In this paper, we propose an improved online self-organizing radial basis function neural network (IOS-RBF) modeling method which can dynamically add or merge hidden neuron while tuning the parameters. First, the initial center of the network is determined using the sample output clustering method, which is able to utilize the prior information contained in the output data. Second, a new neuron self-organization adjustment strategy is proposed, which is able to dynamically optimize the network structure according to the generalization ability of the network and the correlation between the nodes, and then, the network parameters are trained using the sliding window Levenberg-Marquardt (LM) algorithm to accelerate the network convergence speed. Finally, the effectiveness of IOS-RBF is demonstrated by two benchmark simulation experiments.
This paper proposes a novel, improved discrete particle swarm optimization algorithm (IDPSO) to solve the multiobjective flexible job shop scheduling problem (MOFJSP) under dynamic perturbations. Firstly, the idea of variable neighborhood search (VNS) is fused to design three neighborhood structures to enhance the ability of the discrete particle swarm algorithm to jump out of local extremes. Secondly, a non-linear descent inertia weight is designed to help the DPSO algorithm better balance the global and local search capabilities. Furthermore, this paper redesigns the individual optimal as well as global optimal particle update methods of the algorithm for the characteristics of the MOFJSP problem. Then, an external archival solution set of elite strategies is designed to store the non-dominated solutions in the algorithm's solution process. Finally, the effectiveness of this algorithm is verified on the Kacem dataset; compared with other algorithms, the results show that this algorithm has a strong performance in finding the best.
For the problems of long planning paths and slow convergence speed of the ant colony optimization (ACO) in mobile robot path planning, this paper improves the ant colony optimization and introduces the artificial potential field method (APF). Firstly, the initial pheromone concentration is differentiated to make the ants more directional at the early stage of the algorithm. Second, the heuristic factor in the heuristic function is dynamically adjusted to optimize the pheromone update strategy so that the algorithm can effectively distinguish the high-quality paths from the common ones. Finally, the distance parameter is added to the repulsion function of APF, the improved ant colony algorithm is fused with APF, and a fusion algorithm based on the improved ant colony algorithm-artificial potential field method (IACO-APF) is proposed. In the simulation experiments, the trajectories are optimized by the B-spline curve method, and the experiments show that the IACO-APF algorithm can obtain better solutions in a shorter time compared with the traditional ant colony algorithm and other algorithms mentioned in the literature.
KEYWORDS: Control systems, Systems modeling, Matrices, Data modeling, Design and modelling, Detection and tracking algorithms, Process modeling, Modeling, Complex systems, Signal filtering
The predictive control of Wiener systems involving modeling nonlinearities is of great importance for analyzing industrial processes. In this paper, a data-driven predictive control method for Wiener systems is proposed. The methods used in this paper are based on some common simple linear algebra tools such as least squares and QR decomposition, where no complex solution process is involved, resulting in high estimation accuracy and numerical robustness. the data-driven predictive controller is designed according to the subspace predictor. The simulation results reveal that the proposed method are more accuracy and stable in Wiener model predictive control.
In this paper, a novel self-organizing radial basis function neural network (RBFNN)-based nonlinear model predictive control (RBF-NMPC) scheme is proposed. First, the RBFNN is initialized on the training data using clustering and extreme learning machine (ELM) algorithms, and it serves as a dynamic predictor of an unknown plant. In addition, an adaptive growing and merging strategy is utilized in the neural network so that the RBFNN can automatically adjust its structure. Second, an improved Levenberg-Marquardt (LM) algorithm with a fixed time window is used to increase convergence speed while tuning network parameters. Then, the optimal control signal is calculated by gradient method. Finally, the validity of the developed method is demonstrated by a simulation of continuous stirred tank reactor system.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.