Abstract: The term Soft Computing (SC) represents the combination of emerging problem-solving technologies such as Fuzzy Logic (FL), Probabilistic Reasoning (PR), Neural Networks (NNs), and Genetic Algorithms (GAs). Each of these technologies provide us with complementary reasoning and searching methods to solve complex, real-world problems. After a brief description of each of these technologies, we will analyze some of their most useful combinations, such as the use of FL to control GAs and NNs parameters; the application of GAs to evolve NNs (topologies or weights) or to tune FL controllers; and the implementation of FL controllers as NNs tuned by backpropagation-type algorithms.

Abstract: Genetic Algorithms are a class of powerful, robust search techniques based on genetic inheritance and the Darwinian metaphor of “Natural Selection”. These algorithms maintain a finite memory of individual points on the search landscape known as the “population”. Members of the population are usually represented as strings written over some fixed alphabet, each of which has a scalar value attached to it reflecting its quality or “fitness”. The search may be seen as the iterative application of a number of operators, such as selection, recombination and mutation, to the population with the aim of producing progressively fitter individuals. These operators are usually static, that is to say that their mechanisms, parameters, and probability of application are fixed at the beginning and constant throughout the run of the algorithm. However, there is an increasing body of evidence that not only is there no single choice of operators which is optimal for all problems, but that in fact the optimal choice of operators for a given problem will be time-variant i.e. it will depend on such factors as the degree of convergence of the population. Based on theoretical and practical approaches, a number of authors have proposed methods of adaptively controlling one or more of the operators, usually invoking some kind of “meta-learning” algorithm, in order to try and improve the performance of the Genetic Algorithm as a function optimiser. In this paper we describe the background to these approaches, and suggest a framework for their classification, based on the learning strategy used to control them, and what facets of the algorithm are susceptible to adaptation. We then review a number of significant pieces of work within the context of this setting, and draw some conclusions about the relative merits of various approaches and promising directions for future work.

Abstract: Software agents help people with time-consuming activities. One relatively unexplored area of application is that of agents that buy and sell on behalf of users. We recently conducted a real-life experiment in creating an agent marketplace, using a slighly modified version of the Kasbah system. Approximately 200 participants intensively interacted with the system over a one-day (six-hour) period. This paper describes the set-up of the experiment, the architecture of the electronic market and the behaviours of the agents. We discuss the rationale behind the design decisions and analyze the results obtained. We conclude with a discussion of current experiments involving thousands of users interacting with the agent marketplace over a long period of time, and speculate on the long-range impact of this technology upon society and the economy.

Abstract: Intelligent agent technology is a rapidly developing area of research However, in reality, there is a truly heterogeneous body of work being carried out under the ‘agent’ banner In this paper, software agent technology is introduced by briefly overviewing the various agent types currently under investigation by researchers

Abstract: Until the late 1980s research in many-valued logic (MVL) focussed on theoretical issues in proof theory, algebra, expressivity, axiomatizability and, on the applicative side, discrete function minimization and simplification The first papers with practical implementation of deduction systems in mind came up in paraconsistent/annotated logic programming [18, 76] and in automated theorem proving [55, 108] A partial survey of the results up to 1993 is contained in [58] In the past five years deduction methods for MVL got more and more refined Recent results can match those in classical theorem proving with respect to depth and attention to detail They are not confined to mimicking improvements of deduction invented in classical logic, rather, specifically non-classical strategies are started to being pursued As can be seen from the references list of this article, there is considerable activity in MVL deduction which is why we felt justified in writing this survey Needless to say, we cannot give a general introduction to MVL in the present context For this, we have to refer to general treatments such as [153, 53, 93]

Abstract: It is by now a cliche that there is no one, universally accepted, definition of intelligent agent technology, but a number of loosely related techniques. Yet there are certain themes that appear common to agent-based systems, and, correspondingly, certain problems that must be addressed and overcome by all agent system builders. The aim of this paper is to briefly survey the tools and techniques that can be used to address these common issues, and that hence form a substrate for software agent systems. This paper begins with a review of agent communication languages, focusing particularly on the emerging standard known as KQML. Then a thumbnail sketch of various programming languages for building agent-based systems is presented, and there follows a discussion on support for ontologies, which allow agents to communicate using commonly defined terms and concepts. Then other computing infrastructure support for agent-based systems is considered, in particular, the use of client/server architectures and distributed object frameworks. Finally, some general comments and conclusions are presented.

