Problem understanding through landscape theory

In order to understand the structure of a problem we need to measure some features of the problem. Some examples of measures suggested in the past are autocorrelation and fitness-distance correlation. Landscape theory, developed in the last years in the field of combinatorial optimization, provides mathematical expressions to efficiently compute statistics on optimization problems. In this paper we discuss how can we use landscape theory in the context of problem understanding and present two software tools that can be used to efficiently compute the mentioned measures.

[1]  Enrique Alba,et al.  Elementary landscape decomposition of the 0-1 unconstrained quadratic optimization , 2011, Journal of Heuristics.

[2]  Peter F. Stadler,et al.  Laplacian Eigenvectors of Graphs , 2007 .

[3]  Christian Bierwirth,et al.  Landscape Regularity and Random Walks for the Job-Shop Scheduling Problem , 2004, EvoCOP.

[4]  Enrique Alba,et al.  Exact Computation of the Fitness-Distance Correlation for Pseudoboolean Functions with One Global Optimum , 2012, EvoCOP.

[5]  Christian M. Reidys,et al.  Combinatorial Landscapes , 2002, SIAM Rev..

[6]  B. Drossel.  Biological evolution and statistical physics , 2001, cond-mat/0101409.

[7]  Enrique Alba,et al.  A Methodology to Find the Elementary Landscape Decomposition of Combinatorial Optimization Problems , 2011, Evolutionary Computation.

[8]  Enrique Alba,et al.  Local Optima Networks, Landscape Autocorrelation and Heuristic Search Performance , 2012, PPSN.

[9]  J. Wesley Barnes,et al.  The theory of elementary landscapes , 2003, Appl. Math. Lett..

[10]  Vassilis Zissimopoulos,et al.  Autocorrelation Coefficient for the Graph Bipartitioning Problem , 1998, Theor. Comput. Sci..

[11]  Enrique Alba,et al.  Autocorrelation measures for the quadratic assignment problem , 2012, Appl. Math. Lett..

[12]  Marc Schoenauer,et al.  Rigorous Hitting Times for Binary Mutations , 1999, Evolutionary Computation.

[13]  Enrique Alba,et al.  Exact computation of the expectation curves of the bit-flip mutation using landscapes theory , 2011, GECCO '11.

[14]  Lov K. Grover.  Local search and the local structure of NP-complete problems , 1992, Oper. Res. Lett..

[15]  Leonardo Vanneschi,et al.  A Study of Fitness Distance Correlation as a Difficulty Measure in Genetic Programming , 2005, Evolutionary Computation.

[16]  Eric Angela,et al.  On the landscape ruggedness of the quadratic assignment problem , 2001 .

[17]  E. Weinberger,et al.  Correlated and uncorrelated fitness landscapes and how to tell the difference , 1990, Biological Cybernetics.

[18]  Vassilis Zissimopoulos,et al.  On the landscape ruggedness of the quadratic assignment problem , 2001, Theor. Comput. Sci..

[19]  Terry Jones,et al.  Fitness Distance Correlation as a Measure of Problem Difficulty for Genetic Algorithms , 1995, ICGA.

[20]  L. Darrell Whitley,et al.  A Tractable Walsh Analysis of SAT and its Implications for Genetic Algorithms , 1998, AAAI/IAAI.

[21]  Andrew M. Sutton,et al.  Understanding elementary landscapes , 2008, GECCO '08.

[22]  Andrew M. Sutton,et al.  A polynomial time computation of the exact correlation structure of k-satisfiability landscapes , 2009, GECCO '09.

[23]  Peter F. Stadler,et al.  Correlation length, isotropy and meta-stable states , 1997 .

[24]  Andrew M. Sutton,et al.  Partial neighborhoods of elementary landscapes , 2009, GECCO '09.

[25]  P. Stadler.  Landscapes and their correlation functions , 1996 .

[26]  Andrew M. Sutton,et al.  Mutation rates of the (1+1)-EA on pseudo-boolean functions of bounded epistasis , 2011, GECCO '11.