EP1627354A2 - System and method of predictive modeling for managing decisions for business enterprises - Google Patents
System and method of predictive modeling for managing decisions for business enterprisesInfo
- Publication number
- EP1627354A2 EP1627354A2 EP04751132A EP04751132A EP1627354A2 EP 1627354 A2 EP1627354 A2 EP 1627354A2 EP 04751132 A EP04751132 A EP 04751132A EP 04751132 A EP04751132 A EP 04751132A EP 1627354 A2 EP1627354 A2 EP 1627354A2
- Authority
- EP
- European Patent Office
- Prior art keywords
- business
- infrastructure
- components
- dynamic characteristics
- predictive model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
Definitions
- a business infrastructure includes technological as well as non-technological components.
- Technological components include hardware and software components that implement an underlying information technology (IT) infrastructure for one or more business processes.
- Non-technological components include the organizational structure and operation of the business ente ⁇ rise and manual business processes.
- the architectures of business infrastructures are generally required to handle varying degrees of workload and priorities under imposed business constraints.
- the present invention provides a system and method of predictive modeling for managing decisions for business enterprises.
- a system and method of predictive modeling includes (i) providing a description of mfrastructure components that support one or more business solutions in a business enterprise; (ii) from the description, generating a predictive model that mathematically expresses dynamic characteristics and behavior of the infrastructure components; (iii) generating performance metrics from the predictive model for each of the infrastructure components; and (iv) translating the performance metrics into enterprise decision metrics or indicators that correspond to service, performance and financial states of the business enterprise.
- the predictive model mathematically expresses the dynamic characteristics and behavior of each infrastructure component as including direct effects from a first set of infrastructure components and indirect effects from a second set of infrastructure components.
- the predictive model is generated by, for each infrastructure component, (i) generating a mathematical expression of the dynamic characteristics and behavior of the infrastructure component from direct effects of the first set of infrastructure components and (ii) perturbing the mathematical expression of the dynamic characteristics and behaviors of the infrastructure component by including indirect effects of the second set of infrastructure components.
- the predictive model includes a number of layers including an enterprise layer that translates performance metrics generated from the predictive model into enterprise decision metrics or indicators that correspond to the service, performance and financial states of the business enterprise.
- the predictive model can also include business, application, and system layers.
- the business layer mathematically expresses the dynamic characteristics and behaviors of business processes that support each business solution.
- the application layer mathematically expresses the dynamic characteristics and behaviors of software components that support one or more of the business processes in the business layer.
- the system layer mathematically expresses the dynamic characteristics and behaviors of hardware components that support one or more of the software components in the application layer. Performance metrics can be generated for each of the infrastructure components at the business, application, and system layers of the predictive model.
- a system and method of predictive modeling includes (i) providing a description of infrastructure components that support one or more business solutions; (ii) from the description, generating a predictive model that mathematically expresses dynamic characteristics and behavior of the infrastructure components.
- a mathematical expression is generated for each ii frastructure component that expresses the dynamic characteristics and behavior of the infrastructure component based on direct effects of the first set of infrastructure components.
- Each mathematical expression is then perturbed by including indirect effects of the second set of infrastructure components.
- the predictive model includes a plurality of layers. For each of the one or more business solutions, business, application, and system layers of the predictive model are generated from the description. Each layer mathematically expresses the dynamic characteristics and behavior of each infrastructure component associated with the layer as including direct effects from a first set of infrastructure components and indirect effects from a second set of infrastructure components; and generating performance metrics for each of the infrastructure components at each layer.
- Fig. 1 is a high level diagram illustrating a method of predictive modeling for managing decisions of a business enterprise according to one embodiment
- Fig. 2 is a schematic diagram that illustrates a system of predictive modeling for managing decisions for business enterprises according to one embodiment
- Fig. 3 is a diagram illustrating a set of parameters that describe the enterprise of a business infrastructure according to one embodiment
- Figs. 4A, 4B and 4C are diagrams illustrating a method for providing the descriptive input that represents a business enterprise and its businesses according to one embodiment
- Fig. 5 is a diagram illustrating an underlying infrastructure of an enterprise business solution according to one embodiment
- Figs. 6A-6H are diagrams that represent methods for providing descriptive input regarding a business infrastructure according to one embodiment
- Fig. 7 is a conceptual diagram illustrating the layers represented in a predictive model according to one embodiment
- Fig. 8A-8F are diagrams illustrating user interfaces that express performance metrics at business, application, and system infrastructure layers of a predictive model according to one embodiment
- Figs 9A-9D is a diagram illustrating a table that describes a set of enterprise decision metrics according to one embodiment
- Fig. 10 is a diagram illustrating combinations of enterprise decision metrics to generate general indicators of the health of a business enterprise according to one embodiment
- Fig. 11 is a schematic diagram illustrating direct and indirect perturbation effects on infrastructure components within a business enterprise according to one embodiment
- Fig. 12 is a diagram illustrating a method of improved predictive modeling using perturbation theory according to one embodiment.
- Fig. 13 is a diagram of an operating environment for embodiments of the invention.
- a business executive of an enterprise may make decisions that can affect its success. Unfortunately, many times these decisions are made based on a partial view of the business infrastructure. Wrong decisions are likely to have a negative impact on the success of the enterprise.
- Examples of such decisions include a decision to (i) offer a new business solution to customers, (ii) merge with another company to increase geographic presence, (iii) acquire the assets of another company, (iv) enter into a service level agreement (SLA) with guaranteed levels of service, (iv) set the premium on an insurance policy based on risk of a business failure.
- SLA service level agreement
- Each of these decisions require a strong understanding of the dynamic characteristics and behavior of the business infrastructure in order to appreciate the impact of a decision on the service, performance and financial state of a business enterprise.
- the present invention is directed to a system and method of predictive modeling for managing decisions for a business enterprise.
- performance metrics of a business infrastructure are generated from a predictive model and then translated into enterprise decision metrics that correspond to the service, performance and financial states of the business ente ⁇ rise.
- a business executive can evaluate, support, and monitor the effect of decisions on the ente ⁇ rise based on different designs and scenarios of a business infrastructure.
- the accuracy of the predictive modeling is improved by mathematically expressing the dynamic characteristics and behavior of each infrastructure component as a result of direct and indirect effects of the infrastructure components impacting one another.
- Perturbation theory is a branch of mathematics that can be used to express such behavior.
- Fig. 1 is a high level diagram illustrating a method of predictive modeling for managing decisions of a business ente ⁇ rise according to one embodiment.
- the method includes providing a description of the business infrastructure 20 that supports one or more business solutions.
- the description 20 includes parameters that represent the ente ⁇ rise, business and technological components of the business infrastructure.
- a predictive model 20 is constructed that mathematically expresses the dynamic characteristics and behaviors of the infrastructure components.
- the dynamic characteristics and behaviors include time delays, locks and contentions associated with each infrastructure component over time.
- ente ⁇ rise decision metrics 30 are generated for diagnostics, action planning and achievement evaluation.
- the ente ⁇ rise decision metrics translate performance metrics of the business infrastructure in terms that a business executive can understand.
- 'the ente ⁇ rise decision metrics may represent the impact of a particular business infrastructure in terms of profitability, productivity, growth, and risk.
- Decision monitoring and management 40 is based on the decision metrics 30 and may involve modifying the design of the business infrastructure at any level (i.e., ente ⁇ rise, business, technology) or evaluating the business infrastructure in different scenarios for stress and component level sensitivity testing.
- Fig. 2 is a schematic diagram that illustrates a system of predictive modeling for managing decisions for business ente ⁇ rises according to one embodiment.
- the system includes an input module 60, a model construction module 70, a component model library 75, a metric calculation module 80, and an output module 90.
- the input module 60 receives descriptive input 10 from an input device, a network, or a storage device.
- the descriptive input 10 includes parameters that represent the components, characteristics, constraints, drivers and scenarios of a business infrastructure.
