| dc.description.abstract | In this project, bin-packing has been analysed as a way of better utilizing
resources in a continuous integration system in order to increase its efficiency. The
worst-fit, best-fit, next-fit and first-fit algorithms have been investigated in conjunction
with an optimal job scheduling approach. The many differences between bin-packing
and software testing inspired a solution combining several approaches. The ROST
algorithm, short for Resource-Optimized-Software-Testing algorithm was developed
to be adapted to software testing. The ROST algorithm is a combination of an
optimal job scheduling approach, the worst-fit and the best-fit algorithm where the
algorithm changes based on the current bin capacity. The results show that it is just
as good or better than the other approaches it has been compared to, and
significantly better than the most common solution in continuous integration. | en_US |
