Fine-grained Differential Harmony Search algorithm

Abstract

A novel Fine-grained Differential Harmony Search algorithm (FDHS) is presented in this paper. The new algorithm incorporates differential mutation scheme with the pitch adjustment operator of Harmony Search (HS) algorithm. Meanwhile a fine-grained evaluation strategy is adopted inside pitch adjustment on every dimension instead of construction completion. The innate self-adaptive feature of differential mutation operator makes it demonstrate better exploitation ability than fixed-step-size method. While, fine-grained strategy overcomes the interference among dimensions throughout evaluation process to a large extent. The experiments conducted on typical benchmark functions show that the proposed FDHS algorithm demonstrates better convergent speed and solution precision than other HS variants with differential mutation operator.