Are infinite behaviours of parallel system schemata necessary?
A behaviour of a parallel system schema is a sequence of suitably choosen events occuring in progress of a computation induced by the schema. There is an evidence that any "resonable" parallel system schema is describable by means of elementary-recursive formulas with finite behaviours as operand values. There are some features usually inexpressible in any recursive way but nonetheless expressible by formulas with finite behaviours as operand values. In their number: deadlock, strong form of fairness and starvation. A weak, more adequate, form of fairness and starvation seem to exemplify features expressible by formulas with operands assuming merely infinite behaviours as their values. It turns out, however, the later formulas have their equivalent counterparts with finite behaviours as operand values. The paper discusses the issues in detail.
KeywordsMutual Exclusion Information Processing Letter Weak Fairness Synchronization Rule Infinite Behaviour
Unable to display preview. Download preview PDF.
- L. Czaja, A specification of parallel problems, Information Processing Letters, 8(4) (1979) 162–167Google Scholar
- L. Czaja, Parallel system schemas and their relation to automata, Information Processing Letters, 10(3) (1980) 153–158Google Scholar
- L. Czaja, Deadlock and fairness in parallel schemas: a set-theoretic characterization and decision problems, Information Processing Letters, 10 (4,5) (1980) 234–239Google Scholar
- D. König, Über eine Schlussweise aus dem Endlichen ins Umendliche, Acta Litt. Ac. Sci. Hung. Fran. Josep. 3 (1927), 121–130Google Scholar