Generating linear temporal logic formulas for complex pattern -based specifications
Formal verification techniques such as theorem proving, runtime verification, and model checking have been shown to improve program dependability. These techniques check the correctness of the system against specifications written in a formal language. Many of the verification tools, such as the highly acclaimed model checker SPIN, use Linear Temporal Logic (LTL) to specify system properties. While LTL is highly expressive and supports specification of software properties such as safety and liveness, it requires a high level of mathematical sophistication to write, read, and validate properties. There have been a number of research efforts to facilitate the generation of formal specifications in LTL using the notion of property patterns and scopes. These efforts, however, lack the general and consistent formalization of these patterns and scopes when using multiple propositions. The dissertation research formalized the definitions of patterns and scopes including those that involve more than one proposition, and it defined an approach to automate the generation of LTL formulas for a wide range of complex software properties. The research verified the correctness of generated LTL formulas through formal proofs, and it defined techniques for testing the correctness of the formulas and for supporting their validation by visually representing the formulas using traces of computations. The work is significant because it supports the generation of provably correct LTL specifications that can be used by a variety of formal verification techniques, and it provides an approach for clarifying and understanding properties in particular those related to sequential and concurrent properties that are characteristic of reactive systems.
Salamah, Salamah Ibrahim, "Generating linear temporal logic formulas for complex pattern -based specifications" (2007). ETD Collection for University of Texas, El Paso. AAI3262910.