Article: Application of evolutionary neural model in path optimisation of simulation system Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.39 pp.92 - 113 Abstract: This research introduces an evolving neural model that combines learning from neural networks with evolutionary computation techniques to optimise paths in simulation systems. The suggested approach overcomes the shortcomings of traditional algorithms when faced with dynamic, complicated situations by merging structural and parametric optimisation. To improve flexibility, convergence speed, and generalisation performance, the model uses an actor-critic reinforcement learning scheme, evolutionary field optimisation, and neural architecture search. Analyses of experimental data show that the suggested method is more efficient, stable, and accurate than more conventional methods like Q-learning and DQN. Path planning, risk minimisation, and live system simulation are three areas where the results show evolving neural models could improve intelligent decision-making. Inderscience Publishers - linking academia, business and industry through research

Title: Application of evolutionary neural model in path optimisation of simulation system

Authors: Meng Xu; Cheng Xu; Luyao Pei; Bo Wang; Zhixuan Lv

Addresses: Energy Development Research Institute, China Southern Power Grid, Guangzhou, Guangdong, 510700, China ' Energy Development Research Institute, China Southern Power Grid, Guangzhou, Guangdong, 510700, China ' Energy Development Research Institute, China Southern Power Grid, Guangzhou, Guangdong, 510700, China ' Energy Development Research Institute, China Southern Power Grid, Guangzhou, Guangdong, 510700, China ' Energy Development Research Institute, China Southern Power Grid, Guangzhou, Guangdong, 510700, China

Abstract: This research introduces an evolving neural model that combines learning from neural networks with evolutionary computation techniques to optimise paths in simulation systems. The suggested approach overcomes the shortcomings of traditional algorithms when faced with dynamic, complicated situations by merging structural and parametric optimisation. To improve flexibility, convergence speed, and generalisation performance, the model uses an actor-critic reinforcement learning scheme, evolutionary field optimisation, and neural architecture search. Analyses of experimental data show that the suggested method is more efficient, stable, and accurate than more conventional methods like Q-learning and DQN. Path planning, risk minimisation, and live system simulation are three areas where the results show evolving neural models could improve intelligent decision-making.

Keywords: topics covered include digital twins; computational intelligence; evolutionary algorithms; EAs; evolutionary neural networks; path optimisation; simulation systems; and actor-critical models.

DOI: 10.1504/IJICT.2025.149789

International Journal of Information and Communication Technology, 2025 Vol.26 No.39, pp.92 - 113

Received: 30 Aug 2025
Accepted: 14 Sep 2025
Published online: 12 Nov 2025 *

Title: Application of evolutionary neural model in path optimisation of simulation system

Keep up-to-date