Analysis of multiple assets in view of uncertainties

• US 20080120148A1
• Filed: 05/01/2006
• Published: 05/22/2008
• Est. Priority Date: 04/29/2005
• Status: Active Grant

• Current Assignee: Landmark Graphics Corporation (Halliburton Company)
• Original Assignee: Landmark Graphics Corporation (Halliburton Company)
• Inventors: Narayanan, Keshav, Heath, David, Cullick, Alvin Stanley
• Granted Patent: US 8,209,202 B2
• Time in Patent Office: Days
• US Class Current: 705/7
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/01 Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/063 Operations research, analysis or management
• CPC Class Codes: G06Q 10/0635 Risk analysis of enterprise or organisation activities

Abstract

A client-server based system for building and executing flows (i.e., interconnected systems of algorithms). The client allows a user to build a flow specification and send the flow specification to the server. The server assembles the flow from the flow spec and executes the flow. A flow may be configured to analyze the impact (e.g., the financial impact) of a number of uncertainties associated with a plurality of assets. Uncertainty variables are used to characterize the uncertainties associated with the assets. An uncertainty variable associated with one asset may be functionally dependent on an uncertainty variable associated with another asset.

Claims

1. A method comprising:

(a) receiving first information specifying a set of uncertainty variables for a plurality of assets and specifying a functional relationship between a first and a second of the uncertainty variables, wherein the first uncertainty variable is associated with a first of the assets, wherein the second uncertainty variable is associated with a second of the assets;

(b) generating values for each of the uncertainty variables, wherein said generating includes generating a value for the first uncertainty variable, and computing a value for the second uncertainty variable from the value of the first uncertainty variable based on the functional relationship;

(c) determining for each of the assets a corresponding input data set using at least a corresponding subset of the uncertainty variable values, wherein said determining includes;

determining a first input data set corresponding to the first asset using at least the value of the first uncertainty variable; and

determining a second input data set corresponding to the second asset using at least the value of the second uncertainty variable;

(d) for each of the assets, invoking execution of a corresponding set of one or more algorithms, wherein each set of one or more algorithms operates on the corresponding input data set to generate a corresponding output data set;

(e) performing (b), (c) and (d) a plurality of times to generate a plurality of output data sets for each asset;

(f) computing one or more statistics for each of the assets based on the corresponding plurality of output data sets;

(g) generating resultant data based at least partially on the statistics of the assets;

(h) displaying an indication of the resultant data on a display device.

2. The method of claim 1, wherein the assets are assets related to the exploration and production of one or more of oil and gas.

3. The method of claim 1 wherein the set of one or more algorithms for each asset are selected by a user.

4. The method of claim 1, wherein (d) includes:

distributing the input data sets of the assets to one or more remote computers for remote execution of the corresponding sets of algorithms; and

receiving the output data sets from the one or more remote computers.

5. The method of claim 1, wherein said first information also specifies one or more correlations among the uncertainty variables, wherein said generation of values for each of the uncertainty variables respects the one or more specified correlations.

6. The method of claim 5, wherein the uncertainty variables include a third uncertainty variable and a fourth uncertainty variable associated with different ones of the assets, wherein the one or more correlations include a correlation between the third and fourth uncertainty variables.

7. The method of claim 1, wherein (c) includes selecting one of two or more models for incorporation into one of the input data sets based on the value of a third of the uncertainty variables.

8. The method of claim 1, wherein (c) includes:

selecting one of two or more models based on the value of a third of the uncertainty variables;

selecting one of two or more submodels of the selected model based on the value of a fourth of the uncertainty variables;

wherein the selected submodel is usable to determine data to be incorporated into one of the input data sets.

9. The method of claim 1 further comprising:

(i) receiving second information specifying a plurality of decision variables;

(j) generating values for the decision variables, wherein the decision variable values are used to determine the input data sets for the assets;

(k) computing a value of a global objective from the one or more statistics of each of the assets;

executing an optimizer in order to determine one or more sets of values for the decision variables, wherein said executing the optimizer includes performing at least (j), (e), (f) and (k) a number of times, wherein the optimizer is configured to search for a maximum or minimum of the global objective over at least a portion of a space defined by the decision variables.

10. The method of claim 9, wherein the optimizer is selectable by a user from a set of supported optimizers.

11. The method of claim 9, wherein said displaying an indication of the resultant data includes displaying a graphical representation of at least a subset of the one or more sets of values for the decision variables.

12. The method of claim 9, wherein said second information also specifies one or more constraints on the decision variables, wherein said generation of values for the decision variables respects the one or more constraints.

13. The method of claim 9, wherein said second information also specifies one or more functional dependencies between the decision variables, wherein said generation of values for the decision variables respects said one or more functional dependencies between the decision variables.

14. The method of claim 9, wherein each of the decision variables has an associated set of attainable values representing possible outcomes of a corresponding decision, wherein each of the uncertainty variables has an associated set of attainable values.

15. The method of claim 9 further comprising:

(l) computing a value of an auxiliary function from the one or more statistics of each of the assets;

wherein said performing the optimization includes performing (j), (e), (f), (k) and (l) said number of times.

16. The method of claim 15, wherein the optimizer is configured to search for said maximum or minimum of the global objective subject at least to a constraint on a functional combination of the global objective and the auxiliary function.

17. The method of claim 15, wherein said displaying an indication of the resultant data on a display device includes:

displaying a plot of the global objective value versus auxiliary function value.

