Baseline performance in multi-project scheduling for SMEs: evidence from microsoft project and MPSPLib

Small and medium-sized enterprises (SMEs) often manage multiple concurrent projects with limited resources but rely on general-purpose commercial project scheduling software due to affordability and accessibility constraints. While the research community has advanced powerful heuristic algorithms validated on benchmarks, the performance of commercial tools in complex multi-project environments has received little systematic evaluation. This paper evaluates Microsoft Project 2024, considered as a representative example of a widely available general-purpose scheduling tool, on the full MPSPLib benchmark suite (140 multi-project instances) to quantify its scheduling capabilities and limitations. We assess performance from two complementary time-based perspectives: portfolio-level efficiency, measured through total makespan (TMS), and project-level impact, measured through average project delay (APD). We compare the schedules produced by Microsoft Project’s resource-leveling engine with the best-known results from the literature and find systematic overruns with distributional differences across portfolio configurations. These results are not intended as a critique of Microsoft Project, which was never designed as a high-performance optimizer, but as evidence of the performance gap between a widely adopted commercial tool and state-of-the-art heuristics. The findings provide SMEs with empirical insight into the performance of default resource-leveling procedures in multi-project environments and highlight opportunities to transfer academic advances into commercially usable features.