Fitness Probability Distribution of Bit-Flip Mutation

被引:30
作者
Chicano, Francisco [1 ]
Sutton, Andrew M. [2 ]
Whitley, L. Darrell [3 ]
Alba, Enrique [1 ]
机构
[1] Univ Malaga, Dept Lenguajes & Ciencias Computac, E-29071 Malaga, Spain
[2] Univ Jena, Fak Math & Informat, D-07745 Jena, Germany
[3] Colorado State Univ, Dept Comp Sci, Ft Collins, CO 80523 USA
关键词
Bit-flip mutation; evolutionary algorithms; landscape theory; combinatorial optimization; randomized algorithms; COMPUTATION;
D O I
10.1162/EVCO_a_00130
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Bit-flip mutation is a common mutation operator for evolutionary algorithms applied to optimize functions over binary strings. In this paper, we develop results from the theory of landscapes and Krawtchouk polynomials to exactly compute the probability distribution of fitness values of a binary string undergoing uniform bit-flip mutation. We prove that this probability distribution can be expressed as a polynomial in p, the probability of flipping each bit. We analyze these polynomials and provide closed-form expressions for an easy linear problem (Onemax), and an NP-hard problem, MAX-SAT. We also discuss a connection of the results with runtime analysis.
引用
收藏
页码:217 / 248
页数:32
相关论文
共 26 条
[1]  
[Anonymous], 2007, Lecture Notes in Mathematics
[2]  
[Anonymous], 2010, SOCS
[3]   Exact computation of the expectation surfaces for uniform crossover along with bit-flip mutation [J].
Chicano, Francisco ;
Whitley, Darrell ;
Alba, Enrique .
THEORETICAL COMPUTER SCIENCE, 2014, 545 :76-93
[4]  
Chicano F, 2011, GECCO-2011: PROCEEDINGS OF THE 13TH ANNUAL GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, P2027
[5]   A Methodology to Find the Elementary Landscape Decomposition of Combinatorial Optimization Problems [J].
Chicano, Francisco ;
Whitley, L. Darrell ;
Alba, Enrique .
EVOLUTIONARY COMPUTATION, 2011, 19 (04) :597-637
[6]  
Doerr B, 2013, GECCO'13: PROCEEDINGS OF THE 2013 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, P1589
[7]  
Feinsilver P, 2005, RECENT ADVANCES IN APPLIED PROBABILITY, P115, DOI 10.1007/0-387-23394-6_5
[8]   Rigorous Hitting Times for Binary Mutations [J].
Garnier, Josselin ;
Kallel, Leila ;
Schoenauer, Marc .
EVOLUTIONARY COMPUTATION, 1999, 7 (02) :173-203
[9]   LOCAL SEARCH AND THE LOCAL-STRUCTURE OF NP-COMPLETE PROBLEMS [J].
GROVER, LK .
OPERATIONS RESEARCH LETTERS, 1992, 12 (04) :235-243
[10]   Towards an analytic framework for analysing the computation time of evolutionary algorithms [J].
He, J ;
Yao, X .
ARTIFICIAL INTELLIGENCE, 2003, 145 (1-2) :59-97