CN109588054A - Accurate and detailed modeling using distributed simulation engine to the system with large complicated data set - Google Patents

Abstract

It is a kind of for the collection of fully integrated business impact data, the analysis of the data and analysis-driven business decision and the system of the generation of both analysis-driven simulations of spare candidate business actions.The commercial operating systems can be used for based on past and the Current commercial data retrieved in subsidiary company and from the multiple external sources being pre-programmed into system and predicting to formulate the result of candidate business decision.Then operation creates the model of actor using the simulation of the data and predefined parameter.Also the numerical estimation of taking risks of candidate decision is calculated.

Description

Using distributed simulation engine to the accurate of the system with large complicated data set and Detailed modeling

Cross reference relevant to application

The application be submit on July 8th, 2016 it is entitled " for automate capture and analyze for reliable business speculation System (the SYSTEM FOR AUTOMATED CAPTURE AND ANALYSIS OF of the business information of prediction of result BUSINESS INFORMATION FOR RELIABLE BUSINESS VENTURE OUTCOME PREDICTION) " the U.S. The PCT application of patent application serial number 15/206,195 and requirement are enjoyed its priority and are submitted on June 18th, 2016 Entitled " system (the SYSTEM for automating capture and analyzing the business information for reliable business speculation prediction of result FOR AUTOMATED CAPTURE AND ANALYSIS OF BUSINESS INFORMATION FOR RELIABLE BUSINESS VENTURE OUTCOME PREDICTION) " U.S. Patent Application Serial Number 15/186,453 PCT application And require to enjoy its priority, the entire specification of each of which is incorporated herein by way of being cited in full text herein.

Technical field

The invention belongs to use the field of computer system in the works in business information management, operation and prediction.Specifically, The component of system integrate business information and operation data, complex data analysis and using the data, preprogrammed command and parameter with And the function of machine learning reliably prognosis modelling can be answered such as applied to the large size of business speculation result herein with creating one kind Miscellaneous system and calculate predicted path capital risk commercial operating systems.

Background technique

It is applicable for business decision person from omniscient as what is found on business equipment over the past several decades Sensor the resource or can get from third party's resource, detailed cause and effect data and business processes monitoring software Finance, operation, foundation structure, the amount of risk management and philosophy information extended to wherein data and surmounted company's warp Reason follows all these information and determines and explains and in given business environment to the energy that data can be applied to make significant purposes The degree of power.In other words, bursting out much for corporate decision person or the applicable business relevant information of policymaker team It has exceeded and needs most using it to follow completely or reliably using the ability of those of its people.It can not identify important trend or change Be able to that timely mode knows that information resulted in that the client of highly-visible faced in the past few years for NETFLIX^TM, FACEBOOK^TM, and UPS^TMCut-off, only enumerate several.

Several research and development of business software are had existed, are determined originating from streaming or the analysis of automated business data or business The purpose of plan process.PLANATIR^TMSoftware is provided so that the mode in mass data, DATABRICKS is isolated^TMCustom analysis is provided Service, ANAPLAN^TMFinancial shock calculation service is provided, and exists and slows down the identification of business data relevance, analyze the data and quotient The other software resource of some characteristic aspects of industry decision automation, but these schemes can not handle the list than entire task A characteristic aspect is more.This implies the institute in as decision-making process of one of the variable as the data from a software package The technology used generally has to significant ground and is manually changed into be introduced into software to be used to analyze next time, if there is suitable Software.The step does not use human resources sufficiently and has introduces the potentially possible of mistake at crucial process point Property.

Had made great progress in accurate system's modeling and simulation field.As people will quickly speculate, quotient is predicted Industry system or congenial, any novel system or congenial authentic simulation are of virtually very high potentiality to save a large amount of moneys This, both currency and manpower can usually be run to much to be completed a little faster than true process, and if does not meet accident Inefficient or risk, or even if original scheme utter failure, can quickly and efficiently make for design parameter and The change of hypothesis, until find optimal and engineered solution or plan to lie on the table, it is all these that there is minimum of resources to use And capital loss.When calculation processing ability is increased, more conventional modeling and analogy method such as system dynamic, wherein The actor of same type such as automobile or truck will be with their descriptive datas, such as travel speed or in simulation system The weight of middle high aggregation, and then those aggregation actor types each will indicate simulate interaction during such Type；Or discrete event simulation, wherein being handled by and being usually regarded as continuous event and being divided into series of discrete subevent Simulation executes the effect of overlength process to show to target or target group, and such as tripping is into hospital emergency room: " opening transport Vehicle doorn ", " patient is put into vehicle ", " drive vehicle 0.4 mile to patient door corner " ... is " in healthy computing system " collect not by the payment of insurance covering ", and " patient leaves doctor by " being checked by doctor " ... for middle record patient health insurance information " ... Institute ", wherein the even described step can be segmented further: having discovered that highly useful recovery.Such as people it is to be appreciated that The significantly rising of computing capability causes to be acted on behalf of or actor has in simulations by allowing each target and may mark There is the single model of own and promises the simulation engine much more accurately modeled.The ability allows people to test identical These differences cause the progress of simulation and the effect of output between type individual, but may also require that the significant more processing time With data bandwidth to complete and be also commonly used in the simulation than System Dynamic Model smaller scale.

There is currently applicable multiple open source simulation engines.DEUS, discrete event simulator engine；OM-NET++, separately One discrete event simulator engine；Siafu, the simulation engine based on agency；And with high scalability based on discrete thing The simulator engine of part.All these provided objects are subjected to the variation range from limited scalability and deployable.Such as with On other business datas are captured and converted with what provided object was proposed, these simulation engines as total business analysis with Another isolated service in forecasting system, wherein the data from other services must usually reformat first and with defensive position Dynamic input.Further many accurate but executable simulation is dependent on multiple analogy methods, system dynamic, discrete event and institute Based on agency, depending on minimum accuracy needed for simulating the part that is just being demonstrated and generating reliable output.

A kind of fully integrated system is needed, from many various Source Search business relevant informations, identifies and analyzes high capacity Data, are converted into the useful format of business and then using the data to drive highly integrated scalable simulation engine, can With the combination using the example based on event in system dynamic, discrete event and dry run so that acquisition is more useful and smart True data, and the needs stored for analysis.The multi-method simulation engine forms " commercial operation system using remaining module System ".

Summary of the invention

Therefore, inventor has been developed that a kind of distributed system is large-scale to having using distributed simulation engine to be used for And the system of complex data collection carries out accurate and detailed modeling.System further uses the result of business information analysis to optimize It makes business decision and allows using latest data and machine prediction algorithm simulation between two parties interactive action path.Specifically, system A part be applied to following field, reliably predict the result in different business decision paths and by using most current sensor Device data are to the simulation that each decision path is in progress and for the risk of each Tactic selection ensemble prediction business value, special programming Decision defined parameters and can be using business environment data and then showing the number with the most useful format of author for simulation According to.

Preferred embodiment in accordance with the present invention, one kind is for using distributed simulation engine to large-scale and complex data The system of collection is accurate and the system that models in detail includes: business data search engine, is stored in the memory of computing device simultaneously It operates on the processor of computing device；Business data analysis engine is stored in the memory of computing device and operates in meter On the processor for calculating device；And plan automatically and numerical estimation module of taking risks, it is stored in one of multiple computer installations Memory in and operate on its processor.The result of the action analog module is stored in the memory of one or more computing devices In and operate on its processor, wherein business information search engine: from the multiple business related datas of multiple Source Search；Directly Ground receives multiple analysis parameters and control command from human-computer interface device or from one or more order and control storage device, with And the retrieved information of storage accumulation is for by data analysis engine processing or scheduled data time-out.Business information analysis Engine retrieves multiple data types from business information search engine, and is based on analyzing author in current active by business processes Middle illustrated special objective and demand execute multiple analytic functions and transformation to data detected.Automatically plan and take risks Numerical estimation module: using the data performed by business information analysis engine analyze and convert result, with from it is multiple come The order and ginseng that author is analyzed using supplementary data and the study of any current active special machine, from business processes in source Number is used as performed by business information analysis engine together to be formulated Current commercial plan and risk status report The result of data analysis and transformation analyzes appointing for author using supplementary data, from business processes with from multiple sources The input that current task special command and parameter are collected together and from machine learning algorithm is with defeated to first terminal user Send business decision path simulation and numerical value support of commercially taking risks.The result of the action analog module: retrieval is drawn by business information analysis At least part for holding up the result of performed data analysis and transformation, retrieves the initial data from business information search engine It is at least a piece of, using from plan automatically and take risks numerical estimation module input multiple parameters, using obtain to execute The information of the prognosis modelling of business speculation or business decision process path and by simulation author's use at least one mould Intend example combination system dynamic approach, discrete event method or analogy method and original initialization based on the method for agency As a result, use movement profile data and configuration parameter group with create the computer based model of real world item with It is simulated.

