Catálogo de publicaciones - libros

Inspired by the mechanism of small-world phenomenon, some small-world optimization operators, mainly including the local short-range searching operator and random long-range searching operator, are constructed in this paper. And a new optimization algorithm, Small-World Optimization Algo-rithm (SWOA) is explored. Compared with the corresponding Genetic Algorithms (GAs), the simulation experiment results of some complex functions optimization indicate that SWOA can enhance the diversity of the population, avoid the prematurity and GA deceptive problem to some extent, and have the high convergence speed. SWOA is shown to be an effective strategy to solve complex tasks.

Chaotic optimization is a new optimization technique. For image segmentation, conventional chaotic sequence is not fit to two-dimension gray histogram because it is proportional distributing in [0,1]×[0,1]. In order to generate a chaotic sequence can be used to the optimization processing of image segmentation method in two-dimension gray histogram, we propose an chaotic sequence generating method based on Arnold chaotic system and Bézier curve generating algorithm. Simulation results show that the generated sequence is pseudorandom. The most important characteristic of this chaotic sequence is that its distribution is approximately inside a disc whose center is (0.5,0.5) , this characteristic indicates that the sequence is superior to the Arnold chaotic sequence in image segmenting. Based on the extended chaotic sequence generating method, we study the two-dimension Otsu’s image segmentation method using chaotic optimization. Simulation results show that the method using the extended chaotic sequence has better segmentation effect and lower computation time than the existed two-dimension Otsu’s method.

Time-varying wavelets are highly desired in exploiting the nonstationarity of signals. However, it is difficult to hold the perfect reconstruction (PR) and regularity properties simultaneously in the construction of time-varying wavelets. This paper proposes a simple method to construct time-varying orthogonal wavelets based on the lattice structure of two-channel paraunitary (PU) filter banks, in which both the PR and orthogonality properties are well preserved. The regularity conditions imposed on the lattice structure are expressed in terms of the lattice coefficients and the wavelet filter banks are obtained by using an optimization technique. Then the time-varying orthogonal wavelets can be constructed by the lattice structure formulation for time-varying filter banks. Design examples show that this method is of great flexibility and effectiveness.

It is important to acquire accurate knowledge of traffic matrices of networks for many traffic engineering or network management tasks. Direct measurement of the traffic matrices is difficult in large scale operational IP networks. One approach is to estimate the traffic matrices statistically from easily measured data. The performance of the statistical methods is limited due to they rely on the limited information and require large amount of computation, which limits the convergence of such computation. In this paper, we present an alternative approach to traffic matrix estimation. This method uses assignment model . The model is based on the link characters and includes a fast algorithm. The algorithm combines statistical and optimized tomography. The algorithm is evaluated by simulation and the simulation results show that our algorithm is robust, fast, flexible, and scalable.

In conventional filter banks, the sampling factors are restricted to rational numbers and frequency partition is always rather inflexible, stemming from the fact that certain constraint on each subband position is always placed. In this paper, we present a class of -channel nonuniform filter banks with arbitrary sampling factors including integer, rational, and even irrational numbers. Consequently, the frequency partitioning in the proposed filter bank is much more flexible, which is very attractive in many applications.

The variance minimization dual adaptive control approach with unknown parameters for stochastic system is studied in this paper. First,the problem is proposed for solving the optimal dual control law. Furthermore, the differential equation is transferred into state space model, the suboptimal dual control law is obtained through DUL algorithm. Finally, the example is given to verify the method developed in this paper. It is shown that the method is effective and practical.

The least-cost multicast routing with delay constrained is an NP-Complete problems, to deal with which, a Multi-Agent Immune Clonal Selection Algorithm based multicast routing (MAICSA) is proposed in this paper. MAICSA combines the characteristic of Multi-Agent with the search strategy of Immune Clonal Selection Algorithm. To compare with the conventional Genetic Algorithm (GA), MAICSA overcomes the most serious drawbacks, such as slow convergence rate and “prematurity”. The experimental results show that MAICSA has faster astringency and higher precision than traditional GA, MAGA (Multi-Agent multicast routing based on Genetic Algorithm) and MAIA (Multi-Agent multicast routing based on Immune Algorithm).

With the rapid development of genome projects, clustering of gene expression data is a crucial step in analyzing gene function and relationship of conditions. In this paper, we put forward an estimation of distribution algorithm for fuzzy clustering gene expression data, which combines estimation of distribution algorithms and fuzzy logic. Comparing with sGA, our method can avoid many parameters and can converge quickly. Tests on real data show that EDA converges ten times as fast as sGA does in clustering gene expression data. For clustering accuracy, EDA can get a more reasonable result than sGA does in the worst situations although both methods can get the best results in the best situations.

This paper presents a novel approach, called MWPAlign Maximum Weighted Path approach to multiple ALIGNment, to perform global multiple alignment of DNA sequences. In our method, de Bruijn graph is used to describe input sequences information. As a result, a consensus-finding problem can be transformed to a maximum weighted path problem of the graph. MWPAlign gets almost linear computation speed of multiple sequences alignment problem. Experimental results show that the proposed algorithm is feasible, and for large number of sequences with lower mutation rate 5.2%, MWPAlign generates better alignment and has a lower computation time as compared to CLUSTALW, T-Coffee and HMMT.

Computer-aided radiotherapy planning within a clinically acceptable time has the potential to improve the therapeutic ratio by providing the optimized and customized treatment plans for the tumor patients. In this paper, a hybrid method is proposed to accelerate the beam angle optimization (BAO) in the intensity modulated radiotherapy (IMRT) planning. In this hybrid method, the genetic algorithm (GA) is used to find the rough distribution of the solution, i.e., to give the initial pheromone distribution for the following ant colony system (ACS) optimization. Then, the ACS optimization is implemented to find the precise solution of the BAO problem. The comparisons of the optimization on a clinical nasopharynx case with GA, ACS and the hybrid method show that the proposed algorithm can obviously improve the computation efficiency.