Automated program flaw finding using simulated annealing

TLDR

One of the major costs in a software project is the construction of test‑data. This paper outlines a generalised test‑case data generation framework based on optimisation techniques. The framework incorporates multiple testing criteria for functional and non‑functional properties and is applied to testing specification failures and exception conditions. Case studies demonstrate the efficiency and effectiveness of this dynamic optimisation‑based test‑data generation approach.

Abstract

One of the major costs in a software project is the construction of test-data. This paper outlines a generalised test-case data generation framework based on optimisation techniques. The framework can incorporate a number of testing criteria, for both functional and non-functional properties. Application of the optimisation framework to testing specification failures and exception conditions is illustrated. The results of a number of small case studies are presented and show the efficiency and effectiveness of this dynamic optimisation-base approach to generating test-data.