, though quickened simulation speed very widely, still far away from the demand of real-time.By 1 hour emulation of Huaihe River Flooding, if do not adopt parallel computation, may need 30 days time of emulation, adopt parallel computation, still need 2 days time of emulation, still lag behind the flood accident progress of reality far away.
Simplify the method for realistic model,, reduced simulation accuracy, often decision-making or conclusion because precision makes the mistake inadequately though accelerate simulation velocity to a certain extent.
Therefore, existing emulation technology awaits improving and development.
Summary of the invention
The purpose of this invention is to provide a kind of automatic real-time emulation and parallel method thereof based on simulation knowledge base, technical matters to be solved is at above-mentioned existing emulation technology defective, by simulation knowledge base storage preparation incident and the existing emulation technology of utilization in advance the preparation incident is carried out the expection simulation result that emulation obtains.
Technical scheme of the present invention is as follows:
A kind of automatic real-time emulation and parallel method thereof based on simulation knowledge base, described simulation knowledge base comprises: preparation event base, expection simulation result storehouse, mapping table, emulation rule base; Its method may further comprise the steps:
A, compile the preparation incident, sort out and it is added the preparation event base;
B, compile emulation rule in the existing emulation technology, sort out and it is added the emulation rule base;
C, call described emulation rule the preparation incident is carried out emulation, simulation result is sorted out added expection simulation result storehouse, and its corresponding expected results of described preparation incident is set up mapping relations;
When D, the generation of real incident, the preparation incident of match retrieval according to described mapping relations, finds expected results, and with its real-time simulation result output as described real incident.
Described method wherein, also comprises step:
E, when described real-time simulation result undesirable, then described real incident is added described preparation event base as the preparation incident, and described preparation incident and expected results thereof extended in the described simulation knowledge base, run into similar real incident next time and realize automatic real-time emulation to make things convenient for.
Described method wherein, in the described steps A, is done corresponding renewal to its corresponding simulation result and their mapping relations when the preparation event base upgrades.
Described method wherein, among the described step C, is done corresponding renewal to the pairing incident of this result and their mapping relations when the simulation result storehouse is upgraded.
Described method, wherein, among the described step D, emulation transfers nonreal time simulation to automatic real-time emulation based on simulation knowledge base.
Described method, wherein, the structure of described simulation knowledge base, renewal, and incident emulation parallelization carry out.
A kind of automatic real-time emulation and parallel method thereof provided by the present invention based on simulation knowledge base, owing to adopted by simulation knowledge base storage preparation incident and the existing emulation technology of utilization in advance the preparation incident carried out the expection simulation result output that emulation obtains, overcome the defective in the existing emulation technology, satisfied the demanding application demand of real-time, and simulation accuracy height of the present invention is difficult for causing decision-making or the conclusion that makes mistake according to the simulation result work.
Description of drawings
Fig. 1 is the automatic real-time emulation schematic diagram based on simulation knowledge base of the present invention;
Fig. 2 is a simulation knowledge base structural drawing of the present invention;
Fig. 3 is the overall plan synoptic diagram of the automatic real-time emulation based on simulation knowledge base of the present invention;
Fig. 4 is the overall parallel synoptic diagram of the automatic real-time emulation based on simulation knowledge base of the present invention;
Fig. 5 is the structure and renewal overall plan synoptic diagram of simulation knowledge base of the present invention;
Fig. 6 is the structure of simulation knowledge base of the present invention and the overall parallel process flow diagram of renewal;
Fig. 7 is the automatic real-time emulation overall plan schematic flow sheet based on simulation knowledge base of the present invention;
Fig. 8 is the totally parallel process flow diagram of the automatic real-time emulation based on simulation knowledge base of the present invention.
Embodiment
Below in conjunction with accompanying drawing, will make a more detailed description each preferred embodiment of the present invention.
The present invention is based on the automatic real-time emulation of simulation knowledge base and the core inventive point of parallel method thereof and be by simulation knowledge base storage preparation incident and the existing emulation technology of utilization in advance the preparation incident to be carried out emulation, the expection simulation result that obtains is exported as the automatic real-time emulation result.When real incident takes place, from simulation knowledge base, find the preparation incident that is complementary with real incident, when expection simulation result during with real incident simulation result coupling, then its expection simulation result is exported as real incident simulation result, as shown in Figure 1.
