RDEx-CSOP: Feasibility-Aware Reconstructed Differential Evolution with Adaptive epsilon-Constraint Ranking

2026-03-27

AI Summary

This work addresses constrained single-objective numerical optimization under limited evaluation budgets, where balancing the maintenance of feasible solutions and the convergence speed toward the optimum is critical. The proposed RDEx-CSOP algorithm innovatively integrates feasibility-aware reconstructed differential evolution, a success-history-driven parameter adaptation mechanism, an exploitation-oriented hybrid local search, and a time-varying ε-constraint handling strategy. Furthermore, it introduces the U-score multidimensional evaluation framework to comprehensively assess algorithmic performance. Evaluated on the CEC 2025 CSOP benchmark suite, RDEx-CSOP achieves the highest aggregate score and the best average ranking across 28 test functions, demonstrating significantly enhanced solution efficiency and constraint-handling capability.

Abstract

Constrained single-objective numerical optimisation requires both feasibility maintenance and strong objective-value convergence under limited evaluation budgets. This report documents RDEx-CSOP, a constrained differential evolution variant used in the IEEE CEC 2025 numerical optimisation competition (C06 special session). RDEx-CSOP combines success-history parameter adaptation with an exploitation-biased hybrid search and an ε-constraint handling mechanism with a time-varying threshold. We evaluate RDEx-CSOP on the official CEC 2025 CSOP benchmark using the U-score framework (Speed, Accuracy, and Constraint categories). The results show that RDEx-CSOP achieves the highest total score and the best average rank among all released comparison algorithms, mainly through strong speed and competitive constraint-handling performance across the 28 benchmark functions.

Problem

• constrained optimization
• numerical optimisation
• feasibility maintenance
• objective convergence
• evaluation budget

Innovation

• Differential Evolution
• ε-constraint handling
• adaptive parameter control
• constrained optimization
• hybrid search