Optimizing software integration by considering integration test complexity and test effort

Software integration testing is often a bottleneck in the development process. Selecting the next component to integrate often depends heavily on integrators expertise or solely on the integration schedule. This may lead to an increasing number of stubs and makes testing more difficult. In this work we present a novel metric to calculate the test complexity of a certain integration order and provide an approach for optimizing it with respect to the integration test complexity and the integration test effort using simulated annealing.