The present invention at first makes up simulation knowledge base, and simulation knowledge base comprises: preparation event base, expection simulation result storehouse, mapping table, emulation rule base, as shown in Figure 2.
Wherein prepare event base comprise the 1st class preparation event base, the 2nd class preparation event base ... M1 class preparation event base, the 1st class preparation event base comprise 1_1 class preparation event base, 1_2 class preparation event base ... 1_N prepares event base, so analogize, can constantly expand and refinement as required.
Wherein expect the simulation result storehouse comprise the 1st class expection simulation result storehouse, the 2nd class expection simulation result storehouse ... M class expection simulation result storehouse.The 1st class expection simulation result storehouse comprise 1_1 class expection simulation result storehouse, 1_2 class expection simulation result storehouse ... 1_N class expection simulation result storehouse, so analogize, can constantly expand and refinement as required.
Wherein mapping table comprise the 1st class mapping table, the 2nd class mapping table ... M class mapping table, the 1st class mapping table comprise 1_1 class mapping table, 1_2 class mapping table ... 1_N class mapping table, so analogize, can constantly expand and refinement as required.
Wherein the emulation rule base comprise the 1st class emulation rule base, the 2nd class emulation rule base ... M class emulation rule base, the 1st class emulation rule base comprise 1_1 class emulation rule base, 1_2 class emulation rule base ... 1_N class emulation rule base, so analogize, can constantly expand and refinement as required.
After simulation knowledge base made up, it can upgrade, and can utilize simulation knowledge base that real incident is carried out automatic real-time emulation as shown in Figure 3 simultaneously.
The renewal of simulation knowledge base be asynchronous, loose coupling based on simulation knowledge base to the automatic real-time emulation of real incident.When promptly not carrying out based on simulation knowledge base the automatic real-time emulation of real incident, simulation knowledge base can upgrade; When real incident was carried out automatic real-time emulation, simulation knowledge base also can upgrade; When simulation knowledge base upgrades, can carry out automatic real-time emulation to real incident; When simulation knowledge base upgrades, also can carry out automatic real-time emulation to real incident.Simultaneously simulation knowledge base is mutually related again with the automatic real-time emulation that carries out to real incident.The renewal that is simulation knowledge base can cause the variation in preparation event base and expection simulation result storehouse, and then can cause the variation to the automatic real-time emulation result of the real incident relevant with renewal; Simultaneously based on simulation knowledge base to the automatic real-time emulation result of real incident when undesirable, can automatically real incident be added the preparation event base as the preparation incident, and generate corresponding expection simulation result by emulation rule in the simulation knowledge base and add expection simulation result storehouse, and will prepare incident and expected results mapping relations adding mapping table, thereby reach the purpose of simulation knowledge base self study, self-perfection.Its flow process is as follows:
The A step: make up simulation knowledge base;
The B step: as required, simulation knowledge base is upgraded;
The C step:, real incident is carried out automatic real-time emulation based on simulation knowledge base.
After analogue system was provided with the parallelization operation, simulation knowledge base can make up concurrently, and behind the structure, simulation knowledge base can upgrade concurrently, can utilize simulation knowledge base that real incident is carried out automatic real-time emulation concurrently simultaneously.The overall of automatic real-time emulation based on simulation knowledge base walks abreast flow process as shown in Figure 4, and its operating process is as follows:
The A step: distribute all structure tasks, make up tasks in parallel ground according to each and make up simulation knowledge base;
The B1 step: when needs upgrade simulation knowledge base, distribute all and upgrade the simulation knowledge base task, concurrently simulation knowledge base is upgraded;
The B2 step: when needs carry out automatic real-time emulation to real incident, distribute all real incident automatic real-time emulation tasks, concurrently real incident is carried out automatic real-time emulation.
