A novel reinforcement learning based Heap-based optimizer

Xuesen Ma, Zhineng Zhong, Yangyu Li, Dacheng Li, Yan Qiao

Abstract: Highlights•Proposed a novel Heap-based optimizer based on reinforcement learning.•It is proposed to control the search scope of the search agent through reinforcement learning to achieve balanced exploitation and exploration.•A self-learning strategy and a convergence strategy based on similar search directions speed up the algorithm’s convergence.•The proposed algorithm exhibits commendable attributes, including rapid convergence, high accuracy, and robust stability.