Using hyper-heuristic to select leader and archiving methods for many-objective problems

Abstract

Multi-objective Particle Swarm Optimization (MOPSO) is a promising meta-heuristic to solve Many-Objective Problems (MaOPs). Previous works have proposed different leader and archiving methods to tackle the challenges caused by the increase in the number of objectives, however, selecting the most appropriate components for a given problem is not a trivial task. Moreover, the algorithm can take advantage by using a variety of methods in different phases of the search. To deal with those issues, we adopt the use of hyper-heuristics, whose concept emerges for dynamically selecting components for effectively solving a problem. In this work, we use a simple hyper-heuristic to select leader and archiving methods during the search. Unlike other studies, our hyper-heuristic is guided by the R2 indicator due to its good measuring characteristics and low computational cost. Experimental studies were conducted to validate the new algorithm where its performance is compared to its components individually and to the state-of-the-art MOEA/D-DRA algorithm. The results show that the new algorithm is robust, presenting good results in different situations.