R1: A rule-based configurer of computer systems

McDermott, John

doi:10.1016/0004-3702(82)90021-2

Cited by 732 publications

(193 citation statements)

References 1 publication

Supporting

Mentioning

189

Contrasting

Unclassified

Order By: Relevance

“…It consists of only a small number of rule types and is significantly different from the earlier systems [55,56] developed at Carnegie-Mellon University. 8 ILOG is a run-of-the-mill knowledge-intensive expert system.…”

Section: Log7mentioning

confidence: 99%

“…The suggestion of 32-64 processors for the multiprocessor is derived as a result of measurements and simulations done for many large production-system programs [39,43,55,56]. Because of the small number of productions affected per change to working memory and because of the variance in the processing required by the affected productions, simulations for most production-system programs show that using more than 32 processors does not yield any additional speed-up.…”

Section: Architecture Of the Production-system Machinementioning

confidence: 99%

“…They have been used extensively to understand the nature of intelligence--in cognitive modeling, in the study of problem-solving systems, and in the study of learning systems [2,44,45,62,73,74,92]. They have also been used extensively to develop large expert systems spanning a variety of applications in areas including computer-aided design, medicine, configuration tasks, oil exploration [11, 14,39,41,42,55,56,81]. Production-system programs, however, are computation intensive and run quite slowly.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Parallel production systems

Ishida¹

1994

Parallel, Distributed and Multiagent Production Systems

View full text Add to dashboard Cite

i AbstractProduction systems (or rule-based systems) are widely used in Artificial Intelligence for modeling intelligent behavior and building expert systems. Most production system programs, however, are extremely computation intensive and run quite slowly. The slow speed of execution has prohibited the use of production systems in domains requiring high performance and real-time response. This thesis explores the role of parallelism in the high-speed execution of production systems.On the surface, production system programs appear to be capable of using large amounts of parallelismmit is possible to perform match for each production in a program in parallel. The thesis shows that in practice, however, the speed-up obtainable from parallelism is quite limited, around 10-fold as compared to initial expectations of 100-fold to 1000-fold. The main reasons for the limited speed-up are: (1) there are only a small number of productions that are affected (require significant processing) per change to working memory; (2) there is a large variation in the processing requirement of these productions; and (3) the number of changes made to working memory per recognizeact cycle is very small. Since the number of productions affected and the number of workingmemory changes per recognize-act cycle are not controlled by the implementor of the production system interpreter (they are governed mainly by the author of the program and the nature of the task), the solution to the problem of limited speed-up is to somehow decrease the variation in the processing costof affected productions. The thesis proposes a parallel version of the Rete algorithm which exploits parallelism at a very fine grain to reduce the variation. It further suggests that to exploit the fine-grained parallelism, a shared-memory multiprocessor with 32-64 high performance processors is desirable. For scheduling the fine-grained tasks consisting of about 50-100 instructions, a hardware task scheduler is proposed.The thesis presents simulation results for a large set of production systems exploiting different sources of parallelism. The thesis points out the features of existing programs that limit the speed-up obtainable from parallelism and suggests solutions for some of the bottlenecks. The simulation results show that using the suggested multiprocessor architecture (with individual processors performing at 2 MIPS), it is possible to obtain execution speeds of about 12000 working-memory element changes per second. This corresponds to a speed-up of 10-fold over the best known sequential implementation using a 2 MIPS processor. This performance is significantly higher than that obtained by other proposed parallel implementations of production systems.

show abstract

Section: Log7mentioning

confidence: 99%

Section: Architecture Of the Production-system Machinementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Parallel production systems

Ishida¹

1994

Parallel, Distributed and Multiagent Production Systems

View full text Add to dashboard Cite

show abstract

“…Along with many other researchers [60], we believe that building cooperative problem-solving systems and interactive knowledge media is at least as important a goal as building autonomous thinking machines. The major difference between classical expert systems, such as MYCIN [6] and Rl [45], and cooperative problem-solving systems involves the roles of the human and computer. Most expert systems ask the user for input, make all decisions and then return an answer.…”

Section: Cooperative Problem Solvingmentioning

confidence: 99%

The role of critiquing in cooperative problem solving

Fischer

Lemke

Mastaglio

et al. 1991

ACM Trans. Inf. Syst.

121

View full text Add to dashboard Cite

Cooperative problem-solving systems help users design solutions themselves as opposed to having solutions designed for them. Critiquing-presenting a reasoned opinion about a user's product or action-is a major activity of a cooperative problem-solving system. Critics make the constructed artifact "talk back" to the user. Conditions under which critics are more appropriate than autonomous expert systems are discussed. Critics should be embedded in integrated design environments along with other components, such as an argumentative hypertext system, a specification component, and a catalog. Critics support learning as a by-product of problem solving. The major subprocesses of critiquing are goal acquisition, product analysis, critiquing strategies, adaptation capability, explanation and argumentation, and advisory capability. The generality of the critiquing approach is demonstrated by discussing critiquing systems developed in our group and elsewhere. Limitations of many current critics include their inability to learn about specific user goals and their intervention strategies.

show abstract

“…One of the earliest and most successful expert systems introduced in an industrial environment was a product configurator, when -in the early 1980's -Digital Corp. developed the R1/XCON (McDermott, 1982) system for automating the configuration process for their complex computer systems. Although R1/XCON was one of the first systems of that kind, two typical aspects in the context of configuration systems have not changed since then.…”

Section: Introductionmentioning

confidence: 99%

Research Issues in Knowledge-Based Configuration

Jannach

Mass Customization Information Systems in Business

View full text Add to dashboard Cite

Knowledge-based configuration systems have made their way into industrial practice. Nowadays, all major vendors of configuration systems rely on some form of declarative knowledge representation and intelligent search techniques for solving the core configuration problem, due to the inherent advantages of that technology: On the one hand, changes in the business logic (configuration rules) can be accomplished more easily because of the declarative and modular nature of the knowledge bases while on the other hand highly-optimized, domain independent problem solving algorithms are available for the task of constructing valid configurations.Still, the development has not come to an end as -in a world that becomes increasingly automated and wired together -constantly new challenges for the development of intelligent configuration systems come in: Web-based configurators are being made available for large heterogeneous user groups, the provision of mass-customized products requires the integration of companies along a supply chain, configuration and reconfiguration of services becomes an increasingly important issue, just to name a few.This chapter gives an overview on these current and future research issues in the domain of knowledge-based configuration technology and thus summarizes the state of the art, recent achievements, novel approaches, and open challenges in the field.

show abstract

R1: A rule-based configurer of computer systems

Cited by 732 publications

References 1 publication

Parallel production systems

Parallel production systems

The role of critiquing in cooperative problem solving

Research Issues in Knowledge-Based Configuration

Contact Info

Product

Resources

About