SDADE: A state-aware disturbance adaptive differential evolution algorithm

doi:10.1016/j.knosys.2026.117086

SDADE: A state-aware disturbance adaptive differential evolution algorithm

2026-09-16

0

PRE

AI

DOI：10.1016/j.knosys.2026.117086

Original

Original request for help

Save

Abstract

En 中文

Differential Evolution (DE) variants often suffer from premature convergence when population diversity collapses. To address this issue, this paper proposes a State-Aware Disturbance Adaptive Differential Evolution (SDADE) algorithm that distinguishes two premature-convergence states: local-optimum trapping and dimensional stagnation. SDADE integrates three coordinated components: a D E / r a n d − t o − p b e s t / 1 mutation refinement, an MDV-based diversity-collapse indicator, and state-dependent perturbation operators. Comparative experiments against six strong baselines, including JADE, SHADE, l-SHADE, jSO, NL-SHADE-RSP, and CMA-ES, show that SDADE achieves the best overall ranking. In a traffic-signal-timing case, SDADE demonstrates better performance than Webster, Empirical Preset, and Equal Split under the same traffic demand and four-phase structure.Specially, SDADE reduces the average delay to 62.87 s/veh, corresponding to reductions of 3.8%, 22.6%, and 74.8% compared with Webster, Empirical Preset, and Equal Split, respectively.

Keywords:

Differential evolution

Global optimization

Mutation strategy

State-aware mechanism

SDADE: A state-aware disturbance adaptive differential evolution algorithm

Abstract

Journal

Organization

Cited Papers