Computational complexity of elitist population-based evolutionary algorithms : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Computer Science at Massey University, Palmerston North, New Zealand

Evolutionary Algorithms (EAs) are a modern heuristic algorithm that have proven efficiency on a large number of real-life problems. Despite the rich history of applications understanding of both how and why EAs work is lagging far behind. This is especially true for one of the main components of EAs, that is hypothesized by many to underlie their efficiency: population. The first problem considered in this thesis is the introduction of a recombination operator, K Bit-Swap (KBS) and its comparison to mainstream operators, such as mutation and different types of crossover. A vast amount of statistical evidence is presented that shows that EAs using KBS outperform other algorithms on a whole range of problems. Two problems are selected for a deep theoretical analysis: OneMax and Royal Roads. The main problem of modeling EAs that use both population and a pool of parents is the complexity of the structures that arise from the process of evolution. In most cases either one type of species is considered or certain simple assumptions are made about fitness of the species. The main contribution of this thesis is the development of a new approach to modeling of EAs that is based on approximating the structure of the population and the evolution of subsets thereof. This approach lies at the core of the new tool presented here, the Elitism Levels Traverse Mechanism that was used to derive upper bounds on the runtime of EAs. In addition, lower bounds were found using simpler assumptions of the underlying distribution of species in the population.The second important result of the approach is the derivation of limiting distributions of a subset of the population, a problem well-known in areas such as epidemiology. To the best of the author's knowledge, no such findings have been published in the EA community so far.

[1]  Dirk Sudholt,et al.  Computational complexity of evolutionary algorithms, hybridizations, and swarm intelligence , 2008 .

[2]  Michael D. Vose,et al.  A Critical Examination of the Schema Theorem , 1993 .

[3]  Günter Rudolph,et al.  Convergence of non-elitist strategies , 1994, Proceedings of the First IEEE Conference on Evolutionary Computation. IEEE World Congress on Computational Intelligence.

[4]  James P. Crutchfield,et al.  Statistical Dynamics of the Royal Road Genetic Algorithm , 1999, Theor. Comput. Sci..

[5]  Günter Rudolph,et al.  Convergence analysis of canonical genetic algorithms , 1994, IEEE Trans. Neural Networks.

[6]  Michael D. Vose,et al.  The simple genetic algorithm - foundations and theory , 1999, Complex adaptive systems.

[7]  Günter Rudolph,et al.  Theory of Evolutionary Algorithms: A Bird's Eye View , 1999, Theor. Comput. Sci..

[8]  Magnus Rattray,et al.  THE DYNAMICS OF A GENETIC ALGORITHM FOR A SIMPLE LEARNING PROBLEM , 1996 .

[9]  Colin R. Reeves,et al.  Genetic Algorithms: Principles and Perspectives: A Guide to Ga Theory , 2002 .

[10]  Stephen R. Marsland,et al.  Convergence of a Recombination-Based Elitist Evolutionary Algorithm on the Royal Roads Test Function , 2011, Australasian Conference on Artificial Intelligence.

[11]  Stephen R. Marsland,et al.  The K-bit-swap: a new genetic algorithm operator , 2010, GECCO '10.

[12]  Stephen Marsland,et al.  Derivation of Upper Bounds on Optimization Time of Population-Based Evolutionary Algorithm on a Function with Fitness Plateaus Using Elitism Levels Traverse Mechanism , 2012, ArXiv.

[13]  Carsten Witt,et al.  An Analysis of the (µ+1) EA on Simple Pseudo-Boolean Functions , 2004, GECCO.

[14]  Prügel-Bennett,et al.  Analysis of genetic algorithms using statistical mechanics. , 1994, Physical review letters.

[15]  Pietro Simone Oliveto,et al.  Analysis of population-based evolutionary algorithms for the vertex cover problem , 2008, 2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence).

[16]  Ingo Wegener,et al.  Randomized local search, evolutionary algorithms, and the minimum spanning tree problem , 2004, Theor. Comput. Sci..

[17]  J. Crutchfield,et al.  Metastable evolutionary dynamics: Crossing fitness barriers or escaping via neutral paths? , 1999, Bulletin of mathematical biology.

[18]  Anabela Simões,et al.  Transposition: A Biological-Inspired Mechanism to Use with Genetic Algorithms , 1999, ICANNGA.

[19]  Xin Yao,et al.  Time complexity of evolutionary algorithms for combinatorial optimization: A decade of results , 2007, Int. J. Autom. Comput..

[20]  Thomas Jansen,et al.  A building-block royal road where crossover is provably essential , 2007, GECCO '07.

[21]  Ingo Wegener,et al.  Real Royal Road Functions for Constant Population Size , 2003, GECCO.

[22]  John G. Kemeny,et al.  Finite Markov chains , 1960 .

[23]  Stephen R. Marsland,et al.  Convergence Properties of (μ + λ) Evolutionary Algorithms , 2011, AAAI.

[24]  Ingemar Nåsell,et al.  The quasi-stationary distribution of the closed endemic sis model , 1996, Advances in Applied Probability.

[25]  Carsten Witt,et al.  Tight Bounds on the Optimization Time of the (1+1) EA on Linear Functions , 2011, ArXiv.

[26]  I. Nåsell,et al.  On the quasi-stationary distribution of the stochastic logistic epidemic. , 1999, Mathematical biosciences.

[27]  J. Rosenthal.  A First Look at Rigorous Probability Theory , 2000 .

[28]  David H. Wolpert,et al.  No free lunch theorems for optimization , 1997, IEEE Trans. Evol. Comput..

[29]  Stephen R. Marsland,et al.  Convergence Properties of Two (μ+λ) Evolutionary Algorithms on OneMax and Royal Roads Test Functions , 2011, IJCCI.

[30]  G. Unter Rudolph.  Convergence Rates of Evolutionary Algorithms for a Class of Convex Objective Functions , 1997 .

[31]  Joe Suzuki,et al.  A Markov chain analysis on simple genetic algorithms , 1995, IEEE Trans. Syst. Man Cybern..

[32]  David J. Marchette,et al.  Modeling computer virus prevalence with a susceptible-infected-susceptible model with reintroduction , 2004, Comput. Stat. Data Anal..

[33]  Brian W. Kernighan,et al.  An Effective Heuristic Algorithm for the Traveling-Salesman Problem , 1973, Oper. Res..

[34]  Dirk Sudholt,et al.  The impact of parametrization in memetic evolutionary algorithms , 2009, Theor. Comput. Sci..

[35]  Günter Rudolph,et al.  Convergence of evolutionary algorithms in general search spaces , 1996, Proceedings of IEEE International Conference on Evolutionary Computation.

[36]  XI FachbereichInformatik.  Finite Markov Chain Results in Evolutionary Computation: a Tour D'horizon , 1998 .

[37]  Melanie Mitchell,et al.  The royal road for genetic algorithms: Fitness landscapes and GA performance , 1991 .

[38]  Aram Ter-Sarkisov,et al.  Elitism Levels Traverse Mechanism for the derivation of upper bounds on unimodal functions , 2012, 2012 IEEE Congress on Evolutionary Computation.

[39]  Hao Wang,et al.  Introduction to Genetic Algorithms in Electromagnetics , 1995 .

[40]  Thomas Worsch.  Lower and upper bounds for (sums of) binomial coefficients , 1994 .

[41]  Pietro Simone Oliveto,et al.  Simplified Drift Analysis for Proving Lower Bounds in Evolutionary Computation , 2008, Algorithmica.