Batch size estimate

In this work we analyze the main batch resolution algorithms. We particularly focus on the tree-based class to underline how their efficiency depends on the batch size. In fact, batch size is a critical parameter when using smart resolution strategies that take advantage this information to improve resolution efficiency. The dissertation will continue with the analysis of noteworthy techniques available in literature for the batch size estimate: in fact, original papers pay attention on the resolution process and leave the estimate problem in the background. Finally we propose and analyze GEGA, an estimate algorithm particularly good in terms of estimate accuracy over time taken by the estimate process.