摘要:

Abstract:A genetic algorithm‐based (GA‐based) system (GAPOLE) is used to evolve a self‐learning, self‐optimising control strategy for a typical inherently unstable, dynamic system—a simulated pole‐cart system. The dynamics of the system are unknown to GAPOLE. The only information for evaluating performance is a signal indicating that the pole‐cart system is out of control. This presents a genuinely difficult credit assignment problem. We present some evidence which shows that GAPOLE compares well with the best alternative methods, but it is noteworthy that it is most robust. It is argued that maintaining a population of partial solutions offers some advantages: GA‐based algorithms can deliver more than one good solution at a time; they are able to adapt better in complex changing conditions. Results characterising the performance of the method as population size is varied are also presented. The results show that GAPOLE performed best with a population size of 300. Therefore it is suggested that this parameter, like other GA parameters, may have to be tailored to a particular application. This appears to contradict an earlier claim that a population size in the range of 60–110 is optimal for genetic algorithm‐based applications (Grefenstette 1

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00021.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY

摘要:

Abstract:The KADS methodology (Schreiber et al. 1993; Tansley&Hayball 1993) and its successor, CommonKADS (Wielinga et al. 1992) have proved to be very useful approaches for modelling the various transformations involved between eliciting knowledge from an expert and encoding this knowledge in a computer program. These transformations are represented in a series of models. While it is widely agreed that these methods are excellent approaches from a theoretical viewpoint, the documentation provided concentrates on defining what models should be produced, with only general guidance on how the models should be produced. This has the advantage of making KADS and CommonKADS widely applicable, but it also means that considerable training and experience is required to become proficient in them. This paper reviews three projects that investigated the feasibility of producing specific guidance for certain decisions which are required when using KADS or CommonKADS to develop a knowledge‐based system. Guidance was produced for the identification of the generic task addressed by a knowledge‐based system; for the selection of appropriate AI techniques for implementing the analysed knowledge; and for selecting a suitable tool for implementing the system. Each set of guidance was encoded in its own knowledge‐based system, which was itself developed with the assistance of KADS or CommonKADS. These projects therefore both studied and applied KADS and CommonKADS in order to produce knowledge‐based guidance for knowledge engineers. The projects showed that it was feasible to produce heuristic guidance which could be understood, applied and occasionally overridden by knowledge engineers. The guidance provides reasonably experienced knowledge engineers with a framework for making the key decisions required by CommonKADS, in the same way that CommonKADS provides knowledge engineers with a framework for representing knowledge. The projects also produced some new insights about CommonKADS domain modelling and about the process of task identif

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00022.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY

摘要:

Abstract:Expert systems have traditionally been designed with one particular purpose in mind, whether problem‐solving, critiquing etc. When the system is intended for users with varying background knowledge it becomes necessary to provide different facilities for different groups of users, which in turn puts demands on how the knowledge must be represented. This paper describes an industrial prototype—Protein Purification Advisor (PPA)—in the area of protein purification, and how a suitable choice of knowledge representation has allowed us to use the same knowledge for a number of purposes including those mentioned above. Particular interest is laid on the topic of how to represent the heuristic knowledge and how to use it in actual decision

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00023.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY

摘要:

Abstract:The recent surge of interest in connectionist models arose through the availability of high speed parallel supercomputers and the advent of new learning algorithms. The computations performed on concurrent architectures are less costly than similar ones performed on sequential machines. In this paper, the design and implementation of a parallel version of fuzzy ARTMAP (Carpenter et al. 1992), which encompasses both neural and fuzzy logic, is discussed. Fuzzy ARTMAP is a supervised learning algorithm utilising two fuzzy ART modules and an associated mapping network. A simplified version of fuzzy ARTMAP (SFAM) was designed by incorporating a simplification of the match tracking concept on unsupervised fuzzy ART paradigms. The proposed simplified version consists of only one fuzzy ART module and an associated mapping network. A parallel fuzzy ARTMAP (PFAM) algorithm is then designed and implemented on a hypercube simulator (iPSC). The algorithm is parallelised for any architecture and, with the exception of issues related to communications, the implementation remains the same on any type of parallel machine. PFAM enjoys the advantage of reduced training time that makes the algorithm a successful candidate for applications that require both online testing and training. Such applications can range from underwater sonar detection and chemical plant processing control to nuclear reactor process control, flexible manufacturing and systems analysis.

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00024.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00025.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00026.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00027.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00028.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00030.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY

ISSN:0266-4720    

DOI:10.1111/j.1468-0394.1995.tb00031.x

出版商:Blackwell Publishing Ltd    

年代:1995

数据来源: WILEY