Test Coverage Optimization for Large Code Problems

Because running all previous tests for the regression testing of a system is time-consuming, the size of a test suite of the system must be reduced intelligently with adequate test coverage and without compromising its fault detection capability. Five algorithms were designed for reducing the size of test suites where two metrics, test's function reach ability and function's test intensity, were defined. Approaches to the algorithm CW-NumMin, CW-CostMin, or CW-CostCov-B are the safe-mode of test case selection with full-modified function coverage, while the CW-CovMax algorithm is of non-safe mode, which was performed under time restriction. In this study, the most efficient algorithm could reduce the cost (time) of a test suite down to 1.10%, on the average, over the MPLS area of Cisco IOS.