Multiobjective Pump Scheduling Optimization Using Harmony Search Algorithm (HSA) and Polyphonic HSA

Harmony Search Algorithm (HSA) is a metaheuristic method that has attracted the scientific interest since its first presentation in 2001. It is a music inspired method, imitating the music creation process in order to find optimal solutions in complicated problems. HSA’s successful application on single – objective optimization problems has resulted to an increasing interest in the implementation of HSA towards multiobjective optimization. The authors have adjusted HSA in order to deal successfully with multi-criteria water management problems. This adjustment has resulted to the creation of Multiobjective – HSA (MO-HSA). In addition, they have designed the multiobjective variant Polyphonic-HSA (Poly-HSA), which is inspired by the independent development of different voices in music and borrows elements from Swarm Intelligence and the single-objective variant Global-Best HSA. In the first part of this paper, both methods are presented in detail. Moreover, the performance of the proposed Algorithms is evaluated using standard multiobjective test – functions. ZDT and DTLZ multiobjective tests have been chosen and indicators such as Hypervolume, C – metric and diversity metric – Δ have been used to measure the convergence to the optimal front and the diversity of the solutions obtained by the proposed methods. In the second part, MO-HSA and Poly-HSA have been introduced towards the optimization of a pump scheduling problem. The objectives considered are water supply, pumping cost, electric power peak demand and pump maintenance cost. Both methods converged to non-dominated fronts and provided excellent results which are presented in 3d figures, indicating that these methods can be effectively used in multiobjective water management problems.     

Modelling of grate combustion in a medium scale biomass furnace for control purposes

A new mathematical model for the grate combustion of biomass has been derived from physical considerations. Various models for grate combustion can already be found in the literature. Usually their intention is to simulate the real situation in a furnace as precisely as possible. Hence they are very detailed, typically consisting of many partial differential equations. However, because of their complexity they are useless for control purposes. The new model is very simple, consisting of only two ordinary differential equations, which makes it particularly suitable as a basis for model based control strategies. To verify the model, experiments were performed at a pilot scale furnace with horizontally moving grate. The pilot plant is a downscaled version (180 kW_th) of a typical medium scale furnace in terms of geometry and instrumentation. Comparison of the measured and calculated values shows good agreement.     

Spectroscopic studies of the binding of bilirubin by ligandin and aminoazo-dye-binding protein A

Ligandin and aminoazo-dye-binding protein A both bind bilirubin at a single site. Quantitative studies of the interactions using difference spectrophotometry show that at pH 7.0, protein A binds the tetrapyrrole with an association constant (K) greater than or equal to 2 X 10(7) litre/mol, whereas binding by ligandin is slightly weaker (K = 7 X 10(6) litre/mol) at this pH. The protein-bilirubin complexes give rise to absorption and fluorescence spectra quite different from those of unbound bilirubin and also to large Cotton effects. It appears that on binding to both proteins, the ligand is forced into a rigid twisted configuration in a hydrophobic environment. Ligandin and protein A resemble serum albumin in their interactions with bilirubin.     

A data locality methodology for matrix–matrix multiplication algorithm

Matrix-Matrix Multiplication (MMM) is a highly important kernel in linear algebra algorithms and the performance of its implementations depends on the memory utilization and data locality. There are MMM algorithms, such as standard, Strassen–Winograd variant, and many recursive array layouts, such as Z-Morton or U-Morton. However, their data locality is lower than that of the proposed methodology. Moreover, several SOA (state of the art) self-tuning libraries exist, such as ATLAS for MMM algorithm, which tests many MMM implementations. During the installation of ATLAS, on the one hand an extremely complex empirical tuning step is required, and on the other hand a large number of compiler options are used, both of which are not included in the scope of this paper. In this paper, a new methodology using the standard MMM algorithm is presented, achieving improved performance by focusing on data locality (both temporal and spatial). This methodology finds the scheduling which conforms with the optimum memory management. Compared with (Chatterjee et al. in IEEE Trans. Parallel Distrib. Syst. 13:1105, 2002; Li and Garzaran in Proc. of Lang. Compil. Parallel Comput., 2005; Bilmes et al. in Proc. of the 11th ACM Int. Conf. Super-comput., 1997; Aberdeen and Baxter in Concurr. Comput. Pract. Exp. 13:103, 2001), the proposed methodology has two major advantages. Firstly, the scheduling used for the tile level is different from the element level’s one, having better data locality, suited to the sizes of memory hierarchy. Secondly, its exploration time is short, because it searches only for the number of the level of tiling used, and between (1, 2) (Sect. 4) for finding the best tile size for each cache level. A software tool (C-code) implementing the above methodology was developed, having the hardware model and the matrix sizes as input. This methodology has better performance against others at a wide range of architectures. Compared with the best existing related work, which we implemented, better performance up to 55% than the Standard MMM algorithm and up to 35% than Strassen’s is observed, both under recursive data array layouts.     

