Affiliations: [a] CWI, P.D. Box 4079, 1009 AB Amsterdam, The Netherlands | [b] Department of Computer Science & Systems McMasler University, Hamilton, Ont. L8S 4K1, Canada
Note: [*] The research of Jan Rutten was partially supported by ESPRIT project 415: Parallel Architectures and Languages for Advanced Information Processing – a VLSI-directed approach.
Note: [**] The research of Jeffery Zucker was supported by the National Science Foundation under grant no. DCR-8504296 and by a grant from the Natural Sciences and Engineering Research Council of Canada.
Abstract: In the semantic framework of metric process theory, we undertake a general investigation of fairness of processes from two points of view: (1) intrinsic fairness of processes, and (2) fair operations on processes. Regarding (1), we shall define a “fairification” operation on processes called Fair such that for every (generally unfair) process p the process Fair(p) is fair, and contains precisely those paths of p that are fair. Its definition uses systematic alternation of random choices. The second part of this paper treats the notion of fair operations on processes: suppose given an operator on processes (like merge, or infinite iteration), we want to define a fair version of it. For the operation of infinite iteration we define a fair version, again by a “fair scheduling” technique.
Keywords: Fairness, semantic domains of metric processes, fair infinite iteration, alternation of random choices
