Test Data Combination Strategy for Effective Test Suite Generation

Abstract

It is important to reduce the size of a test suite in order to reduce the cost and time in software testing. Teat data combination strategy is important because the size of a test suite is different by test data combination strategies even if their fault detection rates are equal. But in many cases, combination strategies include redundancy because they do not consider relationship between inputs and outputs of software under test. Also, some combination strategies considering relationship between inputs and outputs do not reduce the size of a test suite enough. In this paper, we propose test data combination strategy that improves existing test data combination strategy considering relationship between inputs and outputs. Also, we compared our strategy with existing test data combination strategies in robot software, and the experimental result shows the advantage of our algorithm.