Simulation knowledge base can make up and upgrades in the time of automatic real-time emulation reality incident, and simulation knowledge base is divided into preparation event base, expection simulation result storehouse, preparation incident and mapping table, the emulation rule base of expecting between the simulation result.When making up simulation knowledge base, at first make up library structure, add data then in the storehouse, it comprises: preparation incident, emulation rule, expection simulation result, preparation incident and expection simulation result mapping relations.The preparation incident reflects preparation event base and expection simulation result storehouse mapping relations with the mapping table of expection simulation result, stores the result that the emulation rule is carried out emulation in the preparation course of event emulation rule base in the preparation event base in the expection simulation result storehouse.After simulation knowledge base makes up and finishes, at any time can upgrade wherein library structure or database data, but note the completeness and the consistance of data when revising, note the relation between four storehouses, for example in the preparation event base, add a preparation incident, must call in the emulation rule base emulation rule this preparation incident is carried out emulation, and in expection simulation result storehouse, add the expection simulation result of this preparation incident correspondence, in mapping table, add this preparation incident simultaneously and expect mapping (enum) data between the simulation result.The structure of simulation knowledge base with upgrade overall plan as shown in Figure 5, its concrete steps are as follows:
The A step: create library structure;
The B step: parallelization is upgraded;
The B step comprises following parallelization step:
The B1 step: when structure need be upgraded, then upgrade library structure;
The B2 step: when needs collection possibility incident, or need in the preparation event base, increase incident the time, the preparation event base is sorted out and added to the preparation incident that increases, obtain up-to-date preparation event base, utilize the emulation rule that the preparation incident that increases is carried out emulation then, obtain increasing preparation incident simulation result, then its simulation result is sorted out and is added expection simulation result storehouse, obtain up-to-date expection simulation result storehouse, and, obtain up-to-date mapping relations table increasing foundation mapping between preparation incident and its simulation result;
The B3 step: when needs upgrade the emulation rule base, then upgrade the emulation rule base;
B4 step: when needs are revised in the preparation event base incident, in the preparation event base, former preparation incident is replaced with amended preparation incident, obtain up-to-date preparation event base, utilize in the emulation rule base emulation rule that amended preparation incident is carried out emulation then, obtain amended preparation incident simulation result, again amended preparation incident simulation result is sorted out and is added expection simulation result storehouse, obtain up-to-date expection simulation result storehouse, and between amended preparation incident and its simulation result, set up mapping, obtain up-to-date mapping relations table;
B5 step: when needs from the preparation event base during deletion event, this incident of deletion in the preparation event base, obtain up-to-date preparation event base, from expection simulation result storehouse, delete the expection simulation result of this preparation incident correspondence then, obtain up-to-date expection simulation result storehouse, delete this preparation incident and its expected results mapping relations again, obtain up-to-date mapping relations table.
After the analogue system parallelization is provided with, make up simulation knowledge base at first concurrently, upgrade library structure or database data then concurrently.The overall of the structure of simulation knowledge base and renewal walks abreast flow process as shown in Figure 6, and its idiographic flow step is as follows:
The A step: make up library structure as required;
The B step: accept and distribute all simulation knowledge bases to make up and updating task, concurrently each task is handled.Its treatment step is: B1 step, B2 step, B3 step, B4 step, the 5th step;
The B1 step: when needs upgrade library structure, then upgrade library structure;
B2 step: when the need collection may incident or needed to increase incident in the preparation event base, distribute the preparation incident that all need increase;
The B2a step: the preparation event base is sorted out and added to the preparation incident that will increase, and obtains up-to-date preparation event base.
