Nachol Chaiyaratana

Multi-objective Co-operative Co-evolutionary Genetic Algorithm

This paper presents the integration between two types of genetic algorithm: a multi-objective genetic algorithm (MOGA) and a co-operative co-evolutionary genetic algorithm (CCGA). The resulting algorithm is referred to as a multi-objective co-operative coevolutionary genetic algorithm or MOCCGA. The integration between the two algorithms is carried out in order to improve the performance of the MOGA by adding the co-operative co-evolutionary effect to the search mechanisms employed by the MOGA. The MOCCGA is benchmarked against the MOGA in six different test cases. The test problems cover six different characteristics that can be found within multi-objective optimisation problems: convex Pareto front, non-convex Pareto front, discrete Pareto front, multi-modality, deceptive Pareto front and non-uniformity in the solution distribution. The simulation results indicate that overall the MOCCGA is superior to the MOGA in terms of the variety in solutions generated and the closeness of solutions to the true Pareto-optimal solutions. A simple parallel implementation of MOCCGA is described. With an 8-node cluster, the speed up of 2.69 to 4.8 can be achieved for the test problems.

Thalassaemia classification by neural networks and genetic programming

This paper presents the use of a neural network and a decision tree, which is evolved by genetic programming (GP), in thalassaemia classification. The aim is to differentiate between thalassaemic patients, persons with thalassaemia trait and normal subjects by inspecting characteristics of red blood cells, reticulocytes and platelets. A structured representation on genetic algorithms for non-linear function fitting or STROGANOFF is the chosen architecture for genetic programming implementation. For comparison, multilayer perceptrons are explored in classification via a neural network. The classification results indicate that the performance of the GP-based decision tree is approximately equal to that of the multilayer perceptron with one hidden layer. But the multilayer perceptron with two hidden layers, which is proven to have the most suitable architecture among networks with different number of hidden layers, outperforms the GP-based decision tree. Nonetheless, the structure of the decision tree reveals that some input features have no effects on the classification performance. The results confirm that the classification accuracy of the multilayer perceptron with two hidden layers can still be maintained after the removal of the redundant input features. Detailed analysis of the classification errors of the multilayer perceptron with two hidden layers, in which a reduced feature set is used as the network input, is also included. The analysis reveals that the classification ambiguity and misclassification among persons with minor thalassaemia trait and normal subjects is the main cause of classification errors. These results suggest that a combination of a multilayer perceptron with a blood cell analysis may give rise to a guideline/hint for further investigation of thalassaemia classification.

Effects of diversity control in single-objective and multi-objective genetic algorithms

This paper covers an investigation on the effects of diversity control in the search performances of single-objective and multi-objective genetic algorithms. The diversity control is achieved by means of eliminating duplicated individuals in the population and dictating the survival of non-elite individuals via either a deterministic or a stochastic selection scheme. In the case of single-objective genetic algorithm, onemax and royal road R1 functions are used during benchmarking. In contrast, various multi-objective benchmark problems with specific characteristics are utilised in the case of multi-objective genetic algorithm. The results indicate that the use of diversity control with a correct parameter setting helps to prevent premature convergence in single-objective optimisation. Furthermore, the use of diversity control also promotes the emergence of multi-objective solutions that are close to the true Pareto optimal solutions while maintaining a uniform solution distribution along the Pareto front.

DNA fragment assembly using an ant colony system algorithm

This work presents the use of an ant colony system algorithm in a DNA (deoxyribonucleic acid) fragment assembly. The assembly problem is a combinatorial optimisation problem where the aim of the search is to find the right order and orientation of each fragment in the fragment ordering sequence that leads to the formation of a consensus sequence. In this paper, an asymmetric ordering representation is proposed where a path cooperatively generated by all ants in the colony represents the search solution. The optimality of the fragment layout obtained is then determined from the sum of overlap scores calculated for each pair of consecutive fragments in the layout. Two types of assembly problem are investigated: single-contig and multiple-contig problems. The simulation results indicate that in single-contig problems, the performance of the ant colony system algorithm is approximately the same as that of a nearest neighbour heuristic algorithm. On the other hand, the ant colony system algorithm outperforms the nearest neighbour heuristic algorithm when multiple-contig problems are considered.

Variable-length haplotype construction for geneߝgene interaction studies

Genetic epidemiology is a research field that aims to identify genetic polymorphisms that are involved in disease susceptibility. In this article, a variable-length haplotype construction for gene-gene interaction (VarHAP) technique is proposed. The technique will involve nonparametric classification where haplotypes inferred from multiple single nucleotide polymorphism (SNP) data are the classifier inputs.

Wireless LAN access point placement using a multi-objective genetic algorithm

This paper presents the use of a multi-objective genetic algorithm (MOGA) for solving an access point placement problem in a wireless LAN. The aim is to maximize signal coverage over the interested area. The problem has been formulated as a multi-objective optimization problem where the decision variables are derived from the locations of the access points in the target area. The objectives consist of the number of access points and the average SNR over the whole area. The major advantage of using MOGA is that multiple optimal placement configurations for different numbers of access points can be obtained from a single run. A set of solutions provides more alternatives to the network designer. The simulation results indicate that the MOGA is capable of generating a placement result which is superior to that produced using standard placement techniques. In addition, the result assessment has been confirmed using statistical and analyzed data profile.

