Multi-objective optimization design and evaluation of power transmission and transformation projects based on game theory combination weighting and improved NSGA-III

Abstract Traditional multi-objective optimization in power transmission and transformation projects suffers from static reference point settings, inefficient constraint handling, and subjective evaluation weighting. To address these challenges, this study proposes a collaborative optimization–evaluation framework integrating an improved Non-dominated Sorting Genetic Algorithm III (INSGA-III) with a game-theoretic combined weighting approach. An adaptive reference point mechanism, guided by population distribution entropy, dynamically regulates reference density to enhance convergence and diversity across six conflicting objectives: lifecycle cost, reliability, short-circuit current, voltage stability, electromagnetic impact, and land occupation. A feasibility-first constraint strategy embeds power flow, capacity, and N−1 safety criteria directly into environmental selection, ensuring engineering validity. For solution evaluation, entropy, Criteria Importance Through Intercriteria Correlation (CRITIC), and analytic hierarchy process (AHP) weights are integrated via Shapley-value cooperative games, generating balanced, stable indicator weights. Validation on a 500 kV project shows the INSGA-III achieves 22.5% lower inverse generational distance (IGD), 14.0% higher hypervolume (HV), and 39.3% lower generational distance (GD) than the baseline, with 91% feasible solutions. The combined weighting maintains ranking stability under ±15% perturbations. The proposed framework effectively bridges theoretical optimization and engineering decision-making, offering a robust paradigm for scientific planning of complex power infrastructure in the context of new-type power systems.