DID YOU KNOW? DTIC has over 3.5 million final reports on DoD funded research, development, test, and evaluation activities available to our registered users. Click HERE
to register or log in.
Using the Parka Parallel Knowledge Representation System. Version 3.2
Technical research rept.
MARYLAND UNIV COLLEGE PARK INST FOR SYSTEMS RESEARCH
Pagination or Media Count:
This document is the user manual for the Parka Knowledge Representation System, version 3.2 Parka 3. The purpose of this document is to describe the features of the Parka system and to provide some examples of how knowledge can be represented using the Parka language. Detail on the research issues addressed by Parka and the Parallel Server operation can be found in other documents described below. Parka has much in common with typical frame-based knowledge representation systems. What distinguishes Parka is its ability to perform certain types of very useful inferences faster than any existing system we know of. This speed is accomplished through the use of massively parallel hardware, such as the Connection Machine CM-2 and CM-5 supercomputer made by Thinking Machines Corporation. If a parallel machine is not available, Parka can still be run in serial mode. If a CM is available as a parallel server, certain inferences will be done more quickly in parallel. Parka has been developed by the Parallel Understanding Systems A.I. Research Group, under the direction of Prof. James Hendler, in the Department of Computer Science at the University of Maryland at College Park. The original version of Parka was designed and implemented by Mat Evett and Lee Spector. Bill Andersen designed and implemented Parka 3. Some aspects of the Parka 3 language design were suggested by Brian Kettler. Sean Luke implemented a graphical browser and graphical query window for Parka 3. Merwyn Taylor assisted in the testing of Park 3.
APPROVED FOR PUBLIC RELEASE