An Improved Monarch Butterfly Optimization Algorithm based on RNA Computing

Abstract

As a new natural heuristic algorithm, the MBO(monarch butterfly optimization) algorithm simulates the migration behavior of monarch butterfly. Meta heuristic algorithm has significant and effective performance when working on high-dimensional nonlinear problems. Aiming at the problems of the original monarch butterfly algorithm such as weak search ability and unable to jump out of local optimization, an ameliorated monarch butterfly optimization algorithm (RNA-MBO) established on RNA computing is proposed. Firstly, RNA operator is introduced to seize the equilibrium of the algorithm in local and global searching ability during the migration and cross calculation of the monarch butterfly; secondly, the top 1/4 population with the best ranking is selected to perform transposition, neck ring and replacement operation to increase disturbance force, It can replace the elite retention strategy in the standard MBO algorithm and improve its computational efficiency. Finally, the performance of three MBO and improved algorithms is tested and analyzed based on 23 test functions of 30 dimensions; RNA-MBO and two classical algorithms are tested and analyzed based on 8 test functions of 50 dimensions. The consequence of simulation reveal that their comprehensive performance of RNA-MBO algorithm is the best of all the other MBO optimization algorithms and classic algorithms, It has better stability and optimization ability, and has more advantages in solving complex multi-dimensional optimization problems.