Using Fuzzy Preference Orderings in theta-Dominance with Application to Health Monitoring of Li-Ion Batteries

A Genetic Fuzzy Model of the State of Health of a Li-Ion battery is developed where both outputs of the system and its first derivative with respect to the stored charge are approximated. This approximation is a viable diagnosis technique to detect cell degradation in modern Li-Ion battery technologies. The model is fitted to data by means of a specialization of the theta-Dominance Evolutionary Algorithm, that alters the prioritization of the individuals in the selection stage. An specific operator is used which complements Pareto Non-Dominance levels with a partial order at each level thus models that are potentially better have a reproductive advantage. An empirical study is performed where the results of different multi and many-objectives genetic algorithms are assessed for this problem.