Comparing interactive evolutionary multiobjective optimization methods with an artificial decision maker

Abstract Solving multiobjective optimization problems with interactive methods enables a decision maker domain expertise to direct the search for most preferred trade-offs preference information and learn about problem. There are different methods, it is important compare them find best-suited one solving problem in question. Comparisons real makers expensive, artificial (ADMs) have been proposed simulate humans basic testing before involving makers. Existing ADMs only consider type of information. In this paper, we propose ADM-II, which tailored assess several evolutionary able handle types We two phases solution processes, i.e., learning separately, so that ADM-II generates ways each reflect nature phases. demonstrate how can be applied problems. also an indicator performance methods.