Abstract: Evolutionary programming (EP) has been successfully applied to many parameter optimization problems. We propose a mean mutation operator, consisting of a linear combination of Gaussian and Cauchy mutations. Preliminary results indicate that both the adaptive and non-adaptive versions of the mean mutation operator are capable of producing solutions that are as good as, or better than those produced by Gaussian mutations alone. The success of the adaptive operator could be attributed to its ability to self-adapt the shape of the probability density function that generates the mutations during the run.

Abstract: This paper describes a distributed system of intelligent agents, Jasper, for performing information tasks over the Internet World Wide Web (WWW) on behalf of a community of users. Jasper can summarise and extract keywords from WWW pages and can share information among users with similar interests automatically. Jasper provides agents which can retrieve relevant WWW pages quickly and easily. A Jasper agent holds a profile of its user, based on observing their behaviour and learning more about their interests as the system is used.

Abstract: There is presented a mathematical background under the name pseudo-analysis for treating problems with uncertainty, nonlinearity and optimization in soft computing. There are given some significant examples of applications in fuzzy logics, fuzzy sets, optimization, decision theory, fuzzy neural networks, parallel processing and control theory by Hamilton–Jacobi equations.

Abstract: Description of individuals in ill-structured domains produces messy data matrices. In this context, automated classification requires the management of those kind of matrices. One of the features involved in clustering is the evaluation of distances between individuals. Then, a special function to calculate distances between individuals partially simultaneously described by qualitative and quantitative variables is required. In this paper properties and details of the metrics used by Klass in this situation is presented - Klass is a clustering system oriented to the classification of ill-structured domains which implements an adapted version of the reciprocal neighbors algorithm; it also takes advantage of any additional information that an expert can provide about the target concepts.

Abstract: Successful enterprises are built on change. Increasingly, businesses operate in a rapidly evolving environment where the response to changing markets may of necessity be measured in hours and days instead of months and years. Responsiveness and adaptability will be the hallmarks of business success. BT is strategically placed as both a major potential facilitator of this change, as well as benefiting from its technology. This paper describes how agent-based process management systems can provide powerful tools for managing the enterprise of the future. It explores recent work combining distributed computing technology with autonomous software agent techniques for business process management, and argues that these represent a viable supplement and even an alternative to existing workflow management systems. This is supported by the results of a number of projects, including ADEPT, BeaT and a number of other related schemes, which are exploring how leading edge technology can improve the way business processes are managed. This paper provides a vision of how agent-based process management systems can support the needs of the ‘virtual’ enterprise of the future and the integration of the information systems of small to medium-sized enterprises (SME).

Abstract: The system described in this paper (MORSE — movie recommendation system) makes personalised film recommendations based on what is known about users' film preferences. These are provided to the system by users rating the films they have seen on a numeric scale. MORSE is based on the principle of social filtering. The accuracy of its recommendations improves as more people use the system and as more films are rated by individual users. MORSE is currently running on BT Laboratories' World Wide Web (WWW) server. A full evaluation, described in this paper, was carried out after over 500 users had rated on average 70 films each. Also described are the motivation behind the development of MORSE, its algorithm, and how it compares and contrasts with related systems.

Abstract: In this paper proof theory of many-valued logic is connected with areas outside of logic, namely, linear optimization and computer aided logic design. By stating these not widely-known connections explicitly, I want to encourage interaction between the mentioned disciplines. Once familiar with the others’ terminology, I believe that the respective communities can greatly benefit from each other.

