Improvements to the design of interactiveevolutionarymultiobjective Algorithms (iEMOAs) are unlikely without quantitative assessment of their behaviour in realistic settings. Experiments with human decision-makers (...
详细信息
ISBN:
(纸本)9781450383509
Improvements to the design of interactiveevolutionarymultiobjective Algorithms (iEMOAs) are unlikely without quantitative assessment of their behaviour in realistic settings. Experiments with human decision-makers (DMs) are of limited scope due to the difficulty of isolating individual biases and replicating the experiment with enough subjects, and enough times, to obtain confidence in the results. Simulation studies may help to overcome these issues, but they require the use of realistic simulations of decision-makers. Machine decision-makers (MDMs) provide a way to carry out such simulation studies, however, studies so far have relied on simple utility functions. In this paper, we analyse and compare two stateof-the-art iEMOAs by means of a MDM that uses a sigmoid-shaped utility function. This sigmoid utility function is based on psychologically realistic models from behavioural economics, and replicates several realistic human behaviours. Our findings are that, on a variety of well-known benchmarks with two and three objectives, the two iEMOAs do not consistently recover the most-preferred points. We hope that these findings provide an impetus for more directed design and analysis of future iEMOAs.
This paper evaluates the applicability of different multi-objectiveoptimization methods for environmentally conscious supply chain design. We analyze a case study with three objectives: costs, CO2 and fine dust (also...
详细信息
This paper evaluates the applicability of different multi-objectiveoptimization methods for environmentally conscious supply chain design. We analyze a case study with three objectives: costs, CO2 and fine dust (also known as PM - Particulate Matters) emissions. We approximate the Pareto front using the weighted sum and epsilon constraint scalarization methods with pre-defined or adaptively selected parameters, two popular evolutionary algorithms, SPEA2 and NSGA-II, with different selection strategies, and their interactive counterparts that incorporate Decision Maker's (DM's) indirect preferences into the search process. Within this case study, the CO2 emissions could be lowered significantly by accepting a marginal increase of costs over their global minimum. NSGA-II and SPEA2 enabled faster estimation of the Pareto front, but produced significantly worse solutions than the exact optimization methods. The interactive methods outperformed their a posteriori counterparts, and could discover solutions corresponding better to the DM preferences. In addition, by adjusting appropriately the elicitation interval and starting generation of the elicitation, the number of pairwise comparisons needed by the interactiveevolutionary methods to construct a satisfactory solution could be decreased. (C) 2016 Elsevier Ltd. All rights reserved.
暂无评论