An Improved Enzyme Action Optimizer for Parameter Identification of the Lorenz Chaotic System

Accurate parameter identification of chaotic systems remains challenging because of their sensitivity to initial conditions, multimodal error landscapes, and the tendency of conventional optimizers to become trapped in non-optimal regions. This study introduces an improved enzyme action optimizer (IEAO), an enhanced variant of EAO, to improve convergence reliability and exploitation capability for chaotic parameter estimation. The proposed IEAO integrates sinusoidal adaptive scheduling, randomized enzyme concentration, dual-candidate generation, and lightweight local refinement to support a balanced exploration–exploitation process. The Lorenz system is used as a representative benchmark to evaluate the proposed algorithm. Numerical results obtained under the adopted simulation settings indicate that IEAO estimates the Lorenz parameters with very small residual errors and improved convergence behaviour compared with the baseline EAO and selected competing algorithms, including recent optimizers and previously reported chaotic-identification methods. Statistical, convergence, sensitivity, and computational analyses support the effectiveness of the proposed modifications for this benchmark. Online identification simulations also show that IEAO can track abrupt parameter variations with stable transient behaviour in the noise-free numerical scenario. These findings suggest that IEAO is a promising optimization approach for chaotic parameter identification, while broader validation on additional chaotic systems and noisy measurements remains necessary.