18. A computer-readable memory medium configured to store program instructions, wherein the program instructions are executable to implement:

(a) receiving first information specifying a set of uncertainty variables for a plurality of assets and specifying a functional relationship between a first and a second of the uncertainty variables, wherein the first uncertainty variable is associated with a first of the assets, wherein the second uncertainty variable is associated with a second of the assets;

(b) generating values for each of the uncertainty variables, wherein said generating includes generating a value for the first uncertainty variable, and computing a value for the second uncertainty variable from the value of the first uncertainty variable based on the functional relationship;

(c) determining for each of the assets a corresponding input data set using at least a corresponding subset of the uncertainty variable values, wherein said determining includes;

determining a first input data set corresponding to the first asset using at least the value of the first uncertainty variable; and

determining a second input data set corresponding to the second asset using at least the value of the second uncertainty variable;

(d) for each of the assets, invoking execution of a corresponding set of one or more algorithms, wherein each set of one or more algorithms operates on the corresponding input data set to generate a corresponding output data set;

(e) performing (b), (c) and (d) a plurality of times to generate a plurality of output data sets for each asset;

(f) computing one or more statistics for each of the assets based on the corresponding plurality of output data sets;

(g) generating resultant data based at least partially on the statistics of the assets;

(h) displaying an indication of the resultant data on a display device.

28. The method of claim 18 further comprising:

displaying a graphical representation of progress of the optimization.

19. A computer system comprising:

a memory configured to store program instructions;

at least one processor configured to read and execute the program instructions, wherein, said program instructions are executable to implement;

(a) receiving first information specifying a set of uncertainty variables for a plurality of assets and specifying a functional relationship between a first and a second of the uncertainty variables, wherein the first uncertainty variable is associated with a first of the assets, wherein the second uncertainty variable is associated with a second of the assets;

(b) generating values for each of the uncertainty variables, wherein said generating includes generating a value for the first uncertainty variable, and computing a value for the second uncertainty variable from the value of the first uncertainty variable based on the functional relationship;

(c) determining for each of the assets a corresponding input data set using at least a corresponding subset of the uncertainty variable values, wherein said determining includes;

determining a first input data set corresponding to the first asset using at least the value of the first uncertainty variable; and

determining a second input data set corresponding to the second asset using at least the value of the second uncertainty variable;

(d) for each of the assets, invoking execution of a corresponding set of one or more algorithms, wherein each set of one or more algorithms operates on the corresponding input data set to generate a corresponding output data set;

(e) performing (b), (c) and (d) a plurality of times to generate a plurality of output data sets for each asset;

(f) computing one or more statistics for each of the assets based on the corresponding plurality of output data sets;

(g) generating resultant data based at least partially on the statistics of the assets;

(h) displaying an indication of the resultant data on a display device.

20. The computer system of claim 19, wherein (d) includes:

distributing the input data sets of the assets to one or more remote computers for remote execution of the corresponding sets of algorithms; and

receiving the output data sets from the one or more remote computers.

21. A method comprising:

(a) receiving first user input specifying two or more algorithms and specifying one or more connections between the two or more algorithms, wherein each of the connections is a connection between an output of one of the algorithms and the input of another of the algorithms;

(b) receiving second user input specifying one or more data structures consistent with the two or more algorithms;

(c) operating on the one or more data structures in response to third user input in order to build one or more models for the two or more algorithms, wherein the one or more models include one or more uncertainty variables and one or more decision variables;

(d) receiving fourth user input specifying an optimizer; and

(e) performing an optimization using the user-specified optimizer in order to optimize an objective function, wherein said performing the optimization includes;

(e1) generating one or more values of the one or more decision variables respectively;

(e2) invoking a plurality of executions of the two or more algorithms according to the one or more connections, wherein each of the executions operates on a corresponding set of one or more instantiated models determined by the one or more models, the one or more decision variable values and a corresponding set of one or more instantiated values of the one or more uncertainty variables;

(e3) computing one or more statistics based on output of the two or more algorithms from said plurality of executions;

(e4) computing a value of the objective function from the one or more statistics; and

(e5) repeating (e1) through (e4) a number of times.

22. The method of claim 21, wherein (a)-(d) are performed on a first computer, the method further comprising:

sending information specifying the two or more algorithms, information specifying the one or more connections, information specifying the one or more statistics, information specifying the optimizer and information specifying the global objective to a server computer, wherein (e) is performed on the server computer.

23. The method of claim 21, wherein the server computer performs (e2) by:

(e2.1) generating a set of one or more instantiated values of the one or more uncertainty variables;

(e2.2) determining one or more instantiated models from the one or more models and the set of one or more instantiated values;

(e2.3) sending the one or more instantiated models to a remote computer, wherein the remote computer performs one of said plurality of executions of the two or more algorithms on the one or more instantiated models; and

repeating (e2.1) through (e2.3) in order to distribute the plurality of executions to a set of one or more remote computers.

24. The method of claim 23 further comprising:

receiving user input specifying a dispatcher, wherein the dispatcher is a program that controls said repeating of(e2.1) through (e2.3).

25. The method of claim 21 further comprising:

receiving user input specifying one or more auxiliary functions, each auxiliary function being a corresponding functional combination of a corresponding subset of the one or more statistics;

receiving user input specifying one or more constraints involving the one or more auxiliary functions and the global objective;

wherein said performing the optimization respectively the one or more constraints involving the one or more auxiliary functions and the global objective.

26. The method of claim 21, wherein the one or more algorithms include algorithms that simulate processes associated with the exploration, development or production of oil or gas.

27. The method of claim 21 further comprising:

receiving fifth user input specifying the one or more statistics to be computed on the output of the two or more algorithms; and

receiving sixth user input specifying a functional combination of the one or more statistics in order to define the objective function.