According to another embodiment of the present invention, a kind of for the fully-integrated collection of business impact data, the data The system of the generation of the analysis-driven simulation of analysis and analysis-driven type business decision and spare candidate business decision includes: quotient Industry data retrieval engine is stored in the memory of computing device 2 and operates on its processor.The system of claim 1, In be stored in the memory of computing device and operate in the business information search engine on its processor and use for human-machine interface The port of mouthful device, business related data input in its at least part, and with Current commercial speculative activities alternate item into The relevant order of row and parameter input it at least in another part.Wherein automation plan and numerical estimation module use of taking risks The future of data before having analyzed based on Current commercial decision is at least reliably predicted based on the statistical analysis of information theory As a result.Automation plan and numerical estimation module of taking risks take risks Numerical Principle using at least Monte Carlo heuristic model with can The future of data takes risks numeric number before estimating having analyzed based on Current commercial decision by ground.Automation plan and load wind Dangerous numerical estimation module is commercially being determined based on the data storage service of figure with efficiently storing and handling using special designing The large data structure created during plan interpretation of result.Automation plan and numerical estimation module of taking risks are with Operation control Function, the operation for allowing that single parameter sets is used to run with single iteration, and including successive ignition and there are termination rule The operation of predefined parameter set then when analysis result needed for obtaining to stop executing.Some operations are with similar batch mode Off-line operation and other operations are with interactive mode on-line operation, after wherein user is used for based on the result input of iteration before The parameter of continuous iteration is until the analysis result being pre-designed terminates execution.At least one analog module include for disaster factors, The module of tender spots, contractual specifications obligation and financial capital loss.Automation plan and numerical estimation module of taking risks act on At least one computer based model is to modify it before simulation.

According to another embodiment of the present invention, it is a kind of for fully integrated collection business impact data, analyze the data and The system for generating the analysis-driven simulation of the business decision and spare candidate business decision of analysis-driven includes: business information retrieval Engine is stored in the memory of computing device and operates on its processor, using the port for human-computer interface device, quotient Industry related data inputs in its at least part, and order relevant to the carry out of Current commercial speculative activities alternate item and parameter are defeated Enter it at least in another part.Automation plan and numerical estimation module of taking risks are using at least based on the statistical of information theory Analysis is reliably to predict the future outcomes of data before having analyzed based on Current commercial decision.Automation plan and number of taking risks Value estimation block takes risks Numerical Principle using at least Monte Carlo heuristic model reliably to estimate based on Current commercial decision Analyzed before futures of data take risks numeric number.Automation plan and numerical estimation module of taking risks are set using special Meter based on the data storage service of figure with efficiently store and manipulate created during business decision interpretation of result it is big Type data structure.Automation plan and numerical estimation module of taking risks allow with Operation control function using single parameter Gather the operation that runs with single iteration, and the parameter including successive ignition and with termination rules predetermined set set Operation, with when obtain needed for analysis result when stop execute.Some operations are with similar batch mode off-line operation, other works Industry is with interactive mode on-line operation, and wherein user is used for the parameter of successive iterations based on the result input of iteration before, until pre- The analysis result of design, which terminates, to be executed.

Preferred embodiment in accordance with the present invention, it is a kind of for the fully integrated collection of business impact data, point of the data The method of analysis and the generation of analysis-driven business decision and the analysis-driven simulation of spare candidate business decision comprising steps of A) it uses and is stored in the memory of computing device and operates in the reception business decision of the client access interface on its processor Parameter and target；B) it uses and is stored in the memory of computing device and operates in the business data search engine on its processor From the multiple business datas of multiple Source Search；C) it uses and is stored in the memory of one of multiple computing devices and runs at which Reason device on the result of the action analog module from can apply business data creation real-world objects simulation model；D) prediction is predetermined Business decision or business speculation candidate as a result, and to decision or congenial being won by using the result of the action analog module It simulates and estimates the numerical value of taking risks for being attached to each candidate.

Detailed description of the invention

Detailed description of the invention several embodiments of the invention, and be used to explaining according to the embodiment hair together with specification Bright principle.It would be recognized by those skilled in the art that specific embodiment shown in the drawings is merely exemplary, and not anticipate It is limiting the scope of the invention.

Fig. 1 is the figure of the exemplary architecture of commercial operating systems according to an embodiment of the present invention.

Fig. 2 is the figure of the exemplary architecture of automation plan service cluster and correlation module according to an embodiment of the present invention.

Fig. 3 is the figure of the exemplary architecture of the result of the action analog module and correlation module according to an embodiment of the present invention.

Fig. 4 is that the setting of three kinds of the result of the action analog modules operation according to an embodiment of the present invention and the flow chart of execution are said It is bright.

Fig. 5 is to depict the figure of the primary processing location of single component of the result of the action analog module simulation.

Fig. 6 is to illustrate the flow chart of centralization event queue timing according to an embodiment of the present invention.

Fig. 7 is to illustrate the flow chart of time step-by-step movement queue timing according to an embodiment of the present invention.

Fig. 8 is to illustrate the flow chart of conservative event-driven queue timing according to an embodiment of the present invention.

Fig. 9 is to illustrate the flow chart of optimistic event-driven queue timing according to an embodiment of the present invention.

Figure 10 is to illustrate to use the box of the exemplary hardware architecture of the computing device in various embodiments of the present invention Figure.

Figure 11 is the side for illustrating each embodiment according to the present invention for the example logic framework of client terminal device Block diagram.

Figure 12 is the exemplary frame of the client for illustrating each embodiment according to the present invention, server and external service The block diagram of structure arrangement.

Figure 13 is another box for illustrating to be used in the exemplary hardware architecture of computing device in each embodiment of the present invention Figure.

Figure 14 is offline single operation according to an embodiment of the present invention and automation plan service cluster operation is run multiple times Flow chart.

Figure 15 is online single operation according to an embodiment of the present invention and automation plan service cluster operation is run multiple times Flow chart.

Specific embodiment

Inventor has already envisaged for and has put into practice a kind of for using distributed simulation engine to large-scale and load The system of data set carries out system that is accurate and modeling in detail.

Inventor have already envisaged for and put into practice it is a kind of for fully integrated capture and analyzing cause forecast and decision make and The system and method for the business information of simulation.

One or more different inventions can be described in this application.Further, for one of the invention described herein Or it is multiple, many alternative embodiments can be described；It should be understood that these show only for illustrating illustrative purpose.The implementation Example is not intended to the restriction of any idea.One or more of the invention can extensively be suitable for many embodiments, such as from this It discloses clearly visible.In general, with sufficient datail description embodiment so that those skilled in the art implement one of the invention or It is multiple, and it should be understood that can use other embodiments, and structure can be made, logic, software, electrical changed with other Become and without departing from the range of specific invention.It therefore, it would be recognized by those skilled in the art that can be using various modifications example and standby Example is selected to implement one or more of the invention.It is referred to one or more embodiments or attached drawing describes of the invention one or more A special feature, the embodiment form a part of this disclosure, and wherein show the present invention by the mode of explanation One or more specific embodiments.It should be appreciated, however, that this feature be not exclusively for use in one or more specific embodiments or Referring to which depict in the attached drawing of one or more specific embodiments.The disclosure is not all embodiments of one or more inventions Verbal description, be also not necessarily present the list of the one or more features of the invention in all embodiments.

The title of beginning and the patent application of provided section is in the present patent application only for convenient, and does not answer It is regarded as limiting the disclosure in any way.

The device being in communication with each other need not be continuously in communication with each other, unless clearly in addition specified.In addition, the device being in communication with each other It can directly or indirectly be communicated by the medium of one or more logics or physics.

The explanation of embodiment with the several components being in communication with each other, which is not implied by, needs all these components.On the contrary, can With the one or more extensive various possible embodiments for describing various optional features to illustrate of the invention, and in order to completeer In terms of illustrating one or more features of the invention entirely.It similarly, although can be with successively sequence description process steps, method step Suddenly, algorithm etc. can usually configure these processes, method and algorithm with alternative sequential working, unless particularly making phase Counter-statement.In other words, the arbitrary sequence of the step of can describing in the present patent application or sequence itself indicate requirement with The sequence executes step.The step of process, can actually execute in any order.Further, it is possible to simultaneously execute Step, although description or hint are sequentially generation (for example, because a step is described as after another step).This Outside, illustrated process is not implied by the explanation of process by its description in the accompanying drawings to remove other variations and repair it Change example, be not implied by illustrated step or its arbitrary steps be for one or more of the invention it is necessary, it is also not dark Show that illustrated process is preferred.In addition, usually to describe step primary for each embodiment, but this does not imply that they must Beard and hair is raw primary or they can occur once in each run or executive process, method or algorithm.In some embodiments Or it can be omitted some steps in some situations, or step can be executed in given embodiment or situation more than one.

When describing single device or when article, it is obvious that single device or article can be substituted and use is more than One device or article.Similarly, when describing more than one device or article, it is obvious that can be instead of being more than one A device or article and use single device or article.

It can be by being not explicitly described as having the function of that other one or more devices of this or feature alternatively embody The function or feature of device.Therefore, one or more other embodiments of the invention need not include device itself.

Described herein or reference technology and mechanism will describe in the singular sometimes for clear.However, should infuse Meaning, specific embodiment include the successive ignition of technology or repeatedly showing unless in addition marking for mechanism.Process specification in attached drawing Or box is construed as representation module, section or code section comprising for specific logical function or step in implementation process Rapid one or more executable instructions.Alternative embodiment include in the scope of embodiments of the invention, wherein for example, Function can be executed other than shown or described sequence, including substantially simultaneously or with reverse order, depended on involved Function, as will be recognized by those possessing ordinary skill.

Definition

As used herein, " swimming lane " is to receive simultaneously distributor in time series sensing data and divide for holding Communication channel between the data storage mechanism for the data time series sensing data matched.Swimming lane can be between two Mobile special, limited data volume.For example, single swimming lane can reliably carry the equivalence from 10 sensors in 5 seconds It in the data that 5 second datas are worth and has included into data storage, this is its capacity.It attempts to place using a swimming lane from 6 Received 5 seconds value data of a sensor will lead to data degradation.

