Execution of a canonical rules based agent

• US 7,970,724 B1
• Filed: 12/22/2006
• Issued: 06/28/2011
• Est. Priority Date: 12/22/2006
• Status: Active Grant

• Current Assignee: Callahan Cellular LLC (Intellectual Ventures LLC)
• Original Assignee: Curen Software Enterprises LLC (Intellectual Ventures LLC)
• Inventors: Hauser, Robert R.
• Primary Examiner(s): Gaffin; Jeffrey A
• Assistant Examiner(s): COUGHLAN, PETER D
• Application Number:
  US11/645,301
• Time in Patent Office: 1,649 Days
• Field of Search: None
• US Class Current: 706/47
• CPC Class Codes: G06N 5/025 Extracting rules from data

Abstract

A method for executing an agent utilizing a rule engine and a set of canonical rules in an execution environment comprising collecting an initial set of data, asserting the initial data into a working memory, executing the rule engine utilizing the working memory and the set of canonical rules and firing an applicable rule by the rule engine.

Claims

1. A computer-implemented method comprising:

collecting an initial set of data;

storing the initial set of data in a memory;

receiving previously-completed agent outcomes for a domain;

dynamically removing rules from a set of canonical rules based on the previously-completed agent outcomes to provide a pruned set of canonical rules;

executing a rule engine utilizing the memory and the pruned set of canonical rules; and

causing the rule engine to trigger an implementation of one or more rules of the pruned set of canonical rules.

2. The computer-implemented method of claim 1, wherein the collecting comprises collecting the initial set of data from one or more of an execution environment or an accessible database.

3. The computer-implemented method of claim 1, wherein the collecting comprises collecting the initial set of data from at least one of a machine to machine interface or a human to machine interface.

4. The computer-implemented method of claim 1, further comprising:

receiving rule statistics from a local rule repository, wherein the rule statistics are collected from the rule engine and the rule engine is associated with an agent; and

wherein the dynamically removing rules from the set of canonical rules further comprises dynamically removing rules from the set of canonical rules based on the rule statistics to provide the pruned set of canonical rules.

5. The computer-implemented method of claim 1, wherein causing the rule engine to fire the one or more rules of the pruned set of canonical rules is in response to an addition of data into the memory.

6. The computer-implemented method of claim 1, further comprising, responsive to causing the rule engine to trigger the implementation of the one or more rules of the pruned set of canonical rules, causing the rule engine to add additional data into the memory.

7. The computer-implemented method of claim 5, further comprising, responsive to causing the rule engine to trigger the implementation of the one or more rules of the pruned set of canonical rules, causing an agent to add additional data into the memory.

8. The computer-implemented method of claim 1, further comprising, responsive to causing the rule engine to trigger the implementation of the one or more rules of the pruned set of canonical rules, causing a modification or deletion of data in the memory.

9. The computer-implemented method of claim 8, further comprising, responsive to causing the rule engine to trigger the implementation of the one or more rules of the pruned set of canonical rules, causing the rule engine to modify or delete data in the memory.

10. The computer-implemented method of claim 8, further comprising, responsive to causing the rule engine to trigger the implementation of the one or more rules of the pruned set of canonical rules, causing an agent to modify or delete data in the memory.

11. A tangible computer readable medium having stored thereon, computer-executable instructions that, in response to execution by a computing device, cause the computing device to perform a method comprising:

collecting an initial set of data;

storing the initial set of data in a memory in an execution environment;

receiving previously-completed agent outcomes for the execution environment;

filtering a set of canonical rules, based on the previously-completed agent outcomes, to create a filtered set of canonical rules by removing from the set of canonical rules one or more canonical rules that are not associated with a desirable outcome in the execution environment;

executing a rule engine on the initial set of data utilizing the working memory and the filtered set of canonical rules;

causing the rule engine to trigger an implementation of a rule of the filtered set of canonical rules; and

receiving a notification that the rule was fired.

12. The tangible computer readable medium of claim 11, wherein the method further comprises:

receiving rule statistics from a local rule repository, wherein the rule statistics are collected from the rule engine and the rule engine is associated with an agent; and

wherein the filtering the set of canonical rules further comprises filtering the set of canonical rules based on the rule statistics to create the filtered set of canonical rules.

13. The tangible computer readable medium of claim 11, wherein the method further comprises:

collecting additional data in response to the causing the rule engine to trigger the implementation of the rule of the filtered set of canonical rules; and

inserting the additional data into the memory.

14. The tangible computer readable medium of claim 11, wherein the method further comprises:

causing, in response to the causing the rule engine to trigger the implementation of the rule of the filtered set of canonical rules, one or more of modification or deletion of data in the memory.

15. The tangible computer readable medium of claim 11, wherein the notification is configured to notify an agent that one or more implementations of the rule were triggered.

16. The tangible computer readable medium of claim 15, wherein the method further comprises:

causing the agent to collect additional data and assert the additional data as part of the memory.

17. The tangible computer readable medium of claim 15, wherein the method further comprises:

causing the agent to inform another agent that one or more implementations of the rule were fired.

18. A system, comprising:

a tangible computer-readable memory that is configured to store an agent in a first execution environment;

a computing device that is configured to be communicably coupled to the computer-readable memory, wherein the agent is configured to be executed by the computing device to;

receive a request;

collect an initial set of data;

store the initial set of data into a memory;

receive previously-completed agent outcomes for a second execution environment;

filter a set of canonical rules, based on the previously-completed agent outcomes, to create a filtered set of canonical rules by removing from the set of canonical rules one or more canonical rules that are not associated with a desirable outcome in the second execution environment;

compile the filtered set of canonical rules to create a compiled filtered canonical rule set, one or more rules of the filtered set of canonical rules being retrieved from the second execution environment;

cause the agent to create a rule engine;

execute the rule engine utilizing the memory and the compiled filtered canonical rule set to cause one or more implementations of a rule of the compiled filtered canonical rule set to be triggered;

process a notification that the one or more implementations of the rule were triggered; and

send a response to the request.

19. The system of claim 18, further comprising a second agent stored in the computer-readable memory and configured to initiate the request.

20. The system of claim 18, wherein the agent is further configured to be executed by the computing device to send a preliminary response to the request in response to the notification that the one or more implementations of the rule were fired.

21. A software architecture embodied on one or more tangible computer-readable media and executable by a computing device, the software architecture comprising:

an agent configured to receive a canonical rule set in response to a request for the canonical rule set generated by an application, wherein the canonical rule set has one or more survival rules configured to determine a lifespan of the agent;

a rule compiler configured to receive the canonical rule set from the agent and compile the canonical rule set into a compiled rule set; and

a rule engine configured to receive the compiled rule set and to trigger one or more of the survival rules of the compiled rule set to terminate the agent, wherein the agent is further configured to, responsive to termination, halt the rule engine, record one or more historical rule statistics, and store one or more recorded historical rule statistics in a rule repository, wherein the historical rule statistics comprise data relating to the triggering of one or more rules of the canonical rule set by the rule engine.

22. The architecture of claim 21, wherein the agent is further configured to cause creation of the rule engine.

23. The architecture of claim 21, wherein the agent is further configured to receive the canonical rule set from the rule repository, and wherein the rule repository resides in an execution environment that is remote to a different execution environment in which the agent is executable.

24. The architecture of claim 21, wherein the agent is further configured to be executable in a plurality of different execution environments.

25. The architecture of claim 24, wherein the agent is further configured to encode itself with the canonical rule set to generate a transferable form of the agent, the transferable form of the agent being transferable to one or more of the plurality of different execution environments.