ABLE agents

The Intelligent Agents console for iSeries™ Navigator works with Agent Building and Learning Environment (ABLE) agents.

ABLE agents are Java™ objects capable of automating tasks through the use of rule-based reasoning and learning certain behaviors over time by using data mining algorithms contained in the ABLE component library. ABLE is a Java framework and toolkit used for building multiagent intelligent autonomic systems, and provides specific support for developing agents that work with the iSeries Navigator Intelligent Agent platform and console. Intelligent agents developed using ABLE can have the following capabilities:

• Learn from experience and predict future states
• Analyze metric data using classification and clustering algorithms to detect complex states and diagnose problems
• Interface with other autonomic components via web services
• Reason using domain-specific Java application objects
• Use powerful machine reasoning, including: Boolean forward and backward chaining, predicate logic (Prolog), Rete'-based pattern match, and fuzzy systems
• Have autonomous (proactive) behavior and goals
• Correlate events into situations, reason, and take actions

The ABLE toolkit contains several examples of how to design your own agent, and an iSeries template agent is included that you can use as a model when developing your own agent. To create an agent that can be fully managed from the console, the agent should extend AbleEServerDefaultAgent.