Decomposition-Based Approach for Solving Large Scale Multi-objective Problems

Decomposition is a well-established mathematical programming technique for dealing with multi-objective optimization problems (MOPs), which has been found to be efficient and effective when coupled to evolutionary algorithms, as evidenced by MOEA/D. MOEA/D decomposes a MOP into several single-objective subproblems by means of well-defined scalarizing functions. It has been shown that MOEA/D is able to generate a set of evenly distributed solutions for different multi-objective benchmark functions with a relatively low number of decision variables (usually no more than 30). In this work, we study the effect of scalability in MOEA/D and show how its efficacy decreases as the number of decision variables of the MOP increases. Also, we investigate the improvements that MOEA/D can achieve when combined with coevolutionary techniques, giving rise to a novel MOEA which decomposes the MOP both in objective and in decision variables space. This new algorithm is capable of optimizing large scale MOPs and outperforms MOEA/D and GDE3 when solving problems with a large number of decision variables (from 200 up to 1200).

[1]  Marco Laumanns,et al.  Scalable Test Problems for Evolutionary Multiobjective Optimization , 2005, Evolutionary Multiobjective Optimization.

[2]  Jouni Lampinen,et al.  GDE3: the third evolution step of generalized differential evolution , 2005, 2005 IEEE Congress on Evolutionary Computation.

[3]  Carlos A. Coello Coello,et al.  A Study of Multiobjective Metaheuristics When Solving Parameter Scalable Problems , 2010, IEEE Transactions on Evolutionary Computation.

[4]  Xiaodong Li,et al.  Cooperative Co-evolution for large scale optimization through more frequent random grouping , 2010, IEEE Congress on Evolutionary Computation.

[5]  Marco Laumanns,et al.  Performance assessment of multiobjective optimizers: an analysis and review , 2003, IEEE Trans. Evol. Comput..

[6]  Lothar Thiele,et al.  Multiobjective evolutionary algorithms: a comparative case study and the strength Pareto approach , 1999, IEEE Trans. Evol. Comput..

[7]  Qingfu Zhang,et al.  MOEA/D: A Multiobjective Evolutionary Algorithm Based on Decomposition , 2007, IEEE Transactions on Evolutionary Computation.

[8]  Kalyanmoy Deb,et al.  A fast and elitist multiobjective genetic algorithm: NSGA-II , 2002, IEEE Trans. Evol. Comput..

[9]  Kenneth A. De Jong,et al.  Cooperative Coevolution: An Architecture for Evolving Coadapted Subcomponents , 2000, Evolutionary Computation.

[10]  Enrique Alba,et al.  A comparative study of the effect of parameter scalability in multi-objective metaheuristics , 2008, 2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence).

[11]  Carlos A. Coello Coello,et al.  Use of cooperative coevolution for solving large scale multiobjective optimization problems , 2013, 2013 IEEE Congress on Evolutionary Computation.

[12]  Xin Yao,et al.  Multilevel cooperative coevolution for large scale optimization , 2008, 2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence).

[13]  Lothar Thiele,et al.  The Hypervolume Indicator Revisited: On the Design of Pareto-compliant Indicators Via Weighted Integration , 2007, EMO.

[14]  P. Raven,et al.  BUTTERFLIES AND PLANTS: A STUDY IN COEVOLUTION , 1964 .

[15]  Kenneth A. De Jong,et al.  A Cooperative Coevolutionary Approach to Function Optimization , 1994, PPSN.

[16]  Andries Petrus Engelbrecht,et al.  A Cooperative approach to particle swarm optimization , 2004, IEEE Transactions on Evolutionary Computation.

[17]  Hitoshi Iba,et al.  Enhancing differential evolution performance with local search for high dimensional function optimization , 2005, GECCO '05.