Benchmarking CHC on a New Application: The Software Project Scheduling Problem

In this article we analyze the behavior and scalability of the CHC algorithm over a benchmark of instances of the software project scheduling problem. Our goal is to analyze the performance of the CHC algorithm when solving realistic NP-hard combinatorial problems and test whether its previously reported high performance on similar problems also holds on this one. We perform a preliminary study to obtain a suitable configuration of the parameters in the algorithm. After choosing the configuration, we show the results for the problem instances in the benchmark. To give a reference on how CHC performs and scales, its results are compared against those of a GA. We conclude that CHC outperforms GA in large problem instances. Moreover, CHC produces promising results for the software project scheduling problem domain, and could be used by practitioners.