Distributed Episodic Exploratory Planning (Deep)

Abstract : DEEP is a mixed-initiative decision support system that utilizes past experiences to suggest courses of action (COAs) for new situations. It was designed as a distributed multi-agent system, using agents to maintain and exploit the experiences of individual commanders, as well as to transform suggested past plans into potential solutions for new problems. The commander, through the agent, can view and modify the contents of the shared repository. Agents interact through a common knowledge repository, represented by blackboard, selected because of its opportunistic reasoning capabilities and implemented in Java for platform independence. Java was chosen for ease of development and integration with other projects. Research also included investigations into various scalability software suites and frameworks, as well as different database management systems. Hibernate, an object/relational persistence and query service, was chosen for interaction with the database. Comprehensive testing revealed the Java Distributed Blackboard was limited only by system resources and network bandwidth. Thus, its architecture is well suited for dealing with ill-defined, complex situations such as military planning.