This paper presents a fuzzy satisfaction-based intelligent framework for early-stage multiobjective sizing of a buck DC–DC converter under uncertain operating conditions. Lightweight closed-form estimators are used to evaluate inductor current ripple, output voltage ripple, and efficiency, including an explicit decomposition of ripple into capacitive and ESR-induced components to distinguish capacitance-dominated and ESR-dominated regimes. Engineering targets for ripple, efficiency, and passive size/cost pressure are mapped to reproducible piecewise membership functions and aggregated into a bounded overall satisfaction score using a weighted geometric operator; alternative non-compensatory and OWA-type aggregators are considered for sensitivity analysis. The resulting nonconvex design problem is solved via a compact two-stage derivative-free strategy that combines global screening with an interpretable Takagi–Sugeno (TSK) rule-based refinement layer, which generates bounded, physics-consistent updates of the design variables and supports rapid feasibility restoration followed by preference-driven tuning. Uncertainty in operating conditions and parameter drift is addressed through scenario evaluation and worst-case or average-case aggregation of satisfaction, linking the fuzzy decision objective to robust scenario design. Numerical studies for a 24 ± 4 V to 12 V converter illustrate regime-dependent adaptation: in low-ESR conditions, ripple improvement is driven mainly by capacitance/frequency adjustments, while in high-ESR conditions, the rule base shifts corrections toward inductor and frequency choices that reduce ESR-dominated ripple.
Hinov et al. (2026) studied this question.