Obviousness

The obviousness analysis under 35 U.S.C. § 103 requires determining whether "the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains". This involves identifying relevant prior art, determining the scope and content of that prior art, identifying the differences between the claimed invention and the prior art, and then assessing whether a person of ordinary skill in the art (PHOSITA) would have been motivated to combine the prior art references to achieve the claimed invention.

The filing date of US Patent 6851115 is January 5, 1999. Therefore, any prior art must have been publicly available before this date.

The patent itself identifies several areas of prior art in its Background of the Invention section, including:

• Networked computing models
• Distributed object approach (DOOP)
• Mobile objects (mobile agents)
• Blackboard architectures
• Agent-based software engineering

It also explicitly mentions shortcomings in prior agent-based technologies, particularly regarding complex goals, facilitator strategies, and integration of human agents through natural language. The patent also notes limitations of the initial version of SRI International's Open Agent Architecture (OAA) technology, which supported only limited compound goals, fixed formats for conjoined/disjoined sub-goals, hard-wired parallel goal solving, single parameter sets for entire lists, and did not adequately address scalability.

A person having ordinary skill in the art in 1999 would likely have a background in computer science, software engineering, and artificial intelligence, with an understanding of distributed systems, object-oriented programming, and agent-based paradigms.

Here's an analysis of potential obviousness combinations:

Combination 1: Agent-Based Systems with Facilitators + Blackboard Architectures + Knowledge Query and Manipulation Language (KQML)

Prior Art References:

• Agent-based systems with facilitators: The patent acknowledges that "Several agent-based projects have helped to evolve the notion of facilitation." SRI's Open Agent Architecture (OAA), mentioned in the patent and described in a 1999 paper by Martin, Cheyer, and Moran, specifically details a framework where facilitators broker communication and cooperation among distributed agents. This includes the idea that facilitators match requests with agent capabilities and that the requesting agent doesn't need to know the identities or locations of other agents.
• Blackboard architectures: The patent describes blackboard architectures as typically allowing "multiple processes to communicate by reading and writing tuples from a global data store." It also notes their flexibility for problem-solving by dynamic communities of distributed processes. Blackboard systems were a known artificial intelligence approach dating back to the 1970s and '80s, where a common knowledge base (the "blackboard") is iteratively updated by specialist knowledge sources to solve a problem. Knowledge sources operate independently and communicate exclusively through the blackboard.
• Knowledge Query and Manipulation Language (KQML): The patent references KQML as an existing approach to interagent communication, specifically in its discussion of ICL's conversational protocol. KQML was developed in the early 1990s as a language and protocol for communication among software agents and knowledge-based systems, enabling agents to interact with intelligent systems and coordinating interactions through "communication facilitators." KQML defines "performatives" for operations agents perform on each other's knowledge and goal stores.

Motivation for Combination and Obviousness:

A PHOSITA would have been motivated to combine these elements to address the acknowledged limitations of prior agent-based systems. The patent itself highlights that "existing agent-based technologies and architectures are typically very limited in the extent to which agents can specify complex goals or influence the strategies used by the facilitator."

1. Facilitator-based Coordination (Agent-based systems with facilitators): The core concept of a facilitator coordinating agent interactions, matching requests to capabilities, and providing transparent delegation was well-established in agent-based systems like OAA prior to the patent's filing.
2. Flexible Problem Solving (Blackboard architectures): Integrating a blackboard-style global data store, as described in blackboard architectures, with a facilitator-based agent system would be a natural step to enhance flexibility and dynamic interaction. The patent notes that a facilitator "may also provide a global data store for its client agents, allowing them to adopt a blackboard style of interaction." This suggests that the concept was already considered. The known advantages of blackboard architectures for "flexible framework for problem solving by a dynamic community of distributed processes" would motivate their integration into agent systems seeking to solve complex problems.
3. Rich Communication and Goal Expression (KQML): While the patent distinguishes ICL from KQML in terms of expressiveness, the fundamental idea of a structured interagent communication language with performatives for queries and actions, and the use of facilitators for knowledge sharing, was present in KQML. A PHOSITA, observing the limitations of prior agent systems in specifying complex goals, would be motivated to develop a more expressive communication language building upon the concepts of KQML to enable more nuanced goal expressions and interaction protocols. The patent acknowledges that KQML includes performatives like ask_all or ask_one for satisfying queries. The idea of extending such a language to handle more complex goal expressions (e.g., compound goals with logical connectors and nesting) would be a logical progression for a PHOSITA trying to improve agent collaboration on intricate tasks.

