PARS: A Process Algebra with Resources and Schedulers MohammadReza Mousavi, Michel Reniers, Twan Basten, Michel Chaudron Eindhoven University of Technology, Post Box 513, NL-5600 MB, Eindhoven, The Netherlands Email: {m.r.mousavi,m.a.reniers,a.a.basten,m.r.v.chaudron}@tue.nl Abstract. In this paper, we introduce a dense time process algebraic formalism with support for specification of (shared) resource require- ments and resource schedulers. The goal of this approach is to facili- tate and formalize introduction of scheduling concepts into process al- gebraic specification using separate specifications for resource requiring processes, schedulers and systems composing the two. The benefits of this research are twofold. Firstly, it allows for formal investigation of scheduling strategies. Secondly, it provides the basis for an extension of schedulability analysis techniques to the formal verification process, facil- itating the modelling of real-time systems in a process algebraic manner using the rich background of research in scheduling theory. 1 Introduction Scheduling theory has a rich history of research in computer science dating back to the 60’s and early 70’s. Process algebras have been studied as a formal the- ory of system design and verification since about the same time. These theories have remained separate until recently some connections have been investigated. However, combining scheduling theory in a process algebraic design still involves many theoretical and practical complications. In this paper, building upon pre- vious attempts in this direction, we propose a process algebra for the design of scheduled real-time systems called PARS (for Process Algebra with Resources and Schedulers). Previous attempts to incorporate scheduling algorithms in pro- cess algebra either did not have an explicit notion of schedulers such as that of [3,12,13] (thus, coding the scheduling policy in the process specification) or scheduling is treated for restricted cases such as those of [4,10] (that only support single-processor scheduling). Our approach to modelling scheduled systems is depicted in Figure 1. Process specification (including aspects such as causal relations of actions, their timing and resource requirements) is separated from specification of schedulers. System level specification consists of applying schedulers to process specifications, on the one hand, and composing scheduled systems, on the other hand. A distin- guishing feature of our process algebra is the possibility of specifying schedulers as process terms (similar to resource-consuming processes). Another advantage of the proposed approach is the separation between process specification and