You are on page 1of 2

RANKING SELECTION A very crucial part of genetic algorithm is the selection method or selection mechanism.

Basically, selection method is a process of determining which genes will be carried over to the next generation. For this to be as efficient as possible, the better individuals must be preferred over weaker ones [1]. This is somewhat the culmination of Charles Darwins Survival of the Fittest. In genetic algorithm, there is an ample amount of selection mechanisms to choose from. In this research, ranking selection is one of the inspirations as to how the individuals were chosen. From the name itself, ranking selection ranks the population, from best to worst, according to its fitness values. An additional selection probability is computed but this time, according to their ranks and not their fitness value [2]. Super genotypes or genes that have dominating fitness values are not always chosen [3]. This would mean that they cannot take over in the process. This makes rank selection an explorative technique since the solution would not converge prematurely [4]. In some researches, the exploratory nature of ranking selection maintained a diverse population and did not compromise its performance in terms of computation time and solution quality [4, 5]. Looking at the time complexity of this two-step selection mechanism (ranking/sorting of individuals and computing of selection probability), it would average at O(n log n) considering that a good sorting algorithm is adopted (worst case must be O(n log n)) and the selection probability assignment clocks between O(n) to O(n2) [6].

SOURCES: 1 A Review of Selection Methods In Genetic Algorithm. R. Sivaraj, T. Ravichandran; http://www.ijest.info/docs/IJEST11-03-05-190.pdf 2 Genetic Algorithm Performance with Different Selection Strategies in Solving TSP. Noraini Mohd Razali, John Geraghty; http://www.iaeng.org/publication/WCE2011/WCE2011_pp11341139.pdf 3 The GENITOR Algorithm and Selection Pressure: Why Rank-Based Allocation of Reproductive Trials is Best. Darrel Whitley; http://pdf.aminer.org/000/310/952/the_genitor_algorithm_and_selection_pressure_why_rank_ba sed_allocation.pdf 4 Blending Roulette Wheel Selection & Rank Selection in Genetic Algorithms. Rakesh Kumar, Jyotishree; http://www.ijmlc.org/papers/146-C00572-005.pdf 5 Genetic Algorithm Performance with Different Selection Methods in Solving Multi-Objective Network Design Problem. R. O. Oladele, J. S. Sadiku; http://research.ijcaonline.org/volume70/number12/pxc3887848.pdf 6 A Comparative Analysis of Selection Schemes Used in Genetic Algorithms. David E. Goldberg, Kalyanmoy Deb; http://www.cse.unr.edu/~sushil/class/gas/papers/Select.pdf

Other source not mentioned: 1. http://cisjournal.org/journalofcomputing/archive/vol3no7/vol3no7_8.pdf

You might also like