Catálogo de publicaciones - libros

Based on the study of Particle Swarm Optimization (PSO) on the mechanism of information communion, a new adaptive method of PSO is presented in this paper. This new adaptive method is to avoid the particles getting into local best solution during the optimization. By applying Adaptive Particle Swarm Optimization (APSO) to optimize transition sample values in FIR filter, the maximum stop band attenuation is obtained. The simulations of designing low-pass FIR have been done and the simulation results show that APSO is better than PSO not only in the optimum ability but also in the convergence speed.

The evolutionary granular computing model (EGCM) combining evolutionary computing and granular computing techniques is introduced in this paper. The model presents a new approach to simulate the cognition of human beings that can be viewed as the evolutionary process through the automatic learning from data sets. The information granule, which is the building block of cognition in EGCM, can be synthesized and created by the basic operators. It also can form the granules network by linking each other among granules. With learning from database, the system can evolve under the pressure of selection. The EGCM creates a dynamic model that can adapt to the environment.

PID optimal parameters selection have been extensively studied, in order to improve some strict performance requirements for complex systems. Ziegler-Nichols methods give estimated values for these parameters based on the system’s transient response. Therefore, a fine tuning of these parameters is required to improve the system’s behavior. In this work, genetic programming is used to optimize the three parameters , and , after been tuned by Ziegler-Nichols method, to control a high-order process, a large time delay plant and a highly non-minimum phase process. The results were compared to some other tuning methods, and showed to be promising.

A new approach for dealing with a huge number of cutting pattern combinations encountered in two-dimensional Cutting Stock Problem (CSP) is described. Firstly, cutting patterns are produced according to a novel cutting method LF(Lease Fit) algorithm which can effectively cuts a sequence of small rectangular pieces from a big stock, heuristically maximizing the stock’s utilization ratio. Then Genetic Algorithm (GA) is applied to search for a near optimal solution which consists of many patterns namely a pattern combination. To evaluate the combination’s fitness, LP (Linear Programming) algorithm is used in polynomial time without bringing about much error. The performance and efficiency are justified by numerical experiments.

The complete genomes of living organisms have provided much information on their phylogenetic relationships. In the past few years, we proposed three alternative methods to model the noise background in the composition vector of protein sequences from a complete genome. The first method is based on the frequencies of the 20 kinds of amino acids appearing in the genome and the multiplicative model. The second method is based on the iterated function system model in fractal geometry. The last method is based on the relationship between a word and its two sub-words in the theory of symbolic dynamics. Here we introduce these methods. The complete genomes of prokaryotes and eukaryotes are selected to test these algorithms. Our distance-based phylogenetic tree of prokaryotes and eukaryotes agrees with the biologists’ “tree of life” based on the 16S-like rRNA genes in a majority of basic branches and most lower taxa.

Cellular automata (CA) have been accepted as a good evolutionary computational model for the simulation of complex physical systems. They have been used for various applications, such as parallel processing computations and number theory. In the meanwhile, elliptic curve cryptosystems (ECC) are in the spotlight owing to their significantly smaller parameters. The most costly arithmetic operation in ECC is division, which is performed by multiplying the inverse of a multiplicand. Thus, this paper presents an evolutionary hardware architecture for division based on CA over GF(2) in ECC. The proposed architecture has the advantage of high regularity, expandability, and a reduced latency based on periodic boundary CA. The proposed architecture can be used for the hardware design of crypto-coprocessors.

Surface roughness is one of the essential quality control processes that the carried out to ensure that manufactured parts conform to specified standards and influences the functional characteristics of the work-piece such as fatigue, fracture resistance and surface friction. The most widely used surface finish parameter in industry is the average surface roughness (R) and is conventionally measured by using a stylus type instrument, which has a disadvantage that it requires direct physical contact and may not represent the real characteristics of the surface. Alternately, surface roughness monitoring techniques using non – contact methods based on computer vision technology [1] are becoming popular. In this paper, an evolvable hardware (EHW) configuration using Xilinx Virtex xvc1000 architecture to perform adaptive image processing i.e. noise removal and improve the accuracy of measurement of surface roughness is presented.

Investigate how the level of rationality of power suppliers impacts on equilibrium. First fictitious play was established to electricity market. Then a leaning model Price-deviation-adjust (PD-adjust) was proposed, which inherits main characters of the fictitious play but in a lower rationality because of poor information. An interesting phenomenon is observed in numerical simulations: the errors coming from lower rationality of the agents can be reinforced and often bring the agents extra profits rather than loss, and eventually drive the market to enter an unstable state from the stable equilibrium one. The conclusion is a set of game models identified by a rationality variable should be introduced to understand the electricity market better.

In this paper, a novel evolutionary algorithm based on adaptive multiple fitness functions and adaptive objective space division for multiobjective optimization is proposed. It can overcome the shortcoming of those using the weighted sum of objectives as the fitness functions, and find uniformly distributed solutions over the entire Pareto front for non-convex and complex multiobjective programming. First, we divide the objective space into multiple regions with about the same size by uniform design adaptively, then adaptively define multiple fitness functions to search these regions, respectively. As a result, the Pareto solutions found on each region are adaptively changed and eventually are uniformly distributed over the entire Pareto front. We execute the proposed algorithm to solve five standard test functions and compare performance with that of four widely used algorithms. The results show that the proposed algorithm can generate widely spread and uniformly distributed solutions over the entire Pareto front, and perform better than the compared algorithms.

Based on the Antibody Clonal Selection Theory and the dynamic process of immune response, a novel Immune Forgetting Multiobjective Optimization Algorithm (IFMOA) is proposed. IFMOA incorporates a Pareto-strength based antigen-antibody affinity assignment strategy, a clonal selection operation, and a technique simulating the progress of immune tolerance. The comparison of IFMOA with other two representative methods, Multi-objective Genetic Algorithm (MOGA) and Improved Strength Pareto Evolutionary Algorithm (SPEA2), on different test problems suggests that IFMOA extends the searching scope as well as increasing the diversity of the populations, resulting in more uniformly distributing global Pareto optimal solutions and more integrated Pareto fronts over the tradeoff surface.