B2b step: utilize in the emulation rule base emulation rule that the preparation incident that increases is carried out emulation, obtain increasing preparation incident simulation result;
The B2c step: the simulation result of the preparation incident that will increase is sorted out and is added expection simulation result storehouse, obtains up-to-date expection simulation result storehouse;
The B2d step: between the preparation incident that increases and its simulation result, set up mapping, obtain up-to-date mapping relations table;
The B3 step: when needs upgrade the emulation rule base, then upgrade the emulation rule base concurrently;
B4 step: when needs are revised in the preparation event base incident, distribute the preparation incident that all need be revised;
The B4a step: in the preparation event base, former preparation incident is replaced with amended preparation incident, obtain up-to-date preparation event base;
B4b step: utilize in the emulation rule base emulation rule that amended preparation incident is carried out emulation, obtain amended preparation incident simulation result;
The B4c step: amended preparation incident simulation result is sorted out and added expection simulation result storehouse, obtain up-to-date expection simulation result storehouse;
The B4d step:, obtain up-to-date mapping relations table with setting up mapping between amended preparation incident and its simulation result;
B5 step:, distribute the preparation incident that all need be deleted when in the needs deletion preparation event base during incident;
The B5a step: this preparation incident of deletion in the preparation event base obtains up-to-date preparation event base;
The B5b step: the expection simulation result of this preparation incident correspondence of deletion from expection simulation result storehouse obtains up-to-date expection simulation result storehouse;
B53 step: delete linking between this preparation incident and its expected results, obtain up-to-date mapping relations table.
When the present invention has new real incident to carry out emulation, at first retrieval and the preparation incident of mating in the preparation event base in simulation knowledge base, the preparation event base has a lot of word banks, word bank can have word bank (multistage word bank) again, therefore, relatively the time, need to determine match retrieval in which preparation incident word bank, according to mapping table the pairing expection simulation result of the preparation incident of matching degree maximum is returned then, as real incident automatic real-time emulation result according to real event property.If it is undesirable that this result is judged as, then real incident is added the expected event storehouse as expected event automatically, and utilize the emulation rule that this expected event is carried out emulation, and with its simulation result adding expection simulation result storehouse, simultaneously should preparation incident and this expected results mapping relations adding mapping table.Based on the automatic real-time emulation overall plan of simulation knowledge base as shown in Figure 7, its concrete steps are as follows:
The A step: accept real incident;
The B step: in the preparation event base, retrieve preparation incident with real event matches;
C step: when matching degree meets the requirements, the expection simulation result that then should the preparation incident be shone upon is exported as real incident simulation result;
The D step: undesirable when matching degree, real incident is increased in the simulation knowledge base, and therefore starts the renewal engine of simulation knowledge base, obtain real incident nonreal time simulation result.
The expection simulation result of after parallelization is provided with, retrieving, mate, shine upon, utilize emulation rule emulation preparation incident to generate can the parallelization operation.Based on the totally parallel flow process of the automatic simulation of simulation knowledge base as shown in Figure 8, concrete steps are as follows:
The A step: accept real incident, distribute all real incidents, concurrently each real incident is handled, treatment step is A1 step, A2 step;
The A1 step: will be distributed to all relevant preparation incident word banks to the match retrieval task of present reality incident, concurrently each relevant preparation incident word bank is for further processing, obtains preparing in the event base and prepare incident and matching degree thereof with the present reality event matches;
The A1a step: distribute all preparation incidents in the current preparation incident word bank, concurrently each preparation incident is for further processing, obtain preparing incident and matching degree thereof with the present reality event matches in the current preparation incident word bank;
The A1a1 step: calculate real incident and preparation event matches degree, obtain present reality incident and preparation event matches degree;
A2a step: when matching degree meets the requirements, the expection simulation result that then should the preparation incident be shone upon is exported as real incident simulation result;
The A2b step: undesirable when matching degree, then real incident is increased in the simulation knowledge base, and therefore starts the renewal engine of simulation knowledge base, obtain the nonreal time simulation result of real incident.
Above embodiment has used simulation knowledge base, makes the required time of emulation not rely on simulation scale and complexity, has accelerated simulation velocity greatly, makes emulation and reality synchronous, and is ahead of reality.Even and existing emulation technology adopts simplified model and parallel calculating method, can not reach live effect, generally be used for afterwards assessment and causality analysis, lost the most crucial value of emulation.Above embodiment utilizes simulation knowledge base when real incident does not take place, and is simulated the expection simulation result and is stored in the expection simulation result storehouse by the preparation incident, carries out real-time simulation for real incident and uses.Emulation rule in the above embodiment simulation knowledge base is existing emulation technology, has therefore both exceeded existing emulation technology, has made full use of existing emulation technology again.
