Policy Governed Software Agent System & Method of Operation

Abstract

A system and method for deploying software agents in a policy-governed environment. The use of over-arching policies to control the actions and interactions of the software agents preserves desirable agent autonomy without allowing unwanted and potentially unpredictable (or at least difficult-to-predict) agent behavior. The system allows the agents to be deployed in a high-assurance environment.

Description

CROSS-REFERENCES TO RELATED APPLICATIONS[0001]This non-provisional patent application claims the benefit, pursuant to 3 C.F.R. §1.53, of an earlier-filed provisional application. The provisional application was filed on Jul. 16, 2014. It was assigned Ser. No. 62 / 025,035. It listed the same inventors.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT[0002]Some of the work underlying this application has been federally sponsored.MICROFICHE APPENDIX[0003]Not Applicable.BACKGROUND OF THE INVENTION[0004]1. Field of the Invention[0005]The present invention pertains to the field of computer hardware and software. More specifically, the invention comprises a policy-governed system for software agents and a method for controlling the operations thereof.[0006]2. Description of the Related Art[0007]Automated data collection and dissemination systems now make it possible to accumulate and distribute vast quantities of information. In order to be useful, this information must ultima...

AI Technical Summary

Benefits of technology

This patent describes a system and method for using software agents in a policy-governed environment. This approach ensures that these agents act within their desired bounds and prevents them from causing unexpected or unpredictable functions. The policy interpretation and enforcement is performed by the agent system itself, reducing the burden on the agent programmers and ensuring individual agent compliance with the policies. The integration of a policy management system with the agent system allows for the deployment of these agents in high-assurance environments.

Problems solved by technology

Unfortunately, the modern problem is not that of having too little information but rather the problem of being inundated with too much information.

It is simply not possible for a human operator to sift through such a daunting volume in order to find die tiny amount of relevant information.

Moreover, because the specifies of how goals are to he accomplished by the agent are relatively unspecified, the behavior and consequences of agent actions may be much more unpredictable that the behavior and results of traditional software whose execution is more predictably algorithmic in nature.

For this reason, people may be uncomfortable delegating such activities Lo agents.

The nature of such agents can cause concern regarding security, privacy, and control.

Images