中文摘要：

高效的区间Pareto支配比较对于提高区间多目标进化优化算法性能至关重要.针对现有的区间多目标进化优化采用单一区间数比较的不足,提出基于混合比较策略的区间多目标进化优化算法.深入分析区间数μ比较和可能度P比较策略的优劣,提出融合这两种方法的混合比较策略和基于该混合策略的NSGA-II算法.该算法在典型多目标区间函数和含区间不确定性的煤矿井下射频识别阅读器布局中的应用,验证了所提出的混合区间比较策略的有效性.

英文摘要：

Ranking strategies among interval values are more critical for obtaining superior Pareto front with better spread,distribution and approximation. Most current interval evolutionary multi-objective optimizations（EMOs） adopt only one interval ranking method, which is difficult to entirely cover the interval information. Accordingly, an interval EMO with the improved hybrid ranking strategy is proposed, in which two different interval comparison metrics, i.e., μ and P are complements. Then, Pareto dominance for μ ⊕ P ranking is defined and employed to the powerful NSGA-II algorithm for optimizing interval multi-objective problems. The proposed algorithm is applied to benchmark functions and then further to RFID location in underground mine circumstances, and its outstanding performance is experimentally demonstrated.