From above embodiment as can be seen, the present invention is operated at ordinary times, the preparation incident that simulation knowledge base may take place in the future by emulation when real incident does not take place, and will prepare incident, preparation event base with and the regular expanding and updating of emulation in simulation knowledge base.And existing emulation technology only when real incident takes place, is busy with emulation, and the artificial resource that is useful on is in idle condition when not taking place, so the present invention utilizes resource more fully.
Should be understood that the description of above-mentioned specific embodiment is comparatively detailed, can not therefore be interpreted as the restriction to scope of patent protection of the present invention, scope of patent protection of the present invention should be as the criterion with claims.
Claims (6)
1, a kind of automatic real-time emulation and parallel method thereof based on simulation knowledge base, described simulation knowledge base comprises: preparation event base, expection simulation result storehouse, mapping table, emulation rule base; Its method may further comprise the steps:
A, compile the preparation incident, sort out and it is added the preparation event base;
B, compile emulation rule in the existing emulation technology, sort out and it is added the emulation rule base;
C, call described emulation rule the preparation incident is carried out emulation, simulation result is sorted out added expection simulation result storehouse, and its corresponding expected results of described preparation incident is set up mapping relations;
When D, the generation of real incident, the preparation incident of match retrieval according to described mapping relations, finds expected results, and with its real-time simulation result output as described real incident.
2, method according to claim 1 is characterized in that, also comprises step:
E, when described real-time simulation result undesirable, then described real incident is added described preparation event base as the preparation incident, and described preparation incident and expected results thereof extended in the described simulation knowledge base, run into similar real incident next time and realize automatic real-time emulation to make things convenient for.
3, method according to claim 1 is characterized in that, in the described steps A, when the preparation event base upgrades its corresponding simulation result and their mapping relations is done corresponding renewal.
4, method according to claim 1 is characterized in that, among the described step C, when the simulation result storehouse is upgraded the pairing incident of this result and their mapping relations is done corresponding renewal.
5, method according to claim 1 is characterized in that, among the described step D, emulation transfers nonreal time simulation to automatic real-time emulation based on simulation knowledge base.
6, method according to claim 1 is characterized in that, the structure of described simulation knowledge base, renewal, and incident emulation parallelization carry out.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2008100664810A CN101266628A (en) | 2008-04-02 | 2008-04-02 | Automatic real-time emulation and its paralleling method based on emulated knowledge library |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2008100664810A CN101266628A (en) | 2008-04-02 | 2008-04-02 | Automatic real-time emulation and its paralleling method based on emulated knowledge library |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101266628A true CN101266628A (en) | 2008-09-17 |
Family
ID=39989042
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2008100664810A Pending CN101266628A (en) | 2008-04-02 | 2008-04-02 | Automatic real-time emulation and its paralleling method based on emulated knowledge library |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101266628A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102004827A (en) * | 2010-11-08 | 2011-04-06 | 大连理工大学 | Design method of inter-satellite communication system of formation small satellites based on Petri net |
| CN102289496A (en) * | 2011-08-22 | 2011-12-21 | 北京航空航天大学 | Wireless cognitive network knowledge base constructing method based on Bayesian network |
| CN102855278A (en) * | 2012-07-23 | 2013-01-02 | 电信科学技术研究院 | Simulation method and system |
| CN103324704A (en) * | 2013-06-17 | 2013-09-25 | 深圳先进技术研究院 | Method and system for dynamically updating knowledge base |
| CN108389104A (en) * | 2018-01-31 | 2018-08-10 | 口碑(上海)信息技术有限公司 | A kind of emulation verification method and device of network activity |
