HCI Bibliography Home | HCI Journals | About IJMMS | Journal Info | IJMMS Journal Volumes | Detailed Records | RefWorks | EndNote | Hide Abstracts
IJMMS Tables of Contents: 020304050607080910111213141516171819202122

International Journal of Man-Machine Studies 12

Editors:B. R. Gaines; D. R. Hill
Publisher:Academic Press
Standard No:ISSN 0020-7373; TA 167 A1 I5
Links:Table of Contents
  1. IJMMS 1980 Volume 12 Issue 1
  2. IJMMS 1980 Volume 12 Issue 2
  3. IJMMS 1980 Volume 12 Issue 3
  4. IJMMS 1980 Volume 12 Issue 4

IJMMS 1980 Volume 12 Issue 1

Editorial BIB 1
  E. H. Mamdani
Industrial Applications of Fuzzy Logic Control BIBA 3-10
  P. Martin Larsen
Fuzzy logic control projects related to the Technical University of Denmark are presented, and common theoretical and practical problems experienced during their realization are discussed. A detailed description is given of two rotary cement kiln control projects, which resulted in one of the first successful test runs on a full scale industrial process. It is concluded that further investigations are required with respect to the applicability of structural programming, and stability problems in fuzzy control systems.
An Analysis of Human Decision-Making in Cement Kiln Control and the Implications for Automation BIBA 11-23
  I. G. Umbers; P. J. King
This paper describes the investigation into the control skills of cement kiln operators with a view to modelling these skills by a fuzzy algorithm. The operators' strategies were studied by observing and making a detailed record of the operators' behaviour as they controlled the kiln. Analysis of the observational record suggested that many aspects of the operators' control behaviour could be modelled by a threshold logic controller. An improvement to the threshold model was attempted by comparing the operators' performance with the performance produced by a set of prescriptive control rules. Although the prescriptive rules were found to be an inadequate model of the kiln operators, analysis indicated that the rules had the potential to produce better kiln control. The implementation of the prescriptive rules as a fuzzy controller is discussed.
Computerised Ordered Metric Scaling BIBA 25-33
  D. J. Mulhall
This paper describes an automated method for scaling subjective judgements. The variables to be measured are defined either by the experimenter or the respondent and the personal intensity of each these is then scaled by comparison of pairs of predefined adjectival phases.
   A FORTRAN IV program is describe which produces questionnaires specific to particular investigations. From 1 to 15 variables may be measured on ordered metric scales varying in length from 4 points to 14 points. The same program evaluates the responses indicating whether or not they are consistent. The advantages of this type of scaling are discussed and it is argued that the method incorporates features which a new generation of psychological tests will need to adopt if they are to be applied to a dynamic model of man rather than the static model which has tended hitherto to dominate psychometrics.
On Foundations of Reasoning with Uncertain Facts and Vague Concepts BIBA 35-62
  Peter Schefe
"Fuzzy sets theory" and "fuzzy logic" based on the former have become of rapidly increasing interest. The foundations, however, are still disputed. Especially, the definitions of some set operations and logical connectives appear to be somewhat arbitrary. The relations of "fuzziness" to "probability" and "possibility" are not yet clear. This paper contains an outline of a probabilistic foundation of multi-valued ("fuzzy") reasoning. The fundamental concept is "agreement probability". It is shown that some undesirable consequences of "fuzzy logic", e.g. that tautologies of propositional calculus are not preserved can be avoided. A proposal for alternative definitions of "degree of membership" and operations on membership-graded sets is given.
   "Fuzziness" is interpreted as a subjectivistic concept, i.e. subjective uncertainty pertaining to the truth of a proposition. An important consequence thereof is that, from a graded agreement associated with a conjecture, an agreement degree pertaining to its negation cannot be computed.
   According to this foundation Shortliffe's model of medical diagnosis is reviewed as an application paradigma. There is no fundamental disagreement with Shortliffe's interpretation.
   However, Zadeh's "linguistic modelling" is shown to be inadequate. "Fuzziness" of linguistic concepts is interpreted as uncertainty of the applicability of a predicate in a given situation. This leads to the conclusion that the definition of derived concepts, especially, of hedged expressions referring to continuous scales, cannot be modelled using Zadeh's fuzzy set operations. Experimental findings of Hersh & Caramazza support an alternative interpretation.
   Particularly, Zadeh's conjecture that truth values can be equated with membership degrees is shown to be inadequate. Alternative interpretations of the linguistic phenomena considered and of the sorites paradox are given. Especially, the meta-linguistic character of the phenomena is emphasized. It is argued that "vagueness" and "uncertainty" should be clearly distinguished as well as "possibility" and "applicability". Suggestions are made how underlying measuring scales and orderings of objects are used in reasoning processes involving vague concepts.
Knowledge Acquisition by Encoding Expert Rules versus Computer Induction from Examples: A Case Study Involving Soybean Pathology BIBA 63-87
  R. S. Michalski; R. L. Chilausky
In view of growing interest in the development of knowledge-based computer consulting systems for various problem domains, the problems of knowledge acquisition have special significance. Current methods of knowledge acquisition rely entirely on the direct representation of knowledge of experts, which usually is a very time and effort consuming task. The paper presents results from an experiment to compare the above method of knowledge acquisition with a method based on inductive learning from examples. The comparison was done in the context of developing rules for soybean disease diagnosis and has demonstrated an advantage of the inductively derived rules in performing a testing task (which involved diagnosing a few hundred cases of soybean diseases).
Semantics of Implication Operators and Fuzzy Relational Products BIBA 89-116
  Wyllis Bandler; Ladislav J. Kohout
After a brief discussion of the need for fuzzy relation theory in practical systems work, the paper explains the new triangle products of relations and the sort of results to be expected from them, starting from a crisp situation. The asymmetry of these products, in contrast to correlation, is noted as essential to the investigation of hierarchial dependencies. The panoply of multi-valued implication operators, with which the fuzzification of these products can be accomplished, is presented, and a few of their properties noted. Then, most importantly, a checklist paradigm is given, by which entirely new light is thrown upon the semantics of these operators, connecting them, in a unified way, with measures which might be made upon more refined data. Using a well-known psychological test in an actual situation, so that the finer structure is in fact available, a comparison is made between a checklist measure and several of the operator values, showing the interrelationship concretely. Finally, some products and their interpretations are presented, using further real-world data.

