Catálogo de publicaciones - libros

Continuum topology optimization with stresses constraints is not only important but also complicated. Because stress is a local quantity, a large number of constraints must be considered, and the complication of optimization algorithm and sensitivity analysis is also increased. A global stress constraints method based on ICM (Independent Continuous Mapping) method, which takes minimum weight and strain energy of structure with multi-load-case as design objective and constraint respectively, is suggested in this paper. Dual quadratic programming is applied to solve the optimal model for continuum structure established in this paper. As a result, the number of constraints is reduced and the local optimal solution of the weight is also avoided. Two numerical examples are discussed, and their results show that the present method is effective and efficient.

Structural topological optimization is to seek the best path of transmitting forces for structures. The topological optimization under multiple loading cases involves the balance of many paths of transmitting forces. Based on the Independent Continuous Mapping (ICM) method, the optimization problem under multiple loading cases is solved under three conditions, namely local constraints, global constraints, and their combination. In this paper, local constraints are firstly analyzed by envelope method and average method. Secondly, global constraints are uniformly calculated with mathematical programming (MP). Thirdly, local and global constraints are processed by synthesizing above two methods. Finally, the results are compared. From the present numerical examples, it is shown that the envelope method and the MP or their combination can be used to efficiently model and accurately simulate the topological optimization problem under multiple loading cases.

To solve the complicated non-linear mode-sorting problem of protein secondary structure prediction, the chapter proposed a new method based on radial basis function neural networks and learning from evolution. It also discussed the influence of data selection and structure design on the performance of the networks. The results indicate that this method is feasible and effective.

Our problem is about a routing of a vehicle with pickup and delivery of product with time window constraints. This problem requires to be attended with large scale instances (nodes ≥ 100). A strong active time window exists (≥90%) with a large factor of amplitude (≥75%). The problem is NP-hard and for that reason the application of an exact method is limited by the computational time. We propose a hybrid methodology which offers good solutions in computational times that do useful its application.

An improved Immune Genetic Algorithm was proposed to solve the Flow Shop Problem. Basing on the standard Immune Genetic Algorithm, the vaccination technique and a novel method for calculating the affinity between the antibodies was added. Finally the algorithm was test with standard benchmark problems, and the experiment result shows the validity of the algorithm.

A discrete particle swarm optimization (PSO)-based algorithm for travelling salesman problem (TSP) is presented by re-designing the ‘subtraction’ operator. To accelerate the convergence speed, a crossover eliminating technique is also added. Numerical results of some benchmark instances show that the size of solved problems could be increased by using the proposed algorithm compared with those of the existing PSO-based algorithms.

In this paper, we propose a search method for finding a simple Nash Equilibrium in a 2-player game and execute it on 24 classes of games. The result shows that our algorithm performs better than the classical Lemke–Howson algorithm.

Both GA and PSO are typical evolution algorithm with their own advantages. In this paper, a new evolution algorithm is introduced based on GA and PSO. The total population are divided into several tribes, which one Hero individual and several common particles are included in each tribe. The movement velocity of each hero is calculated by global peak point and local best point among this tribe just like PSO. Other common particles will search neighbourhood of hero using recombination method of GA. Hero algorithm will converge fast and escape from local peak inheriting advantages of GA and PSO. These conclusions are proven from experiment of some familiar Benchmark functions.

The goal of this article is to develop an efficient scheduling method based a proposed novel discrete particle swarm optimization (PSO) for the job-shop scheduling problem (JSSP). This paper also introduces a novel concept for distance and velocity of particles in the PSO to pave the way for the JSSP. The proposed method effectively exploits the capabilities of distributed and parallel computing systems. Simulation results show that the proposed algorithm can obtain high quality solutions for typical benchmark problems.

This paper focuses on the visual servo control for an uncalibrated robotic arm with an eye-in-hand camera. Without a priori knowledge of the robotic arm’s kinematic model or camera calibration, the control system can track a moving object through visual feedback. Two methods are proposed to resolve the output constraint of control. The joint angle feedback is used to correct the calculating values of the joint angles in the first control method. The output of controller is re-scaled to ensure that the output constraint of controller is not violated in the second method. The performances of the proposed control methods are illustrated by the computer simulations.