The Distributed Operator Model Architecture

Abstract : Distributed OMAR (D-OMAR) is a new implementation of the Operator Model Architecture (OMAR) designed to operate in a distributed computing environment. In the new network based configuration, Core OMAR operates as a server that provides the OMAR simulator. The developer's interface operates as a client that may connect to any of the instances of Core OMAR operating on the network. As in OMAR, D-OMAR is designed as a simulation environment in which to create human performance models. It also provides the capability to develop the complex multi-tasking agents that are needed in agent based systems. The representation languages: the Simple Frame Language (SFL), the SCORE procedural language, and the rule language are provided to facilitate model development. The developer's interface: including the Simulation Control panel, the Concept Editor, the Procedure Browser, and the Agent and Event Timeline analysis tools has been reimplemented in Java. The architecture for D-OMAR was designed to enable operation in several different middle ware environments. Currently supported middle ware layers include Java RMI, CORBA, and the Defense Modeling and Simulation Organization's High Level Architecture. A series of Combat Air Patrol scenarios, used as test cases to support D-OMAR development, are available as demonstration scenarios.