Money for Nothing: Speeding Up Evolutionary Algorithms Through Better Initialization

被引:24
作者
de Laillevault, Axel de Perthuis [1 ]
Doerr, Benjamin [1 ]
Doerr, Carola [2 ,3 ]
机构
[1] Univ Paris Saclay, Ecole Polytech, Paris, France
[2] CNRS, Paris, France
[3] Univ Paris 06, Paris, France
来源
GECCO'15: PROCEEDINGS OF THE 2015 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE | 2015年
关键词
Theory; Runtime Analysis; Random Restarts; Initialization; BOUNDS; SEARCH; TIME;
D O I
10.1145/2739480.2754760
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
That the initialization can have a significant impact on the performance of evolutionary algorithms (EAs) is a well known fact in the empirical evolutionary computation literature. Surprisingly, it has nevertheless received only little attention from the theoretical community. We bridge this gap by providing a thorough runtime analysis for a simple iterated random sampling initialization. In the latter, instead of starting an EA with a random sample, it is started in the best of k search points that are taken from the search space uniformly at random. Implementing this strategy comes at almost no cost, neither in the actual coding work nor in terms of wall-clock time. Taking the best of two random samples already decreases the Theta(n log n) expected runtime of the (1+1) EA and Randomized Local Search on ONEMAX by an additive term of order root 7n. The optimal gain that one can achieve with iterated random sampling is an additive term of order root n log n. This also determines the best possible mutation-based EA for ONEMAX, a question left open in [Sudholt, IEEE TEC 2013]. At the heart of our analysis is a very precise bound for the maximum of k independent Binomially distributed variables with success probability 1/2.
引用
收藏
页码:815 / 822
页数:8
相关论文
共 14 条
  • [1] [Anonymous], 2011, THEORY RANDOMIZED SE, DOI DOI 10.1142/7438/SUPPL_FILE/7438_CHAP0L.PDF
  • [2] [Anonymous], PARALLEL PROBLEM SOL
  • [3] Auger A., 2011, THEORY RANDOMIZED SE
  • [4] The accuracy of the Gaussian approximation to the sum of independent variates
    Berry, Andrew C.
    [J]. TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY, 1941, 49 (1-3) : 122 - 136
  • [5] Böttcher S, 2010, LECT NOTES COMPUT SC, V6238, P1, DOI 10.1007/978-3-642-15844-5_1
  • [6] Comparing evolutionary algorithms to the (1+1)-EA
    Borisovsky, P. A.
    Eremeev, A. V.
    [J]. THEORETICAL COMPUTER SCIENCE, 2008, 403 (01) : 33 - 41
  • [7] Colin S., 2014, P GECCO 14, P753
  • [8] DOERR B, 2014, P GECCO 14, P1375
  • [9] Esseen C.-G., 1942, astronomi och fysik, V28, P1
  • [10] Hwang H., 2014, CORR