INTERACTIVE MULTIOBJECTIVE TABU/SCATTER SEARCH BASED ON REFERENCE POINT

Abstract

This paper presents multiobjective tabu/scatter search architecture with preference information based on reference points for problems of contin- uous nature. Features of this new version are: its interactive behavior, its deterministic approximation to Pareto-optimality solutions near the refer- ence point, and the possibility to change progressively the reference point to explore different preference regions. The approach does not impose any restrictions with respect to the location of the reference points in the objective space. On 2-objective to 10-objective optimization test problems the modified approach shows its efficacy and efficiency to find an adequate non-dominated set of solutions in the preferred region.

Keywords

• multiple objectives
• metaheuristics
• reference point
• continuous optimization