Abstract: Many important and useful applications for software agents require multiple agents on a network that communicate with each other. Such agents must find each other and perform a useful joint computation without having to know about every other such agent on the network. This paper describes Yenta, a matchmaker system designed to find people with similar interests and introduce them to each other. It describes how the agents that make up the matchmaking system can function in a decentralised fashion, yet can group themselves into clusters which reflect their users' interests. These clusters are then used to make introductions or allow users to send messages to others who share their interests. The algorithm uses referrals from one agent to another in the same fashion that word-of-mouth is used when people are looking for an expert. A prototype of the system has been implemented, and the results of its use are presented.

Abstract: In this paper the fundamentals of Cellular Neural Networks (CNNs) are introduced. Subsequently it is shown that, due to their locally distributed way of exchanging signals, such structures can be used as powerful devices to simulate and to reproduce, in an analog fashion and low cost, complex behaviors, i.e. dynamics commonly encountered in living systems, such as autonomous wave formation and propagation as well as morphogenetical pattern development. In fact it is proven that both of these behaviours can be simulated with CNNs with the same cell structure, and the thoroughly different dynamics can arise only suitably modulating the CNN cell parameters. Therefore a unifying approach to pattern formation and active wave propagation phenomena is presented. The derivation of the complex phenomena is analytically addressed and several simulation results are also reported.

Abstract: Agents have for a while been a key concept in artificial intelligence, but often all that the word refers to is a computational process or task with a capability for autonomous action, either alone or in an artificial society of similar agents. However, the artificial nature of these societies restricts the flexibility of agents to a point where social interaction between people and agents is blocked by significant social and psychological factors not usually considered in artificial intelligence research. This paper argues that to overcome these problems it will be necessary to return to the study of human psychology and interaction, and to introduce the concept of ‘psychological agents.’

Abstract: What is soft computing? What is fuzzy computing? What is the relationship between them? This paper intends to provide clear answers to these questions. We focus on exploring the notions of the fuzzy coordinate system and the related transformations between qualitative and quantitative information. These notions are considered to be the core ideas of fuzzy computing. Then the three novel theories of fuzzy computing and soft computing developed by the first author of this paper, namely, the Falling Shadow Representation of fuzzy theory, the Factors Space theory and the Truth Value Flow Inference theory are introduced.

Abstract: The effectiveness of evolutionary learning depends both on the variation-selection search operations used and on the structure-function relations of the organization to which these operations are applied. Some organizations—in particular those that occur in biology—are more evolution friendly than others. We describe an artificial neuromolecular (ANM) architecture that illustrates the structure-function relationships that underlie evolutionary adaptability and the manner in which these relationships can be represented in computer programs. The ANM system, a brain-like design that combines intra- and interneuronal levels of processing, can be coupled to a variety of pattern recognition-effector control tasks. The capabilities of the model, in particular its adaptability properties, are here illustrated in the context of Chinese character recognition.

Abstract: “The essence of soft computing is that unlike the traditional, hard computing, soft computing is aimed at an accommodation with the pervasive imprecision of the real world. Thus, the guiding principle of soft computing is: ‘...exploit the tolerance for imprecision, uncertainty and partial truth to achieve tractability, robustness, low solution cost and better rapport with reality’. In the final analysis, the role model for soft computing is the human mind.” [1]

Abstract: In this paper we describe the language Milord II. The description is made in terms of computer language concepts and not in terms of the logical semantics underlying it. In this sense the paper complements others in which the focus of the description has been either the object level multi-valued language description, or the reflective component of the architecture, or even the several applications built using it. All the necessary elements to understand how a system programmed in Milord II executes have room in this full description: types, facts, rules, modules, local logics, control strategies, ... Although the description is guided by the Milord II language syntax, this is by no means a user's manual, which would deserve a much longer document, but a language summary description that places all the components of the language in their correct place.

Abstract: We present a semantics for certain Fuzzy Logics of vagueness by identifying the fuzzy truth value an agent gives to a proposition with the number of independent arguments that the agent can muster in favour of that proposition.

