Denis Pallez

Paired comparison-based Interactive Differential Evolution

We propose a system of Interactive Differential Evolution (IDE) based on paired comparisons for reducing user fatigue and evaluate its convergence speed in comparison with Interactive Genetic Algorithms (IGA) and tournament IGA. User interface and convergence performance are central to reducing Interactive Evolutionary Computation (IEC) user fatigue. Unlike IGA and conventional IDE, users of the proposed IDE and tournament IGA do not need to compare whole individuals with each other but rather only to compare pairs of individuals, which largely decreases user fatigue. In this paper, we design a pseudo-IEC user and evaluate another factor, IEC convergence performance, using IEC simulators and show that our proposed IDE converges significantly faster than IGA and tournament IGA, i.e. our proposed method is superior to others from both user interface and convergence performance points of view.

CAD/CAPP Integration using Feature Ontology

In a collaborative computer-supported engineering environment, the interoperation of various applications will need a representation that goes beyond the current geometry-based representation, which is inadequate for capturing semantic information. The primary purpose of this study is to discuss a semantically based information exchange protocol that will facilitate seamless interoperability among current and next generation computer-aided design systems (CAD) and between CAD and other systems that use product data. An ontological approach is described to integrating computer-aided design (CAD) and computer-aided process planning (CAPP). Two commercial software applications are used to demonstrate the approach. This involves the development of a shared ontology and domain specific ontologies in the Knowledge Interchange Format (KIF) language. Domain specific ontologies — which are feature-based — are developed after a detailed analysis of the CAD and the CAPP software. Mapping between the domain ontologies and the shared ontology is achieved by several mapping rules. The approach is validated by using a variety of parts.

Eye-tracking evolutionary algorithm to minimize user fatigue in IEC applied to interactive one-max problem

In this paper, we describe a new algorithm that consists in combining an eye-tracker for minimizing the fatigue of a user during the evaluation process of Interactive Evolutionary Computation. The approach is then applied to the Interactive One-Max optimization problem.

Comparative analysis of multi-objective evolutionary algorithms for QoS-aware web service composition

Graphical abstractDisplay Omitted HighlightsMost existing approaches reduce the multi-objective optimization problem of QoS-aware web service composition to a single-objective problem using scalarization.No comparative study with several multi-objective algorithms has been found in the literature for this specific problem.Several EMOAs are tested: NSGA-II, SPEA2, POSDE, GDE3, and MOEA/D.Differential evolution (DE) algorithms - in particular GDE3 - yields the best results on this specific problem for several scenarios, also having the lowest time complexity. Web service composition combines available services to provide new functionality. The various available services have different quality-of-service (QoS) attributes. Building a QoS-optimal web service composition is a multi-criteria NP-hard problem. Most of the existing approaches reduce this problem to a single-criterion problem by aggregating different criteria into a unique global score (scalarization). However, scalarization has some significant drawbacks: the end user is supposed to have a complete a priori knowledge of its preferences/constraints about the desired solutions and there is no guarantee that the aggregated results match it. Moreover, non-convex parts of the Pareto set cannot be reached by optimizing a convex weighted sum. An alternative is to use Pareto-based approaches that enable a more accurate selection of the end-user solution. However, so far, only few solutions based on these approaches have been proposed and there exists no comparative study published to date. This motivated us to perform an analysis of several state-of-the-art multi-objective evolutionary algorithms. Multiple scenarios with different complexities are considered. Performance metrics are used to compare several evolutionary algorithms. Results indicate that GDE3 algorithm yields the best performances on this problem, also with the lowest time complexity.

Adaptive MOEA/D for QoS-based web service composition

QoS aware service composition is one of the main research problem related to Service Oriented Computing (SOC). A certain functionality may be offered by several services having different Quality of Service (QoS) attributes. Although the QoS optimization problem is multiobjective by its nature, most approaches are based on single-objective optimization. Compared to single-objective algorithms, multiobjective evolutionary algorithms have the main advantage that the user has the possibility to select a posteriori one of the Pareto optimal solutions. A major challenge that arises is the dynamic nature of the problem of composing web services. The algorithms performance is highly influenced by the parameter settings. Manual tuning of these parameters is not feasible. An evolutionary multiobjective algorithm based on decomposition for solving this problem is proposed. To address the dynamic nature of this problem we consider the hybridization between an adaptive heuristics and the multiobjective algorithm. The proposed approach outperforms state of the art algorithms.