Detecting purely epistatic multi-locus interactions by an omnibus permutation test on ensembles of two-locus analyses

BackgroundPurely epistatic multi-locus interactions cannot generally be detected via single-locus analysis in case-control studies of complex diseases. Recently, many two-locus and multi-locus analysis techniques have been shown to be promising for the epistasis detection. However, exhaustive multi-locus analysis requires prohibitively large computational efforts when problems involve large-scale or genome-wide data. Furthermore, there is no explicit proof that a combination of multiple two-locus analyses can lead to the correct identification of multi-locus interactions.ResultsThe proposed 2LOmb algorithm performs an omnibus permutation test on ensembles of two-locus analyses. The algorithm consists of four main steps: two-locus analysis, a permutation test, global p-value determination and a progressive search for the best ensemble. 2LOmb is benchmarked against an exhaustive two-locus analysis technique, a set association approach, a correlation-based feature selection (CFS) technique and a tuned ReliefF (TuRF) technique. The simulation results indicate that 2LOmb produces a low false-positive error. Moreover, 2LOmb has the best performance in terms of an ability to identify all causative single nucleotide polymorphisms (SNPs) and a low number of output SNPs in purely epistatic two-, three- and four-locus interaction problems. The interaction models constructed from the 2LOmb outputs via a multifactor dimensionality reduction (MDR) method are also included for the confirmation of epistasis detection. 2LOmb is subsequently applied to a type 2 diabetes mellitus (T2D) data set, which is obtained as a part of the UK genome-wide genetic epidemiology study by the Wellcome Trust Case Control Consortium (WTCCC). After primarily screening for SNPs that locate within or near 372 candidate genes and exhibit no marginal single-locus effects, the T2D data set is reduced to 7,065 SNPs from 370 genes. The 2LOmb search in the reduced T2D data reveals that four intronic SNPs in PGM1 (phosphoglucomutase 1), two intronic SNPs in LMX1A (LIM homeobox transcription factor 1, alpha), two intronic SNPs in PARK2 (Parkinson disease (autosomal recessive, juvenile) 2, parkin) and three intronic SNPs in GYS2 (glycogen synthase 2 (liver)) are associated with the disease. The 2LOmb result suggests that there is no interaction between each pair of the identified genes that can be described by purely epistatic two-locus interaction models. Moreover, there are no interactions between these four genes that can be described by purely epistatic multi-locus interaction models with marginal two-locus effects. The findings provide an alternative explanation for the aetiology of T2D in a UK population.ConclusionAn omnibus permutation test on ensembles of two-locus analyses can detect purely epistatic multi-locus interactions with marginal two-locus effects. The study also reveals that SNPs from large-scale or genome-wide case-control data which are discarded after single-locus analysis detects no association can still be useful for genetic epidemiology studies.

Multi-objective Optimisation by Co-operative Co-evolution

This paper presents the integration between a co-operative co-evolutionary genetic algorithm (CCGA) and four evolutionary multi-objective optimisation algorithms (EMOAs): a multi-objective genetic algorithm (MOGA), a niched Pareto genetic algorithm (NPGA), a non-dominated sorting genetic algorithm (NSGA) and a controlled elitist non-dominated sorting genetic algorithm (CNSGA). The resulting algorithms can be referred to as co-operative co-evolutionary multi-objective optimisation algorithms or CCMOAs. The CCMOAs are benchmarked against the EMOAs in seven test problems. The first six problems cover different characteristics of multi-objective optimisation problems, namely convex Pareto front, non-convex Pareto front, discrete Pareto front, multi-modality, deceptive Pareto front and non-uniformity of solution distribution. In contrast, the last problem is a two-objective real-world problem, which is generally referred to as the continuum topology design. The results indicate that the CCMOAs are superior to the EMOAs in terms of the solution set coverage, the average distance from the non-dominated solutions to the true Pareto front, the distribution of the non-dominated solutions and the extent of the front described by the non-dominated solutions.

Compressed-objective genetic algorithm

A strategy for solving an optimisation problem with a large number of objectives by transforming the original objective vector into a two-objective vector during survival selection is presented. The transformed objectives, referred to as preference objectives, consist of a winning score and a vicinity index. The winning score, a maximisation criterion, describes the difference of the number of superior and inferior objectives between two solutions. The minimisation vicinity index describes the level of solution clustering around a search location, particularly the best value of each individual objective, is used to encourage the results to spread throughout the Pareto front. With this strategy, a new multi-objective algorithm, the compressed-objective genetic algorithm (COGA), is introduced. COGA is subsequently benchmarked against a non-dominated sorting genetic algorithm II (NSGA-II) and an improved strength Pareto genetic algorithm (SPEA-II) in six scalable DTLZ benchmark problems with three to six objectives. The results reveal that the proposed strategy plays a crucial role in the generation of a superior solution set compared to the other two techniques in terms of the solution set coverage and the closeness to the true Pareto front. Furthermore, the spacing of COGA solutions is very similar to that of SPEA-II solutions. Overall, the functionality of the multi-objective evolutionary algorithm (MOEA) with preference objectives is effectively demonstrated.

Hybridisation of neural networks and genetic algorithms for time-optimal control

This paper presents the use of neural networks and genetic algorithms in time-optimal control of a closed-loop robotic system. Radial-basis function networks are used in conjunction with PID controllers in an independent joint position control to reduce tracking errors. The results indicate that using neural network controllers is more effective than using the trajectory pre-shaping scheme, reported in early literature. Subsequently, a genetic algorithm with a weighted-sum approach and a multi-objective genetic algorithm (MOGA) are used to solve a multi-objective optimisation problem related to time-optimal control. The results indicate that the MOGA is the best method in terms of the Pareto front coverage while the genetic algorithm with a weighted-sum approach is more effective in terms of finding the best individual according to the weighted-sum criteria. As a result of using both neural networks and genetic algorithms in this application, an idea of a task hybridisation between neural networks and genetic algorithms for use in a control system is also effectively demonstrated.