- the input module 60 passes the descriptive input 10 to the construction module 70 that generates a predictive model 20 of the business infrastructure from the description.
- the construction module 70 is coupled to a component model library 75.
- the library 75 includes a number of premodeled components with each component model mathematically expressing the dynamic characteristics and behavior of a particular infrastructure component.
- the descriptive input 10 may include component index values that enable the construction module 70 to select the particular component models from the library 75.
- the descriptive input 10 may also include parameter values that can be input as variables to the mathematical expressions of the selected component models. These parameter values may specify particular configurations of the components (e.g., partitioning of hardware servers) or external drivers and constraints (e.g., business workload events).
- the descriptive input 10 also identifies relationships between infrastructure components in order to combine the mathematical expressions of dependent infrastructure components. For example, the response time of an application component to a request depends on the speed of the supporting processor.
- the construction module 70 then passes the predictive model 20 onto the metric calculation module 80 in order to generate the ente ⁇ rise decision metrics 30 from the predictive model 20.
- the decision metrics 30 are then forwarded to the output module 90, which provides the decision metrics 30 to an output device, a network or a storage device.
- the output module provides the ente ⁇ rise decision metrics 30 to a display device for the designer of the business infrastructure.
- the ente ⁇ rise decision metrics 30 represent the service, performance, and financial state of the business ente ⁇ rise.
- Fig. 3 is a diagram illustrating a set of parameters that describe the ente ⁇ rise of a business infrastructure according to one embodiment.
- the descriptive parameters of the ente ⁇ rise include parameters representing volumes and dynamics 110, ente ⁇ rise and business organization 112, time windows corresponding to peaks and valleys in business activity 114, evolution and disruptive conditions 116, competition benchmarks and pressure 118, technological and non- technological characteristics 120, service quality (e.g., SLA) 122, cost including operations and maintenance costs 124, assets and cycles 126, and constraints on margin and planning 128.
- SLA service quality
- such ente ⁇ rise parameters are provided as descriptive input 10 to the input module 40. Some of these ente ⁇ rise parameters are subsequently passed as input to the model construction module 50 in order to generate the predictive model 20, including the ente ⁇ rise and business organization 112, technological and non-technological characteristic parameters 120, and cost including operations and maintenance costs 124, for example. Other ente ⁇ rise parameters are forwarded to the metric calculation module 80 in order to represent the conditions for a particular scenario, including volumes and dynamics 110 and time windows 114, for example.
- Figs. 4 A, 4B and 4C are diagrams illustrating a method for providing the descriptive input that represents a business ente ⁇ rise and its businesses according to one embodiment.
- Fig. 4A includes a graphical representation 150 of a business ente ⁇ rise that is organized into a central headquarters 160 and remote business units 162, 164, 166, 168.
- this representation can be generated by any UML (Unified Modeling Language) graphics tool, for example, by "dragging and dropping" icons that represent the business units into the graphical representation and identifying communication flows between the units with arrow icons.
- UML Unified Modeling Language
- Specific parameters can be associated with each business unit by selecting one of the business unit icons and entering data (e.g., values, component models, links to other infrastructure components) through a sub-user interface 155 associated with a selected business unit.
- data e.g., values, component models, links to other infrastructure components
- the sub-user interface can be triggered for display to the user by "double-clicking" on a particular business unit icon.
- a business ente ⁇ rise may be organized in a hierarchical manner that includes multiple levels of organizational granularity.
- Figs. 4 A through 4C illustrate descriptive input that represents the hierarchical organization of a business ente ⁇ rise according to one embodiment.
- these graphical representations can be generated by any UML (Unified Modeling Language) graphics tool with specific parameters being entered through sub-user interfaces, for example.
- Fig. 4B includes a graphical representation of a departmental organization within a business unit 162.
- the departmental organization of business unit 162 includes a human resource department 172, a customer support department 174, and an IS department 178. Each department, in turn, implements a number of business solutions. As shown in Fig. 4C, the human resources department 172 handles employment programs 182 and employee benefits programs 184.
- the resulting predictive model can determine performance and ente ⁇ rise decision metrics across each business unit, department, and business solution.
- a business executive can analyze the effects of particular decision on the business ente ⁇ rise at the different organizational levels.
- Specific parameters associated with a business solution may include technological characteristic parameters 120 (Fig. 3).
- the technological characteristic parameter indicates that the business solution includes technological components
- the business solution can be further represented by an underlying information technology (IT) infrastructure.
- IT information technology
- Fig. 5 is a diagram illustrating an underlying infrastructure of an ente ⁇ rise business solution according to one embodiment.
- the employee benefits program 184 and the employment program 182 of Fig. 4C may each be implemented by three business processes 190a, 190b, and 190c and business processes 220a, 220b and 220c, respectively.
- Each set of business processes is supported by a set of technological components.
- business process 190a is implemented by software component 200a, which is further supported by hardware component 210.
- Software components can be shared among different business processes and across business solutions.
- software component 230a supports business processes 190c and 220a, which implement different business solutions 182 and 184.
- the business process 220c is a manual business process that does not have a technological infrastructure and thus can be represented as a manual delay. For pu ⁇ oses of clarity, only one software component and one hardware component are shown as supporting each business process. However, it should be understood that each business process may be implemented by any number of hardware and software components and the predictive modeling tool is able to express and model each component and their interactions among each other.
- the descriptive input for representing each of the business, application and system (i.e., hardware/network) infrastructure components can be input to the predictive modeling system in a number of ways.
- the business processes can be input through a graphical user interface by "dragging and dropping" icons that represent business processes and their interactions.
- Specific parameters can be associated with each business process by selecting one of the business process icons and entering data (e.g., values, application component models, links to other infrastructure components) through a sub-user interface associated with a selected business unit.
- the sub-user interface can be triggered for display to the user by "double-clicking" on a particular business process icon.
- one or more software component models can be selected from a library of component models 75 to implement the business process.
- These software component models may mathematically express of the dynamic characteristics and behavior of accounting programs, billing programs, and other programs.
- component models that represent the hardware component supporting the software components can be selected. Specific parameters can also be provided that describe particular configurations of the hardware and software components.
- Figs. 6A-6H are diagrams that represent methods for providing descriptive input regarding a business infrastructure according to one embodiment.
- descriptive input for the business, application, and system infrastructures, refer to U.S. Application Serial No. 10/014,317, filed October 26, 2001, entitled “System and Method for Improving Predictive Modeling of an Information System," the entire contents of which are inco ⁇ orated herein by reference.
- a predictive model 20 is generated that mathematically expresses the dynamic characteristics and behavior of the infrastructure components individually and in combination.
- Fig. 7 is a conceptual diagram illustrating the layers represented in a predictive model according to one embodiment.
- the predictive model 20 represents the business infrastructure as a number of layers including an ente ⁇ rise layer 310, a business layer 320, an application layer 330, and a system layer 340.
- each layer models the dynamic characteristics and behavior of its components individually and collectively in terms of probabilities for delays due to processing, conflicts, contentions and locks.
- Each layer has an effect on the dynamic characteristics and behavior expressed in the other layers as indicated by the arrows.
- the model can be used to predict the service, performance and financial state of a business infrastructure components at each layer.
- the business workload events generated at the ente ⁇ rise layer dictate the number of jobs or tasks to be performed at the business layer.
- the volume of jobs or tasks translates to a number of requests/responses by software components at the application layer, which in turn translates into I/O transactions at the system infrastructure layer.
- the rate at which the business events can be processed depends on a number of factors including the architectural design of the business infrastructure. Specifically, delays occur at each layer due to processing, contentions for resources, locks, and conflicts.
- Fig. 8A-8F are diagrams illustrating user interfaces that express performance metrics at business, application, and system infrastructure layers of a predictive model according to one embodiment.
- each infrastructure component is associated with a financial cost.
- infrastructure components at the application and system layers can be associated with operational and maintenance costs.
- the financial costs can include salaries and capital costs, for example.
- a revenue can be associated with processed business events.
