A Heterogeneous Distributed Approach for Many-Objective Optimization

The necessity to optimize simultaneously multiple objectives arises in different fields. Multi-Objective Evolutionary Algorithms (MOEAs) have been widely used to solve these problems. However, when the number of objectives is greater than three, these problems are called many-objective problems and their resolution poses major challenges to researchers in the field. Researchers have investigated different approaches and proposed various algorithms. However, none of them can outperform the others in all problems. Although, the characteristics of MOEAs can be collaboratively combined improving the search ability. These remarks advocate using a Heterogeneous Distribution framework. The framework uses the Island-model to share information among the MOEAs. This model allows collaboration by migrating solutions between the islands. In this study, the framework is evaluated using two state-of-the-art MOEAs: NSGA-III and MOEA/D-STM. Besides, it investigates two different strategies of communication: synchronous and asynchronous. Likewise, a set of experiments is made to evaluate the framework using a set of Multi-objective problems, quality indicators, and statistical tests. The results indicate that the collaboration improves the convergence and diversity of the algorithms in most studied problems. Moreover, the version using synchronous communication presents better results, and we suggest using this version when the computational cost of the collaborating MOEAs is different.

[1]  Kalyanmoy Deb,et al.  An Evolutionary Many-Objective Optimization Algorithm Using Reference-Point-Based Nondominated Sorting Approach, Part I: Solving Problems With Box Constraints , 2014, IEEE Transactions on Evolutionary Computation.

[2]  Xin Yao,et al.  Stochastic Ranking Algorithm for Many-Objective Optimization Based on Multiple Indicators , 2016, IEEE Transactions on Evolutionary Computation.

[3]  Qingfu Zhang,et al.  An Evolutionary Many-Objective Optimization Algorithm Based on Dominance and Decomposition , 2015, IEEE Transactions on Evolutionary Computation.

[4]  Qingfu Zhang,et al.  Stable Matching-Based Selection in Evolutionary Multiobjective Optimization , 2014, IEEE Transactions on Evolutionary Computation.

[5]  Hisao Ishibuchi,et al.  Performance of Decomposition-Based Many-Objective Algorithms Strongly Depends on Pareto Front Shapes , 2017, IEEE Transactions on Evolutionary Computation.

[6]  Jun Zhang,et al.  A hybrid evolutionary algorithm with dual populations for many-objective optimization , 2016, 2016 IEEE Congress on Evolutionary Computation (CEC).

[7]  Zhang Jun,et al.  Parallel multi-strategy evolutionary algorithm using massage passing interface for many-objective optimization , 2016 .

[8]  Xin Yao,et al.  Two_Arch2: An Improved Two-Archive Algorithm for Many-Objective Optimization , 2015, IEEE Transactions on Evolutionary Computation.

[9]  Xin Yao,et al.  Many-Objective Evolutionary Algorithms , 2015, ACM Comput. Surv..

[10]  Qingfu Zhang,et al.  Distributed evolutionary algorithms and their models: A survey of the state-of-the-art , 2015, Appl. Soft Comput..

[11]  Qingfu Zhang,et al.  The performance of a new version of MOEA/D on CEC09 unconstrained MOP test instances , 2009, 2009 IEEE Congress on Evolutionary Computation.