As used herein, " first swimming lane " is the transmission capacity of two or more the true swimming lanes transparent for request process Required logical combination.It can initiate wherein unit time received data amount and be expected the sensing with time change height homogeneity Device is studied to use first swimming lane.Using above example, single true swimming lane can transmit and merge being worth for 5 seconds for 10 sensors Data and have no data degradation, suddenly will be so that being from the sensing data that 13 sensors receive input in 5 second interim System two swimming lane member swimming lanes of creation with 10 sensing datas of the receiving standard in a true swimming lane, and second, it is transparent 3 sensing data surpluses in the true swimming lane of addition, however without changing data acceptance logic, because of data receiver and distribution Device will pellucidly add additional true swimming lane.

Conceptual framework

Fig. 1 is the exemplary architecture of commercial operating systems 100 according to an embodiment of the present invention.Special data is inputted, System control is made certainly to the client access interface 105 of system and for interacting such as automatic Prediction with system output Plan and plan and alternative path simulation are occurred by the distributed expandable high bandwidth cloud interface 110 of system, using logical With, healthy and strong network application driving interface with the input of the information for curstomer-oriented end and display and data storage 112, take Certainly in embodiment such as, but not limited to MONGODB^TM,COUCHDB^TM,CASSANDRA^TMOr REDIS^TM.From client business model It encloses interior source and is also by the input of cloud interface 110 from many business datas based on cloud service, being analyzed by system System, data reach analysis and the transform component of system, and orientation calculates figure module 155, and high capacity network crawler module 115 is more Dimension time series databases 120 and figure stack services.Orientation calculates figure module 155 from multiple Source Search one Or multiple data flows, multiple sources include but are defined in multiple physical sensors, network-based inquiry table and application form absolutely not, The monitoring of basic electronic setting, masses' sourcing activity and human input apparatus information.Figure module 155 is calculated in orientation Interior, data can be divided into two equivalent streams in dedicated pre-programmed data assembly line 155a, and one of subflow can be sent out It send for batch processing and stores, and another subflow can be reformatted to analyze for transformation pipeline.Then by data Generic converter service module 160 is transmitted to using a part for linear data transformation as analysis, or is transmitted to and can divide Solve transformation of the transformer service module 150 for branch or iteration as a part of analysis.Orientation calculates figure module 155 will All data are expressed as directed graph, wherein transformation is that node and the result message between the transformation edge of figure are sent.Gao Rong Amount webcrawler module 115 is using the pre-programmed web crawlers for being resident multiple servers, and when automatically configuring, it is deployed in net In network scraping frame 115a, example SCRAPY^TM, to from network-based source title and retrieve not by traditional network The data of interest of crawler technology excellent marker.Multidimensional time-series database module 120 is several different types of from can be Large-scale multiple sensors receive data.Module is designed by dynamically pre-allocating network bandwidth and to handle input data Server process channel and adapt to irregular and high capacity surge.For including but is not limited to C++, PERL, PYTHON, and ERLANG^TMExamples of language include that programming packaging allows complicated programmed logic being added to multidimensional time-series database 120 Default function and do not know kernel programming in detail, greatly expand the width of function.When can further analyze by multidimensional Between the data retrieved of sequence database 120 and high capacity network crawler module 115 and figure 155 and related general is calculated by orientation Transformer service 150 and 160 block transforms of decomposable transformer service are the result of task optimization.Alternatively, it generallys use Script prompt information determines important vertex 145a, can will come from multidimensional time-series database and high capacity network crawler module Data be sent to figure stack services module 145, use the figured standard for information flow to be changed to data Change agreement, for example, open figure Internet technology, although the present invention is independent of any one standard.Pass through these steps, figure Shape stack services module 145 indicates data with the graphic form influenced by any predetermined scripts modification 145a, and is deposited It stores up and stores 145b such as GIRAPH in the data based on figure^TMIn or code key numerical value pairing categorical data store REDIS^TMOr RIAK^TM, wherein it is all these to be suitable for storing the information based on figure.

The result of transform analysis process can then with other clients instruction, additional business rules and with also transporting In the automation plan service module 130 of prediction statistical function and machine learning algorithm of the row based on strong information theory 130a It can apply analysis and the relevant practical combination of situational information except data to allow result obtained by being based on current system Rapidly forecast future trend and result and select multiple possible business decisions each.Data are applied using all, from Dynamicization intended services module 130 can propose business decision, most likely for have can land used high certainty most have The result of sharp commercial results.It is closely related in the result obtained using system with automation plan service module, together with auxiliary Help terminal user to make in business decision possible external provided additional information together, the result of the action analog module 125 its have There are the discrete event simulator programming module 125a coupled with terminaloriented user observation and state estimation service 140, service 140 If situation needs be height can scripting 140b, and have game engine 140a with more realistically among consideration To the possible outcome of business decision, this allows business decision person based on one movement of analysis and investigation selection that can currently apply data Course co-pending possible outcome on top of the other.For example, pile line operation control room is it has been reported that height in territory The very small reduction of one section of assembly line Central Plains oil pressure in remote sections.Many people believe problem entirely due to silt up, The fluid sensor that may be failed, other people believe that near upstream pump can suck foreign matter.The correction of both these possibilities It is the sensor for increasing the output of impacted pump hopefully to remove it or be silted up.It must be in next maintenance The sensor of period replacement failure.However, a few peoples feel pressure decline be due to damaged in pipeline, may at this point compared with It is small, but nonetheless, crude oil is in leakage and the remediation option for being silted up sensor or pump is incited somebody to action so that leakage is more serious And hereafter waste the more time.Company can employ contractor except 8 hours, or can rent satellite time to check, but Both being is expensive for possible Sensor Problem, is significantly born although then having significantly less than cleaning oil spilling Face public exposure degree.These Sensor Problems are occurring and commercial operating systems 100 have from their number before According to due to a large amount of column diagrams, nobody really studies them, therefore the alternative 125,140 of run action.Based on it is all can Using data, since other can apply data, the sensor or pump that system prediction is silted up are specifically unlikely to be basic reason, and And contractor is sent.She finds small breach in pipeline.Small cleaning will be used and need to close pipeline and be used to repair, But several ten million dollars are saved.This is only an example of a large amount of possible purposes of commercial operating systems, this field skill Art personnel will be easy to illustrate more.

Fig. 2 is the automation plan service module of embodiment according to the present invention and the exemplary architecture of correlation module Figure.The more detailed view of automation plan service module 130 as shown in fig. 1 is seen herein.Module passes through client-side interface 205 Receive business decision or business speculation candidate and it is relevant currently can application-dependent data and any activity analysis modification life It enables and runs.Module, which can be used for providing, has converted data or operating parameter to the result of the action analog module 125 to run Or transformation is isolated with the one or more actors run in the result of the action analog module 125,320,320b, 320d, 320f Intermediate result data before during dry run sow simulation.It can also make during the simulation of alternative business decision process Support information with a large amount of, can such as, but not limited to influence current decision or it is congenial, via commercial operating systems as whole Body is collected and stored in the analysis for calculating figure module 155 in data storage as multidimensional time-series database 120, orientation The current business conditions of the network-based data retrieval capability of ability and high capacity network crawler module 115, infrastructure, Ongoing speculative state, financial status, market condition and world event, it is all these to can store at one or more In a data storage 220,225, these can leave variable be used as but be not limited to embodiment timing, the method that termination changes, The sequence and timing or the influence instead of another target is acted and selected under present analysis that component part is completed.

Expected movement, which can be, resolves into multiple composition events, or fulfils and act towards risk under analysis, Or the missing of each event is shown by discrete event simulation module 211, then make those each events for based on letter The theoretical statistical analysis 211 of breath can be applied, this allows the criterion using the machine learning obtained from data before in the phase of variation According to similar event analysis current decision event under conditions of the opposite sex；Other than other factors, it can be estimated by uncertainty Calculate module 213 analyze the result of the analysis with further finely tune will be included in the analysis of completion confidence level.It sets Certainty by be the stochastic variable distribution given for each analysis event weighted calculation.It extracts and pushes away in dynamical system Manage calculated in module 214 with from wherein client commercial operation as the microenvironment of the more expensive stage of regional economy or into one It walks out of the anything of the successful angle of client business equally complicated system under the analysis involved by business speculation Event at least part of effect expection, module 214 used other than other tool algorithms based on Shannon entropy, breathe out Special benefit entropy and mutual information Correlation Theory.

Any business bring up in plan or new business speculation above all through on top of the other select a decision and It is placed in the amount of the business value under risk.The usual numerical value is money, but it is also possible to competitive position, operational efficiency or is based on Client relationship, such as: this can be to maintain it is older, may the slightly client relationship management system of irregular working effect Fruit, instead of replacing it with 1,000 4 million dollars and deposit, more a quarters.Automation plan service module has special based on covering Using the ability of the result of each decision of programming prediction unit amount, this allows to estimate number of taking risks Carlow heuristic model 216 Single " state " of value.The amount of computing capability needed for being very difficult to the expected one or more for completing the analysis of these business decisions, It can change huge according to demands of individuals and usually be run simultaneously with several alternate items.Therefore the present invention is designed for It operates on expansible cluster 215 in distributed, modularization and extendible scheme, such as but is not limited to, Amazon The offer object of AWS.Similarly, these analysis operations can permit many hours to complete, and many clients are expected Fall into a long wait it is simple " if ... what if " option, this will not affect that their recent commercial operation, and other Client can require urgent decision situation, and wherein they need alternate item fastly as far as possible.This presence by job queue and mention For allowing implementing analysis operation under one down to urgent multiple priority.Change by more imaginary analysis operations Become in more urgent situation, also can be used change job priority during job queue 218 is run priority-based and There is no progress lost.

