Optimising an evolutionary algorithm for scheduling

Urquhart, Neil B.; Chisholm, Ken; Paechter, Ben

Abstract

This paper examines two techniques for setting the parameters of an evolutionary Algorithm (EA). The example EA used for test purposes undertakes a simple scheduling problem. An initial version of the EA was tested utilising a set of parameters that were decided by basic experimentation. Two subsequent versions were compared with the initial version, the first of these adjusted the parameters at run time, the second used a set of parameters decided on by running a meta-EA. The authors have been able to conclude that the usage of a meta-EA allows an efficient set of parameters to be derived for the problem EA.

Citation

Urquhart, N. B., Chisholm, K., & Paechter, B. (2000). Optimising an evolutionary algorithm for scheduling. In S. Cagnoni, R. Poli, G. D. Smith, D. Corne, M. Oates, E. Hart, …T. C. Fogarty (Eds.), Real-World Applications of Evolutionary Computing: EvoWorkshops 2000: EvoIASP, EvoSCONDI, EvoTel, EvoSTIM, EvoRob, and EvoFlight, Proceedings (307-318). https://doi.org/10.1007/3-540-45561-2_30

• Conference Name: EvoWorkshops 2000: EvoIASP, EvoSCONDI, EvoTel, EvoSTIM, EvoRob, and EvoFlight,
• Conference Location: Edinburgh
• Start Date: Apr 17, 2000
• End Date: Apr 17, 2000
• Publication Date: 2000-04
• Deposit Date: Jun 12, 2009
• Publisher: Springer
• Peer Reviewed: Peer Reviewed
• Pages: 307-318
• Series Title: Lecture Notes in Computer Science
• Series Number: 1803
• Series ISSN: 0302-9743
• Book Title: Real-World Applications of Evolutionary Computing: EvoWorkshops 2000: EvoIASP, EvoSCONDI, EvoTel, EvoSTIM, EvoRob, and EvoFlight, Proceedings
• ISBN: 9783540673538
• DOI: https://doi.org/10.1007/3-540-45561-2_30
• Keywords: Evolutionary algorithm; Scheduling problem; Experimental parameters; meta-EA
• Public URL: http://researchrepository.napier.ac.uk/id/eprint/2714
• Publisher URL: http://dx.doi.org/10.1007/3-540-45561-2_30