Therefore, combining known agent-based systems using facilitators for coordination, with the flexible data sharing and problem-solving mechanisms of blackboard architectures, and enhancing interagent communication through a more expressive language building on KQML's principles, would have been obvious to a PHOSITA seeking to overcome the limitations of existing distributed agent systems in handling complex goals and dynamic interactions.

Combination 2: Agent-Based Systems with Facilitators + PROLOG-like Logic Programming for Goal Expression

Prior Art References:

• Agent-based systems with facilitators: As noted in Combination 1, the concept of agent systems with facilitators was well-established, particularly with SRI's OAA.
• PROLOG and logic programming in distributed systems: The patent explicitly states that the content layer of ICL "preferably supports unification and other features found in logic programming language environments such as PROLOG" and that PROLOG is "suitable for implementing and extending into the content layer of the ICL." PROLOG has a long history, with implementations and research into its use in distributed systems predating the patent's filing. Specifically, the potential for implicit exploitation of parallelism in Prolog was recognized.

Motivation for Combination and Obviousness:

A PHOSITA addressing the challenges of enabling agents to specify and solve "arbitrarily complex goal expressions" would naturally look to logic programming paradigms. The patent itself frames the problem in terms of "arbitrarily complex goal expressions" and then points directly to PROLOG as a suitable foundation.

1. Enhanced Goal Expressiveness (PROLOG): Given the acknowledged limitations of prior agent systems in specifying complex goals, a PHOSITA would be motivated to incorporate a powerful, expressive language for goal representation. PROLOG, as a logic programming language, is inherently designed for expressing complex relationships and solving problems through logical inference, making it an obvious choice for defining and processing "arbitrarily complex goal expressions," including those with logical connectors (AND, OR, NOT) and nesting. The patent's description of compound goals using operators similar to PROLOG (comma for conjunction, semicolon for disjunction, arrow for conditional execution) directly demonstrates this inspiration.
2. Facilitator Interpretation and Delegation: Combining PROLOG-like goal expressions with a facilitator agent would allow the facilitator to leverage its "knowledge base that records the capabilities of a collection of agents" to interpret these complex goals and "construct a goal satisfaction plan." The ability of PROLOG to support unification would further aid the facilitator in matching complex goal patterns to the registered capabilities (solvables) of agents.

Therefore, building an interagent communication language that leverages the expressiveness and logical processing capabilities of PROLOG for complex goal expressions, within an existing facilitator-based agent architecture, would have been obvious to a PHOSITA seeking to enable more sophisticated task delegation and problem-solving in distributed agent systems.

General Obviousness Considerations

• Transparent Delegation and Compound Goals: The patent highlights "transparent delegation" and "facilitator handling of compound goals" as distinguishing features. However, the general concept of facilitators delegating tasks to agents based on capabilities, without the requester needing to know the specifics of the fulfilling agent, was present in OAA. The challenge of handling complex, multi-part requests (compound goals) in distributed systems was a known problem in the field. A PHOSITA would be motivated to enhance existing facilitator mechanisms to better parse, plan, and execute such complex requests, possibly drawing inspiration from distributed planning or problem decomposition techniques.
• Multi-modal User Interfaces: The patent emphasizes "agent-based provision of multi modal interfaces, including natural language." Prior art in human-computer interaction and AI already explored multi-modal input and natural language processing. The application of these technologies within an agent-based framework, where specialized agents handle different modalities and collaborate through a facilitator, would be a natural extension for a PHOSITA aiming to create more intuitive user experiences for distributed systems. For example, the OAA itself was used to implement "Multimodal User Interfaces in the Open Agent Architecture" as detailed in a 1999 paper by Moran et al., and the "Automated Office" application mentioned in the patent.
• Agent Library: The patent describes an agent library providing common infrastructure. The concept of providing libraries or toolkits to simplify the development of distributed systems and agent-based applications was also known in the prior art of software engineering.

In conclusion, while US Patent 6851115 describes a robust and flexible architecture, many of its individual components and the underlying motivations for combining them can be found in the prior art related to distributed computing, agent-based systems, blackboard architectures, and logic programming, particularly through publications related to SRI International's Open Agent Architecture (OAA) which directly lists Adam J. Cheyer and David L. Martin as contributors. A PHOSITA would have been motivated to combine these known elements to create a more capable and user-friendly distributed agent system, especially to address the known limitations of existing systems in handling complex goals and dynamic interactions.