Mutation Strategy Selection Based on Fitness Landscape Analysis: A Preliminary Study

Different algorithms and strategies behave disparately for different types of problems. In practical problems, we cannot grasp the nature of the problem in advance, so it is difficult for the engineers to choose a proper method to solve the problem effectively. In this case, the strategy selection task based on fitness landscape analysis comes into being. This paper gives a preliminary study on mutation strategy selection on the basis of fitness landscape analysis for continuous real-parameter optimization based on differential evolution. Some fundamental features of the fitness landscape and the components of standard differential evolution algorithm are described in detail. A mutation strategy selection framework based on fitness landscape analysis is designed. Some different types of classifiers which are applied to the proposed framework are tested and compared.

[1]  Michael Affenzeller,et al.  A Comprehensive Survey on Fitness Landscape Analysis , 2012, Recent Advances in Intelligent Engineering Systems.

[2]  Liang Zhong,et al.  Self-feedback differential evolution adapting to fitness landscape characteristics , 2019, Soft Comput..

[3]  Rainer Storn,et al.  Differential Evolution – A Simple and Efficient Heuristic for global Optimization over Continuous Spaces , 1997, J. Glob. Optim..

[4]  Mario A. Muñoz,et al.  Landscape characterization of numerical optimization problems using biased scattered data , 2012, 2012 IEEE Congress on Evolutionary Computation.

[5]  Marcus Gallagher,et al.  Sampling Techniques and Distance Metrics in High Dimensional Continuous Landscape Analysis: Limitations and Improvements , 2014, IEEE Transactions on Evolutionary Computation.

[6]  Ponnuthurai Nagaratnam Suganthan,et al.  Problem Definitions and Evaluation Criteria for the CEC 2014 Special Session and Competition on Single Objective Real-Parameter Numerical Optimization , 2014 .

[7]  T. J. Mitchell,et al.  Exploratory designs for computational experiments , 1995 .

[8]  P. N. Suganthan,et al.  Differential Evolution: A Survey of the State-of-the-Art , 2011, IEEE Transactions on Evolutionary Computation.

[9]  Riccardo Poli,et al.  Decomposition of fitness functions in random heuristic search , 2007, FOGA'07.

[10]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[11]  Riccardo Poli,et al.  Information landscapes and problem hardness , 2005, GECCO '05.

[12]  Ponnuthurai N. Suganthan,et al.  Recent advances in differential evolution - An updated survey , 2016, Swarm Evol. Comput..

[13]  Zongming Fei,et al.  Measures in the time and frequency domains for fitness landscape analysis of dynamic optimization problems , 2017, Appl. Soft Comput..

[14]  N. Altman.  An Introduction to Kernel and Nearest-Neighbor Nonparametric Regression , 1992 .

[15]  Jing J. Liang,et al.  Problem Definitions and Evaluation Criteria for the CEC 2005 Special Session on Real-Parameter Optimization , 2005 .

[16]  Leonardo Vanneschi,et al.  Theory and practice for efficient genetic programming , 2004 .

[17]  Sébastien Vérel,et al.  Negative Slope Coefficient: A Measure to Characterize Genetic Programming Fitness Landscapes , 2006, EuroGP.

[18]  R. Storn,et al.  Differential evolution a simple and efficient adaptive scheme for global optimization over continu , 1997 .

[19]  Liang Shen,et al.  A mixed strategy for Evolutionary Programming based on local fitness landscape , 2010, IEEE Congress on Evolutionary Computation.

[20]  Riccardo Poli,et al.  Information landscapes , 2005, GECCO '05.

[21]  Zongming Fei,et al.  Analysis of the similarities and differences of job-based scheduling problems , 2018, Eur. J. Oper. Res..

[22]  M. Friedman.  The Use of Ranks to Avoid the Assumption of Normality Implicit in the Analysis of Variance , 1937 .

[23]  Xin Yao,et al.  Population Evolvability: Dynamic Fitness Landscape Analysis for Population-Based Metaheuristic Algorithms , 2018, IEEE Transactions on Evolutionary Computation.

[24]  Sébastien Vérel,et al.  Where are bottlenecks in NK fitness landscapes? , 2003, The 2003 Congress on Evolutionary Computation, 2003. CEC '03..