Book Review

"System Structure in Engineering; Economic Design and Production," by O. Bjorke and O. I. Franksen BIB 117
  G. Langholz

IJMMS 1980 Volume 12 Issue 2

Cognitive Processes in Design BIBA 119-140
  Ashok Malhotra; John C. Thomas; John M. Carroll; Lance A. Miller
This paper reports on a observational study and 2 experimental studies of the design process deriving from our interest in improving the design of computer software. A model of design is developed from these studies. This consists of 3 interacting processes: goal elaboration, design generation and design evaluation. The experimental studies show that design specifications are often incorrect and incomplete with different designers paying more or less attention to different aspects of the design. Finally, from the design model and the results of the experimental studies, a number of aids to the design process are developed and discussed.
Multi-Attribute Decision-Making Using a Fuzzy, Heuristic Approach BIBA 141-156
  J. Efstathiou; V. Rajkovic
Multi-attribute decision-making is treated in this paper as a special kind of structured human problem solving. Emphasis is placed on the use of the available knowledge about utilities, which is obtained by combining heuristics and traditional aggregation methods. In this way, the problem of partial utilities and their interdependence may be solved.
   A fuzzy approach to DM is described, incorporating linguistic variables, relations and algorithms. It is summarised in a formal model and illustrated by an example.
Pattern Processing and Stability Criteria in Cellular Logic Networks BIBA 157-162
  M. C. Fairhurst; G. P. Goutos
An important factor in the application of dynamic cellular logic networks to pattern processing tasks is the stability of the network, defined in terms of the number of modes of activity which it can exhibit. This paper presents results which resolve a potential conflict in choosing the number of inputs per network element between increasing the network stability and reducing its capacity for discrimination. Implications for pattern processing in practical pattern classifiers and in neural networks are discussed.
On Casual Users BIBA 163-187
  Rodney N. Cuff
Although the computing literature carries many references to "casual users", their requirements have not been subjected to careful study. This paper constructs a detailed profile of such users, with particular reference to their demands of database query systems. It uses this profile to draw up general requirements for query systems which aim principally at casual as distinct from regular and committed users.
   Present-day systems, even those which claim an interest in the casual user, demand too detailed a knowledge of data, organisation and procedure. The assumption that any query system must be powerful (e.g. relationally complete) and suitable for all classes of user leads in practice to an interface which strongly resembles a programming language. This inhibits the casual user without a programming background, who can benefit from more guidance, less emphasis on precision and logic, and a less algorithmic stance from a system which is designed especially for him or her.
Understanding and Debugging Programs BIBA 189-202
  F. J. Lukey
A theory of program understanding and debugging is proposed. The theory is discussed with reference to a computer system that embodies many aspects of the theory. Program understanding is viewed as the construction of program descriptions. The theory's approach to debugging is based on the use of these descriptions. The role of programming knowledge is stressed.
A Man-Machine Approach to Multicriteria Decision Making BIBA 203-215
  Geoffrey G. Roy
Approaches to solving real world decision problems should recognise their multi-objective nature along with the often fuzziness of these objectives. This paper proposes that an approach using the decision maker in an interactive environment has distinct advantages over methods which use models of the decision maker's behaviour to discover "optimum" solutions. The proposed computer-aided approach avoids the problems of quantifying objective functions and allows the decision maker to explore the region of feasible solutions while imposing his own value judgements. The technique is thought to be of value in problem areas where subjectively assessed criteria are likely to play an important role, e.g. welfare and social planning, urban and regional planning, and land use management.