- Figs. 9A-9D are diagrams illustrating tables that describe a set of ente ⁇ rise decision metrics according to one embodiment.
- the decision metrics 30 can include (1) elongation ratio, (2) unit utilization ratio, (3) ceiling threshold, (4) business response time, (5) aging ratio, (6) degradation ratio, (7) non-productive ratio, (8) non-productive ratio, (9) process latency, (10) cost inflation rate, (11) margin erosion rate, (12) total-added-cost ratio, (13) uncovered-value ratio, (14) under-utilization ratio, (15) operational risk ratio, (16) inefficiency ratio, and (17) lifetime contraction ratio.
- Some of these metrics can also represent behavior at the underlying infrastructure layers as well.
- combinations of these metrics can be combined mathematically to provide general indicators of the health of a business ente ⁇ rise as shown in Fig. 10.
- Fig. 10 is a diagram illustrating combinations of ente ⁇ rise decision metrics to generate general indicators of the health of a business ente ⁇ rise according to one embodiment.
- combinations of the ente ⁇ rise business metrics can provide indications of productivity 410, profitability 420, growth 430 and risk 440 associated with a particular business infrastructure.
- the profitability 420 of a business infrastructure can be represented as a combination of elongation ratio, ceiling threshold, business response time, degradation ratio, cost inflation rate, margin erosion rate, total added cost ratio, uncovered value ratio, and under- utilization ratio. Therefore, if the profitability corresponds to a low value, the designer of the business infrastructure may analyze each of the constituent metrics to determine wliich metric(s) are causing the reduction in profitability.
- the normalized performance coefficient is a global indicator of a business infrastructure to meeting business requirements.
- the NPC is the product of the profitability, productivity, and growth indicators divided by risk.
- Other global indicators can include (1) Effective Cost Advantage, (2) Effective Value Advantage, (3) Effective Response Elongation, and (4) Overall Scalability Index.
- Effective Cost Advantage is the percentage of cost inflation (or deflation) that will yield the business growing for at least the same proportion. For example, if cost is increased by 30% due to re-engineering the business ente ⁇ rise, the revenue stream will consequently increase by at least 30%.
- Effective Value Advantage represents the growth in value between two points where major change in revenue results from investing in moving the system dynamics between the two points.
- the investment might change the configuration in short, medium, or long terms but the Effective Value Advantage will represented a normalized parameter that expresses the resultant earnings per share, profit, assets, market mane, and total return to investors.
- Effective Response Elongation is an indicator of the delays that are independent of service requirements. By definition, this indicator is the time to deliver divided by tht time to execute without contention or conflict minus one.
- the Effective Response Elongation can be used to assess the scope of improvement opportunities. When its value is close to zero, there are few improvement opportunities that can take place. If the value is higher than one, serious actions must be taken in order to improve the service and avoid rapid degradation.
- Overall Scalability Index is a measure of the ability of structure to replicate the same cost and performance characteristics each time addition load will be in service.
- the index varies from 1 to 100, where 100 corresponds to full replication and 0 corresponds to no replication.
- a scalability index of 0.85 means that we need 15% addition resource contingencies added to replicate percentages with the new workload.
- a fully scalable system will closely replicate the same workload growth.
- Embodiments of the present invention can be applied in a number of situations that involve decisions that can significantly impact the success of a business ente ⁇ rise.
- embodiments of the invention can be used during outsourcing deals. In a typical outsourcing deal, three stages are involved: (1) a pre-assessment of the client environment, processes, resources and assets;
- the predictive model will show limits at each layer (ente ⁇ rise, business, IT infrastructure) and identify issues to assess during the first stage.
- the predictive model can reveal a true and fact based representation of the client environment, processes, resources and assets. Scenario analysis of the predictive model results in ceilings and therefore contingencies being determined for the following stages of the outsourcing deal.
- the predictive model can also be used to support the engineering effort by defining improvement actions and enhancement trails in order to increase margin and justify business cases for the third stage.
- the accuracy of the predictive modeling is improved by mathematically expressing the dynamic characteristics and behavior of each infrastructure component as a result of direct and indirect effects of the infrastructure components impacting one another.
- Perturbation theory is a branch of mathematics that can be used to express such behavior.
- the metrics that represent the critical behavior and evolution of ente ⁇ rise, business, application and system infrastructures and that support related decisions are continuously perturbed by a variety of effects of different nature. These effects are small if taken separately but can become large if they are taken collectively.
- the metrics associated with each infrastructure component can be impacted significantly through direct and indirect perturbation effect.
- Fig. 11 is a schematic diagram illustrating direct and indirect perturbation effects on infrastructure components within a business ente ⁇ rise according to one embodiment.
- infrastructure component cl interacts directly with components c2 and c3.
- the performance of component cl is directly effected by components c2 and c3.
- component cl does not directly interact with component c4
- component c4 indirectly effects the performance of component cl because it contends for resources on components c2 and c3.
- component el's ability to access those resources are indirectly impacted by component c4.
- component c3 also contends for resources on component c2.
- component c3 has an indirect effect on component cl as well as a direct effect.
- Figs. 9A-9D identify the possible causes of direct and indirect perturbation effects on each of the ente ⁇ rise business metrics.
- Fig. 12 is a diagram illustrating a method of improved predictive modeling using perturbation theory according to one embodiment.
- a mathematical expression of the dynamic characteristics and behavior of the component based on direct effects of a first set of infrastructure components, including the component itself, can be generated.
- Analytically different causes can be expressed either directly or indirectly impacting a perturbing function in proper mathematics that are the best fit to represent such function.
- the mathematical expression may be based on queuing, stochastic or general probability theory to express and determine service impact on response time and cost.
- Deterministic mathematics or set algebra may be used to express the availability of resources to process all requests.
- Mean value analysis or again stochastic processes may also be used to represent latencies and delays for complex communication with external systems.
- the mathematical expression is perturbed by including indirect effects of a second set of infrastructure components.
- Metric calculation can be used at each layer of the predictive model.
- analytically what will be obtained from the lower levels are functions of time that represent different contributions to the perturbation problem at the top level (e.g., ente ⁇ rise layer). Numerically this might be cost contributions to an overall cost inflation for a certain point of time.
- the problem will be reduced to a value of a parameter, such as response time or throughput, but its robustness and validity in time will be largely reduced to a single point.
- this last case is predominant in the current evaluation methods and consequently their inability to guarantee a viable decision as they ignores higher order perturbations that were truncated too early to allow their inclusions.
- conserving an analytical expression in terms of the direct and indirect effects provides greater validity, better accuracy and larger interval of time and more robust solution.
- the perturbation theory approach involves a dynamic system of Lagrange-like partial differential equations that represent the dynamic behavior of a cost function and a solution that will capture both direct and indirect perturbations around a base of the un-perturbed solution.
- (c) O represents the unperturbed value of a metric, or its minimum admitted value for simplicity
- O represents a measure of a perturbed metric due to the direct impact
- 0 represents the indirect perturbation due to the perturbed affect of
- U(t) is a square matrix (K x K) and v(t) is a known vectorial function.
- the matrix is determined by:
- Fig. 13 is a diagram of an operating environment for embodiments of the invention.
- Computer system 610 includes at least processor 611, for processing information according to programmed instructions, memory 612, for storing infonnation and instructions for processor 611, storage system 613, such as a magnetic or optical disk system, for storing large amounts of information and instructions on a relatively long-term basis, and display system 614, such as a computer monitor, for displaying various graphical elements that facilitate user interaction with computer system 610.
- Processor 611, memory 612, storage system 613, and display system 614 are coupled to bus 615, which preferably provides a high-speed means for devices connected to bus 615 to communicate with each other.
- computer system 610 is illustrative, and that alternative systems and architectures may be used with the present invention. It will further be understood that many other devices, such as a network interface (not shown), and a variety of other input and output devices (not shown) may be included in computer system 610. While this invention has been particularly shown and described with references to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims.
