Three Different Ways for Incorporating Preference Information in Interactive Reference Point Based Methods

In this paper, we introduce new ways of utilizing preference information specified bythe decision maker in interactive reference point based methods. A reference point con-sists of desirable values for each objective function. The idea is to take the desires of thedecision maker into account more closely when projecting the reference point onto theset of nondominated solutions. In this way we can support the decision maker in findingthe most satisfactory solutions faster. In practice, we adjust the weights in the achieve-ment scalarizing function that projects the reference point. We identify different casesdepending on the amount of additional information available and demonstrate the caseswith examples. Finally, we summarize results of extensive computational tests that giveevidence of the efficiency of the ideas proposed.