Cao, FangfangServranckx, TomHe, ZhengwenVanhoucke, Mario2025-09-292025-10-082025-10-082025-101094-613610.1007/s10951-025-00835-2https://repository.vlerick.com/handle/20.500.12127/7746We propose a novel approach for sizing the activity buffers in the project by clustering similar activities and allocating the buffers using a unique attribute in each cluster. Since the number of clusters as well as the assignment of attributes to these clusters has an impact on the buffer sizing, the problem is solved using an adapted multifactorial evolutionary algorithm (aMFEA) in which multiple buffer allocation problems (BAPs) are solved simultaneously. Several decoding schemes are compared to improve the synergies between the different BAPs and the evolutionary operators. The results show the added value of the evolutionary components of the aMFEA and show that the proposed approach is superior to existing benchmarking procedures. Furthermore, the solution quality improves with an increasing number of clusters, while the solution quality goes down again as the number of clusters becomes too large. From a practical perspective, this study highlights the need to identify good activity attributes that are linked to the buffer sizing decisions and the importance of activity clustering in order to reduce the time and effort needed for better buffer sizing decisions.enBuffer Allocation ProblemClusteringMultifactorial evolutionary algorithmProject SchedulingJournal of Scheduling1099-142558614