Intelligent computing and applications (LSMS and ICSEE 2010)

Kang Li, Haibo He, Qun Niu
2012 Neural computing & applications (Print)  
Many real-world problems can be formulated as optimization problems, but they are complex and NP hard, and conventional optimization approaches often fail to perform well. Intelligent optimization algorithms have thus become the attractive alternatives. The first five papers cover some new methods and their applications. Zhang et al. investigate two-stage multi-item inventory systems with stochastic demands. Due to the complexity of stochastic inventory optimization in a multi-echelon system,
more » ... w analytical models and effective algorithms exist. In the paper, they establish exact stochastic optimization models and propose an efficient hybrid genetic algorithm (HGA). Monte Carlo method is then used to simulate the actual demand and thus to approximate the long-run average cost. The numerical experiments show that when the variance of stochastic demand increases, echelon policy outperforms installation policy and the proposed heuristic search technique can significantly enhance the search capacity. Wang, Zhao, and Li research the Support Vector Domain Description (SVDD) for the detection of novel data or outliers where the training of a SVDD is a constrained quadratic programing problem. They develop a linear Particle Swarm Optimization for such problems. To overcome the premature convergence, a converging linear PSO is proposed to guarantee at least a local optimum. Experimental results show the efficacy of the proposed method. In order to maximize the amount of final products while reducing the production of by-products in batch processes, Jia, Cheng, and Chiu propose an improved multi-objective particle swarm optimization method based on Pareto-optimal solutions. A novel diversity preservation strategy is used to select the global best and thus enable the convergence and diversity of the Pareto front. Simulation results on benchmark examples and two classical batch processes verify the efficiency and practicability of the proposed algorithm. Xu et al. study the optimal bandwidth scheduling problem for networked two-layer learning control systems where multi-networked feedback control loops share a common communication media. They first formulate a noncooperative game fairness model, taking into account factors such as transmission rate, sampling rate, scheduling patterns, and networked control. Then, an improved shuffled frog leaping algorithm is used to obtain the optimal solutions. They show that the algorithm has a high convergence rate, and simulation results show the effectiveness of the proposed method. Finally, Xu et al. propose a new multi-population cultural differential evolution algorithm. In the algorithm, each of the populations is managed by its private cultural
doi:10.1007/s00521-012-1122-z fatcat:ckg2p5kt3bethmmv3trx7bqbki