Asynchronous Parallel Cartesian Genetic Programming

被引:3
作者
Harter, Adam [1 ]
Tauritz, Daniel R. [1 ]
Siever, William M. [2 ]
机构
[1] Missouri Univ Sci & Technol, Dept Comp Sci, Nat Computat Lab, Rolla, MO 65401 USA
[2] Washington Univ, Dept Comp Sci & Engn, St Louis, MO 63130 USA
来源
PROCEEDINGS OF THE 2017 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE COMPANION (GECCO'17 COMPANION) | 2017年
关键词
Genetic Programming; Asynchronous Parallel Evolution; Cartesian Genetic Programming; Evolutionary Computing;
D O I
10.1145/3067695.3084210
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The run-time of evolutionary algorithms (EAs) is typically dominated by fitness evaluation. This is particularly the case when the genotypes are complex, such as in genetic programming (GP). Evaluating multiple offspring in parallel is appropriate in most types of EAs and can reduce the time incurred by fitness evaluation proportional to the number of parallel processing units. The most naive approach maintains the synchrony of evolution as employed by the vast majority of EAs, requiring an entire generation to be evaluated before progressing to the next generation. Heterogeneity in the evaluation times will degrade the performance, as parallel processing units will idle until the longest evaluation has completed. Asynchronous parallel evolution mitigates this bottleneck and techniques which experience high heterogeneity in evaluation times, such as Cartesian GP (CGP), are prime candidates for asynchrony. However, due to CGP's small population size, asynchrony has a significant impact on selection pressure and biases evolution towards genotypes with shorter execution times, resulting in poorer results compared to their synchronous counterparts. This paper: 1) provides a quick introduction to CGP and asynchronous parallel evolution, 2) introduces asynchronous parallel CGP, and 3) shows empirical results demonstrating the potential for asynchronous parallel CGP to outperform synchronous parallel CGP.
引用
收藏
页码:1820 / 1824
页数:5
相关论文
共 11 条
[1]   Why Asynchronous Parallel Evolution is the Future of Hyper-heuristics: A CDCL SAT Solver Case Study [J].
Bertels, Alex R. ;
Tauritz, Daniel R. .
PROCEEDINGS OF THE 2016 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE (GECCO'16 COMPANION), 2016, :1359-1365
[2]  
Churchill AW, 2013, 2013 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC), P2924
[3]  
Durillo JJ, 2008, 2008 IEEE INTERNATIONAL SYMPOSIUM ON PARALLEL & DISTRIBUTED PROCESSING, VOLS 1-8, P2404
[4]   Analysis of Cartesian Genetic Programming's Evolutionary Mechanisms [J].
Goldman, Brian W. ;
Punch, William F. .
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION, 2015, 19 (03) :359-373
[5]  
Harding S, 2007, GECCO 2007: GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, VOL 1 AND 2, P1021
[6]  
Martin M.A., 2015, Proceedings of the Companion Publication of the 2015 Annual Conference on Genetic and Evolutionary Computation,' GECCO Companion'15, P1429, DOI DOI 10.1145/2739482.2764718
[7]  
Miller J., 2000, Cartesian Genetic Programming, Natural Computing Series
[8]  
ScoS E.O., 2016, P GEN EV COMP C 2016, P845, DOI [10.1145/2908812.2908934, DOI 10.1145/2908812.2908934]
[9]  
Scott EO, 2015, P COMP PUBL 2015 ANN, P1209, DOI [DOI 10.1145/2739482.2768482, 10.1145/2739482.2768482]
[10]   Asynchronous Master/Slave MOEAs and Heterogeneous Evaluation Costs [J].
Yagoubi, Mouadh ;
Schoenauer, Marc .
PROCEEDINGS OF THE FOURTEENTH INTERNATIONAL CONFERENCE ON GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, 2012, :1007-1014