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A system and method is provided for predictive modeling of technical and non-technical components in a business infrastructure that implementing one or more business solutions. According to a first aspect of the invention, performance metrics generated from a predictive model of a business infrastructure are translated into enterprise decision or indicators that correspond to the service, performance and financial states of a business enterprise. As a result, non-technical executives can utilize the enterprise decision metrics or indicators to evaluate, support, and monitor the effect of business decisions, for example, with respect to profitability, productivity, growth, and risk of the business. According to a second aspect of the invention, the accuracy of the predictive modeling is improved by mathematically expressing the dynamic characteristics and behavior of each infrastructure component as a result of direct and indirect effects of the infrastructure components impacting one another. Perturbation theory can be used to express direct and indirect effects.
Description
S YSTEM AND METHOD OF PREDICTIVE MODELING FOR MANAGING DECISIONS FOR BUSINESS ENTERPRISES
RELATED APPLICATIONS
This application is a continuation-in-part of U.S. Application No. 10/014,317, filed October 26, 2001, which is a continuation-in-part of Application No. 09/127,191, filed July 31, 1998 (now U.S. Patent 6,311,144), which claims the benefit of U.S. Provisional Application No. 60/085,350, filed on May 13, 1998. This application also claims the benefit of U.S. Provisional Application No. 60/467,483, filed May 2, 2003. The entire teachings of the above applications are incorporated herein by reference.
BACKGROUND
Business infrastructures are designed to support business enterprises and the solutions they provide. A business infrastructure includes technological as well as non-technological components. Technological components include hardware and software components that implement an underlying information technology (IT) infrastructure for one or more business processes. Non-technological components include the organizational structure and operation of the business enteφrise and manual business processes. The architectures of business infrastructures are generally required to handle varying degrees of workload and priorities under imposed business constraints.
The design of business infrastructures having such requirements and constraints represents a real challenge. Most existing methodologies, tools and techniques concentrate on static, partial descriptions of business infrastructures. Dynamic system behavior is generally unknown until the system is in construction or in operation, thus, limiting the possibilities for improvement. Unacceptable
performance issues may become exacerbated as the system evolves with the addition of new business solutions that must be supported by the architecture.
Furthermore, when the origin of a problem resides in questionable decisions made early in the development process, the cost of improvement could become prohibitive when a redesign of the business infrastructure is required at some level. A tremendous amount of investment may be lost due to the design of unacceptable infrastructures.
SUMMARY
The present invention provides a system and method of predictive modeling for managing decisions for business enterprises.
According to a first aspect of the invention, a system and method of predictive modeling is provided that includes (i) providing a description of mfrastructure components that support one or more business solutions in a business enterprise; (ii) from the description, generating a predictive model that mathematically expresses dynamic characteristics and behavior of the infrastructure components; (iii) generating performance metrics from the predictive model for each of the infrastructure components; and (iv) translating the performance metrics into enterprise decision metrics or indicators that correspond to service, performance and financial states of the business enterprise.
According to particular embodiments, the predictive model mathematically expresses the dynamic characteristics and behavior of each infrastructure component as including direct effects from a first set of infrastructure components and indirect effects from a second set of infrastructure components. In one embodiment, the predictive model is generated by, for each infrastructure component, (i) generating a mathematical expression of the dynamic characteristics and behavior of the infrastructure component from direct effects of the first set of infrastructure components and (ii) perturbing the mathematical expression of the dynamic characteristics and behaviors of the infrastructure component by including indirect effects of the second set of infrastructure components.
According to particular embodiments, the predictive model includes a number of layers including an enterprise layer that translates performance metrics generated from the predictive model into enterprise decision metrics or indicators that correspond to the service, performance and financial states of the business enterprise.
Preferably, the predictive model can also include business, application, and system layers. The business layer mathematically expresses the dynamic characteristics and behaviors of business processes that support each business solution. The application layer mathematically expresses the dynamic characteristics and behaviors of software components that support one or more of the business processes in the business layer. The system layer mathematically expresses the dynamic characteristics and behaviors of hardware components that support one or more of the software components in the application layer. Performance metrics can be generated for each of the infrastructure components at the business, application, and system layers of the predictive model.
According to a second aspect of the invention, a system and method of predictive modeling is provided that includes (i) providing a description of infrastructure components that support one or more business solutions; (ii) from the description, generating a predictive model that mathematically expresses dynamic characteristics and behavior of the infrastructure components. To generate the predictive model, a mathematical expression is generated for each ii frastructure component that expresses the dynamic characteristics and behavior of the infrastructure component based on direct effects of the first set of infrastructure components. Each mathematical expression is then perturbed by including indirect effects of the second set of infrastructure components.
According to particular embodiments, the predictive model includes a plurality of layers. For each of the one or more business solutions, business, application, and system layers of the predictive model are generated from the description. Each layer mathematically expresses the dynamic characteristics and behavior of each infrastructure component associated with the layer as including direct effects from a first set of infrastructure components and indirect effects from a
second set of infrastructure components; and generating performance metrics for each of the infrastructure components at each layer.
BRIEF DESCRIPTION OF THE DRAWINGS
The foregoing and other objects, features and advantages of the invention will be apparent from the following more particular description of preferred embodiments of the invention, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.
Fig. 1 is a high level diagram illustrating a method of predictive modeling for managing decisions of a business enterprise according to one embodiment;
Fig. 2 is a schematic diagram that illustrates a system of predictive modeling for managing decisions for business enterprises according to one embodiment;
Fig. 3 is a diagram illustrating a set of parameters that describe the enterprise of a business infrastructure according to one embodiment;
Figs. 4A, 4B and 4C are diagrams illustrating a method for providing the descriptive input that represents a business enterprise and its businesses according to one embodiment;
Fig. 5 is a diagram illustrating an underlying infrastructure of an enterprise business solution according to one embodiment;
Figs. 6A-6H are diagrams that represent methods for providing descriptive input regarding a business infrastructure according to one embodiment;
Fig. 7 is a conceptual diagram illustrating the layers represented in a predictive model according to one embodiment;
Fig. 8A-8F are diagrams illustrating user interfaces that express performance metrics at business, application, and system infrastructure layers of a predictive model according to one embodiment;
Figs 9A-9D is a diagram illustrating a table that describes a set of enterprise decision metrics according to one embodiment;
Fig. 10 is a diagram illustrating combinations of enterprise decision metrics to generate general indicators of the health of a business enterprise according to one embodiment;
Fig. 11 is a schematic diagram illustrating direct and indirect perturbation effects on infrastructure components within a business enterprise according to one embodiment;
Fig. 12 is a diagram illustrating a method of improved predictive modeling using perturbation theory according to one embodiment; and
Fig. 13 is a diagram of an operating environment for embodiments of the invention.
DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT
A business executive of an enterprise may make decisions that can affect its success. Unfortunately, many times these decisions are made based on a partial view of the business infrastructure. Wrong decisions are likely to have a negative impact on the success of the enterprise.
Examples of such decisions include a decision to (i) offer a new business solution to customers, (ii) merge with another company to increase geographic presence, (iii) acquire the assets of another company, (iv) enter into a service level agreement (SLA) with guaranteed levels of service, (iv) set the premium on an insurance policy based on risk of a business failure. Each of these decisions require a strong understanding of the dynamic characteristics and behavior of the business infrastructure in order to appreciate the impact of a decision on the service, performance and financial state of a business enterprise.
The present invention is directed to a system and method of predictive modeling for managing decisions for a business enterprise. According to one aspect of the invention, performance metrics of a business infrastructure are generated from a predictive model and then translated into enterprise decision metrics that correspond to the service, performance and financial states of the business enteφrise. As a result, a business executive can evaluate, support, and monitor the effect of decisions
on the enteφrise based on different designs and scenarios of a business infrastructure.
According to another aspect of the invention, the accuracy of the predictive modeling is improved by mathematically expressing the dynamic characteristics and behavior of each infrastructure component as a result of direct and indirect effects of the infrastructure components impacting one another. Perturbation theory is a branch of mathematics that can be used to express such behavior.
