Grammar Systems:

 Recent Results and Perspectives (Foreword)

  Erzsébet CSUHAJ-VARJÚ

 On July 26-27, 1996, a workshop with the title Grammar Systems: Recent Results and Perspectives was held in Budapest, at the Computer and Automation Research Institute of the Hungarian Academy of Sciences{. The aim of the meeting was to provide a forum for exchanging ideas about the state of the art of research in the area of grammar systems and preview the trends and perspectives. The presented talks and the fruitful informal discussions resulted in, among other things, the present volume.

 " Grammar systems" is a recent active field of formal language theory, providing syntactic models, frameworks and tools for describing and studying (the behaviour of) multi-agent systems at the symbolic level. Several scientific areas have motivated and influenced the developments in this theory: distributed and decentralized artificial intelligence, distributed and parallel computing, artificial life, molecular computing, robotics, ecology, sociology, etc. Computer networks, parallel and distributed computer architectures, distributed and cooperative text processing, natural language processing are candidates for possible applications.

 Roughly speaking, a grammar system (the term "grammar" is used here in a general sense) consists of several language identifying devices (language processors or linguistic agents) that jointly develop a common symbolic environment (usually, a string or a finite set of strings) by applying string manipulating operations to it. The symbolic environment can be shared by the components of the system, or it can be given in the form of a collection of separated sub-environments, each belonging to a language processor. At any moment in time, the state of the system is represented by the current string describing the environment (collection of strings of the sub-environments). The functioning of the system is realized by changes in its states. Depending on the variant of multi-agent systems that the actual grammar system represents, in addition to performing derivation steps, the language processors are allowed to communicate with each other. Usually, this is done by exchange of strings that can be data (for example, sentential forms in derivation) or programs (productions or coded form of some operation). The behaviour of the grammar system can be characterized by the set of sequences of environmental states following each other, starting from an initial state, or by the set of all environmental states originating from an initial state and satisfying some criteria (final states).

 Grammar systems are both computational and language identifying devices, capturing several phenomena characteristic for multi-agent systems: cooperation, distribution, communication, parallelism, emergent behaviour, etc.

 Gyenizse Pal 1997. Május 21. Szerda 13:01:34 MET DST