An analysis of network and resource indicators for resource-constrained project scheduling problem instances

In the past decades, resource on resource-constrained project scheduling problem (RCPSP) has grown rapidly, resulting in an overwhelming amount of solution procedures that provide (near)-optimal solutions a reasonable time. Despite rapid progress, little is still known what makes instance hard to solve. Inspired by previous research study shown even small instances with only up 30 activities sometimes solve, current provides analysis data used academic literature. More precisely, it investigates ability four well-known indicators predict hardness RCPSP instance. The introduces new equivalence concept show might have very different values for their without changing any possible this based theorems and search algorithm transforms existing into equivalent more compact resources. This illustrates use misleading. set computational experiment than 10,000 instances, newly constructed are not original but also often better predicting instances. It suggested further compare results obtained from dataset.