Fig. 1 is a high level diagram illustrating a method of predictive modeling for managing decisions of a business enteφrise according to one embodiment. The method includes providing a description of the business infrastructure 20 that supports one or more business solutions. Preferably, the description 20 includes parameters that represent the enteφrise, business and technological components of the business infrastructure. From this description, a predictive model 20 is constructed that mathematically expresses the dynamic characteristics and behaviors of the infrastructure components. The dynamic characteristics and behaviors include time delays, locks and contentions associated with each infrastructure component over time. From the predictive model 20, enteφrise decision metrics 30 are generated for diagnostics, action planning and achievement evaluation. In particular, the enteφrise decision metrics translate performance metrics of the business infrastructure in terms that a business executive can understand. For example, 'the enteφrise decision metrics may represent the impact of a particular business infrastructure in terms of profitability, productivity, growth, and risk. Decision monitoring and management 40 is based on the decision metrics 30 and may involve modifying the design of the business infrastructure at any level (i.e., enteφrise, business, technology) or evaluating the business infrastructure in different scenarios for stress and component level sensitivity testing.
By utilizing predictive modeling in decision management processes, business executives can determine whether a proposed business infrastructure will meet the requirements and conditions of the business solutions prior to implementation. Standard methodologies provide no method for validating business infrastructure designs during early stages of development.
Fig. 2 is a schematic diagram that illustrates a system of predictive modeling for managing decisions for business enteφrises according to one embodiment. The system includes an input module 60, a model construction module 70, a component model library 75, a metric calculation module 80, and an output module 90.
The input module 60 receives descriptive input 10 from an input device, a network, or a storage device. Preferably, the descriptive input 10 includes parameters that represent the components, characteristics, constraints, drivers and scenarios of a business infrastructure.
The input module 60 passes the descriptive input 10 to the construction module 70 that generates a predictive model 20 of the business infrastructure from the description. According to one embodiment, the construction module 70 is coupled to a component model library 75. The library 75 includes a number of premodeled components with each component model mathematically expressing the dynamic characteristics and behavior of a particular infrastructure component. The descriptive input 10 may include component index values that enable the construction module 70 to select the particular component models from the library 75. The descriptive input 10 may also include parameter values that can be input as variables to the mathematical expressions of the selected component models. These parameter values may specify particular configurations of the components (e.g., partitioning of hardware servers) or external drivers and constraints (e.g., business workload events). The descriptive input 10 also identifies relationships between infrastructure components in order to combine the mathematical expressions of dependent infrastructure components. For example, the response time of an application component to a request depends on the speed of the supporting processor.
The construction module 70 then passes the predictive model 20 onto the metric calculation module 80 in order to generate the enteφrise decision metrics 30 from the predictive model 20. The decision metrics 30 are then forwarded to the output module 90, which provides the decision metrics 30 to an output device, a network or a storage device. In one embodiment, the output module provides the enteφrise decision metrics 30 to a display device for the designer of the business
infrastructure. The enteφrise decision metrics 30 represent the service, performance, and financial state of the business enteφrise.
Fig. 3 is a diagram illustrating a set of parameters that describe the enteφrise of a business infrastructure according to one embodiment. In particular, the descriptive parameters of the enteφrise include parameters representing volumes and dynamics 110, enteφrise and business organization 112, time windows corresponding to peaks and valleys in business activity 114, evolution and disruptive conditions 116, competition benchmarks and pressure 118, technological and non- technological characteristics 120, service quality (e.g., SLA) 122, cost including operations and maintenance costs 124, assets and cycles 126, and constraints on margin and planning 128.
Referring back to Fig. 2, such enteφrise parameters are provided as descriptive input 10 to the input module 40. Some of these enteφrise parameters are subsequently passed as input to the model construction module 50 in order to generate the predictive model 20, including the enteφrise and business organization 112, technological and non-technological characteristic parameters 120, and cost including operations and maintenance costs 124, for example. Other enteφrise parameters are forwarded to the metric calculation module 80 in order to represent the conditions for a particular scenario, including volumes and dynamics 110 and time windows 114, for example.
Figs. 4 A, 4B and 4C are diagrams illustrating a method for providing the descriptive input that represents a business enteφrise and its businesses according to one embodiment. For example, Fig. 4A includes a graphical representation 150 of a business enteφrise that is organized into a central headquarters 160 and remote business units 162, 164, 166, 168. In one embodiment, this representation can be generated by any UML (Unified Modeling Language) graphics tool, for example, by "dragging and dropping" icons that represent the business units into the graphical representation and identifying communication flows between the units with arrow icons. Specific parameters can be associated with each business unit by selecting one of the business unit icons and entering data (e.g., values, component models, links to other infrastructure components) through a sub-user interface 155 associated
with a selected business unit. For example, the sub-user interface can be triggered for display to the user by "double-clicking" on a particular business unit icon.
Subsequent lower levels of enteφrise organization may also be represented as needed. For example, a business enteφrise may be organized in a hierarchical manner that includes multiple levels of organizational granularity. Figs. 4 A through 4C, in combination, illustrate descriptive input that represents the hierarchical organization of a business enteφrise according to one embodiment. As in Fig. 4A, these graphical representations can be generated by any UML (Unified Modeling Language) graphics tool with specific parameters being entered through sub-user interfaces, for example. In particular, Fig. 4B includes a graphical representation of a departmental organization within a business unit 162. For example, the departmental organization of business unit 162 includes a human resource department 172, a customer support department 174, and an IS department 178. Each department, in turn, implements a number of business solutions. As shown in Fig. 4C, the human resources department 172 handles employment programs 182 and employee benefits programs 184.
By modeling the hierarchical organization of a business enteφrise, the resulting predictive model can determine performance and enteφrise decision metrics across each business unit, department, and business solution. Thus, a business executive can analyze the effects of particular decision on the business enteφrise at the different organizational levels.
Specific parameters associated with a business solution may include technological characteristic parameters 120 (Fig. 3). For example, if the technological characteristic parameter indicates that the business solution includes technological components, the business solution can be further represented by an underlying information technology (IT) infrastructure.
Fig. 5 is a diagram illustrating an underlying infrastructure of an enteφrise business solution according to one embodiment. For example, the employee benefits program 184 and the employment program 182 of Fig. 4C may each be implemented by three business processes 190a, 190b, and 190c and business processes 220a, 220b and 220c, respectively.
Each set of business processes is supported by a set of technological components. For example, business process 190a is implemented by software component 200a, which is further supported by hardware component 210. Software components can be shared among different business processes and across business solutions. For example, software component 230a supports business processes 190c and 220a, which implement different business solutions 182 and 184. The business process 220c is a manual business process that does not have a technological infrastructure and thus can be represented as a manual delay. For puφoses of clarity, only one software component and one hardware component are shown as supporting each business process. However, it should be understood that each business process may be implemented by any number of hardware and software components and the predictive modeling tool is able to express and model each component and their interactions among each other.
The descriptive input for representing each of the business, application and system (i.e., hardware/network) infrastructure components can be input to the predictive modeling system in a number of ways. For example, the business processes can be input through a graphical user interface by "dragging and dropping" icons that represent business processes and their interactions. Specific parameters can be associated with each business process by selecting one of the business process icons and entering data (e.g., values, application component models, links to other infrastructure components) through a sub-user interface associated with a selected business unit. For example, the sub-user interface can be triggered for display to the user by "double-clicking" on a particular business process icon.
For each business process, one or more software component models can be selected from a library of component models 75 to implement the business process. These software component models may mathematically express of the dynamic characteristics and behavior of accounting programs, billing programs, and other programs.
Likewise, for each software component, component models that represent the hardware component supporting the software components can be selected. Specific
parameters can also be provided that describe particular configurations of the hardware and software components.
