Runtime Analysis of Non-elitist Populations: From Classical Optimisation to Partial Information

被引:46
作者
Dang, Duc-Cuong [1 ]
Lehre, Per Kristian [1 ]
机构
[1] Univ Nottingham, Sch Comp Sci, Jubilee Campus,Wollaton Rd, Nottingham NG8 1BB, England
基金
英国工程与自然科学研究理事会;
关键词
Runtime; Drift analysis; Evolutionary algorithms; Non-elitism; Fitness-levels; Partial evaluation; DRIFT ANALYSIS; HITTING-TIME; EVOLUTIONARY; ALGORITHMS; BOUNDS;
D O I
10.1007/s00453-015-0103-x
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
Although widely applied in optimisation, relatively little has been proven rigorously about the role and behaviour of populations in randomised search processes. This paper presents a new method to prove upper bounds on the expected optimisation time of population-based randomised search heuristics that use non-elitist selection mechanisms and unary variation operators. Our results follow from a detailed drift analysis of the population dynamics in these heuristics. This analysis shows that the optimisation time depends on the relationship between the strength of the selective pressure and the degree of variation introduced by the variation operator. Given limited variation, a surprisingly weak selective pressure suffices to optimise many functions in expected polynomial time. We derive upper bounds on the expected optimisation time of non-elitist evolutionary algorithms (EA) using various selection mechanisms, including fitness proportionate selection. We show that EAs using fitness proportionate selection can optimise standard benchmark functions in expected polynomial time given a sufficiently low mutation rate. As a second contribution, we consider an optimisation scenario with partial information, where fitness values of solutions are only partially available. We prove that non-elitist EAs under a set of specific conditions can optimise benchmark functions in expected polynomial time, even when vanishingly little information about the fitness values of individual solutions or populations is available. To our knowledge, this is the first runtime analysis of randomised search heuristics under partial information.
引用
收藏
页码:428 / 461
页数:34
相关论文
共 36 条
  • [1] [Anonymous], 1964, Elementary inequalities
  • [2] Auger A., 2011, Theory of randomized search heuristics: Foundations and recent developments
  • [3] Blum C, 2012, VARIANTS OF EVOLUTIONARY ALGORITHMS FOR REAL-WORLD APPLICATIONS, P1
  • [4] A New Approach for Analyzing Average Time Complexity of Population-Based Evolutionary Algorithms on Unimodal Problems
    Chen, Tianshi
    He, Jun
    Sun, Guangzhong
    Chen, Guoliang
    Yao, Xin
    [J]. IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART B-CYBERNETICS, 2009, 39 (05): : 1092 - 1106
  • [5] Corus D, 2014, LECT NOTES COMPUT SC, V8672, P912
  • [6] Refined Upper Bounds on the Expected Runtime of Non-elitist Populations from Fitness-Levels
    Dang, Duc-Cuong
    Lehre, Per Kristian
    [J]. GECCO'14: PROCEEDINGS OF THE 2014 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, 2014, : 1367 - 1374
  • [7] Evolution under Partial Information
    Dang, Duc-Cuong
    Lehre, Per Kristian
    [J]. GECCO'14: PROCEEDINGS OF THE 2014 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, 2014, : 1359 - 1366
  • [8] Multiplicative Drift Analysis
    Doerr, Benjamin
    Johannsen, Daniel
    Winzen, Carola
    [J]. ALGORITHMICA, 2012, 64 (04) : 673 - 697
  • [9] Droste S, 2004, LECT NOTES COMPUT SC, V3102, P1088
  • [10] Droste S, 2003, LECT NOTES COMPUT SC, V2723, P909