Abstract: Given a complete and atomic Boolean algebra B, there exists a family taugamma of triangular norms on B such that, under the partial ordering of triangular norms, taugamma is a Boolean algebra isomorphic to B, where gamma is the set of all atoms in B. In other words, as we have shown in this note, every complete and atomic Boolean algebra can be represented by its own triangular norms. What we have not shown in this paper is our belief that taugamma is not unique for B and that, for such a representation, B needs neither to be complete, nor to be atomic.

Abstract: The arrangement principles and design methodology on soft computing for complex control framework of AI control system are introduced. The basis of this methodology is computer simulation of dynamics for mechanical robotic system with the help of qualitative physics and search for possible solutions by genetic algorithms (GA). New approach for direct human-robot communication with natural language (NL) and cognitive graphics is introduced. Active adaptation block which helps to mobile robot to learn a new actions and scripts based on soft computing as fuzzy neural networks, fuzzy control and genetic algorithms are proposed.

Abstract: A fuzzy data broswer for classification and prediction is described and its use demonstrated with several examples. The browser is written in the AI language Fril and provides a friendly user interface for the user to test the performance, see the effect of changes in the rules, visualise the performance and try various different forms of modelling. The rules with their associated fuzzy sets are automatically determined from a learning set of examples given in the form of a database. The fundamental theory of this approach to the automatic extraction of rules from data and the method of inference using these rules to generalise is described in simple terms. The method has wide application to data mining, fuzzy AI modelling, pattern recognition and computing with words.

Abstract: $MV$-algebras were introduced in 1958 by Chang and they are models of Lukasiewicz infinite-valued logic. Chang gives a correspondence between the category of linearly ordered $MV$-algebras and the category of linearly ordered abelian $\ell$-groups. Mundici extended this result showing a categorical equivalence between the category of the $MV$-algebras and the category of the abelian $\ell$-groups with strong unit. In this paper, starting from some definitions and results in abelian $\ell$-groups, we shall study the convergent sequences and the Cauchy sequences in an $MV$-algebra. The main result is the construction of the Cauchy completion $A^{*}$ of an $MV$-algebra $A$. It is proved that a complete $MV$-algebra is also Cauchy complete. Additional results on atomic and complete $MV$-algebras are also given.

Abstract: Topological methods have recently been developed for the classification, analysis, and synthesis of chaotic time series. These methods can be applied to time series with a Lyapunov dimension less than three. The procedure determines the stretching and squeezing mechanisms which operate to create a strange attractor and organize all the unstable periodic orbits in the attractor in a unique way. Strange attractors are identified by a set of integers. These are topological invariants for a two dimensional branched manifold, which is the infinite dissipation limit of the strange attractor. It is remarkable that this topological information can be extracted from chaotic time series. The data required for this analysis need not be extensive or exceptionally clean. The topological invariants: (1) are subject to validation/invalidation tests; (2) describe how to model the data; and (3) do not change as control parameters change. Topological analysis is the first step in a doubly discrete classification scheme for strange attractors. The second discrete classification involves specification of a 'basis set' set of periodic orbits whose presence forces the existence of all other periodic orbits in the strange attractor. The basis set of orbits does change as control parameters change. Quantitative models developed to describe time series data are tested by the methods of entrainment. This analysis procedure has been applied to analyze a number of data sets. Several analyses are described.© (1997) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Abstract: We present a new prover for propositional 3-valued logics, TAS-M3, which is an extension of the TAS-D prover for classical propositional logic. TAS-M3 uses the TAS methodology and, consequently, it is a reduction-based method. Thus, its power is based on the reductions of the size of the formula executed by the F transformation. This transformation dynamically filters the information contained in the syntactic structure of the formula to avoid as much distributions as possible, in order to improve efficiency. In our opinion, this filtering is the key of the TAS methodology which, as shown in this paper, allows the method to be extremely adaptable, because switching to different kinds of logic is possible without having to redesign the whole prover.