Specifications in stochastic process algebra for a robot control problem

Abstract

We present a novel approach to specification of dynamic systems. This approach, a stochastic extension of process algebra, facilitates quantitative, or performance, analysis, in addition to qualitative analysis. For unreliable systems this integrated approach encourages the investigation of the impact of functional characteristics on the performance of the system. Throughout the paper details of the stochastic process algebra are made concrete via an example: a robot control problem. Two specifications are presented of this problem. The first, an idealisation, does not represent the possibility of failures. The second models both failures and recoveries. Each is solved to obtain performance measures for the system.