ActiveEdge® – An Intelligent Distributed Agent Platform

Cougaar Software’s ActiveEdge® is an intelligent agent-based middleware platform for building complex, distributed, intelligent applications. ActiveEdge is a mature, proven commercial product, not a lab prototype we hope to someday commercialize. ActiveEdge has over a dozen companies and 100+ developers that are trained and capable of building applications using the platform. ActiveEdge is built on the Cognitive Agent Architecture (Cougaar)—an open source, distributed agent architecture. ActiveEdge provides all the power of Cougaar and includes key extensions to simplify application development, increase agent functionality, and provide enhanced system capabilities. ActiveEdge is designed to transform data into usable knowledge and provides advanced situational awareness, creating an understandable real-time picture of current enterprise operations. In addition, ActiveEdge provides advanced execution monitoring and collaborative decision support. Cougaar Software’s goal is to provide application developers the platform foundation for building the next generation intelligent system – systems that maintain a representation of a situation that focuses on the current relevance of information to an end-user and that enables intelligent reasoning, secure collaboration, advanced data management, dynamic updates, and custom, rule-based processes.

ActiveEdge uses a distributed intelligent agent architecture based on the human cognitive model of reasoning and planning. This advanced software model captures the way humans observe, reason, plan, and act. It allows you to automate the complex processes humans do everyday more realistically and robustly than other traditional technologies.

ActiveEdge is the first commercial platform to be based on the technologies and methodologies of the Cognitive Grid, the integration of Intelligent Agents, Semantic Web and Grid technologies into a unified platform for intelligent systems. The foundation is intelligent agents, but the addition of the Semantic Web allows the computer to understand the meaning of the concepts it is reasoning over, rather than merely manipulating them as tokens of an unknown language. The Grid technologies, empower both the Service Orientation as well as the distributed resource management and other scalable, distributed features of the architecture.

Intelligent agents, the core of the Cognitive Grid and ActiveEdge, are software entities that operate continuously and autonomously in a particular environment. An agent’s responsibilities are defined by the behaviors that have been built into it, and agents carry out their activities in a flexible and intelligent manner that is responsive to changes in the environment. Agents can adjust behavior dynamically to fit the current situation, determining how their actions and behaviors should change as events change. Agent-based systems represent the next major advancement in network computing and leverage the strengths of object-oriented, peer-to-peer and service-oriented architectures while providing a process-centric design. The real value is that intelligent reasoning occurs at each level of the system to reduce overall system load and increase quality, control, and responsiveness. The key benefits of agent technology come in these areas:

• Dynamic Re-planning – The ability to develop and modify distributed workflows using rules and domain knowledge that is appropriate to the current situation. This benefit allows enterprises to create more accurate and appropriate plans and to react more quickly and appropriately when conditions change.
• Advanced Data Mediation – The ability to gather and process data from multiple diverse sources into a single environment so that it is appropriate for the current situation.
• Situational Awareness – The ability to build and maintain a virtual world representation of the current situation on which intelligent reasoning can occur.
• Collaborative Information Management – The ability to easily share information and coordinate changes across your enterprise.
• Intelligent Reasoning – The ability to emulate the way humans observe, reason, plan, act, and monitor at computer speeds.
• Scalable, Distributed Computing – The ability to handle massive amounts of data across the enterprise while providing more efficient processing.
• Business Process Adaptation/Evolution – The ability to allow significant business changes to be implemented quickly and dynamically by actual users who can easily manage adjustments to the business rules or policies—without engaging consultants to significantly alter their systems. This benefit allows enterprises to be agile and adaptive as conditions change, thus saving valuable costs in process re-engineering.

ActiveEdge is is highly scalable and designed to transform massive amounts of disparate data into situationally grounded actionable knowledge. With our powerful Distributed Data Environment (DDE), ActiveEdge lets you efficiently and reliably gather, integrate, distribute, and manage large volumes of complex data from diverse sources and models. Applying modern Semantic Web concepts and the W3C standard on Web Ontology Language (OWL), The DDE has defined a rich data model, which allows application developers to easily define a semantically rich object model to represent information from available data sources and services. ActiveEdge creates an understandable real-time picture of the current situation as well as provides advanced event management, execution monitoring, and collaborative decision support. Based on modern game theory techniques, situation management (situational knowledge, situational understanding or situational awareness) is one of the most innovative components of the ActiveEdge architecture. It is based on our unique situational reasoning framework (SRF), which incorporates hybrid object-semantic domain model to allow accurate representation of the situation and enable users to more effectively understand and act on the situation.

In SOA environments, intelligent reasoning and planning for data collection, fusion, analysis and results presentation requires the sequential invocation of services with support for mediation and transformation of intermediate results. Performing such a service plan in a fixed environment is easy, dynamically creating a service plan in response to an complex and evolving situation requires advanced reasoning and dynamic planning not found in conventional architectures. Intelligent agents, combined with capabilities like semantic service discovery, enable higher level services that allow SOA environments to be utilized to their fullest potential even under new emerging problems. ActiveEdge provides a user-friendly visualization tool that lets users quickly and easily create task-specific, collaborative applications. In addition, users can modify process flows graphically without changing any code—they can easily ‘drag-and-drop’ components from one application to another. In the middleware layer, the robust ActiveEdge Distributed Data Environment (DDE) offers advanced information management and data mediation capabilities and can support volumes of data from diverse, distributed sources and models, integrating that data into a single environment. The DDE supports secure, selective, and seamless data integration and collaboration among sites in the same organization as well as with employees, partners, and customers at different sites. The ActiveEdge DDE is a state-of-the-art data mediation system that supports advanced mediation and mapping, selective data replication, and intelligent querying using standard and object query constructs as well as higher level, advanced semantic-based queries.

The core services of ActiveEdge systems act on the situational picture to support functional processes and applications. And, these functional applications use the situation and core services to perform functional planning, reasoning, display, and analysis for users. With ActiveEdge, situations are monitored, and, depending on defined events, actions are generated according to user-defined business rules and workflows. Software agents can build and/or adjust these business rules and processes based on rule patterns and their ability to learn from prior actions. The situation management component uses situational and domain agents, or reasoners, to translate an event or group of events into a meaningful situation in the current context and situation. The context includes the set of actions the system is performing, the set of events received, and the actions currently operating on the system. The domain controllers correlate events and reason about the changes to the situation that result, considering not only primary impacts, but also secondary and tertiary impacts.