An analysis of decomposition approaches in multi-objectivization via segmentation

- Contrary to prior evidence when studying the TSP, spatial decompositions can be as competitive as adaptive approaches.
- Decomposition methods that isolate cities within and between neighborhoods are most successful at guiding the search procedure.
- New decompositions of the TSP are analyzed and some are shown to perform better than the previous best methods.
- A progressive multi-objectivization method is proposed and shown to provide competitive performance regardless of computational budget.