Expert Systems on Multiprocessor Architectures. Phase 1
Abstract
The military has a demonstrated need for knowledge-based systems with significantly higher quantitative performance. The current hardware and software architectures for knowledge based systems cannot support such requirements. Based on near-term projections for integrated circuit technologies, it is clear that highly parallel multiprocessor computers consisting of 100s to 1000s of processors and realizing a variety of concurrent architectures can be built. The most promising approach for achieving orders of magnitude improvement in the quantitative performance of knowledge-based systems is by exploiting concurrency on multiprocessor systems. The major issue is whether such computer systems can be effectively used to enhance the performance of knowledge-based systems. Expert systems on Multiprocessor Architectures project is addressing the following questions: (1) can multiprocessor computers be used to achieve significant execution speedup (two to three orders of magnitude) over serial machines for knowledge-based system applications? (2) What are the limiting factors in achieving speedup for such systems? (3) What are appropriate software models and methodologies for programming such systems? (4) What are appropriate hardware architectures for supporting such systems? Given the lack of any formal foundations for studying concurrent knowledge-based systems, the approach taken to answering these questions is empirical rather than theoretical. Keywords: Expert systems; Parallelism; Multiprocessor; Architectures; Blackboard systems; Parallel processing.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1988
- Accession Number
- ADA204907
Entities
People
- Edward A. Feigenbaum
- Robert S. Engelmore
Organizations
- Stanford University