Analyzing human gaze path during an interactive optimization task

Interactive Evolutionary Computation (IEC) community aims at reducing users fatigue during an optimization task involving subjective criteria: a set of graphic potential solutions are simultaneously shown to a user which task is to identify most interesting solutions to the problem he had to solve. Evolutionary operators are applied to user choices expecting to produce better solutions. As traditional IEC ask the user to give a mark to each solution or to explicitly choose bests solutions with a mouse, we propose a new framework that uses in real time gaze information to predict which parts of a screen is more significant for a user. We can therefore avoid the user to explicitly choose which solutions are interesting for him. In this paper, we mainly focus on automatically ordering solutions shown on a screen given a gaze path obtained by an eye-tracker. We applied several supervised learning methods (SVM, neural networks...) on two different experiments. We obtain a formula that predict with 85% user choices. We demonstrate that decisive criterion is time spent on one solution and we show the independency between this formula and the experiment.

Coastal current prediction using CMA evolution strategies

We propose a data-driven evolutionary approach to the modeling of marine currents in the Bay of Monaco. The CMA (Covariance Matrix Adaptation) evolution strategy is used to optimize the parameters of a predictive model that may be used as a surrogate of expensive and time-consuming finite-element simulations. The models obtained are reasonably accurate and easy to interpret.

Social Specialization of Space: Clustering Households on the French Riviera

The aim of this paper is to estimate the extent of social specialization of residential space within the French Riviera metropolitan area. Unlike classical approaches, where social groups are pre-defined through given characteristics of households, our approach determines clusters of households inductively. Socio-demographic characteristics of households are thus measured through 16 different indicators. Clustering is then carried out through the optimization of two distinct criteria. Simulated annealing, simple and multi-objective Genetic Algorithm were adapted for this purpose and has produced pertinent results.

Interactive Intonation Optimisation Using CMA-ES and DCT Parameterisation of the F0 Contour for Speech Synthesis

Expressive speech is one of the latest concerns of text-to-speech systems. Due to the subjectivity of expression and emotion realisation in speech, humans cannot objectively determine if one system is more expressive than the other. Most of the text-to-speech systems have a rather flat intonation and do not provide the option of changing the output speech. We therefore present an interactive intonation optimisation method based on the pitch contour parameterisation and evolution strategies. The Discrete Cosine Transform (DCT) is applied to the phrase level pitch contour. Then, the genome is encoded as a vector that contains 7 most significant DCT coefficients. Based on this initial individual, new speech samples are obtained using an interactive Covariance Matrix Adaptation Evolution Strategy (CMA-ES) algorithm. We evaluate a series of parameters involved in the process, such as the initial standard deviation, population size, the dynamic expansion of the pitch over the generations and the naturalness and expressivity of the resulted individuals. The results have been evaluated on a Romanian parametric-based speech synthesiser and provide the guidelines for the setup of an interactive optimisation system, in which the users can subjectively select the individual which best suits their expectations with minimum amount of fatigue.

Comparing paired comparison-based interactive DE and tournament interactive GA on stained glass design

Tournament Interactive Genetic Algorithm (T-IGA) and Paired Comparison-based Interactive Differential Evolution (PC-IDE) are applied to the design of stained glass windows and the two algorithms with variable length genotype are compared in a context of interactive evolutionary computation. For both methods, stained glass windows are represented by colored 2D Voronoi diagrams, and a specific phenotypic crossover operator allows offspring to inherit visual features from both parents. The two algorithms have been evaluated by two professional stained-glass artists whom use them to create original designs in a controlled experimental setting. The results indicate superiority of PC-IDE, thus confirming previous theoretical results.