Quantitative evaluation of safety-critical software at the early development stage: An interposing logic system software example

Abstract

In order to produce highly reliable software, designers and developers pay particular attention to the early development phases, that is, the requirement phase and the design phase. In this work, a few novel quantitative measures are suggested for estimating how much influence the development methods at the early stage have on the reliability elevation. Three complexity measures are suggested in this work, which are Operation Complexity, Interface Complexity, and Micro-Complexity. These measures are based on the entropy concept. The entropy-based complexities need information on probability distributions. For this purpose, the operational profile for Operation Complexity, the relative frequencies of module couplings for Interface Complexity, and the relative frequencies of module strengths for Micro-Complexity are introduced. We further developed estimation standards for these complexities. In order to demonstrate the utility of these standards, the software design of a digital Interposing Logic System (ILS) of an Engineered Safety Feature Actuation System (ESFAS) in nuclear power plants has been evaluated using the standards presented in this work. The results show that the proposed standards are useful in developing more reliable softwares and minimizing the efforts or costs for coding, testing, etc.