Figs. 6A-6H are diagrams that represent methods for providing descriptive input regarding a business infrastructure according to one embodiment. For more information regarding the descriptive input for the business, application, and system infrastructures, refer to U.S. Application Serial No. 10/014,317, filed October 26, 2001, entitled "System and Method for Improving Predictive Modeling of an Information System," the entire contents of which are incoφorated herein by reference. From the descriptive input, a predictive model 20 is generated that mathematically expresses the dynamic characteristics and behavior of the infrastructure components individually and in combination.
Fig. 7 is a conceptual diagram illustrating the layers represented in a predictive model according to one embodiment. In particular, the predictive model 20 represents the business infrastructure as a number of layers including an enteφrise layer 310, a business layer 320, an application layer 330, and a system layer 340.
Through a system of equations, each layer models the dynamic characteristics and behavior of its components individually and collectively in terms of probabilities for delays due to processing, conflicts, contentions and locks. Each layer has an effect on the dynamic characteristics and behavior expressed in the other layers as indicated by the arrows.
Based on the system of equations, the model can be used to predict the service, performance and financial state of a business infrastructure components at each layer. For example, the business workload events generated at the enteφrise layer dictate the number of jobs or tasks to be performed at the business layer. The volume of jobs or tasks translates to a number of requests/responses by software components at the application layer, which in turn translates into I/O transactions at the system infrastructure layer. Conversely, the rate at which the business events can be processed depends on a number of factors including the architectural design of the business infrastructure. Specifically, delays occur at each layer due to processing, contentions for resources, locks, and conflicts. Such delays percolate up
fro the system infrastructure layer through the application, business and enteφrise layers, and thus, limit the number of business events that can be processed within a certain time period. Fig. 8A-8F are diagrams illustrating user interfaces that express performance metrics at business, application, and system infrastructure layers of a predictive model according to one embodiment.
In order to translate such perfonnance metrics into terms that a business executive can understand, each infrastructure component is associated with a financial cost. For example, infrastructure components at the application and system layers can be associated with operational and maintenance costs. At the business and enteφrise layers, the financial costs can include salaries and capital costs, for example. Similarly, at the enteφrise layer, a revenue can be associated with processed business events. As these financial, service and performance metrics percolate up through the layers, the predictive model is able to model the impact of a business decision.
Figs. 9A-9D are diagrams illustrating tables that describe a set of enteφrise decision metrics according to one embodiment. In particular, the decision metrics 30 can include (1) elongation ratio, (2) unit utilization ratio, (3) ceiling threshold, (4) business response time, (5) aging ratio, (6) degradation ratio, (7) non-productive ratio, (8) non-productive ratio, (9) process latency, (10) cost inflation rate, (11) margin erosion rate, (12) total-added-cost ratio, (13) uncovered-value ratio, (14) under-utilization ratio, (15) operational risk ratio, (16) inefficiency ratio, and (17) lifetime contraction ratio. Some of these metrics can also represent behavior at the underlying infrastructure layers as well. Furthermore, combinations of these metrics can be combined mathematically to provide general indicators of the health of a business enteφrise as shown in Fig. 10.
Fig. 10 is a diagram illustrating combinations of enteφrise decision metrics to generate general indicators of the health of a business enteφrise according to one embodiment. For example, combinations of the enteφrise business metrics can provide indications of productivity 410, profitability 420, growth 430 and risk 440 associated with a particular business infrastructure. For example, the profitability 420 of a business infrastructure can be represented as a combination of elongation
ratio, ceiling threshold, business response time, degradation ratio, cost inflation rate, margin erosion rate, total added cost ratio, uncovered value ratio, and under- utilization ratio. Therefore, if the profitability corresponds to a low value, the designer of the business infrastructure may analyze each of the constituent metrics to determine wliich metric(s) are causing the reduction in profitability. The normalized performance coefficient (NPC) is a global indicator of a business infrastructure to meeting business requirements. According to one embodiment, the NPC is the product of the profitability, productivity, and growth indicators divided by risk. Other global indicators can include (1) Effective Cost Advantage, (2) Effective Value Advantage, (3) Effective Response Elongation, and (4) Overall Scalability Index.
Effective Cost Advantage is the percentage of cost inflation (or deflation) that will yield the business growing for at least the same proportion. For example, if cost is increased by 30% due to re-engineering the business enteφrise, the revenue stream will consequently increase by at least 30%.
Effective Value Advantage represents the growth in value between two points where major change in revenue results from investing in moving the system dynamics between the two points. The investment might change the configuration in short, medium, or long terms but the Effective Value Advantage will represented a normalized parameter that expresses the resultant earnings per share, profit, assets, market mane, and total return to investors.
Effective Response Elongation is an indicator of the delays that are independent of service requirements. By definition, this indicator is the time to deliver divided by tht time to execute without contention or conflict minus one. The Effective Response Elongation can be used to assess the scope of improvement opportunities. When its value is close to zero, there are few improvement opportunities that can take place. If the value is higher than one, serious actions must be taken in order to improve the service and avoid rapid degradation.
Overall Scalability Index is a measure of the ability of structure to replicate the same cost and performance characteristics each time addition load will be in service. The index varies from 1 to 100, where 100 corresponds to full replication
and 0 corresponds to no replication. For example, a scalability index of 0.85 means that we need 15% addition resource contingencies added to replicate percentages with the new workload. A fully scalable system will closely replicate the same workload growth.
Embodiments of the present invention can be applied in a number of situations that involve decisions that can significantly impact the success of a business enteφrise. According to one example, embodiments of the invention can be used during outsourcing deals. In a typical outsourcing deal, three stages are involved: (1) a pre-assessment of the client environment, processes, resources and assets;
(2) a due diligence phase during the preparation of the request for proposals, proposal negotiation and contract definition; and (3) the contract execution. In each one of the above stages a high level of uncertainty exists. Millions of dollars are spent to manage satisfactorily manage the pre-assessment of the client environment and perform the lengthy due diligence and proposal generation, which may result in the threat and real risk of margin erosion during the third stage.
With the embodiments of the invention, the predictive model will show limits at each layer (enteφrise, business, IT infrastructure) and identify issues to assess during the first stage. After collecting the appropriate data regarding the business infrastructure, the predictive model can reveal a true and fact based representation of the client environment, processes, resources and assets. Scenario analysis of the predictive model results in ceilings and therefore contingencies being determined for the following stages of the outsourcing deal. The predictive model can also be used to support the engineering effort by defining improvement actions and enhancement trails in order to increase margin and justify business cases for the third stage.
According to another aspect of the invention, the accuracy of the predictive modeling is improved by mathematically expressing the dynamic characteristics and behavior of each infrastructure component as a result of direct and indirect effects of the infrastructure components impacting one another. Perturbation theory is a branch of mathematics that can be used to express such behavior.
The metrics that represent the critical behavior and evolution of enteφrise, business, application and system infrastructures and that support related decisions are continuously perturbed by a variety of effects of different nature. These effects are small if taken separately but can become large if they are taken collectively. In particular, the metrics associated with each infrastructure component can be impacted significantly through direct and indirect perturbation effect.
Fig. 11 is a schematic diagram illustrating direct and indirect perturbation effects on infrastructure components within a business enteφrise according to one embodiment. As shown, infrastructure component cl interacts directly with components c2 and c3. Thus, the performance of component cl is directly effected by components c2 and c3. Because component cl does not directly interact with component c4, there is no direct effected on component cl by component c4. However, component c4 indirectly effects the performance of component cl because it contends for resources on components c2 and c3. Thus component el's ability to access those resources are indirectly impacted by component c4. Likewise, component c3 also contends for resources on component c2. Thus, component c3 has an indirect effect on component cl as well as a direct effect. For example, Figs. 9A-9D identify the possible causes of direct and indirect perturbation effects on each of the enteφrise business metrics.
While most dynamic perturbing phenomena are generated randomly in the system, there exist sufficient observations to show that their impact follow and might well be represented through deterministic solution where time represent the independent variable for the system of partial differential equations that express the perturbation effects.
Fig. 12 is a diagram illustrating a method of improved predictive modeling using perturbation theory according to one embodiment.
