@inproceedings {pub2871,
	title = {Reference Vector based a posteriori Preference Articulation for Evolutionary Multiobjective Optimization},
	author = {Ran Cheng AND Markus Olhofer AND Yaochu Jin},
	year = {2015},
	month = {May},
	abstract = {The result of a multi objective optimization is usually a set of optimal trade-off solutions for the different criteria. In order to utilize the results, the so called Pareto set, a final decision making process is necessary in most cases in which one single solution has to be selected. In this process a decision maker selects one of the solutions in the set according to his or her preferences and often also based on knowledge gained by observing the generated Pareto surface. Due to the finite number of individual solutions in the Pareto set, the decision maker often faces the problem that the sampling of different trade-off solutions in the preferred region is too dense to identify the ideal solution according to the preferences. This effect becomes a serious problem with increasing number of objectives due to the high dimensionality of the resulting Pareto fronts. In this case an interpolation between best fitting solutions can be generated, with the drawback that the result might be non optimal. Alternatively a time consuming restart of the optimization can be performed in the hope to generate  more solutions in the preferred region.  Since both methods have considerable drawbacks, this paper proposes to use a reference vector based preference articulation (RVPA) method to obtain such additional solutions in preferred regions. After describing the proposed method in detail, experiments are conducted on six benchmark MOPs to assess the performance of the proposed RVPA method. Our empirical results show that, by setting reference vectors in the objective space, the proposed RVPA is able to obtain corresponding solutions in the preferred regions at a low cost. In addition, by setting the reference vectors in a general way, the proposed RVPA method is also able to improve the general quality (convergence and distribution) of the solutions obtained by an MOEA.},
	publisher = {IEEE },
	booktitle = {2015 IEEE Congress on Evolutionary Computation}
}