A generalized equilibrium optimization framework for multi-objective engineering systems

Authors

  • Veerendra Yadav Author
  • Snehal Swapnil Jawahire Author
  • Kathirisetty Nikhila Author
  • Jaspreet Sidhu Author
  • Dr.J.V.G. Rama Rao Author
  • E. Shalini Author
  • Kiran D. Parmar Author

Keywords:

Multi-objective optimization, equilibrium problem, variational inequality, pareto optimality, strong monotonicity, convex analysis, engineering design, projected algorithm.

Abstract

Multi-objective engineering design refers to the necessity of simultaneously optimizing contradicting performance parameters while being subjected to physical, operational, and resource constraint requirements. Most classical methods of scalarization mostly lead to compromise solutions without any sound equilibrium interpretation and convergence guarantees. A generalized equilibrium optimization framework is proposed that enables the representation of vector-valued engineering objectives by a regularized equilibrium operator defined over the closed convex feasible set. In the specific formulation, weighted and achievement-based scalarization methods can be incorporated via variational inequalities, normal cone inclusions, and Karush–Kuhn–Tucker conditions, allowing the definition of Pareto-optimal engineering decisions. Conditions for the existence of equilibrium solutions can be defined based on compactness and continuity conditions, while uniqueness can be proven based on the strict monotonicity of the equilibrium operator. A projected generalized equilibrium algorithm is proposed having stopping criteria based on the residuals and perform a convergence analysis. Finally, the execution of the proposed method is illustrated using three engineering examples, namely thermal equipment sizing, constrained power dispatching, and structural design. The thermal optimization achieved a compromise objective value of 0.218 with an equilibrium residual of 7.4×10⁻⁷, while the power dispatch model maintained exact generation balance with a residual of 8.1×10⁻⁷. The structural design case obtained a compromise value of 0.226 and residual of 6.9×10⁻⁷. According to the convergence analysis, it was revealed that during the 120 iterations the equilibrium objective was reduced from 0.842 to 0.218. Using either weighted projected gradient method or penalty projected gradient method the resulting level attained with the proposed approach was 0.218 achieved in only 120 iterations thus proving the least effective performance of the new framework due to its capacity to provide mathematically provable equilibrium consistency, stable convergence and interpretable trade-offs.

Downloads

Download data is not yet available.

Downloads

Published

2026-09-22

How to Cite

A generalized equilibrium optimization framework for multi-objective engineering systems. (2026). Minimax Theory and Its Applications, 11(2s), 392-408. https://journalmta.com/index.php/jmta/article/view/423