Search

In this paper, a novel multi-objective particle swarm optimization algorithm is proposed based on decomposing the objective space into a number of subregions and optimizing them simultaneously. The subregion strategy has two very desirable properties with regard to multi-objective optimization. One advantage is that the local best in the subregion can effectively guide the particles to Pareto front combining with global best. The other is that it has a better performance on the convergence and diversity of solutions. Additionally, this paper applies min-max strategy with determined weight as fitness functions to multi-objective particle swarm optimization, and there is no additional clustering or niching technique needed. In order to demonstrate the performance of the algorithm, it is compared with MOPSO and DMS-MO-PSO. The results indicate that proposed algorithm is efficient.

In order to make full use of channel, a new dynamic channel allocation algorithm for TD-HSPA system is proposed. The proposed algorithm gives priority to consider the time slot distribution in uplink channels. This paper uses low order modulation coding in uplink channels, but uses high order modulation coding in downlink channels. The transmission rate of uplink and downlink are asymmetric. In his paper, we propose a criterion sharing channel for each other through main and auxiliary frequency when the voice channel is idle. As a result, the system capacity is increased 50% larger than the past. Simulation results show that the proposed algorithm can decrease the call blocking ratio and dropping packet rate of data service, improve the channel utilization efficiency, and increase the number of data users dramatically.

This paper presents a new approach to blind separation of sources using sparse representation in an underdetermined mixture. Firstly, we transform the observations into the new ones within the generalized spherical coordinates, through which the estimation of the mixing matrix is formulated as the estimation of the cluster centers. Secondly, we identify the cluster centers by a new classification algorithm, whereby the mixing matrix is estimated. The simulation results have shown the efficacy of the proposed algorithm.

This paper presents a learning framework for blind source separation (BSS), in which the BSS is formulated as generalized Eigenvalue (GE) problem. Compared to the typical information-theoretical approaches, this new one has at least two merits: (1) the unknown unmixing matrix directly works out from the GE equation without timeconsuming iterative learning; (2) The correctness of the solution is guaranteed. We give out a general learning procedure under this framework. The computer simulation shows validity of our method.

This paper presents a novel recurrent time continuous neural network model for solving eigenvalue and eigenvector problem. The network is proved to be globally convergent to an exact eigenvector of a matrix A with respect to the problem's feasible region. This convergence is called quasi-convergence in the sense of the starting point to be in the feasible set. It also demonstrates that the network is primal in the sense that the network's neural trajectories will never escape from the feasible region when starting at it. By using an energy function, the network's stable point set is guaranteed to be the eigenvector set of the involved matrix. Compared with the existing neural network models for eigenvalue problem, the new model's performance is more effective and more reliable. Moreover, simulation results are given to illustrate further the global convergence and the fundamental validity of the proposed neural network for eigenvalue problem.

A new fitness function is proposed in the paper at first, in which the fitness of an individual is defined by the maximum value of the weighted normalized objectives. In order to get the required weights, the sphere coordinate transformation is used. The fitness constructed in this way can result in a group of uniform search directions in the objective space. By using these search directions, the evolutionary algorithm can explore the objective space uniformly, keep the diversity of the population and find uniformly distributed solutions on the Pareto frontier gradually. The numerical simulations indicate the proposed algorithm is efficient and has a better performance than the compared ones.