| CN108647196A (en) * | 2018-04-16 | 2018-10-12 | 北京航空航天大学 | The artificial intelligence generation method and device of table in digital aircraft simulation report |
| CN109212999A (en) * | 2018-07-20 | 2019-01-15 | 北京航空航天大学 | The intelligent generation method and system of digital satellite emulation operating condition |
| CN116150885A (en) * | 2023-02-14 | 2023-05-23 | 南京友一智能科技有限公司 | Multidisciplinary integration and simulation data management system |
-
2008
- 2008-04-02 CN CNA2008100664810A patent/CN101266628A/en active Pending
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102004827A (en) * | 2010-11-08 | 2011-04-06 | 大连理工大学 | Design method of inter-satellite communication system of formation small satellites based on Petri net |
| CN102289496A (en) * | 2011-08-22 | 2011-12-21 | 北京航空航天大学 | Wireless cognitive network knowledge base constructing method based on Bayesian network |
| CN102855278A (en) * | 2012-07-23 | 2013-01-02 | 电信科学技术研究院 | Simulation method and system |
| CN102855278B (en) * | 2012-07-23 | 2016-03-09 | 电信科学技术研究院 | A kind of emulation mode and system |
| CN103324704A (en) * | 2013-06-17 | 2013-09-25 | 深圳先进技术研究院 | Method and system for dynamically updating knowledge base |
| CN108389104A (en) * | 2018-01-31 | 2018-08-10 | 口碑(上海)信息技术有限公司 | A kind of emulation verification method and device of network activity |
| CN108647196A (en) * | 2018-04-16 | 2018-10-12 | 北京航空航天大学 | The artificial intelligence generation method and device of table in digital aircraft simulation report |
| CN108647196B (en) * | 2018-04-16 | 2022-10-25 | 北京航空航天大学 | Artificial intelligence generation method and device for table in digital aircraft simulation report |
| CN109212999A (en) * | 2018-07-20 | 2019-01-15 | 北京航空航天大学 | The intelligent generation method and system of digital satellite emulation operating condition |
| CN109212999B (en) * | 2018-07-20 | 2020-09-01 | 北京航空航天大学 | Intelligent generation method and system for digital satellite simulation working condition |
| CN116150885A (en) * | 2023-02-14 | 2023-05-23 | 南京友一智能科技有限公司 | Multidisciplinary integration and simulation data management system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101266628A (en) | Automatic real-time emulation and its paralleling method based on emulated knowledge library | |
| Concepcion et al. | DEVS formalism: A framework for hierarchical model development | |
| CN104035751B (en) | Data parallel processing method based on multi-graphics processor and device | |
| CN101819540B (en) | Method and system for scheduling task in cluster | |
| CN102508639B (en) | Distributed parallel processing method based on satellite remote sensing data characteristics | |
| CN101719078A (en) | Parallel computation management-based autonomous navigation simulation and scheduling management system | |
| CN103337041A (en) | System for intelligent decision-making of concrete dam pouring construction based on knowledge engineering and method thereof | |
| CN107609141A (en) | It is a kind of that quick modelling method of probabilistic is carried out to extensive renewable energy source data | |
| CN103412878B (en) | Document theme partitioning method based on domain knowledge map community structure | |
| CN106875320A (en) | The efficient visual analysis method of ship aeronautical data under cloud environment | |
| CN112948123B (en) | Spark-based grid hydrological model distributed computing method | |
| CN103309239A (en) | Multilevel information management and communication method | |
| CN104519112A (en) | Intelligent selecting framework for staged cloud manufacturing services | |
| CN110705716A (en) | Multi-model parallel training method | |
| CN102945198A (en) | Method for characterizing application characteristics of high performance computing | |
| Zaitseva et al. | Dynamic programming method in the tasks of optimal labor capital distribution programs | |
| CN111414961A (en) | Task parallel-based fine-grained distributed deep forest training method | |
| Archibald et al. | Integrating deep learning in domain sciences at exascale | |
| CN109189376B (en) | Artificial intelligence writing method for digital aircraft cluster source code | |
| Bogdanov et al. | Virtual testbed: concept and applications | |
| CN107229234A (en) | The distributed libray system and method for Aviation electronic data | |
| CN102270190B (en) | Computer modeling and solving processing method of complex decision-making problem | |
| CN109214043B (en) | Artificial intelligence writing method for digital aircraft dynamics environment information transmission source code | |
| Honcharenko et al. | Smart Information System for Creating Digital Twins of Construction Project | |
| De Kuyffer et al. | Offshore windmill and substation maintenance planning with distance, fuel consumption and tardiness optimisation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20080917 |