Modelling and optimal design of large-scale, real-time computer systems

The design of large-scale, distributed computer control system requires thorough analysis of application requirements and a `good choice´ for quality measure which can be used in ranking alternatives or as objective in an optimization process. A set of models and algorithms are presented which can solve these problems efficiently. The methods discussed have been implemented on computer code, which uses a modified version of Avizienis model (1980) to evaluate the probabilistic behaviour of the physical processes. The code has been used primarily in the field of nuclear power plants