2.1 Summary of Conclusions

The main objective of this thesis was to develop and analyse a PSO-based MOA for DMOO. However, in order to determine whether the algorithm efficiently solves DMOOPs, benchmark functions are required that are representative of typical real-world problems. Furthermore, performance measures are required to quantify the performance of the algorithm. Chapter 3 provided a comprehensive overview of benchmark functions used for DMOO. The overview highlighted three limitations of current DMOOPs, namely: all benchmark functions have a POS that is defined by linear functions and all decision variables have the same POS, none of the DMOOPs have an isolated POF, and none of the DMOOPs have a deceptive POF. In order to address these shortcomings of currently used DMOOPs, new DMOOPs were proposed that have POSs defined by non-linear functions and where each