See the paper by Chris Bartel: https://bartel.cems.umn.edu/sites/bartel.cems.umn.edu/files/2022-07/bartel.bartel_2022-j.mater_.sci_.pdf.
The idea is to find the lowest energy structure at each composition. A Monte-Carlo approach where you move the lithium around within the TiO lattice might be one approach. There are also "cluster expansion" methods that are claimed to be more efficient but I have not used those.