Benchmarking individual discipline feasible and multidisciplinary feasible architectures for hybrid-electric aircraft design

Multidisciplinary Design Optimization (MDO) plays a central role in the preliminary design of next-generation aircraft, where strong couplings among aerodynamics, propulsion, structures, and mission performance must be handled efficiently. While several MDO architectures have been proposed and analyzed in the literature, quantitative benchmarks assessing their performance on realistic, tightly coupled multidisciplinary problems remain scarce, particularly when different optimization strategies and computational budgets are considered. This aspect becomes particularly relevant when applied to the design of aircraft with non-traditional configurations, for which no established reference constructive realization is currently available. This work presents a systematic benchmark comparison between Individual Discipline Feasible (IDF) and Multidisciplinary Feasible (MDF) architectures applied to the conceptual sizing of a hybrid-electric regional aircraft. Both architectures are coupled with a population-based evolutionary algorithm (NSGA-II) and a gradient-based optimizer (IPOPT), enabling an assessment of solver–architecture interactions under comparable computational budgets. The benchmark evaluates convergence behavior, feasibility management, robustness, and computational efficiency using several performance metrics, including hypervolume, feasibility ratio, cumulative CPU time, and Equivalent Computational Time (ECT) indicators. Results show that, while IDF and MDF exhibit comparable performance when paired with deterministic gradient-based optimization, the MDF architecture consistently outperforms IDF under stochastic global optimization, particularly at limited computational budgets, so achieving faster convergence, higher robustness, and improved feasibility handling. These aspects are particularly relevant for aircraft design problems involving non-conventional configurations, for which no established reference airframe realizations are currently available. Beyond numerical benchmarking, the Pareto-optimal solutions exhibit physically consistent trends, with conventional aircraft configurations obtained when minimizing maximum take-off weight, and highly hybridized designs occurring when fuel consumption is minimized, at the cost of large battery requirements. Overall, the study provides a quantitative and reproducible assessment of MDO architectural choices for hybrid-electric aircraft design, bridging the gap between theoretical comparisons and realistic multidisciplinary applications.