|
|
|
Foundation ::
Artificial Intelligence and Expert Systems ::
HERLA
|
HERLA
Harness for Evaluating Reinforcement Learning Agents
|
 |
Moderators:
Lashon Booker
|
SOURCE CODE AVAILABLE
|
|
|
HERLA (Harness for Evaluating Reinforcement Learning Agents) is a software integration framework within which reinforcement learning agents can be demonstrated and evaluated with respect to simulated environments. It provides for agent to environment interfacing, control, environment visualization, and performance monitoring. It comes with an instantiated set of agents and environments and is extensible to use additional agents and environments. A document is provided that describes its operation, design, configuration and application. Source code is provided to permit tailoring to individual needs.
HERLA hosts C/C++ agents and environments and runs under Unix. Java is used for interfacing to the user.
|
Copyright 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010 by the Mitre Corporation
ALL RIGHTS RESERVED
|
|
More software from The Mitre Corporation
|
|
|
|
|
|