Structuring planning analysis result data can store in general automation plan engine, perform action mark Modeling language (ANML) script with for for which can be used for being prioritized the task modeling towards people and Machine oriented with Reward function is maximized on finite time range 217, or is stored by the data storage 145 based on figure, is depended on Analyze the specification of complexity and runing time.

The result of analysis can be sent to one of the display module at two curstomer-oriented ends, the result of the action analog module 125 or can more visual simulations observations and state estimation module 140, depending on demand and client to the intentional of data It uses.

Fig. 3 is the figure of the exemplary architecture 300 of the result of the action analog module and correlation module according to an embodiment of the present invention. 105 control flow of network application, which is accessed, by the client being programmed in commercial operating systems 305 specifies setting for each simulation It sets and controls.These screens may include since one group of predetermined analog type 305a selection phantom frame and finally newly simulate Option；It checks the progress of multiple lots, can indicate the activity of actor, indicate special event, and control substrate ring Border, other than other simulations task 305d well known by persons skilled in the art；Stream is collected during being sent in simulation process 305b Intermediate result data, digital encryption code key is obtained so as to one or more static for one or more dry runs Assuming that parameter or for manipulate input model data algorithm make routine change, show in addition to well known by persons skilled in the art Except example, to improve fidelity or the serviceability of upcoming plan operation 305e；It reads carefully and thoroughly and checks to simulate from the past and transport Row 305c be obtained and stored in it is multiple with its commercial plan on as a result, if including the parameter in setting, include: risk Prediction, the capital loss being subjected to, the capital gains of acquisition, for the closure of final result expected from planning phase, as a result in Most influential variable, and the hypothesis used in other a large amount of multiple possibilities.Interface also allows to compare from multiple The result parameter of associated analog test starts numerical value and interested hypothesis wherein changing to predict each influence 305c.This hair Bright main advantages are, run multiple associated analogs with the time much more shorter than authenticity test, this is usually simulated It advantage and largely uses the ability of graphical interfaces mechanism setting simulation and then makes integrated business system Other modules, which are mainly handled, saves a large amount of present level related datas and the switched support result for each dry run. Finally, client access network application 105,205,305 provides multiple formats for display data 305f, it can be with range from figure Video or figure magazine are to pure mathematics number, as needed for given point in time simulation and using data and suitable.

Simulation is executed by the result of the action analog module (AOSIM) 320 comprising by AOSIM agent node in embodiment Shown in height expansible, distributed frame: 1 320a of AOSIM worker, 2 320c of AOSIM worker, until AOSIM worker N 320e, the number of node is expected as the scale and complexity of institute's running simulation change, all to simulate master control in the result of the action Under the process control of node, controls and data is introduced in single model actor minor structure 320b, 320d, 320f, And the timing of each simulation steps operation, thus it is possible to vary the periodically mutual trust journey to adapt to simulate interior model actor Degree and simulation with runing time comparison required levels of precision 600,700,800,900, this can with range from transformation mediant According to demand, automation plan service module 130,200 may be used.Reduced data can be then manually entered and be back to In sequence 600, until wherein each actor step ignored and simulation independent with every other actor it is any possible The timed scenario of data dependence effect 900 between actor.In each group of model or actor 320b, 320d, 320f It is interior, it is operated since the distributed performance of AOSIM module can be used as single group, adjacent actor can be in true generation The different actors modeling being modeled in boundary's system.Therefore, as simple, nonexcludability example, if simcity street Corner, first element person can model for worrying the pedestrian to upset to go to work in group 320b, and the second movement Person can model the move vehicle of left-hand bend, and third actor can model the basketball close to identical intersection.Its He can model the condition of the mist or ponding low-lying area actor, more passively act on it with the degree of some variations His actor.It is environment levels under this entire actor structure, monitors the work of each actor, collects and can pass Defeated result is to host 320g for converting and storing.Environment layer can also be transmitted to one or more actors in given situations Control instructions, one of them can be timer command.It collects and by using from multiple available sources by system embodiment 330 The large sample set for the related real world measurement that storage is used for subsequent analog designation, using based on inferencing information theory Statistics, and modeled using exploration as described above, it, can be significant among other technologies well known by persons skilled in the art Influence intrinsic changeability in such as, but not limited to collected data of factor of the result reliability of simulation, measurement inexactness, Simulation setting in unimagined external factor, in calculating parameter data and the difference between real world simulation analog It is different, to reduce the uncertainty of the setting.

What substrate configuration data 370a dry run and serial parameter 370b and strategy and MongoDB GridFS encoded Support document files for simulate setting and initial 370c, all these general utility functions for AOSIM be it is crucial, be stored in Can only be passed through in read-only MongoDB storage 370 and directly by AOSIM APIs MongoDB data storage 350 change with Their integrality is protected, but is used read-onlyly by client access network application to be arranged and initialize the phase in primary simulation Between retrieval parameter.It is also potential big caused by storage in the selected works 360 of MongoDB data storage 350 during dry run Average information is measured, wherein data are actor correlation 360a, actor group correlation 360b, world 360c, mould by genealogical classification Negative effect 360d of the plan weakness of quasi- step for entire result, the identified plan disaster factors during dry run 360e, and any practical plan loss 360f incurred by simulating congenial or business path.Including complete in them A part of 305f is shown as analog result at all these data of state.

Fig. 6 is to illustrate the flow chart 600 of placed in the middleization event queue timing according to an embodiment of the present invention.There are such Situation, wherein the system being modeled is extremely complicated, or the even ability with AOSIM 300 is to from such as but not The data for being limited to following classification execute automatic processing: physical world, computer-physics interaction, society-technology, information network Interaction, individual is integrated and cognitive Inference, it is necessary to check at least some of intermediate result data and can may be by commercial operation system Another module manipulation of system 100, perhaps such as automation plan service module 130, orientation calculate figure module 155, or extremely Few one it is awake in, for before being reintroduced back to simulation 616 for specific simulation steps manually.Before can needing to be based on Step 616 and modify when each actor 601,602,603,604,605,606,607 operation next step in simulations 608,609,610,611,612,613,614 when other times sequence.Under this condition, it is important that next in initiation All data manipulations and series processing 615 are completed before step 617.

Fig. 7 is to illustrate the flow chart 700 of the time step-by-step movement queue timing of embodiment according to the present invention.In the presence of in this way Situation, wherein the system being modeled includes the independent actor of height and almost determines dependent on one at multiple steps Or the result of other multiple actors is to lead to reliable analog result.Under those conditions, it is important to, everything person 701,702,703,704,705,706,708 conation authors in office start to complete each step before next movement (x+1) 717 709、710、711、712、713、714、715。

Fig. 8 is to illustrate the flow chart 800 of conservative event-driven queue timing according to an embodiment of the present invention.In the presence of in this way Situation and may rely at one or more steps wherein the system being modeled includes slightly independent actor The result of one or more actors is to lead to reliable analog result.Under those conditions, it is important that actor 801, 802,803,804,805,806 start before next step (x+1) 815 there are 807 in actor, 808,809,810,811, 812,813, (step x) 814, and therefore when using the timing, when correlation at embodiment determines the step of discussed Individual part person will only execute next step 815 when property is extremely unlikely to be 814.

Fig. 9 is to illustrate the flow chart 900 of optimistic event-driven queue timing according to an embodiment of the present invention.In the presence of in this way Situation, wherein the system being modeled do not include dependent at one or more steps one or more actor result Any self contained function person to obtain authentic simulation as a result, wherein time-constrain and missing update the independent data effect So that needing running simulation as quickly as possible.Under these conditions, each actor 901,902,903,904,905,906,916 Step 907,908,909,910,911,912,913, (x) are independently completed, do not consider to start next step (x+ in actor 1) 915 whether there is correlation 914 before, and therefore when using the timing, each individual part person 915 is as far as possible It is fast to execute 914.

The present invention has no reason not use multiple timings during single simulation, because each scheme becomes appropriate. This speed that can be completed with Optimized Simulated, and ensure to have respected the interior all complementary actor data modes of simulation simultaneously simultaneously And there is being reliably completed for any simulation.

