Due to an increasing interest in solving real-world optimization problems using evolutionary algorithms (EAs), researchers have developed a number of real-parameter genetic algorithms (GAs) in the recent past. In such studies, the main research effort is spent on developing an efficient recombination operator. Such recombination operators use probability distributions around the parent solutions to create an ospring. Some operators emphasize solutions at the center of mass of parents and some around the parents. In this paper, we propose a generic parent-centric recombination operator (PCX) and a steady-state, elite-preserving, scalable, and computationally fast population-alteration model (we called the G3 model). The performance of the G3 model with the PCX operator is investigated on three commonly-used test problems and is compared with a number of evolutionary and classical optimization algorithms including other real-parameter GAs with UNDX and SPX operators, the correlated self-adaptive evolution strategy, the dierential evolution technique and the quasi-Newton method. The proposed approach is found to be consistently and reliably performing better than all other methods used in the study. A scale-up study with problem sizes up to 500 variables shows a polynomial computational complexity of the proposed approach. This extensive study clearly demonstrates the power of the proposed technique in tackling real-parameter optimization problems.

This paper presents a real-coded memetic algorithm that applies a crossover hillclimbing to solutions produced by the genetic operators. On the one hand, the memetic algorithm provides global search (reliability) by means of the promotion of high levels of population diversity. On the other, the crossover hill-climbing exploits the selfadaptive capacity of real-parameter crossover operators with the aim of producing an effective local tuning on the solutions (accuracy). An important aspect of the memetic algorithm proposed is that it adaptively assigns different local search probabilities to individuals. It was observed that the algorithm adjusts the global/local search balance according to the particularities of each problem instance. Experimental results show that, for a wide range of problems, the method we propose here consistently outperforms other real-coded memetic algorithms which appeared in the literature.

In this work we present an improved evolutionary method for inferring S-system model of genetic networks from the time series data of gene expression. We employed Differential Evolution (DE)for optimizing the network parameters to capture the dynamics in gene expression data. In a preliminary investigation we ascertain the suitability of DE for a multimodal and strongly non-linear problem like gene network estimation. An extension of the fitness function for attaining the sparse structure of biological networks has been proposed. For estimating the parameter values more accurately an enhancement of the optimization procedure has been also suggested. The effectiveness of the proposed method was justified performing experiments on a genetic network using different numbers of artificially created time series data.

In this paper, we address the problem of finding gene regulatory networks from experimental DNA microarray data. We focus on the evaluation of the performance of different evolutionary algorithms on the inference problem. These algorithms are used to evolve an underlying quantitative mathematical model. The dynamics of the regulatory system are modeled with two commonly used approaches, namely linear weight matrices and S-systems and a novel formulation, namely H-systems. Due to the complexity of the inference problem, some researchers suggested evolutionary algorithms for this purpose. However, in many publications only one algorithm is used without any comparison to other optimization methods. Thus, we introduce a framework to systematically apply evolutionary algorithms and different types of mutation and crossover operators to the inference problem for further comparative analysis.

In real-coded genetic algorithms, some crossover operators do not work well on functions which have their optimum at the corner of the search space. To cope with this problem, we have proposed a boundary extension methods which allows individuals to be located within a limited space beyond the boundary of the search space. In this paper, we give an analysis of the boundary extension methods from the view point of sampling bias and perform a comparative study on the effect of applying two boundary extension methods, namely the boundary extension by mirroring BEM) and the boundary extension with extended selection (BES). We were able to confirm that to use sampling methods which have smaller sampling bias had good performance on both functions which have their optimum at or near the boundaries of the search space, and functions which have their optimum at the center of the search space. The BES/SD/A (BES by shortest distance selection with aging) had good performance on functions which have their optimum at or near the boundaries of the search space. We also confirmed that applying the BES/SD/A did not cause any performance degradation on functions which have their optimum at the center of the search space. 1.

Abstract not found

Previous studies have proposed simplex crossover (SPX) for real-coded GAs. In this paper, we propose two types of linkage identification for simplex crossover; linkage identification with singlestage evolution (LISS) and linkage identification with multi-stage evolution (LIMS), and perform their comparative study. Results showed LIMS has more stable performance than LISS. I.

This paper presents a new method that improves robustness of Real-Coded Evolu- tionary Algorithms (RCEAs), such as Real- Coded Genetic Algorithms and Evolution Strategies, for function optimization. It is reported that most crossover (or recombination) operators for RCEAs has sampling bias that prevents to find the optimum near the boundary of search space. They like to search the center of search space much more than the other. Therefore, they will not work on functions that have their optima near the boundary of the search space. Although several methods have been proposed to reduce this sampling bias, they could not cancel the whole bias. In this paper, we propose a new method, Toroidal Search Space Conversion (TSC), to remove this sampling bias. TSC converts bounded search space into toroidal one with no parameters. Experimental results show that a RCEA with TSC has higher performance to find the optimum near the boundary of search space and it has improved robustness concerning the relative position of the optimum.

Abstract. Multi-objective problems with parameter interactions can present difficulties to many optimization algorithms. We have investigated the behaviour of Simplex Crossover (SPX), Unimodal Normally Distributed Crossover (UNDX), Parent-centric Crossover (PCX), and Differential Evolution (DE), as possible alternatives to the Simulated Binary Crossover (SBX) operator within the NSGA-II (Non-dominated Sorting Genetic Algorithm II) on four rotated test problems exhibiting parameter interactions. The rotationally invariant crossover operators demonstrated improved performance in optimizing the problems, over a non-rotationally invariant crossover operator. 1

This technical report describes the domain of numerical optimisation and some of the techniques used in solving numerical problems using genetic algorithms. It is organised into three main sections. The first describes the basic numerical optimisation problem considered here. Other definitions exist and in particular this document does not treat constraints, beyond simple bounds on the search space, in much detail. The second section discusses genetic algorithms and some of the techniques involved. The field of genetic algorithms has expanded greatly in recent years and coverage here is obviously limited. In particular, there is little treatment of noisy and dynamic objective functions. The final section discusses parallel implementations of genetic algorithms, considered important for practical application on large-scale problems. Although this document contains some guidance as to algorithm design, implementation details are not provided and the treatment of most issues is presented largely as a (necessarily limited) review of the literature. For a more practical treatment and/or implementation details, refer to the referenced text books (or more recent additions to the literature) or the wide range of Internet resources available. This report is provided “as is” and no guarantee is made as to its accuracy. This document may be freely distributed provided that appropriate acknowledgement is given. The original document can be obtained from