Catálogo de publicaciones - libros

In this paper we propose an immune algorithm () to solve high dimensional global optimization problems. To evaluate the effectiveness and quality of the we performed a large set of unconstrained numerical optimisation experiments, which is a crucial component of many real-world problem-solving settings. We extensively compare the IA against several Differential Evolution (DE) algorithms as these have been shown to perform better than many other Evolutionary Algorithms on similar problems. The DE algorithms were implemented using a range of recombination and mutation operators combinations. The algorithms were tested on 13 well known benchmark problems. Our results show that the proposed IA is effective, in terms of accuracy, and capable of solving large-scale instances of our benchmarks. We also show that the IA is comparable, and often outperforms, all the DE variants, including two Memetic algorithms.

This paper focuses on modified genetic algorithm based on the graph representation and specialized genetic operators. Advantages of changed representation, multi-level organization as well as self-adaptive aspects of the proposed method are described. The results of algorithm usage in optimising skeletal structures are also presented.

Effective scheduling is of great importance to parallel programming environments. The aim is to minimize the completion time of task graphs. The completion time of a task graph is directly affected by the length of its critical path. Hence, the trend of an evolutionary approach for task graph scheduling can be biased towards reduction of the critical path. In this paper, a new genetic scheduling algorithm is presented. The algorithm, in the first priority, minimizes the critical path length of the parallel program task graph and in the second priority minimizes the inter-processor communication time. Thereby, it achieves a better scheduling in comparison with the existing approaches.

The recently discovered Yang-Baxterization process for the quantum double of the dihedral group algebra, is presented keeping on mind the quantum computation. The products resultant from Yang-Baxterization process are interpreted as universal quantum gates using the Bryslinski’s theorem. Results are obtained for two-qubits and two-qutrits gates. Using the Zhang-Kauffman-Ge method (ZKGM), certain Hamiltonians responsible for the quantum evolution of the quantum gates are obtained. Possible physical systems such as anyons systems are mentioned as referents for practical implementation.

This paper introduces a new study in evolutionary computation technique in order to learn optimal configuration of a multilayer neural network. Inspired from thermodynamic perception, the used evolutionary framework undertakes the optimal configuration problem as a Bi-objective optimization problem. The first objective aims to learn optimal layer topology by considering optimal nodes and optimal connections by nodes. Second objective aims to learn optimal weights setting. The evaluation function of both concurrent objectives is founded on an entropy function which leads the global system to optimal generalization point. Thus, the evolutionary framework shows salient improvements in both modeling and results. The performance of the required algorithms was compared to estimations distribution algorithms in addition to the Backpropagation training algorithm.

The paper presents one of additional mechanisms called which can be introduced to the and which can significantly improve the quality of obtained Pareto frontier approximation. The preliminary experimental comparative studies are based on a typical multi-objective problem presenting the most important features of the proposed approach.

The extraction of IF-THEN rules from data is a promising task of data mining including both Artificial Intelligence and Statistics. One of the difficulties encountered is how to evaluate the relevance of the extracted rules? Many authors use statistical interestingness measures to evaluate the relevance of each rule (taken alone). Recently, few research works have done a synthesis study of the existing interestingness measures but their study presents some limits. In this paper, firstly, we present an overview of related works studying more than forty interestingness measures. Secondly, we establish a list of nineteen other interestingness measures not referenced by the related works. Then, we identify twelve semantic properties characterizing the behavior of interestingness measures. Finally, we did a theoretical study of sixty two interestingness measures by outlining their semantic properties. The results of this study are useful to the users of a data-mining system in order to help them to choose an appropriate measure.

A Compact Genetic Algorithm (CGA) is a genetic algorithm specially devised to meet the tight restrictions of hardware-based implementations. We propose a new mutation operator for an elitism-based CGA. The performance of this algorithm, named emCGA, was tested using a set of algebraic functions for optimization. The optimal mutation rate found for high-dimensionality functions is around 0.5%, and the low the dimension of the problem, the less sensitive is emCGA to the mutation rate. The emCGA was compared with other two similar algorithms and demonstrated better tradeoff between quality of solutions and convergence speed. It also achieved such results with smaller population sizes than the other algorithms.

We present an approach for automated generation of proactive aggregation protocols using Genetic Programming. First a short introduction into aggregation and proactive protocols is given. We then show how proactive aggregation protocols can be specified abstractly. To be able to use Genetic Programming to derive such protocol specifications, we describe a simulation based fitness assignment method. We have applied our approach successfully to the derivation of aggregation protocols. Experimental results are presented that were obtained using our own Distributed Genetic Programming Framework. The results are very encouraging and demonstrate clearly the utility of our approach.

This paper proposes an automated quantum circuit synthesis approach, using a genetic algorithm. We consider the circuit as a successive rippling of the so-called gate sections; also, the usage of a database is proposed in order to specify the gates that will be used in the synthesis process. Details are presented for an appropriate comparison with previous approaches, along with experimental results that prove the convergence and the effectiveness of the algorithm.