IJMMS 1980 Volume 12 Issue 3

Interactive Instruction in Solving Fault-Finding Problems -- An Experimental Study BIBA 217-227
  J. B. Brooke; K. D. Duncan; Caroline Cooper
Two experiments are described which investigate the effects of providing information about the diagnostic process during training in fault diagnosis. This process information is presented by an interactive computer program. The experimental results demonstrate that some generalizable strategic learning takes place. This learning is apparently dependent on changes in the problem solving context during training.
Possibility Distributions in the Linguistic Theory of Actions BIBA 229-239
  Marek Nowakowski
In connection with the formal theory of actions (the so-called linguistic theory of actions) formulated by Maria Nowakowska (1973a) some suggestions concerning its broadening have been made. The theory borders on systems theory and decisions theory giving a structural description of predecisional situations. If fuzzy sets theory is introduced to our discussion a great many interesting applications of the theory of actions may be obtained. Application of concepts proposed by L. A. Zadeh (1978) in his theory of possibility leads to interesting interpretations, illustrated by instances taken from engineering psychology and ergonomics where the formal theory of actions acquires a methodological meaning. It is due to the fact that an action in the form of a sequence and configuration of elementary acts is at the same time a form of a "language" used for description and modelling of (among others) man-machine systems.
Synthesis by Rule of Prosodic Features in Word Concatenation Synthesis BIBA 241-258
  S. J. Young; F. Fallside
The quality of speech obtainable by the technique of Word Concatenation Synthesis depends crucially on the accuracy of the pitch and timing contours which need to be computed for each utterance synthesized. A method for the synthesis-by-rule of these prosodic features is described for utterances for which the only information available is a syntactic phrase marker and the lexical structure of each component word.
   Rules and algorithms are presented for the determination of word group boundaries and the placement of stress, accent and nuclei. A timing contour algorithm is described which implements the prominence required for stressed syllables and also generates the stress-timed rhythm of natural English. A corresponding pitch contour algorithm computes appropriate intonation patterns using a minimal set of three tone groups; fall, rise and fall-rise.
The Use of Artificial Intelligence Techniques in Computer-Assisted Instruction: An Overview BIBA 259-282
  Alice Gable; Carl V. Page
One of the major goals of research in Artificial Intelligence is the representation of knowledge so that a computer can solve problems or communicate in a manner which exhibits "common sense". Few programs for computers, including those for education, possess behavior which approaches any facet of the constellation of human skills and knowledge which are imprecisely called "common sense". However, the revolutionary decline in hardware costs now makes it possible to consider economically viable, sophisticated designs for computer-aided instruction systems possessing some of the common sense attributes of a human tutor.
   In this survey we examine, in depth, techniques from Artificial Intelligence that can be used to endow a Computer-Aided Instruction system with approximations to some of the desirable qualities of a human tutor. We consider both techniques which have been proved in prototype systems for Computer-Aided Instruction and some techniques which were originally developed for other purposes.
Aspects of Possibilistic Uncertainty BIBA 283-298
  Ronald R. Yager
We introduce the concept of possibilistic uncertainty. We discuss the mathematical structure associated with this type of uncertainty and give some areas where it is applicable. We discuss some questions of exhaustiveness and exclusiveness of the possibility space. We define the concept of a possibilistic valuation. Then we apply possibility theory to ordinal decision making.
Finite Linearly Ordered Fuzzy Sets with Applications to Decisions BIBA 299-323
  Ronald R. Yager
We introduce S-fuzzy sets as fuzzy sets whose membership grades lie in a finite linearly ordered set. We extend the basic operations from fuzzy sets to these sets. We introduce the shift operation on S fuzzy sets. We then use this new operation to present a method for multiple objective decision making with ordinal information on both membership grades and importance of the objectives.

