The LISP implementation of D-OMAR

This is Open Source!!!

• Simple Frame Language (SFL) for defining the objects in the simulation world.
• Simulation CORE (SCORE) procedure language for defining agents and specifying their behaviors.
• Flavors Expert (FLEX) for defining rule-based behaviors.

OmarL includes a Concept Editor to support the definition of simulation objects. The graphical view of the concept hierarchy is particularly helpful in managing the definitions of the large number of concepts typically required for a complex model. A Procedure Browser provides graphical views of SCORE procedures that map procedure internals, call and who-calls hierarchies, and forward and backward tracking of signal propagation.

The OmarL event-driven simulator can operate in real-time or fast-time mode. The Simulation Control panel provides interactive control of scenario execution. It includes the capabilities to select and initialize the scenario to be run, and manage the running and pausing of the scenario. Data analysis tools take the form of Event and Task Timelines that provide in depth access to the details of agent behaviors.

OmarL's knowledge-languages provide the representational tools with which to model people and the entities in the world that they interact with. Alternatively, they can be used to define the behaviors of software agents supporting a human system user. The agents, human performance models or software agents, operate in a simulation environment running in real-time or fast-time. OmarL supports the notion of an agent whose actions are driven not only by actively seeking to achieve one or more defined goals, but also by reacting to the input and events of the world. The SCORE language include forms to manage the concurrent behaviors produced by these constraints during a simulation run.

The OmarL simulator is an event-driven simulator that manages the activities of the simulation agents. OmarL agents, models of people, vehicles, or workplaces, execute SCORE goals and procedures that define a mix of sequential and parallel behaviors. Dynamically computed prorities are used to resolve conflicts between competing procedures.

Software Requirements

• Allegro Common LISP Version 6.2, 7.0, or 8.0. Allegro LISP can be obtained from Franz.

• Java Development Kit (or JRE) Version 1.5.0. Java can be downloaded for free from Sun.

Connecting an OmarL Model to the External World

At the core of the Cap Object is a "Connection" (supplied by the OmarL system) which serves as a "source" of outgoing Signals generated by the OmarL model and as a "sink" for incoming Signals to be sent to the OmarL model. The Connection object is the Java-end implementation of a socket-based, serializing, message-passing stream through which the LISP-based Core-OMAR module communicates with its foreign language modules. There is also an analogue of this Connection object written in LISP, so you can create an Cap Object in another LISP image, without using Java. There is currently no support for writing a Cap in a language other than Java or LISP.

The Cap object functions as a "wrapper" around the Connection object and serves as an "intermediary" or "translator" between the OmarL system and the external program with which the OmarL system communicates. All details about the mechanisms for communication between OmarL and the external program are encapsulated in the Cap object.

Currently, Caps have been written that can connect to a Jini network, a CORBA Orb (the standard Orb provided with Java), HLA from DMSO, and the CoABS GRID.

Licensing Terms for Using OmarL

Anonymous CVS Repository

GNATS Buglist

Please Help Out!