The focus of this paper is on combination of artificial neural-network (ANN) forecasters with application to the prediction of daily natural gas consumption needed by gas utilities. ANN forecasters can model the complex relationship between weather parameters and previous gas consumption with the future consumption. A two-stage system is proposed with the first stage containing two ANN forecasters, a multilayer feedforward ANN and a functional link ANN. These forecasters are initially trained with the error backpropagation algorithm, but an adaptive strategy is employed to adjust their weights during on-line forecasting. The second stage consists of a combination module to mix the two individual forecasts produced in the first stage. Eight different combination algorithms are examined, they are based on: averaging, recursive least squares, fuzzy logic, feedforward ANN, functional link ANN, temperature space approach, Karmarkar's linear programming algorithm and adaptive mixture of local experts (modular neural networks). The performance is tested on real data from six different gas utilities. The results indicate that combination strategies based on a single ANN outperform the other approaches.

No abstract

This study introduces a new design method for reconfigurable phased arrays using hybrid differential evolution (DE) and enhanced particle swarm optimisation (EPSO) technique. The proposed technique combines DE and enhanced version of standard PSO with improved mechanism that updates velocities and global best solution. In the hybrid algorithm, DE and EPSO are executed in parallel with frequent information sharing to enhance the newly generated population. To demonstrate the effectiveness of the proposed algorithm over each separate algorithm, examples for designing reconfigurable linear and circular antenna arrays with prescribed null directions are presented. Null steering is achieved by position perturbation of array elements in arbitrary directions with minimum sidelobe level change constraint. Another objective is to minimise the number of mobilised elements by introducing elements selection criteria. Simulation results show that the global search ability of the proposed algorithm is improved when compared with DE and EPSO separately.

No abstract

Design of circular arrays (CAs) and hexagonal arrays (HAs) with low sidelobe level (SLL) and high directivity is usually achieved by increasing the number of array elements, which leads to a high undesired mutual coupling. Therefore this study presents an efficient optimisation method and a framework to show how to design multi-ring concentric CAs (CCA) and concentric HAs (CHAs) configurations using a hybrid enhanced particle swarm optimisation and differential evolution (hybrid EPSO/DE) optimisation technique. The presented optimum CCA and CHA have perfect invariant SLL and high directivity with low mutual coupling by keeping the inter-element spacing not less than half a wavelength which is not possible to be achieved in CA and HA arrangements. Different configurations with two-rings, three-rings and four-rings are presented. The rotation angle of outer rings and the complex excitations of array elements are first optimised. Then, this design is further optimised using a time modulation technique by controlling the switch-on times and the phases of elements excitations of the best CCA and CHA array designs. The presented time modulated (TM) concentric CA and TM concentric HA designs attain ultra-low SLL, reduced sideband level and maximised directivity besides reducing the dynamic range of the array excitations.