Technology Portfolio Selection under Variable Budget: A Comparative Analysis of MADM-MODM Pipelines

Technology portfolio selection combines multi-attribute evaluation and constrained optimization, yet the bridge between MADM and MODM remains insufficiently explicit in applied literature. This paper compares five pipelines that differ in the evaluation engine, TOPSIS, ELECTRE III, and TREVO, and in the selection strategy, greedy selection or 0–1 knapsack. The protocol uses synthetic datasets with 20 technologies per seed, including pathological profiles, called Trojan horses, defined by strong imbalance between normative and instrumental value. The results indicate that outranking-based pipelines filter Trojans under all conditions, although with lower budget utilization at high budget levels. Compensatory pipelines and TREVO admit Trojans increasingly as the budget constraint relaxes, with TREVO showing asymmetric behavior across Trojan types, while the Jaccard similarity between some pipelines converges at high budgets. The findings indicate that evaluation architecture and selection strategy produce structural divergence in portfolio composition.