Petri网和代数并发语义

PETRI NETS VS ALGEBRAIC PARALLEL SEMANTICS

Milner认为：“把代数并发性方面的工作与Petri在Petri网上早早取得的丰硕成果联系起来，是一个非常有希望的发展线索．然而由于它们的概念基础并不完全吻合，所以还有困难．”本文即从Milner的图灵奖演说入手，介绍Petri网与л演算相公的若干基本内容和基本观点，并找出Petri网与代数并发语义的根本差别．

It is desirable to build semantics of parallel systems from their syntactic structures.лcalculus is the result of R. Mimer's effort in 1991. He noticed in the course of research that C. A.Petri has worked on concurrency for many years with fruitful achievements. Mimer considers it ameaningful research direction to combine his own work on algebraic semantics with Petri's work onPetri Nets, though it's bound to be a difficult one. In his Turing Award Lecture, Mimer made itclear that there are fundamental differences between his work and Petri's. This paper tries toexplore such differences, starting from a brief introduction to both Petri Nets and л-calculus. Likemost of parallel system models, л-calculus takes as its foundation the interleaving assumption, i. e.the consequence of the parallel execution of events a and b is the same as a and b are executed inarbitrary order. Petri Nets does not accept this assumption in general. To Petri Nets, it is true only ifevents a and b are independent. Different attitudes to interleaving assumption lead to differentsystem structures, different treatments of parallel behavior, and other differences. Examples aredrawn from Mimer's Turing Award Lecture.