نشریه مهندسی برق و مهندسی کامپیوتر ایران
سال ششم شماره 1 (پیاپی 13، بهار 1387)

This research presents new algorithms for subtransmission simultaneous substation and network expansion planning. Given an existing system model, the projected load growth in a target year and various system expansion options, the algorithms find the optimal mix of system expansion options to minimize the cost function subject to various system constraints and single contingencies on lines and transformers.The system expansion options considered include building new subtransmission lines/substations, the capacity to be upgraded and the service area of HV/MV substations. In this research, Genetic Algorithm (GA) with new coding, Ant Colony algorithm (AC) and hybrid Ant Colony and Genetic Algorithm (AC&GA) methods are employed. The optimization results are compared with successive elimination method to demonstrate the performance improvement.

In this paper, a new class of nonlinear thresholding functions with a tunable shape parameter for wavelet-based signal denoising is presented. In addition, a new learning technique for training of thresholding neural network is introduced. Unlike to existing methods, both the shape and the threshold parameters are tuned simultaneously using LMS rule. This permits us to consider the effects of both the threshold and the shape parameters on denoising. The proposed functions are tested in both universal-threshold and subband-adaptive denoising and compared with conventional functions. In addition, to evaluate the proposed training method, several numerical examples are performed. The experimental results obtained from denoising of several standard benchmark signals confirm the efficiency and effectiveness of the proposed methods.

In this paper, two kinds of defects in Golden Delicious apples are recognized: bruise and russet. Russet is divided to two classes: russet in stem-end and russet out of stem-end. Apples are graded into three classes I, II and rejected, according to European standard. To grade the apples, it is necessary to classify apple images into six classes: stem, calyx, bruise, russet in stem-end, russet out of stem-end and healthy. In this method, after pixel-based classification based on RGB color features by a perceptron neural network, correction in classification and stem detection is made. Hue and saturation features are used to correct the image regions classified to bruise. The correction of regions classified to calyx, russet in stem-end and russet out of stem-end is made based on the distance from the gravity center of the stem to the gravity center of each region. This paper presents a new method for defect classification and sub classification of russet to two classes, russet in stem-end and russet out of stem-end. Experimental results of the proposed algorithm show that the correct grading rate of 120 apple images is 81.66%. The grading errors result from misdetection of stem and errors in defect detection.

The main reason of data warehouse systems failure is lack of justification proof. Analysis is an important task for decision about data warehouse creation. In this paper, we present the framework to justify data warehouse based on the input query types. We classify query types and execute them on the databases and data warehouses with different sizes. The query response time and the number of I/O are evaluation parameters. In the experiments, different types of queries have been processed on databases and data warehouses and the results based on time and memory have been compared. These results are presented below:• For answering multidimensional queries and aggregated queries data warehouse systems will be required, • For answering nested queries and join queries, data warehouse system will be useful, • Database systems will be proper for answering simple queries and computational queries. In this work, the tools which can process the above ideas have been produced. The software will take user query and evaluate its process to decide having or not having data warehouses.

This paper describes an approach to adaptive nulling with phased arrays. A genetic algorithm adjusts some of the least significant bits of the beam steering phase-shifters to minimize the total output power of the array. Also, some other criterions such as Mean Square Error and Signal to Interference plus Noise Ratio are used and compared with each other. Using the least significant bits results in small perturbation in the main beam of the radiation pattern and puts the nulls in the direction of the interferences. Double search and weighted mutation are used to reduce the complexity of the algorithm. Also, the performance of genetic algorithm is compared with MPDR which is an optimum technique for beamforming. Finally, it is shown that the genetic algorithm performs superior to MPDR.

In the conventional method for design of waveguide filters, equivalent circuits such as impedance or admittance inverters are used to model the waveguide discontinuities. These inverters are useful for design of filters with narrow bandwidth. In this paper using a genetic algorithm and mode-matching (MM) method, and without using the equivalent circuits, two E-plane and H-plane bandpass waveguide filters are designed. To verify the validity of MM method, the frequency response of filters is obtained by the PML-FDTD method. The two filters are also analyzed by the HFSS software. The results of MM method are in excellent agreement with the results provided by FDTD method and HFSS software. An important characteristic of a waveguide filter is the sensitivity of its frequency response to fabrication errors. In this paper, by using the Monte Carlo statistical method, sensitivity of the frequency response of E-plane and H-plane filters to fabrication errors are determined and compared with each other.

Ensemble learning is an effective machine learning method that improves the classification performance. In this method, the outputs of multiple classifiers are combined so that the better results can be attained. As different classifiers may offer complementary information about the classification, combining classifiers, in an efficient way, can achieve better results than any single classifier. Combining multiple classifiers is only effective if the individual classifiers are accurate and diverse. In this paper, we propose a two-stage method for classifiers combination. In the first stage, by mixture of experts strategy we produce different classifiers and in the second stage by using particle swarm optimization (PSO), we find the optimal weights for linear combination of them. Experimental results on different data sets show that proposed method outperforms the independent training and mixture of experts methods.

In this paper transmission line matrix method is used to simulate the wave propagation through window structures. In this analysis a single window with precise dimensions and also a group of windows are considered. The effect of window’s dimension, window’s material, and angle of incidence are all evaluated in details and in each case, the method has been compared with other numerical results, such as FDTD, ray tracing, and finite element methods. Furthermore, in one case the implementation of TLM method was shown to have a good agreement with the measured values.