At 500, a description of the infrastructure components that support one or more business solutions is provided. This description can be provided in the same manner previously described.
At 510, for each infrastructure component, a mathematical expression of the dynamic characteristics and behavior of the component based on direct effects of a
first set of infrastructure components, including the component itself, can be generated.
Analytically different causes can be expressed either directly or indirectly impacting a perturbing function in proper mathematics that are the best fit to represent such function. For example, the mathematical expression may be based on queuing, stochastic or general probability theory to express and determine service impact on response time and cost. Deterministic mathematics or set algebra may be used to express the availability of resources to process all requests. Mean value analysis or again stochastic processes may also be used to represent latencies and delays for complex communication with external systems.
At 520, for each infrastructure component, the mathematical expression is perturbed by including indirect effects of a second set of infrastructure components.
Metric calculation can be used at each layer of the predictive model. In particular, analytically what will be obtained from the lower levels are functions of time that represent different contributions to the perturbation problem at the top level (e.g., enteφrise layer). Numerically this might be cost contributions to an overall cost inflation for a certain point of time. In this last case the problem will be reduced to a value of a parameter, such as response time or throughput, but its robustness and validity in time will be largely reduced to a single point. Conceptually, this last case is predominant in the current evaluation methods and consequently their inability to guarantee a viable decision as they ignores higher order perturbations that were truncated too early to allow their inclusions. In contrast, conserving an analytical expression in terms of the direct and indirect effects provides greater validity, better accuracy and larger interval of time and more robust solution.
For clarity, this discussion of the perturbation approach is restricted to a first order perturbation (direct and indirect) due to the small magnitudes we manipulate. However, if needed, higher order perturbation solutions can derived following a similar approach.
According to one embodiment, the perturbation theory approach involves a dynamic system of Lagrange-like partial differential equations that represent the
dynamic behavior of a cost function and a solution that will capture both direct and indirect perturbations around a base of the un-perturbed solution. Conceptually, the solution can be expressed with perturbation theory such that any metric X can be expressed in the form: x= x0 + ∑MxMw + ∑N xN ® where X0 is the initial value of a metric (e.g., function or characteristic); XM (d) is the calculated direct impact due to M causes; and XN ω is the calculated indirect impact due to N causes.
In more detail, consider the vector:
O — θ(k) where k = 1, 2, . . . K and where 0^ is a function of time and represents the metrics that describe Coφorate, Financial, Business and Technology engineering characteristics and behavior.
Further consider that:
(c) O represents the unperturbed value of a metric, or its minimum admitted value for simplicity;
( Λ\
O represents a measure of a perturbed metric due to the direct impact
applied on the perturbing function . ; and
0 represents the indirect perturbation due to the perturbed affect of
metrics against each other or the perturbing function -X. due to an external impact.
In general the system of equations that represent the variations can have the form:
— = X w (σ(0) ) + X" (σ(d) ) + X' (σ(i) ) dt
where JY represents basic function. Further assume that: 0 and 0 are vectors representing 0 through different coordinates and that
0 , 0 , and 0 represent the unperturbed values of a metric. Then, the first order direct perturbation is:
dσ _ * dXc (c) (ox (d) (d) (c) ,(0) _ιι(θ) N dt -∑ d<J CTk >c7k )<J +A ^σ >σk >σ k ) (!)
and the first order indirect perturbation is:
This separation seems artificial from theoretical point of view, but it is natural from practical point of view as the origin of perturbation on . and 0 are different.
Next,
s~ ( i) * * *
' *~^ kV a matrix of numerical vectors, γιχ,γι2,...γιmdxe normalization
constants and
15 %2' • • * Xm are ^e Perturbing variables (function in time).
Therefore:
are known functions in time, and can solve the two system equations (1) & (2) in the form:
^ = U(t)σ + v(t) (3) at
where U(t) is a square matrix (K x K) and v(t) is a known vectorial function.
The matrix is determined by:
~ = U(t)Y , (4) at
with Y(t0) = I (5)
where I is a unit matrix and therefore equation (3) becomes:
with X(c) = (Xκ (o)) U is specified in the form
The formula dσ
= U(t)σ dt
forms the system of equations equivalent to the un-perturbed expression:
d ar<Jτ c) - V(c) ^(c) dt Xw(σ )
where the solution Y in equation (4) is known if the partial derivatives of the unperturbed problem is computed with respect to the K integration constants such as
Fig. 13 is a diagram of an operating environment for embodiments of the invention. In particular, a computer system suitable for use with the present invention is described. Computer system 610 includes at least processor 611, for processing information according to programmed instructions, memory 612, for storing infonnation and instructions for processor 611, storage system 613, such as a magnetic or optical disk system, for storing large amounts of information and instructions on a relatively long-term basis, and display system 614, such as a computer monitor, for displaying various graphical elements that facilitate user interaction with computer system 610. Processor 611, memory 612, storage system 613, and display system 614 are coupled to bus 615, which preferably provides a high-speed means for devices connected to bus 615 to communicate with each other.
It will be apparent to one of ordinary skill in the art that computer system 610 is illustrative, and that alternative systems and architectures may be used with the present invention. It will further be understood that many other devices, such as a network interface (not shown), and a variety of other input and output devices (not shown) may be included in computer system 610.
While this invention has been particularly shown and described with references to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims.
Claims
1. A method of predictive modeling, comprising: providing a description of infrastructure components that support one or more business solutions in a business enteφrise; from the description, generating a predictive model that mathematically expresses dynamic characteristics and behavior of the infrastructure components; generating performance metrics from the predictive model for each of the infrastructure components; and translating the performance metrics into enteφrise decision metrics or indicators that correspond to service, performance and financial states of the business enteφrise.
2. The method of claim 1 wherein the predictive model mathematically expresses the dynamic characteristics and behavior of each infrastructure component as including direct effects from a first set of infrastructure components and indirect effects from a second set of infrastructure components.
3. The method of claim 2 wherein generating the predictive model comprises: for each infrastructure component, generating a mathematical expression of the dynamic characteristics and behavior of the infrastructure component from direct effects of the first set of infrastructure components; and for each infrastructure component, perturbing the mathematical expression of the dynamic characteristics and behaviors of the infrastructure component by including indirect effects of the second set of infrastructure components.
The method of claim 1 wherein the predictive model includes a plurality of layers, the method further comprising: from the description, generating an enteφrise layer of the predictive model, the enteφrise layer translating the performance metrics into enteφrise decision metrics or indicators that correspond to the service, performance and financial states of the business enteφrise.
The method of claim 4 further comprising: for each of the one or more business solutions, generating business, application, and system layers of the predictive model from the description, the business layer mathematically expressing dynamic characteristics and behaviors of business processes that support each business solution, the application layer mathematically expressing dynamic characteristics and behaviors of software components that support one or more of the business processes in the business layer, and the system layer mathematically expressing dynamic characteristics and behaviors of hardware components that support one or more of the software components in the application layer; and generating performance metrics for each of the infrastructure components at the business, application, and system layers of the predictive model.
A method of predictive modeling, comprising: providing a description of infrastructure components that support one or more business solutions; from the description, generating a predictive model that mathematically expresses dynamic characteristics and behavior of the infrastructure components, the predictive model being generated by: for each infrastructure component, generating a mathematical expression of the dynamic characteristics and behavior of the infrastructure component based on direct effects of the first set of mfrastructure components; and for each infrastructure component, perturbing the mathematical expression of the dynamic characteristics and behavior of the infrastructure component by including indirect effects of the second set of infrastructure components.
7. The method of claim 6 wherein the predictive model includes a plurality of layers, the method further comprising: for each of the one or more business solutions, generating business, application, and system layers of the predictive model from the description, each layer mathematically expressing the dynamic characteristics and behavior of each infrastructure component associated with the layer as including direct effects from a first set of infrastructure components and indirect effects from a second set of infrastructure components; and generating performance metrics for each of the infrastructure components at each layer.