IJMMS 1980 Volume 12 Issue 4

On the Two-Dimensionality in the Behavioral System Identification Problem, Part I BIBAK 325-340
  Hugo J. J. Uyttenhove
The problem of identifying behavioral (generative) systems from data systems has been addressed within a general systems methodology for empirical investigations by processing the data in such a manner that relational properties which represent the data are discovered. It is shown in these papers that the discovery approach extends itself in two dimensions of systems modelling: (i) the dimension of invariance with respect to supporting variables, and (ii) the dimension of variance with respect to supporting variables. Although a procedural approach may be carried out in the reverse order, in this part we discuss the discovery in the first dimension. The procedures are described previously and provided in an interactive software mode by the author. An evaluation of a set of acceptable behavioral systems is given in terms of approximation and complexity. A comparison is made with the proposed evaluation technique and that of the evaluation of k-state models by the ATOM software, described by Gaines in a problem solving framework.
Keywords: Data systems, Empirical discovery, Invariance relation, Behavior, Approximation, Complexity, Mask evaluation, Problem solving, Systems modelling, k-state Moore models
Programming in the Home of the Future BIBA 341-365
  Bruce Anderson
For computers to be truly useful in the home of the future, they should be programmable by their users. The traditional idea of a program must be abandoned in favour of that of an environment of objects; programming becomes the act of augmenting this environment, and all interaction consists of evaluating expressions in it. Developments in language design and software engineering indicate how this very uniform interface might be possible, though this style of use imposes some difficult requirements. In the right such environment simple programming can yield useful results and lead to greater understanding both of the programming process and of the problem being approached. Such a powerful system could be used in three main areas: the local manipulation of data, including games and simulations; device control, which can be sophisticated and energy-conserving; and to communicate with other systems. This last possibility leads us to discuss several uses including mailing, teleconferencing, and access to databases of various kinds. Some social effects of the widespread availability of computer power are indicated.
Q-Analysis and the Structure of Friendship Networks BIBA 367-378
  Linton C. Freeman
This is a study of a collection of social scientists. It examines the social context out of which they develop close friendship ties and the impact of an extended computer conference on such ties. The perspective of R. H. Atkin (1974) is used to uncover shifts in the structure of ties among the participants.
A Method for the Inference of Non-Recursive Context-Free Grammars BIBA 379-387
  C. Chirathamjaree; Martin H. Ackroyd
A practical method is presented for the automatic generation of a non-recursive context-free grammar (cfg) from a set of strings that the cfg if required to be capable of producing. The method is efficient in computing time by comparison with enumerative methods.
Face-to-Face Guidance of University Computer Users -- I. A Study of Advisory Services BIBA 389-405
  J. L. Alty; M. J. Coombs
This paper reports the results of a quantitative study of the use of guidance systems by university computer users. The study has identified the importance of face-to-face guidance services, particularly those provided by the official advisory service of a computer centre. Data from three other university sites have supported the Liverpool findings and have indicated that they can be generalized across university sites. A study of user attitudes to the official advisory service has revealed that whilst users value the service as a source of sound operational advice they often do not understand the advice given. Moreover, the experience of attending the service is often not a pleasant one. This was recognized as a source of dissonant attitudes which could inhibit the effectiveness of the service. A more detailed analysis revealed that this dissonance was most marked with inexpert users. Such results suggest that present advisory services cater better for the expert user whereas the need of the inexpert user ought to be the principal guiding factor in advisory service provision.
Face-to-Face Guidance of University Computer Users -- II. Characterizing Advisory Interactions BIBA 407-429
  M. J. Coombs; J. L. Alty
This paper reports the results of an investigation carried out to ascertain if face-to-face advisory service interactions in a university computer centre can be regarded as an identifiable category of face-to-face encounter. A further objective of the study was to determine any independent variables affecting the course of an interaction. An analysis of participant goals has revealed a simple 3-part structure and a high degree of uniformity was observed across our sample. Furthermore, studies at three other university sites have supported the Liverpool findings. Two major characteristics of the interactions themselves have been identified -- significant advisor control and the operational nature of the advice give. Quantitative measures for both these characteristics have been defined. Expertise has been identified as the key independent variable in advisory interactions and a re-examination of the data has indicated that advisors exert far more control over inexpert users and are less responsive to their needs. It seems that the very people who would benefit from explanation and sensitive guidance do not receive it and that a real improvement in the efficiency of the advisory service might result from an awareness of this fact and better advisor training.