Fernando G. Lobo

The compact genetic algorithm

Introduces the compact genetic algorithm (cGA) which represents the population as a probability distribution over the set of solutions and is operationally equivalent to the order-one behavior of the simple GA with uniform crossover. It processes each gene independently and requires less memory than the simple GA. The development of the compact GA is guided by a proper understanding of the role of the GAs parameters and operators. The paper clearly illustrates the mapping of the simple GAs parameters into those of an equivalent compact GA. Computer simulations compare both algorithms in terms of solution quality and speed. Finally, this work raises important questions about the use of information in a genetic algorithm, and its ramifications show us a direction that can lead to the design of more efficient GAs.

Parameter Setting in Evolutionary Algorithms

One of the main difficulties of applying an evolutionary algorithm (or, as a matter of fact, any heuristic method) to a given problem is to decide on an appropriate set of parameter values. Typically these are specified before the algorithm is run and include population size, selection rate, operator probabilities, not to mention the representation and the operators themselves. This book gives the reader a solid perspective on the different approaches that have been proposed to automate control of these parameters as well as understanding their interactions. The book covers a broad area of evolutionary computation, including genetic algorithms, evolution strategies, genetic programming, estimation of distribution algorithms, and also discusses the issues of specific parameters used in parallel implementations, multi-objective evolutionary algorithms, and practical consideration for real-world applications. It is a recommended read for researchers and practitioners of evolutionary computation and heuristic methods.

Linkage Learning via Probabilistic Modeling in the Extended Compact Genetic Algorithm (ECGA)

For a long time, genetic algorithms (GAs) were not very successful in automatically identifying and exchanging structures consisting of several correlated genes. This problem, referred in the literature as the linkage-learning problem, has been the subject of extensive research for many years. This chapter explores the relationship between the linkage-learning problem and that of learning probability distributions over multi-variate spaces. Herein, it is argued that these problems are equivalent. Using a simple but effective approach to learning distributions, and by implication linkage, this chapter reveals the existence of GA-like algorithms that are potentially orders of magnitude faster and more accurate than the simple GA.

The parameter-less genetic algorithm in practice

The parameter-less genetic algorithm was introduced a couple of years ago as a way to simplify genetic algorithm operation by incorporating knowledge of parameter selection and population sizing theory in the genetic algorithm itself. This paper shows how that technique can be used in practice by applying it to a network expansion problem. The existence of the parameter-less genetic algorithm stresses the fact that some problems need more processing power than others. Such observation leads to the development of a problem difficulty measure which is also introduced in this paper. The measure can be useful for comparing the difficulty of real-world problems.

A review of adaptive population sizing schemes in genetic algorithms

This paper reviews the topic of population sizing in genetic algorithms. It starts by revisiting theoretical models which rely on a facetwise decomposition of genetic algorithms, and then moves on to various self-adjusting population sizing schemes that have been proposed in the literature. The paper ends with recommendations for those who design and compare adaptive population sizing schemes for genetic algorithms.

Adaptive Population Sizing Schemes in Genetic Algorithms

Summary. This chapter presents a review of adaptive population sizing schemes used in genetic algorithms. We start by briefly revisiting theoretical models which rely on a facetwise design decomposition, and then move on to various self-adjusting population sizing schemes that have been proposed in the literature. For each method, the major advantages and disadvantages are discussed. The chapter ends with recommendations for those who design and compare self-adjusting population sizing mechanisms for genetic and evolutionary algorithms.

Substructural neighborhoods for local search in the bayesian optimization algorithm

This paper studies the utility of using substructural neighborhoods for local search in the Bayesian optimization algorithm (BOA). The probabilistic model of BOA, which automatically identifies important problem substructures, is used to define the structure of the neighborhoods used in local search. Additionally, a surrogate fitness model is considered to evaluate the improvement of the local search steps. The results show that performing substructural local search in BOA significatively reduces the number of generations necessary to converge to optimal solutions and thus provides substantial speedups.

Combining competent crossover and mutation operators: a probabilistic model building approach

This paper presents an approach to combine competent crossover and mutation operators via probabilistic model building. Both operators are based on the probabilistic model building procedure of the extended compact genetic algorithm (eCGA). The model sampling procedure of eCGA, which mimics the behavior of an idealized recombination---where the building blocks (BBs) are exchanged without disruption---is used as the competent crossover operator. On the other hand, a recently proposed BB-wise mutation operator---which uses the BB partition information to perform local search in the BB space---is used as the competent mutation operator. The resulting algorithm, called hybrid extended compact genetic algorithm (heCGA), makes use of the problem decomposition information for (1) effective recombination of BBs and (2) effective local search in the BB neighborhood. The proposed approach is tested on different problems that combine the core of three well known problem difficulty dimensions: deception, scaling, and noise. The results show that, in the absence of domain knowledge, the hybrid approach is more robust than either single-operator-based approach.

Model accuracy in the Bayesian optimization algorithm

Evolutionary algorithms (EAs) are particularly suited to solve problems for which there is not much information available. From this standpoint, estimation of distribution algorithms (EDAs), which guide the search by using probabilistic models of the population, have brought a new view to evolutionary computation. While solving a given problem with an EDA, the user has access to a set of models that reveal probabilistic dependencies between variables, an important source of information about the problem. However, as the complexity of the used models increases, the chance of overfitting and consequently reducing model interpretability, increases as well. This paper investigates the relationship between the probabilistic models learned by the Bayesian optimization algorithm (BOA) and the underlying problem structure. The purpose of the paper is threefold. First, model building in BOA is analyzed to understand how the problem structure is learned. Second, it is shown how the selection operator can lead to model overfitting in Bayesian EDAs. Third, the scoring metric that guides the search for an adequate model structure is modified to take into account the non-uniform distribution of the mating pool generated by tournament selection. Overall, this paper makes a contribution towards understanding and improving model accuracy in BOA, providing more interpretable models to assist efficiency enhancement techniques and human researchers.

Loopy Substructural Local Search for the Bayesian Optimization Algorithm

This paper presents a local search method for the Bayesian optimization algorithm (BOA) based on the concepts of substructural neighborhoods and loopy belief propagation. The probabilistic model of BOA, which automatically identifies important problem substructures, is used to define the topology of the neighborhoods explored in local search. On the other hand, belief propagation in graphical models is employed to find the most suitable configuration of conflicting substructures. The results show that performing loopy substructural local search (SLS) in BOA can dramatically reduce the number of generations necessary to converge to optimal solutions and thus provides substantial speedups.