Sizing optimization of skeletal structures with a multi-adaptive harmony search algorithm

The Harmony Search (HS) algorithm is a popular metaheuristic optimization method that reproduces the music improvisation process in searching for a perfect state of harmony. HS has a remarkable ability in detecting near global optima at low computational cost but may be ineffective in performing local search. This study presents the Multi-Adaptive Harmony Search (MAHS) algorithm for sizing optimization of skeletal structures with continuous or discrete design variables. The main difference between the proposed algorithm and classic HS is the way of choosing and adjusting the bandwidth distance (bw). Furthermore, MAHS dynamically updates the Harmony Memory Consideration Rate (HMCR) and Pitch-Adjusting Rate (PAR) parameters during the search process. The robustness and performance of the MAHS algorithm are evaluated in comparison with literature, and in particular, with well-known HS variants such as Global-best Harmony Search (GHS), Self-Adaptive Harmony Search (SAHS), and Efficient Harmony Search (EHS). Optimization results obtained by the MAHS algorithm confirm the validity of the proposed approach. © 2015 Sharif University of Technology. All rights reserved.