Short-term electric power load forecasting using feedforward neural networks

Abstract: This paper presents the results of a study on short‐term electric power load forecasting based on feedforward neural networks. The study investigates the design components that are critical in power load forecasting, which include the selection of the inputs and outputs from the data, the formation of the training and the testing sets, and the performance of the neural network models trained to forecast power load for the next hour and the next day. The experiments are used to identify the combination of the most significant parameters that can be used to form the inputs of the neural networks in order to reduce the prediction error. The prediction error is also reduced by predicting the difference between the power load of the next hour (day) and that of the present hour (day). This is a promising alternative to the commonly used approach of predicting the actual power load. The potential of the proposed method is revealed by its comparison with two existing approaches that utilize neural networks for electric power load forecasting.     

Detection of chromosomal and plasmid--encoded virulence determinants in Yersinia enterocolitica and other Yersinia spp. isolated from food animals in Greece

The distribution of Yersinia strains in animal reservoirs was examined in 835 food animals (pigs, chickens, sheep, cows) from different Greek departments (Attica, Fthiotida, Viotia and Evia) over a one year period. The isolated strains were characterized with respect to the presence of chromosomal (yst) and plasmid-encoded virulence determinants (virF, yadA) and their antimicrobial susceptibility was tested.In total, Yersiniaspp. were obtained from 9.94% of the 835 food animals at slaughter that were sampled in this study. There was no statistically significant seasonal distribution, nor was any significant departmental distribution observed. From the 83 isolated Yersinia strains, 76 (91,57%) belonged to Y. enterocolitica (58 were of serotype O:3/biotype 4 and 18 strains were non O:3, non O:9), 3 belonged to Y. pseudotuberculosis, 2 to Y. kristensenii and 2 to Y. intermedia. Y. enterocolitica O:3/4 was mainly isolated from the pigs, while Y. enterocolitica non O:3, non O:9 was from the chickens. The strains were grouped into 5 genotypes, with respect to the presence or absence of the virulence genes. A significant predominance of genotype V, the one carrying all the three virulence genes, was observed in the strains isolated from the pigs. Complete susceptibility to most of the 3rd and to the 4th generation cephalosporins and to ciprofloxacin, was observed among the isolates. Remarkable was the association between the presence of each virulence gene separately and resistance to some antimicrobials, a matter of further investigation.     

Eliminating SRS channel depletion in massive WDM systems viaoptical filtering techniques

We demonstrate that the channel depletion due to stimulated Raman scattering in massive wavelength-division-multiplexed (WDM) systems can be eliminated using high-frequency pass filters. These filters, when inserted appropriately into the transmission link, can effectively suppress the SRS power flow from the WDM channels to lower frequency noise. Numerical simulations carried out on WDM systems indicate that the channel depletion penalties can be kept below 0.25 dB even for a total channel power of 2 W     

Effekte zeitdiskreter Ansteuerung von Asynchronmaschinen bei hohen Drehzahlen

When controlling an induction machine, the stator voltage vector is discretized temporally. This leads to cross-coupling in the rotor flux frame as well as to current ripples, deteriorating flux control and estimation performance. This article presents both input and output compensation methods for this effect.