Fig. 4 is the setting and execution of three types the result of the action analog module dry run according to an embodiment of the present invention Flow chart 400.Top plate 410 depicts simple analog, can be used for exploring can occur during commercial operation it is numerous " if ... what if " idea, or wherein little or no be related to predicting capital and the market risk.As can be expected, mould Quasi- setup phase consumes most of steps when initiating primary simulation and executing, author must only etc. and then explanation, It may further result of manipulation 416.Process starts from creative management information 411, and the parameter is used as data to be used herein The boundary of value, by the datum number storage of data to be used according to library origin or multiple origins, if there is more than one source, By commercial operating systems resource level to be used, by timing to be used, by progress indicator to be shown, Yi Jisuo The required display pattern of the simulation of completion can be only for management parameters that are well known by persons skilled in the art, can specify Much larger set of several examples.For any simulation, reliable useful final result integrally depends on will affect the dynamic of simulation The factor of author it is appropriately configured, form wherein they will execute the ambient enviroment of their institute's programming actions, be labeled as herein " world " 412.Must determine, by the world institute's configuration factor carry out constraint and influence under the anticipated conditions of simulation with The matching of those of real world.Of equal importance, actor 413 matches the appropriate of real world item modeling in simulating It sets, again, mould of the authentic data relevant to the behavior of variation in each actor type and actor type in programming Those of real world item must be closely matched under the conditions of quasi-.It is also important that carefully considering the selection of actor type And the number for each type that will include in simulations, because the factor can change result and if true significantly World's ratio then causes conclusion opposite with reality far from those in computer simulation.By the single of particular model (actor) type Example is placed in one or more groups each personal in simulations how behavior but also measurement are from group not only to measure Whole result obtained, this is sometimes useful.By supporting specified 414 actor groups, embodiment allows such.Really Successfully completing for world's authentic simulation can be by more as far as possible using more limited verity World datas operation partial simulation As a result it greatly enhances, or the part of real world data even may be incorporated into the composition part of simulation such as to the world matches Set 412, the selection of actor model characteristic 413 and type, and generate suitable actor group 414.Commercial operating systems To use the potential advantage to provide mechanism, by allowing to be modeled to be allocated as to be attached to automation for operating parameter by listed Inferential statistic and Monte Carlo heuristic algorithm are introduced and support to guide intelligent inclusion and using from business by intended services module The data of other modules of operating system are for improving the course consumed by simulating when operation completes 419.Multiple high It spends at the end of important configuration step, 415 can be created and contain all parameters, module tie-in sale reference and management operating instruction Simulation initialization files.This defines with the format running simulation readable by the result of the action analog module.Then running simulation, It completes any back-end processing and shows as a result, script performance may be manipulated by using there is its game engine 140a and show 140b observation state estimation service 140, or may it is pre-designed for each of the author of simulation with some other formats and It shows.

Centre panel 420 depicts the simulation using finance aggregation, can be used for exploring the loss or increasing of where capital Benefit is possible business decision or speculates.Capital as the same behavior of actor, therefore creates its multiple examples in simulations Waste time and potential computing resource.Therefore capital is handled is used as aggregation, similar in dynamically being simulated based on system Actor, wherein everything person be their real world counter pairs aggregation agency.As can be expected, the setting of simulation Discrete consuming most of steps when initiating primary simulation and executing, author must then simply wait simultaneously then to explain, Ke Nengjin Single stepping result 427.Process starts from creative management information 421, herein side of the parameter as value data ready for use Boundary, the datum number storage of data to be used is according to library origin or multiple origins, if there is more than one source, business ready for use Operating-system resources level, timing ready for use, progress indicator to be shown, and complete the required display of simulation Mode can be and the much much larger set of of management parameters that are well known by persons skilled in the art, can most pushing up several is shown Example.For any simulation, reliable useful final result integrally depends on being appropriately configured for factor, will affect simulation Actor, form wherein they will execute the ambient enviroment of their institute's programming actions, be labeled as " world " 412 herein.It is necessary It is well established that the constraint as practiced by the configuration factor in the world and influence match that of real world under the anticipated conditions of simulation A bit.It is of equal importance, to being appropriately configured of actor 423 of the real world item modeling in simulation, also with regard to each dynamic The authentic data of the behavior of variation must be closely matched under the simulated conditions of programming in author's type and actor type Those of real world item.It is also important that carefully considering the selection of actor type and will include in simulations every The number of a type because the factor can change significantly result and if real world ratio far from computer simulation Those of then cause and the opposite conclusion of reality.Sometimes usefully, the single instance of particular model (actor) type is put It sets and is obtained in one or more groups with not only measuring each personal how in simulations behavior but also measuring from group The whole result obtained.By supporting specified 424 actor groups, embodiment allows such.Can by using as applicable or The composition part such as world that a part of real world data is incorporated into simulation is possibly even configured into 422, actor model Limited verity World data portion in the selection of characteristic 423 and type and the generation of suitable actor group 424 as much Divide the ability for the result for comparing dry run and greatly enhances successfully completing for real world authentic simulation.Commercial operating systems are logical Crossing allows to be modeled to be allocated as to be attached to automation plan service module for operating parameter, by inferential statistic and Meng Teka for listed Lip river heuristic algorithm, which introduces, supports to guide intelligent inclusion and using the data from the other modules of commercial operating systems to work as It is improved when operation completes 429 by the course of simulation cost and the mechanism using the potential advantage is provided.Different from directly retouching above The simple analog 410 stated, it is necessary to specified before running simulation for handling the rule of capital stream and boundary in simultaneously report simulation. This may include project such as, but not limited to total capital budget, and how much money can be tolerated in single or specified number simulation steps This loss, it is higher than other considerations whether income can be placed in some or all of set of simulation, and whether capital can be handed over Brand recognition or confidence level are changed, and can known many more potential selections for those skilled in the art.It is multiple, At the end of the configuration step of high-importance, it is necessary to create 426 and include all parameters, module tie-in sale reference and management operating instruction Simulation initialization files.It is run this defines the format simulated can be read by the result of the action analog module.Then it runs Simulation, any post-processing of completion are simultaneously shown as a result, script may be manipulated by using having its game engine 140a and showing The observation state estimation of energy 140b services 140 and is represented graphically movement, or may be default even for each of simulation author Meter exists with some other formats.

Centre panel 430 depicts the simulation for insurance or risk analysis, can be used for exploring business decision or throwing Machine, wherein since the wholesale loss of market share may have the extensive loss of capital, what infrastructure or capital equipment were invested Lose or lose existing infrastructure or capital equipment be it is possible, other situations well known by persons skilled in the art it In.Insurance simulation, which can be taken, simulates the form for the event that can lead to expensive responsibility expenditure or equipment expenditure, such as exists The medium-and-large-sized oil pipeline failure of hidden area then discharges the crude oil of a large amount of gallons, forces extensive remove with possible based on environment Recovery, or by the imaginabale many other possible examples of those skilled in the art.Capital, behavior identical as actor, Its multiple example is created in simulations, this wastes time and potential computing resource.Therefore it is handled capital as aggregation, Actor is identical in dynamically simulating with the system of being based on, and wherein everything person is the aggregation generation of their real world counter pairs Reason.As can be expected, the setup phase of simulation consumes most of steps when initiating primary simulation execution, and author must be only It waits and then explains, possible further result of manipulation 441.Process starts from creative management information 431, and the parameter can be with herein It is only much much larger set of several examples for the management parameters well known by persons skilled in the art that can specify, such as wait make With data, data to be used datum number storage according to library origin or more origins (if there is more than one source), ready for use Commercial operating systems hierarchical resource, timing ready for use, progress indicator to be shown and the required of simulation is completed Display pattern.For any simulation, reliable useful final result depends on will affect on the whole the factor of the actor of simulation Be appropriately configured, form wherein they will execute the ambient enviroment of their programming actions, be labeled as " world " 432 herein.It must Must it is well established that the constraint as practiced by the configuration factor in the world and influence under the anticipated conditions of simulation with real world that A little matchings.Of equal importance, actor 433 is appropriately configured real-world item modeling in simulation, equally, about The authentic data of the behavior of variation must be fitted closely within programming simulation bar in each actor type and actor type Those of real-world item under part.It is also important that carefully considering selection and each class in simulations including type of action The number of type, because the factor can significantly change result and lead if those in the far super computer simulation of real world ratio Cause the conclusion opposite with reality.Sometimes the single instance of particular model (actor) type is usefully placed in one or more In a group with not only measure in simulation it is each it is personal how the whole result that behavior but also measurement are obtained from group.Embodiment By supporting specified 434 actor groups to allow so.Can by comparing using with can apply limited verity as many A part of real world data is even incorporated into the composition part of simulation such as world's configuration 432, moved by World data The selection of author's model characteristics 433 and type and when the generation of actor group 434 partial simulation operation result and Greatly enhance successfully completing for real world authentic simulation.Commercial operating systems are by allowing using listed simulation member as fortune Row parameter is attached to automation plan service module 439, inferential statistic and Monte Carlo heuristic algorithm is introduced and support instruction intelligence Can in inclusion and using the data of other modules from commercial operating systems to be used for when end of run 439 when improvement by mould Quasi- spent course.Different from the above directly described simple analog 410, it is necessary to specify 435 to be used for before running simulation Handle the rule of capital stream and boundary in simultaneously report simulation.This may include project such as, but not limited to total capital budget, in list How many capital loss can be tolerated in the simulation steps of a or specified number, income is whether in some or all of set of simulation It is placed in higher than other considerations, whether capital can trade brand name recognition or confidence level, and for those skilled in the art It can known many more potential selections.The authentic simulation of risk element is extremely complicated and requires comprising risk simulation Interior includes several additional act persons, such as but is not necessarily limited to, can be with for the model 436 of harm existing under simulated conditions Lead to or be added to one or more category events 437 of flood condition or the beginning of its repercussions, simulates interior everything person's The model 438 of known tender spots, and it is final, specify insurance coverage related with actor present in simulation 440 or letter The responsibility possibility of loan contract together with the potential loss number for Possible event chain to run in currently valid contract The lower leakage of obligation 440.Due to these simulated domains, disaster 436, event occur 437, movement fragility 438 and contract and The extremely complicated property of each of Potential Model 440 is lost, system also utilizes applicable automation plan service module 130 Significant treatment and predictive ability with help include the simulation of these models 449 initialization and progress.Match in multiple high-importances Set at the end of step, can create 426 contain all parameters, module tie-in sale reference and management operating instruction simulation it is initial Change file.This format running simulation limited can be read by the result of the action analog module.Subsequent running simulation, after completion is any It handles and shows as a result, may be by using the observation state with its game engine 140a and displaying manipulation scripting capabilities 140b Estimation service 140 and graphically show movement, or may with for simulate each of author it is pre-designed with some other formats and It shows.Due to the Potential feasibility for multiple accidents that significant amount capital summation can be caused to pay, which can With with the declaration form assets of network analysis insurer to confirm that reserve capital is enough to cover the rational proportion of loss and for single The premium of company is consistent with potential risk and total payoff and terminates 442.

