Optimizing Enterprise Capital Structure Allocation Using the NSGA‐III Algorithm

Abstract

Current enterprise capital structure optimization faces the challenge of conflicting multiple objectives and the difficulty of obtaining an equilibrium solution efficiently. This paper introduces the Nondominated Sorting Genetic Algorithm III to construct a novel multiobjective optimization model for corporate capital structure, uniquely integrating enterprise value maximization, financial risk minimization, and weighted average cost of capital minimization within a single framework. This model also incorporates debt repayment constraints and capital limits. By using a reference point set within the algorithm to guide population search and combining iterative evolution with simulated binary crossover and polynomial mutation, this method achieves a systematic approximation of the Pareto front in high‐dimensional target spaces. Using financial data from selected Chinese listed companies as a sample, this experiment yielded Pareto solutions with mean uniformity of 0.040 (Spacing, SP) and 0.120 (Diversity Δ$$ \Delta $$ ), a convergence index (GD) of 0.0037, and a coverage ratio η of 0.95. Results show that NSGA‐III effectively solves the multiobjective optimization problem of corporate capital structure and provides a comprehensive and balanced allocation solution for decision‐making.