Sampled-data iterative learning control (SILC) for singular systems is addressed for the first time. With the introduction of the constrained relative degree, an SILC algorithm combined with a feedback control law is ...Sampled-data iterative learning control (SILC) for singular systems is addressed for the first time. With the introduction of the constrained relative degree, an SILC algorithm combined with a feedback control law is proposed for singular systems. Convergence of the algorithm is proved in sup-norm, while the conventional convergence analysis is in λ-norm. The final tracking error uniformly converges to a small residual set whose level of magnitude depends on the system dynamics and the sampling-period. Due to inequalities to estimate the level of the existing results of SILC, convergence is guaranteed not only at the sampling instants but on the entire operation interval, so that the inter-sample behavior is guaranteed, which is more practical for real implementation.展开更多
This paper studies the difference algorithm parameters characteristic of the multicast routing problem, and to compare it with genetic algorithms. The algorithm uses the path of individual coding, combined with the di...This paper studies the difference algorithm parameters characteristic of the multicast routing problem, and to compare it with genetic algorithms. The algorithm uses the path of individual coding, combined with the differential cross-choice strategy and operations optimization. Finally, we simulated 30 node networks, and compared the performance of genetic algorithm and differential evolution algorithm. Experimental results show that multi-strategy Differential Evolution algorithm converges faster and better global search ability and stability.展开更多
文摘Sampled-data iterative learning control (SILC) for singular systems is addressed for the first time. With the introduction of the constrained relative degree, an SILC algorithm combined with a feedback control law is proposed for singular systems. Convergence of the algorithm is proved in sup-norm, while the conventional convergence analysis is in λ-norm. The final tracking error uniformly converges to a small residual set whose level of magnitude depends on the system dynamics and the sampling-period. Due to inequalities to estimate the level of the existing results of SILC, convergence is guaranteed not only at the sampling instants but on the entire operation interval, so that the inter-sample behavior is guaranteed, which is more practical for real implementation.
文摘This paper studies the difference algorithm parameters characteristic of the multicast routing problem, and to compare it with genetic algorithms. The algorithm uses the path of individual coding, combined with the differential cross-choice strategy and operations optimization. Finally, we simulated 30 node networks, and compared the performance of genetic algorithm and differential evolution algorithm. Experimental results show that multi-strategy Differential Evolution algorithm converges faster and better global search ability and stability.