Fig. 5 is to depict the figure of the primary processing location of single component of the result of the action analog module.As implied simultaneously It can be the complexity of the expected simulation for giving various business decisions and business speculation progress and result, business software operating system Interior composition multiple modules of the invention can be related to handling.As can be expected, all simulation strategies 501 are present in simulation 502 It is interior, and the true execution of simulation 502 occurs in the result of the action analog module 540.Maintain the real world item for participating in simulation The actor or model assembly 505 of the built modular representation of purpose are mainly associated with the result of the action analog module that it runs herein, But constitute its movement regular 505a and influence boundary can as previously shown 400 and in automation plan service module 130, it is modified before or after dry run in 550, and the decomposable converter of possibility of its function has been described before Service module 150.Similar to still actor ingredient, world's ingredient is not equal to, found simultaneously in the result of the action analog module Execute the parameter of the movement of everything person during affecting simulation, but in this case, can before dry run with And in the single world entities 510 of post-processing of operation beginning and the policing rule of its movement and the influence during execution. The actor group ingredient 515 of the foregoing description is not despite the potential pith of multiple simulations, actually aggregation It is directly interacted with other component during simulation, is also the component of simulation engine, and in automation plan service module 130 With constructed completely in decomposable transformer service module 150,550 and store constitute its, have from world's strategy subassembly The policing rule for the frame that 515a is obtained, wherein the weight and group influence pair of the influence of certain individual part persons can be coordinated Most reliable, the true to nature final result of dry run is generated in the overall strength of simulation.Three components finally listed, disaster mould Type 520, Vulnerability Model 525 and financial modeling 525 be included in the component in all similar libraries in special simulation to enable A part modeling that not more excessively simple simulations execute to the special additional facet of simulation.This be it is extremely important, Due to the gravity of the result in those facets of their effects, this, can for including that their simulation author is very important Can, these components 520,525,530 and managed they how simulation 520a, 525a, 530a in interaction strategy Rule is as the last state of real world conditions current in those fields and data, and really, these three component institutes Continuous can measure that many data provided are subjected to can get data by sensor or network updates.It is anticipated that their construction With maintenance completely by the resume module of commercial operating systems, task is that the high speed of external available data is accumulated with business intentionally Justice processing, is followed by the intelligent integrated of new data Yu already existing relevant historical data, the multidimensional time-series of the foregoing description Data storage 120, orientation calculates figure 155, can decompose transformer service module 150,545 and automation plan services mould Block 550 can be responsible for creating and maintaining accurate useful 530 component of hazard model, Vulnerability Model 525 and financial modeling.

It must state, it is disclosed between special analog component of the invention and specific function (module of embodiment) Combining is schematical and is described as it is believed that it is height that this, which is combined during the execution of analog functuion of the invention, It is possible and fertile.State that the exemplary joint is not construed as meaning that other are combined during simulation is executed with other processes It can not or cannot achieve.They are possible, and for the succinct of smaller influence and understand joint and omit from specification They.The present invention be designed to allow for the calculating between any one function and any other function cooperate may sometimes Carve and pursue most accurate, timely prediction result, and can be needed in some rare cases it is all these be to promote Unforeseen new function some additional programs in combination.

Figure 14 is offline single operation according to an embodiment of the present invention and automation plan service module operation is run multiple times Flow chart.It can be single operation 1410 using the off-line operation operation of automation plan service module comprising one group of mesh Mark and condition and operation special parameters, or be run multiple times, it may include multiple possible targets, multiple conditions and fortune Row special parameters.Use single operation operation 1410 as model, operation start from be in its original state system 130 and One group of business goal 1403 of input.Load 1401 with from each source include WWW, throughout the sensing of client company Device, which includes that the target of physical sensors and logical sensor is relevant, all to be obtained using data and from source known in the art Any data obtained.Before true analysis operation, first verify that 1411 all data mind are normal.Then creation initial construction And instantiate proliferation frame of the planning data structure as subsequent manipulation 1412.The core of analysis 1413 may include using just True 1414 pairs of plan example sowings of parameterized model, execute model, orientation may be used to calculate the ability to transform of Figure 155, connect Be to collect the result of these model analysis 1415 and be followed by subsequent processing those results with by alignment the case where they and design. It can be for the knot of the characteristic uncertainty 1416 business path analysis of further analysis of result inherently such as under input condition Fruit, and the estimation for the alien influence parameter for not including can be in prediction 1417.Finally, including confidence level information Analyze the specification of distribution of results to the service client for initiating analysis 1418.Distribution of results can be related to observation and state estimation Service module 155, and programming is to help any event variation 1404 showed.

The progress of multiple off-line operation operations 1420 is very similar to those above, but allows to analyze business decision, In for the one or more used in analysis input parameter may exist multiple and different numerical value, such as wherein client is wished The implementation time to change one's plans, or do not know condition accurately wherein.The design present invention is to adapt to for art technology Any similar variation known to personnel.From load 1403 known target and retrieve 1401 original states of external data 1402, operation is initialized in automation plan service module 130.Once all parameters of verifying 1421, are defined by the user 1422 for multilinear operation clear stop criterion, this can be with analyzed it is simple as all situations.Also must determine The mode that justice samples system and world parameters, once, each run or some other distributions.Operation phase is run with single Together, it then creates initial construction and instantiates planning data structure as the proliferation frame for being used for subsequent manipulation 1424.It is then right In each 1425 operating procedure 1414,1415,1416,1417 and 1418 of iteration.Finally, all operations are transported in batch post-processing Capable extreme point results, confirmation meet stop criterion, and for the result of 1426 whole service of successful analysis.Then by network analysis The distribution 1427 of whole accounts to client.

Figure 15 is online single operation according to an embodiment of the present invention and automation plan service module operation is run multiple times Flow chart.There are such opportunitys, and the result that wherein operation is once run, which can have, requires client to then manually enter The effect of parameter for next iteration, or wherein determine that next inputting which data set depends on it in order to convenient The result of preceding job run.Under these conditions, the operation of automation plan service module can be with more interactive or online preferential Grade 1530 and 1540 and run, wherein client with negotiation time amount control process.In all scenario 1530,1540, analysis Process is definitely initially as corresponding off-line operation Class1 531,1541.When process consecutive hours expected for successive ignition, very To for simple single type of operation, specific termination condition 1532,1542 is set.Online single operation type is made Industry, operation then run operation 1534 according to offline single and execute, however, replacing at the end in each data input period In generation, is in distribution results and leaves 1418, tests 1535 specific terminal parameters, can be a lack of further data input, and adopt Analysis circulation 1534 is reruned with data 1533 after modification, up to 1535 stop criterions of satisfaction and process exits 1535a.Every Distribution results (not shown) after a period.The function that operation is run multiple times online is very similar, but is related to such as matching offline The definition for the criterion for modification parameter and sample distribution found in object 1423,1543 is closed, to run with for online single 1544,1545,1546 same way of operation controls other iteration, but estimates and adopt after each circulation 1546a, 1546b Sample distribution modification.It when it is present, include into successive iterations 1547 by any modification sample distribution.When estimated all ginsengs When number, circulation exits 1548.

Hardware structure

It is closed in general, the group in hardware or software and hardware can be implemented in technology disclosed herein.For example, they can To implement in operating system nucleus, in discrete consumer process, in the library encapsulation being tied in network application, special On the machine of construction, on specific integrated circuit (ASIC), or on a network interface card.

At least some of software/hardware hybrid embodiment of embodiment disclosed herein can be implemented by being stored in Computer program selective excitation in memory or the programmable network reconfigured are resident machine, and (it is construed as The connected network of interval perceives machine) on.The network equipment can have multiple network interfaces, can configure or be designed to Utilize different types of network communication protocol.Some general frameworks for these machines can be described herein to illustrate thus One or more exemplary schemes of the given unit of function can be implemented.According to specific embodiment, disclosed herein each The feature or function of embodiment it is at least some can be implemented it is general in one or more associated with one or more networks On computer, such as end-user computer system, client computer, network server or other server systems, Mobile computing device (such as tablet computing device, mobile phone, smart phone, laptop computer or other suitable calculating Device), consumer electronics device, music player or any other suitable electronic device, router, wrist-watch or other conjunctions Suitable device, or any combination thereof.In at least some embodiments, the feature or function of each embodiment disclosed herein It is at least some to can be implemented that (such as network query function cloud resides in one or more objects in one or more virtualized computing environments Manage the virtual machine or other suitable virtual environments on computing machine) in.

Referring now to fig. 10, it shows to depict and is suitable for implementing at least part of of disclosed feature or function herein The block diagram of exemplary computing devices 10.Computing device 10 may, for example, be previous paragraphs in listed computing machine it is any One, or can really be executed according to the one or more programs stored in memory based on software or based on hard Any other electronic device of the instruction of part.Computing device 10 may be configured to using for the wired or wireless of the communication Known protocol and with multiple other computing devices such as clients or server in communication network such as wide area network, Metropolitan Area Network (MAN), office It is communicated on domain net, wireless network, internet or any other network.