8. A system of predictive modeling, comprising: an input module providing a description of infrastructure components that support one or more business solutions in a business enteφrise; a construction module generating a predictive model from the description, the predictive model mathematically expressing dynamic characteristics and behavior of the infrastructure components; and a metric calculation module generating performance metrics from the predictive model for each of the infrastructure components and translating the performance metrics into enteφrise decision metrics or indicators that correspond to service, performance and financial states of the business enteφrise.
9. The system of claim 8 wherein the predictive model mathematically expresses the dynamic characteristics and behavior of each infrastructure component as including direct effects from a first set of infrastructure components and indirect effects from a second set of infrastructure components.
10. The system of claim 9 wherein: for each infrastructure component, the construction module generates a mathematical expression of the dynamic characteristics and behavior of the infrastructure component from direct effects of the first set of infrastructure components; and for each infrastructure component, the construction module perturbing the mathematical expression of the dynamic characteristics and behaviors of the infrastructure component by including indirect effects of the second set of infrastructure components.
11. The system of claim 8 wherein the predictive model includes a plurality of layers, the plurality of layers including an enteφrise layer that translates the perfonnance metrics into enteφrise decision metrics or indicators that correspond to the service, performance and financial states of the business enteφrise.
12. The system of claim 11 wherein: the plurality of layers further includes business, application, and system layers; the business layer mathematically expressing dynamic characteristics and behaviors of business processes that support each business solution; the application layer mathematically expressing dynamic characteristics and behaviors of software components that support one or more of the business processes in the business layer; and the system layer mathematically expressing dynamic characteristics and behaviors of hardware components that support one or more of the software components in the application layer; and generating performance metrics for each of the infrastructure components at the business, application, and system layers of the predictive model.
13. A system of predictive modeling, comprising: an input module providing a description of infrastructure components that support one or more business solutions; for each infrastructure component, a construction module generating a mathematical expression of the dynamic characteristics . and behavior of the infrastructure component based on direct effects of the first set of infrastructure components; and for each infrastructure component, the construction module perturbing the mathematical expression of the dynamic characteristics and behavior of the infrastructure component by including indirect effects of the second set of infrastructure components.
14. The system of claim 13 wherein: the predictive model includes a plurality of layers, the plurality of layer including business, application, and system layers, each of the plurality of layers mathematically expressing the dynamic characteristics and behavior of each infrastructure component associated with the layer as including direct effects from a first set of infrastructure components and indirect effects from a second set of infrastructure components; and a metric calculation module generating performance metrics for each of the infrastructure components at each layer.
15. A system of predictive modeling, comprising: means for providing a description of infrastructure components that support one or more business solutions in a business enteφrise; means for generating a predictive model from the description, the predictive model mathematically expressing dynamic characteristics and behavior of the infrastructure components; and means for generating performance metrics from the predictive model for each of the infrastructure components and translating the perfonnance metrics into enteφrise decision metrics or indicators that correspond to service, performance and financial states of the business enteφrise.
16. A system of predictive modeling, comprising: means for providing a description of infrastructure components that support one or more business solutions; for each infrastructure component, means for generating a mathematical expression of the dynamic characteristics and behavior of the infrastructure component based on direct effects of the first set of infrastructure components; and for each infrastructure component, means for perturbing the mathematical expression of the dynamic characteristics and behavior of the infrastructure component by including indirect effects of the second set of infrastructure components.
17. The method of claim 2 wherein the predictive model mathematically expresses the dynamic characteristics and behavior of each infrastructure component using perturbation theory.
8. The system of claim 9 wherein the predictive model mathematically expresses the dynamic characteristics and behavior of each infrastructure component using perturbation theory.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US46748303P | 2003-05-02 | 2003-05-02 | |
PCT/US2004/013603 WO2004100041A2 (en) | 2003-05-02 | 2004-05-03 | System and method of predictive modeling for managing decisions for business enterprises |
Publications (1)
Publication Number | Publication Date |
---|---|
EP1627354A2 true EP1627354A2 (en) | 2006-02-22 |
Family
ID=33435081
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP04751132A Withdrawn EP1627354A2 (en) | 2003-05-02 | 2004-05-03 | System and method of predictive modeling for managing decisions for business enterprises |
Country Status (2)
Country | Link |
---|---|
EP (1) | EP1627354A2 (en) |
WO (1) | WO2004100041A2 (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7657406B2 (en) * | 2005-06-09 | 2010-02-02 | Intepoint, Llc | Multi-infrastructure modeling system |
US10102055B1 (en) | 2016-03-22 | 2018-10-16 | EMC IP Holding Company LLC | Data driven converged infrastructure components evaluation |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6311144B1 (en) * | 1998-05-13 | 2001-10-30 | Nabil A. Abu El Ata | Method and apparatus for designing and analyzing information systems using multi-layer mathematical models |
-
2004
- 2004-05-03 EP EP04751132A patent/EP1627354A2/en not_active Withdrawn
- 2004-05-03 WO PCT/US2004/013603 patent/WO2004100041A2/en active Application Filing
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6311144B1 (en) * | 1998-05-13 | 2001-10-30 | Nabil A. Abu El Ata | Method and apparatus for designing and analyzing information systems using multi-layer mathematical models |
Also Published As
Publication number | Publication date |
---|---|
WO2004100041A8 (en) | 2005-01-27 |
WO2004100041A2 (en) | 2004-11-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7389211B2 (en) | System and method of predictive modeling for managing decisions for business enterprises | |
USRE50192E1 (en) | Predictive risk assessment in system modeling | |
US11030551B2 (en) | Predictive deconstruction of dynamic complexity | |
US20200175439A1 (en) | Predictive Risk Assessment In Multi-System Modeling | |
Castellanos et al. | ibom: A platform for intelligent business operation management | |
US6990437B1 (en) | Systems and method for determining performance metrics for constructing information systems | |
Wickboldt et al. | A framework for risk assessment based on analysis of historical information of workflow execution in IT systems | |
US20020069102A1 (en) | Method and system for assessing and quantifying the business value of an information techonology (IT) application or set of applications | |
JP5961445B2 (en) | Business results trade-off simulator | |
US20180052872A1 (en) | Data cleansing and governance using prioritization schema | |
Orta et al. | Decision-making in IT service management: a simulation based approach | |
US20090112668A1 (en) | Dynamic service emulation of corporate performance | |
Ruiz et al. | Using simulation-based optimization in the context of IT service management change process | |
KR100989494B1 (en) | Process Management Support System and Simulation Method | |
Cimino et al. | A general simulation framework for supply chain modeling: state of the art and case study | |
Varma et al. | A framework for addressing stochastic and combinatorial aspects of scheduling and resource allocation in pharmaceutical R&D pipelines | |
Suzuki et al. | Simulation based process design: Modeling and applications | |
US20080300837A1 (en) | Methods, Computer Program Products and Apparatus Providing Improved Selection of Agreements Between Entities | |
Prakash et al. | ARP–GWO: an efficient approach for prioritization of risks in agile software development | |
Zhang et al. | A simulation approach for evaluation and improvement of organisational planning in collaborative product development projects | |
Ciavotta et al. | Architectural design of cloud applications: A performance-aware cost minimization approach | |
Wu et al. | An intent-driven daas management framework to enhance user quality of experience | |
WO2004100041A2 (en) | System and method of predictive modeling for managing decisions for business enterprises | |
Lozovik et al. | Improvement of the methods of assessing the influence of external factors on the strengths and weaknesses of the enterprise | |
Wirthlin et al. | Defense acquisition system simulation studies |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
17P | Request for examination filed |
Effective date: 20051129 |
|
AK | Designated contracting states |
Kind code of ref document: A2 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LI LU MC NL PL PT RO SE SI SK TR |
|
DAX | Request for extension of the european patent (deleted) | ||
17Q | First examination report despatched |
Effective date: 20101112 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: EXAMINATION IS IN PROGRESS |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN |
|
18D | Application deemed to be withdrawn |
Effective date: 20171114 |