Building the CLMPST 2023 Schedule Using Phylogenetic Algorithms

ROFFÉ, ARIEL JONATHAN

Capital Federal

Congreso; 17th CLMPST; 2023

There are a number of existing approaches to schedule building/optimization in the literature (Oude Vrielink et al., 2019). In this presentation, we describe how the CLMPST 2023 schedule was built using phylogenetic algorithms.In a sense, the problem is similar to the one faced by phylogeneticists when trying to reconstruct ancestry relationships between a set of species (see, e.g., Goloboff, 2022). The number of possible solutions (i.e., trees/cladograms in one case, possible schedules in the other) becomes astronomical when considering an interesting number of organisms/papers. There are also conflicting requirements (i.e., primary homology statements that group the organisms differently; the requirements mentioned above) that can be weighted differently (e.g., by a priori and/or a posteriori weighting schemes). Thus, a number of heuristic algorithms used by phylogeneticists (random addition sequences, swapping methods, etc.) to find optimal trees were applied to the schedule-building problem. These algorithms will be described in more detail and shown to work during the presentation.The main idea for this work came from a formal reconstruction of cladistic/parsimony methods, a classical methodology in the philosophy of science (Roffé, 2020). Looking at these methods in an abstract fashion highlighted the similarities mentioned above. Thus, the utilization of those methods’ techniques will also serve as an illustration of the advantages that formal reconstructions can bring to the fore.