In one embodiment, computing device 10 includes one or more central processing unit (CPU) 12, one or more Interface 15, and one or more buses 14 (such as peripheral components interconnect (PCI) bus).When in suitable software or firmware control When acting under system, CPU 12 can be responsible for implementing special function associated with the function of the computing device of particular arrangement or machine Energy.For example, at least one embodiment, can configure or design computing device 10 for use as CPU 12 is utilized, locally deposits The server system of reservoir 11 and/or remote memory 16 and interface 15.In at least one embodiment, CPU can be made 12 execute the one or more of different type function and/or operation under the control of software module or component, for example, it can be with Including operating system and any suitable application software, driver etc..

CPU 12 may include one or more processors 13, such as, Intel, ARM from microprocessor, One processor of Qualcomm and AMD series.In some embodiments, processor 13 may include the hardware of special designing Such as specific integrated circuit (ASIC), electrically erasable programmable read-only memory (EEPROMs), field programmable gate array (FPGAs) etc., for controlling the operation of computing device 10.In a special embodiment, local storage 11 is (such as non-volatile Random access storage device (RAM) and/or read-only memory (ROM), including such as one or more levels cache) it can also be formed A part of CPU 12.However, it is possible to which there are many different modes to couple the memory to system 10.Memory 11 can be used for Various purposes, such as, capture and/or storing data, programming instruction etc..Should further it know, CPU 12 can be One of various system on chip (SOC) type of hardware, may include additional hardware such as memory or graph processing chips, Such as Qualcomm SNAPDRAGON^TMOr Samsung EXYNOS^TMCPU, if this field becomes more and more common, such as For in mobile device or integrating device.

As used herein, term " processor " is not limited only to be used as processor, mobile processor involved in this field Or those of microprocessor integrated circuit, but be broadly directed to microcontroller, microcomputer, programmable logic controller (PLC), specially With integrated circuit and any other programmable circuit.

In one embodiment, interface 15, which provides, is used as network interface card (NICs).In general, NICs control is in computer network Data packet sends and receives on network；Other kinds of interface 15 can be supported for example used in computing device 10 outside other Peripheral equipment.It is that Ethernet interface, Frame Relay Interface, cable interface, DSL interface, token ring connect among the interface that can be provided Mouth, graphic interface etc..It is furthermore possible to also provide various types interface, such as, universal serial bus (USB), serially, with Too net, FIREWIRE^TM,THUNDERBOLT^TM, PCI, parallel, radio frequency (RF), BLUETOOTH^TM, near-field communication (such as using close Field magnet), 802.11 (WiFi), frame relay, TCP/IP, ISDN, fastethernet interface, gigabit ethernet interface, serially ATA (SATA) or outside SATA (ESATA) interface, high resolution multimedia interface (HDMI), Digital Visual Interface (DVI), mould Quasi- or digital audio interface, asynchronous transfer mode (ATM) interface, HSSI High-Speed Serial Interface (HSSI) interface, point of sale (POS) connects Mouthful, light data distribution formula interface (FDDIs) etc..In general, the interface 15 may include being suitable for and suitable media communication Physical port.In some cases, they can also include independent processor (such as special audio or video processor, such as High-fidelity A/V hardware interface is commonly used in this field), and in some cases, volatibility and or nonvolatile memory (such as RAM).

Although system described and shown above is illustrated for implementing the one or more calculating of the present invention described herein One specific framework of device 10 is feature and technology described herein can be implemented thereon at least part of unique absolutely not Device architectures.It is, for example, possible to use the frameworks with one or arbitrary number processor 13, and the processor 13 may exist In single device or it is distributed among arbitrary number device.In one embodiment, single processor 13 processing communication and Laying computation is routed, and discrete private communication processor can be provided in other embodiments.It in various embodiments, can be with Implement different types of feature or function in a system in accordance with the invention comprising client terminal device (such as operation client The board device or smart phone of software) and server system (all server systems as described in more detail below).

Unrelated with network equipment configuration, system of the invention can be using one or more memories or memory module (such as remote memory chunking 16 and local storage 11), be configured to storage for general-purpose network operations data, Program instruction, either other information (or above any combination) related with the function of embodiment described herein.For example, journey Sequence instruction may include operating system and/or one or more application or control its execution.Memory 16 or memory 11, 16 can also configure to storing data structure, configuration data, encryption data, legacy system operation information or described herein Any other dedicated or general non-program information.

Because the information and program instruction can be used to implement one or more systems or method described herein, at least Some network equipment embodiments may include non-provisional machine-readable storage media, such as can configure or be designed to store Program instruction, status information etc. are for executing various operations described herein.The non-provisional machine-readable storage media Example includes but is not limited to magnetic media such as hard disk, floppy disk and tape；Optical medium such as CD-ROM disk；Magnetic-light medium is all If CD and specialized configuration are to the hardware device for storing and executing program instructions, such as ROM device (ROM), Flash memory (as common in mobile device and integrated system), solid-state drives (SSD) and can be in single hardware device " mixing SSD " storage device of the physical unit of middle combination solid-state and hard drive, memristor memory, random access storage device (RAM), etc..It should know, which can be (being such as soldered on motherboard or another for entirety and non-removable The RAM hardware module that is integrated in electronic device outside) or they can be removable such as pluggable flash memory mould Block (such as " thumb drives " or other removable media designed for fast exchange physical storage device), " hot-swappable " is hard Optical storage disc or other this removable media, and the whole and removable storage can be removed in dish driving or solid-state driving Medium can be utilized interchangeably.The example of program instruction includes the object code that can be such as generated by compiler, such as may be used It, such as can be by such as JAVA with the machine code generated by assembler or linker^TMCompiler generates and Java can be used The syllabified code that virtual machine or equivalent execute, or the more advanced code comprising that can be executed by computer using interpreter File (such as with Python, Perl, Ruby, Groovy or script of other any scripting languages).

In some embodiments, the system according to the present invention can be implemented on distinct computing systems.Referring now to fig. 11, Block diagram is shown, the exemplary frame of one or more embodiments or its component on stand alone computer system is depicted Structure.Computing device 20 includes processor 21, can run and perform the one or more functions or application of the embodiment of the present invention Software, such as client application 24.Processor 21 can execute computations under the control of operating system 22, behaviour Make the WINDOWS of system such as Microsoft^TMOperating system version, Mac OS/X or the iOS operating system of apple are some various (SuSE) Linux OS, the ANDROID of Google^TMOperating system etc..In many cases, one or more public services 23 can Can operate in system 20, and can help for providing public service to client application 24.Service 23 can be such as It is WINDOWS^TMIt services, any other class used in the user's space public service or operating system 21 in Linux environment Type public service framework.Input unit 28 can be any type for being suitable for receiving user's input, including such as keyboard, touch Screen, microphone (such as voice input), mouse, touch pads, trace ball, or any combination thereof.Output device 27 can be suitable For providing any type of output to one or more users, no matter being that remotely-or locally, and can wrap for system 20 Include one or more screens for example for visual output, loudspeaker, printer, or any combination thereof.Memory 25 can be Random access storage device with arbitrary structures known in the art and framework is soft to run to be used by such as processor 21 Part.Storage device 26 can be to be deposited for any magnetism, optics, machinery, memristor or the electricity of storing data in digital form Storage device (such as those described above).The example of storage device 26 includes flash memory, magnetic hard disk drives, CD-ROM, and/or similar 's.

In some embodiments, system of the invention can be implemented on distributed computing network, such as with arbitrary number The network of mesh client and/or server.Referring now to fig. 12, block diagram is shown, is depicted in distributed computing net Implement at least part of exemplary architecture 30 according to the system in the embodiment of the present invention on network.According to embodiment, can provide Arbitrary number client 33.Each client 33 can run the software of client part for carrying out the present invention；Client It may include all systems 20 as shown above.It is furthermore possible to also provide arbitrary number server 32 is for handling from one or more The received request of a client 33.Client 33 and server 32 can be in communication with each other via one or more electric networks 31, Electric network can be internet, wide area network, mobile telephone network (such as CDMA or GSM Cellular Networks in various embodiments Network), wireless network (such as WiFi, Wimax etc.) or local area network (or any network topology known in the art really；This hair Bright any one network topology preferred not on any other network) any one.Network 31 can be used it is any Know network protocol and implement, including such as wiredly and/or wirelessly agreement.

In addition, in some embodiments, when needing to obtain additional information or being related to the extra data about specific call Server 32 can service 37 with call outside.Communication with external service 37 can for example be sent out via one or more networks 31 It is raw.In various embodiments, external service 37 may include that network that is related to hardware device or being itself mounted at thereon is enabled Services or functionalities.For example, implement in the embodiment on smart phone or other electronic devices in wherein client application 24, Client application 24, which can obtain, to be stored in the cloud in server system 32 or is being deployed in specific enterprise or user's house property One or more on external service 37 on information.

