Implementation of Quasi-Simplex local search on MOEA/D

Several real world problems can be modelled as a multi-objective optimization problem (MOP) and demand efficient algorithms to be solved. Some MOPs can not be solved by exact methods efficiently and, therefore, approximation algorithms, such as Multi-Objective Evolutionary Algorithms (MOEAs), are used to solve these problems. Multi-Objective Evolutionary Algorithm based on Decomposition (MOEA/D) has obtained very good results in various MOPs and is one of the most influential MOEAs. This paper focuses on improving the MOEA/D performance by introducing a local search (Quasi-Simplex) component. The CEC 2009 Multi-Objective Competition Benchmark is used to analyse the performance of the proposed algorithm and investigate the importance of its parameters. The empirical results obtained are encouraging.