Evolutionary Population Dynamic Mechanisms for the Harmony Search Algorithm


  • Seyedeh Zahra Mirjalili, Shelda Sajeev, Ratna Saha, Nima Khodadadi, Seyed Mohammad Mirjalili, Seyedali Mirjalili

  • Download Paper

  • 02 September 2022

Evolutionary algorithms have been widely adopted in science and industry for optimizing challenging problems mainly due to their black box nature and high local optima avoidance. As popular soft computing techniques, they benefit from several stochastic operators, including but not limited to, selection, recombination, mutation, elitism, population diversity, and population dynamics. Among such operators, some have been extensively used and analyzed in different algorithms, while others are yet to be explored in different algorithms. This motivated our attempts to integrate Evolutionary Population Dynamics (EPD) in the Harmony Search (HS) algorithm. EPD is an evolutionary mechanism that excludes and/or replaces a set of the poor solutions in each generation and prevents them from reducing the quality of other solutions. EPD has been used in three different ways in HS to impact 10%, 30%, or 50% of the population to see its impact on the performance of this algorithm. It was observed that 10% is a reasonable portion of the population in HS to improve its performance on IEEE Congress of Evolutionary computation (CEC) test functions, which effectively mimic challenging real-world optimization problems.


Leave a Comment