In some embodiments of the invention, client 33 or server 32 (or both) can be used can locally or One or more service-specifics or utensil remotely across the deployment of one or more networks 31.For example, one or more databases 34 can be used or be related to by one or more embodiments of the invention.It should be appreciated by those skilled in the art that database 34 can To be arranged with various frameworks extensively and using extensive various data access and manipulating mechanism.For example, in various embodiments one A or multiple databases 34 may include using the relational database system of structured query language (SQL), and other can wrap It includes alternate data memory technology such as this field and is referred to as those of " NoSQL " (such as Hadoop Cassandra, Google BigTable, etc.).It in some embodiments, can be with deformation data library used according to the invention framework such as towards the number of column According to library, the database of interior storage, Cluster Database, distributed data base, or even can be with flat file used according to the invention Data warehousing.Those skilled in the art should know, can be as suitably using any group known or Future Data library technology It closes, unless the arrangement of particular database technology or particular component is exclusively used in particular embodiment in this.In addition, should know, such as Term " database " used herein can be related to physical database machine, the clusters of machines as single database system, Or the logical data base in entire data base management system.Unless particular meaning is exclusively used in the given use of term " database " On the way, it should it is construed as to imply that any of these meanings of word, it is all these to be interpreted as those skilled in the art for term " number According to library " obvious meaning.

Similarly, most of embodiments of the invention can use one or more security systems 36 and configuration system 35. Safety and configuration management are public information technology (IT) and network function, and each some amount usually with any IT or network System is associated.It should be appreciated by those skilled in the art that this field now or future known to any configuration or secure subsystem It unrestricted can be used in combination with the embodiment of the present invention, unless by the description particular/special requirement spy of any specific embodiments Different safety 36 or configuration system 35 or scheme.

Figure 13 shows the exemplary overview of the computer system 40 of any various positions as can be used for distributed throughout system. It is that can execute code to handle the example of any computer of data.Computer system 40 can be made it is each modification and Change and without departing from the wider wide scope of system and method disclosed herein.Central processing unit (CPU) 41 is connected to always Line 42, bus be also connected to memory 43, nonvolatile memory 44, display 47, input/output (I/O) unit 48 and Network interface card (NIC) 53.I/O unit 48 can be generally attached to keyboard 49, pointing device 50, hard disk 52 and it is real-time when Clock 51.NIC 53 is connected to network 54, can be internet or local network, and wherein local network can have or not have To the connection of internet.Be again shown as the component of system 40 is the electricity for being connected to main alternating current (AC) power supply 46 in this example Source unit 45.The battery that may exist and widely known but be not practicable for current system disclosed herein is not shown With many other devices and modification of the special innovative function of method.Should know, can combine it is some or all of shown in Component, such as in various integrated applications, such as Qualcomm or Samsung system on chip (SOC) device, or no matter what When as long as it can be adapted for multiple performances or function being incorporated into single hardware device (for example, in mobile device such as intelligence It can phone, video game console, the navigation in interior computer system such as automobile or multimedia system or other are integrated hard In part device).

In various embodiments, the function of system or method for carrying out the present invention can be distributed in arbitrary number client Among end and/or server component.Such as, it is possible to implement various software modules are used to execute the various function combined with the present invention Can, and the implementation that the module can be various is to operate in server and/or client.

The possibility for knowing above-mentioned each embodiment is modified range by those skilled in the art.Therefore, the present invention is wanted by right It asks and its equivalent form limits.

Claims (18)

1. a kind of for carrying out accurate to the system with large-scale and complex data collection using distributed simulation engine and building in detail The system of mould, comprising:

Business data search engine is stored in the memory of computing device and operates on the processor of the computing device；

Business data analysis engine is stored in the memory of computing device and operates on the processor of the computing device； And

Automation plan and numerical estimation module of taking risks, are stored in the memory of one or more computing devices and operate in On the processor of one or more of computing devices；

The result of the action analog module is stored in the memory of one or more computing devices and operates in one or more of On the processor of computing device；

Wherein, the business information search engine:

(a) from the multiple business related datas of multiple Source Search；

(b) from human-computer interface device or from one or more order and control storage device directly receive multiple analysis parameters with Control command；And

(c) the retrieved information of storage accumulation is by data analysis engine processing or scheduled data time-out；

The wherein business information analysis engine:

(d) multiple data types are retrieved from the business information search engine；And

(e) based on the special objective and demand illustrated in current active by business processes engine analysis author to number detected According to the multiple analytic functions of execution and transformation；

Wherein the automation plan and take risks numerical estimation module:

(f) using the analysis of the data as performed by the business information analysis engine and the result converted and from multiple sources Using supplementary data and any current active special purpose machinery study, from business processes analysis author order and parameter Together to illustrate Current commercial plan and risk status report；And

(g) using the analysis of the data as performed by the business information analysis engine and the result converted and from multiple sources Using supplementary data, any current active specific command from business processes analysis author and parameter and from machine The input that learning algorithm is collected is together to be delivered to first eventually for business decision path simulation and numerical value support of commercially taking risks End subscriber；

Wherein, the result of the action analog module:

(h) at least part for the result analyzed and converted by the data that the business information analysis engine executes is retrieved；

(i) at least a piece of of the initial data from the business information search engine is retrieved；

(j) using the multiple parameters inputted from the automation plan and numerical estimation module of taking risks；

(k) using obtained information to execute the prognosis modelling in business speculation or business decision progress path and result such as by simulating Author is primitively initialized using analogy method, and the analogy method is at least one simulative example combined system dynamic side Method, discrete event method or the method based on agency；

(l) use movement profile data and configuration parameter group with create the computer based model of real-world item with It is acted in the simulation.

2. system according to claim 1, wherein described to be stored in the memory of computing device and operate in calculating dress Business information search engine on the processor set uses the port for human-computer interface device, and business related data is input to institute State at least part of port, and described in order relevant to the alternative carry out of Current commercial speculative activities and parameter be input to At least another part of port.

3. system according to claim 1, wherein the automation plan and numerical estimation module of taking risks are using at least Reliably predict to be currently based on the future outcomes of data before having analyzed of business decision based on the statistical analysis of information theory.

4. system according to claim 1, wherein the automation plan and numerical estimation module of taking risks are using at least Monte Carlo heuristic model takes risks Numerical Principle reliably to estimate data before being currently based on having analyzed of business decision Future takes risks numeric number.

5. system according to claim 1, wherein it is special that the automation plan and numerical estimation module of taking risks use Design is created during business decision interpretation of result based on the data storage service of figure with efficiently storing and manipulating Large data structure.

6. system according to claim 1, wherein the automation plan and numerical estimation module of taking risks are with operation Control function allows the operation run in single iteration using single parameter sets, and including successive ignition and using eventually Only the set of the predefined parameter set of criterion is to stop the operation executing when analysis result needed for obtaining.

7. system according to claim 6, wherein some operations are run offline with similar batch-mode, other operations exist Line is run with interactive mode, wherein the input of the result user based on iteration before is used for the parameter of successive iterations, until pre-designed Analysis result terminate execute.

8. system according to claim 1, wherein at least one described simulation includes for disaster, fragility, contract rule The model of definition business and financial capital loss.

9. system according to claim 1, wherein the automation plan and take risks numerical estimation module act on to To modify the model before the simulation on a few computer based model.

10. it is a kind of for the collection of fully integrated business impact data, the analysis of the data and analysis-driven business decision and The method of the generation of the analysis-driven simulation of spare candidate's business decision, comprising steps of

A) it is accessed using the client for being stored in the memory of computing device and operating on the processor of the computing device Interface business decision parameter and target；

B) it is examined using the business data for being stored in the memory of computing device and operating on the processor of the computing device Index is held up from the multiple business datas of multiple Source Search；

C) using the business data point for being stored in the memory of computing device and operating on the processor of the computing device The business data retrieved is transformed to the form for helping that business decision is supported to support by analysis engine；

D) it uses and is stored in the memory of one or more computing devices and operates in one or more of computing devices The result of the action analog module on processor is from the simulation model that can apply business data creation real world objects；

E) predict that scheduled business is determined by simulating the decision or congenial winning using the result of the action analog module Plan or the result of business speculation candidate and estimation are attached to the numerical value of taking risks of each candidate.

11. described to be stored in the memory of computing device and operate in calculating according to the method described in claim 10, wherein Business information search engine on the processor of device uses the port for human-computer interface device, and data relevant to business are defeated Enter at least part of the port, and order relevant to the carry out of Current commercial speculative activities alternate item and parameter input To at least another part of the port.

12. according to the method described in claim 10, wherein, the automation plan and numerical estimation module of taking risks are using extremely Few statistical analysis based on information theory is reliably to predict to be currently based on the following knots of data before having analyzed of business decision Fruit.

13. according to the method described in claim 10, wherein, the automation plan and numerical estimation module of taking risks are using extremely Few Monte Carlo heuristic model takes risks Numerical Principle reliably to estimate data before the analysis for being currently based on business decision Future take risks the number of numerical value.

14. according to the method described in claim 10, wherein, the automation plan and numerical estimation module of taking risks use spy That very designs is created during business decision interpretation of result based on the data storage service of figure with efficiently storing and manipulating Large data structure.

15. according to the method described in claim 10, wherein, the automation plan and numerical estimation module of taking risks have work Industry control function is allowed the operation operated in single iteration using single parameter sets, and including multiple iteration and had The set of the predefined parameter set of stop criterion is to stop the operation executing when analysis result needed for obtaining.

16. according to the method for claim 15, wherein some operations are with similar batch mode off-line operation, other operations With interactive mode on-line operation, wherein user is used for the parameter of successive iterations until pre-designed based on the result input of iteration before It analyzes result and terminates execution.

17. according to the method described in claim 10, wherein, at least one simulation includes being used for disaster, fragility, contractual specifications Obligation and the model of financial capital loss.

18. system according to claim 10, wherein the automation plan and numerical estimation module of taking risks act on At least one computer based model before the simulation to modify the model.