In this paper a modification of the parallel machines scheduling problem with machines capacity for the purpose of running software tests in a cloud environment is considered. The goal function to minimise is the sum of the makespan and the time needed to obtain the schedule. The processing times of tests are unknown, but can be estimated with the use of the history of execution of previous tests. A mathematical model of the problem is presented and two solving methods, adapted Largest Processing Time algorithm and Simulated Annealing metaheuristics are implemented and compared in a computer experiment using real-life data. The results indicate that the Simulated Annealing method is better up to around 14,000 tests despite its longer running time. Results also show that, when paired with a simple method of estimation of software tests duration, both solving methods are fairly robust, providing schedules with similar quality to the ones obtained without estimation.
CITATION STYLE
Rudy, J. (2020). Algorithm-Aware Makespan Minimisation for Software Testing Under Uncertainty. In Advances in Intelligent Systems and Computing (Vol. 987, pp. 435–445). Springer Verlag. https://doi.org/10.1007/978-3-030-19501-4_43
Mendeley helps you to discover research relevant for your work.