A Search-Based Approach to Generate MC/DC Test Data for OCL Constraints

Abstract

Automated generation of test data is an important and challenging activity in Model-based Testing. This typically requires solving of constraints, written in Object Constraint Language (OCL), specified on models in order to obtain solutions that can be used as test data. Test data generation techniques in the literature discuss various coverage criteria for test generation to achieve a sufficient level of coverage. One of the recommended criteria is modified condition/decision coverage (MC/DC) that is a requirement of different safety standards, such as DO-178C. In this paper, we propose a search-based strategy that utilizes case-based reasoning (CBR) to reuse the already generated test data and generate new test data that provides MC/DC coverage of OCL constraints. To evaluate the performance of the proposed approach in solving MC/DC constraints, we perform an empirical evaluation using AVM without CBR, AVM with CBR, and use Random Search (RS) as a baseline for comparison. We use 84 OCL constraints from four case studies belonging to different domains with varying size and complexity. The experimental results show that our proposed strategy of reusing already generated test data is better as compared to generating test data without using previous test data.