Representing the Finite calculus in Multi Interaction Nets Concurrency Interaction Non determinism
نویسنده
چکیده
We extend the Interaction Nets of Lafont with some non determinism capabilities and then show how to implement the nite monadic calculus in that system Introduction The calculus of Milner et al is one of the most popular theoretical tools for the investigation of concurrent computations Its popularity is due to its con ceptual simplicity yet great expressive power However similar to the calculus the calculus leaves some of the basic low level components of computation im plicit namely the distribution of values and synchronization expressed as the global process of substitution Uncovering this ner computational structure and dispensing with the important role that syntactic entities like names play in calculus is the goal of this paper Interaction Nets IN of Lafont are a novel model of parallel compu tation simple and elegant Their essential properties as relating to parallelism are the locality of interaction and the simplicity of the rewriting process We in troduce a version of IN extended with non determinism Multi Interaction Nets and translate faithfully the calculus to MIN Our translation is similar to the nets of Milner and the Interaction Diagrams of Parrow but we represent blocking synchronization in a distributed manner without using boxes We also compare our construction to the Concurrent Combinators of Honda and Yoshida a Interaction Nets and Non Determinism Interaction Nets IN were introduced by Lafont as a simple and elegant model of parallel computation inspired by Linear Logic s Proof Nets and de signed to be useful both as an abstract programming language and as a useful intermediate representation Supported by a University of Alberta Dissertation Scholarship INs are graphs consisting of nodes agents of certain types and undirected edges connecting them The points of contact of edges and nodes are called ports and every node type has a particular signature of ports In conventional INs every node type has exactly one principal port denoted with a bold triangle and every port hosts exactly one edge Computation in INs is governed by interaction rules of the form
منابع مشابه
Embedding the Finitary Pi-Calculus in Differential Interaction Nets
We propose a translation of a finitary (that is, replicationfree) version of the monadic localised pi-calculus into the purely exponential part of promotion-free differential interaction nets. This embedding is a simulation of reduction. Since the introduction of Linear Logic by Girard in 1986, it was clear to many logicians and computer scientists that some deep connection between this new log...
متن کاملOn differential interaction nets and the pi-calculus
We propose a translation of a finitary (that is, replication-free) version of the pi-calculus into promotionfree differential interaction net structures, a linear logic version of the differential lambda-calculus (or, more precisely, of a resource lambda-calculus). For the sake of simplicity only, we restrict our attention to a monadic version of the pi-calculus, so that the differential intera...
متن کاملMultiport Interaction Nets and Concurrency
We consider an extension of Lafont’s Interaction Nets, called Multiport Interaction Nets, and show that they are a model of concurrent computation by encoding the full π-calculus in them. We thus obtain a faithful graphical representation of the π-calculus in which every reduction step is decomposed in fully local graph-rewriting rules.
متن کاملFormal Modeling of Temporal Interaction Aspects in Multi-Agent Systems
Multi-agent interaction protocols play a crucial role in multi-agent systems (MAS) development. They are used to manage and to control interactions among several autonomous agents in a MAS. Their formal specification, as well as their verification, constitute an essential task for the design of MAS applications. Several approaches have been proposed to formally represent agent interaction proto...
متن کاملPi-nets: interaction nets for pi-calculus
π-calculus is a framework that aims to describe concurrent calculations through a formal definition of processes. Originally, π-calculus is defined by a formal language and a set of reduction rules, much in the spirit of λ-calculus. Our aim is to provide a graphical representation of π-calculus using multi-wired interaction nets, in the spirit this time of proof-nets of linear logic. 1 π-calcul...
متن کامل