Search

A hybrid method based on modified augmented Lagrangian multiplier and composite evolutionary algorithm is proposed to solve constrained optimization problems. The basic steps of the proposed method are comprised of an outer iteration, in which the Lagrangian multipliers and various penalty parameters are updated using a first-order update scheme, and an inner iteration, in which a nonlinear optimization of the modified augmented Lagrangian function with simple bound constraints is implemented by composite evolutionary algorithm, which combining several effective crossover operators with some suitable mutation operators. Numerical simulation results show that the proposed method can yield better solutions than those reported in the literature for most problems.

An active set truncated-Newton algorithm (ASTNA) is proposed to solve the large-scale bound constrained sub-problems. The global convergence of the algorithm is obtained and two groups of numerical experiments are made for the various large-scale problems of varying size. The comparison results between ASTNA and the subspace limited memory quasi-Newton algorithm and between the modified augmented Lagrange multiplier methods combined with ASTNA and the modified barrier function method show the stability and effectiveness of ASTNA for simultaneous optimization of distillation column.

A potential reduction algorithm is proposed for optimization of a convex funnction subject to linear constraints. At each step of the algorithm, a system of linear equations is solved to get a search direction and the Armijo's rule is used to determine a stepsize. It is proved that the algorithm is gobally convergent. Computational results are reported.

Active set method and gradient projection method are currently the main approaches for linearly constrained convex programming. Interior-point method is one of the most effective choices for linear programming. In the paper a predictor-corrector interior-point algorithm for linearly constrained convex programming under the predictor-corrector motivation was proposed. In each iteration, the algorithm first performs a predictor-step to reduce the duality gap and then a corrector-step to keep the points close to the central trajectory. Computations in the algorithm only require that the initial iterate be nonnegative while feasibility or strict feasibility is not required. It is proved that the algorithm is equivalent to a level-1 perturbed composite Newton method. Numerical experiments on twenty-six standard test problems are made. The results show that the proposed algorithm is stable and robust.

By using Fukushima’s differentiable merit function, Taji, Fukushima and Ibaraki have given a globally convergent modified Newton method for the strongly monotone variational inequality problem and proved their method to be quadratically convergent under certain assumptions in 1993. In this paper a hybrid method for the variational inequality problem under the assumptions that the mapping F is continuously differentiable and its Jacobian matrix ΔF (x) is positive definite for all x∈S rather than strongly monotone and that the set S is nonempty, polyhedral, closed and convex is proposed. Armijo-type line search and trust region strategies as well as Fukushima’s differentiable merit function are incorporated into the method. It is then shown that the method is well defined and globally convergent and that, under the same assumptions as those of Taji et al., the method reduces to the basic Newton method and hence the rate of convergence is quadratic. Computational experience show the efficiency of the proposed method.

A trust region algorithm is proposed for box constrained nonlinear optimization. At each step of the algorithm a quadratic model problem in a box is minimized. Global convergence and quadratic convergence rate to a strong local minimizer are given. Computational results are presented to show the efficiency of the algorithm

The slow convergence speed and the lack of effective constraint handling strategies are the major concerns when applying genetic algorithms (Gas) to constrained optimization problem. An improved genetic algorithm was proposed by dividing population into three parts: optimal subpopulation, elitists subpopulation and spare subpopulation. We applied genetic algorithm on three subpopulations with different evolutionary strategies. Isolation of optimal subpopulation was to improve convergence speed. Population diversity was kept by spare subpopulation setting and aperiodically decreasing of the size of optimal subpopulation. Gene segregation was carried out by crossover operation between optimal subpopulation and spare subpopulation. Combination of penalty function method and the strategy of elitists preservation by setting elitists subpopulation was used to constraint handling. Some numerical tests have been made and the results show that the algorithm is effective.

Recently Particle Swarm Optimization (PSO) algorithm gained popularity and employed in many engineering applications because of its simplicity and efficiency. The performance of the PSO algorithm can further be improved by using hybrid techniques. There are various hybrid PSO algorithms published in the literature where researchers combine the benefits of PSO with other heuristics algorithms. In this paper, we propose a cooperative line search particle swarm optimization (CLS-PSO) algorithm by integrating local line search technique and basic PSO (B-PSO). The performance of the proposed hybrid algorithm, examined through four typically nonlinear optimization problems, is reported. Our experimental results show that CLS-PSO outperforms basic PSO.

Production and injection rates are the most abundant data available in any injection-production project. One can analyze these data to obtain valuable and useful information about interwell connectivity. The oil reservoir is considered to be an input-output system with the injection rates as the input and production rates as the output. A state space model is developed to infer the interwell connectivity only from the injection and production data on a reservoir. A synthetic field with five injectors and four producers is used to test the model. The simulation results show that the long time dependent behavior between injectors and producers can be satisfactorily captured by the proposed state space model.

A new one-step-ahead predictive control algorithm for a family of nonlinear MISO systems is presented. First, least squares support vector machine (LSSVM) with the quadratic multinomial function kernel is used to obtain the off-line nonlinear systemspsila model. Then, the analytical equation which includes the unknown optimal input increment and the known inputs/outputs is derived according to the predictive control mechanism. And finally the analytical equation is solved through numerical algorithm. The simulation result show the effectiveness and advantage of the one-step-ahead predictive control algorithm.

A kind of product-truth-flowing fuzzy control algorithm is presented in this paper. Firstly, the steps of the algorithm are listed in detail. The algorithm holds advantage that fuzzy reasoning process is intuitionistic and its computation is simple. Secondly, based on the algorithm, a constant-temperature control system is designed with MAX7651 MCU, Pt100, MOC3041. Finally, practical control result shows that the performance of control system is excellent, its rise-time is about 21 minutes, the overshot of temperature is only 0.4 degC, and real-time temperature waves between -0.3degC and +0.3degC at expectation value.

Recently, variants of hybrid PSO algorithms are presented, but hybrid Optimization technique combining classical local line search with PSO is less focused on. In this paper, we proposed a new hybrid cooperative evolutionary algorithm by integrating local line search technique and basic PSO. The performance of our algorithm tested using four typical nonlinear optimization problems is reported and compared with that of basic PSO. Results show that performance of CLS-PSO algorithm is much better than basic PSO, and has considerable steady and robust.

In this paper, a novel fuzzy control algorithm is presented for conventional DISO(double-input & single-output) fuzzy controller. The new algorithm is based on potentially-inherited variable universe. By defining an auxiliary fuzzy singleton set and a fuzzy singleton transition function, the operation of variable universe is greatly simplified at the consequent part of inference. So, the algorithm's computation is significantly reduced, comparing with the method of directly-inherited variable universe or the method of integral variable universe. Finally, the convergence property of this control algorithm is proved theoretically.

A combined method of expression feature extraction with particle swarm optimization (PSO) and independent component analysis (ICA) is proposed. The basic ICA algorithm is used to derive the independent base vector from the expression images. To decrease the computing complexity, the dimension of the expression image is reduced, and then PSO algorithm is applied to process expression data set to get the best optimal solution set. Finally, hidden Markov model is used to validate the correctness and validity of the algorithm. The experiments in the expression database show faster way of expression features extraction based on correct rate of expression recognition.