Muge Komurcu

The Wind Driven Optimization Technique and its Application in Electromagnetics

A new type of nature-inspired global optimization methodology based on atmospheric motion is introduced. The proposed Wind Driven Optimization (WDO) technique is a population based iterative heuristic global optimization algorithm for multi-dimensional and multi-modal problems with the potential to implement constraints on the search domain. At its core, a population of infinitesimally small air parcels navigates over an

Wind Driven Optimization (WDO): A novel nature-inspired optimization algorithm and its application to electromagnetics

N

Stub-loaded inverted-F antenna synthesis via Wind Driven Optimization

-dimensional search space following Newtons second law of motion, which is also used to describe the motion of air parcels within the earths atmosphere. Compared to similar particle based algorithms, WDO employs additional terms in the velocity update equation (e.g., gravitation and Coriolis forces), providing robustness and extra degrees of freedom to fine tune. Along with the theory and terminology of WDO, a numerical study for tuning the WDO parameters is presented. WDO is further applied to three electromagnetics optimization problems, including the synthesis of a linear antenna array, a double-sided artificial magnetic conductor for WiFi applications, and an E-shaped microstrip patch antenna. These examples suggest that WDO can, in some cases, out-perform other well-known techniques such as Particle Swarm Optimization (PSO), Genetic Algorithm (GA) or Differential Evolution (DE) and that WDO is well-suited for problems with both discrete and continuous-valued parameters.

Adaptive Wind Driven Optimization

Nature-inspired techniques such as the Genetic Algorithm (GA) [1], Ant Colony Optimization (ACO) [2], and Particle Swarm Optimization (PSO) [3] have been shown to be some of the most effective global optimization strategies. Consequently, these techniques are currently in widespread use throughout the scientific and engineering communities. In this paper, we introduce a new type of global optimization algorithm that is inspired by the motion of wind in the Earths atmosphere. We call this new nature-inspired technique Wind-Driven Optimization (WDO). WDO is a population based iterative heuristic global optimization technique for multidimensional problems. A population of infinitesimally small air parcels are distributed throughout an N-dimensional problem space and assigned random velocities such that the positions of air parcels are updated at each iteration based on the physical equations that govern large-scale atmospheric motion.

Characterisation of pollutant sources in Istanbul with PM10 and EU directives

A novel technique to design and optimize a low profile wire antenna with high broadside gain is introduced. The desired antenna properties are achieved by loading an electrically long inverted-F antenna (IFA) with two optimal wire stubs along the body of the antenna. The best lengths and locations of the stubs as well as other antenna design parameters were selected by a new nature-inspired technique called Wind Driven Optimization (WDO). The results of this work demonstrate that an optimized stub-loaded inverted-F antenna (SLIFA) can provide not only lower profile than that of a conventional IFA but also an enhanced gain of 8.2 dB at broadside where the IFA has a null.

Multiobjective Adaptive Wind Driven Optimization

In this paper, we propose two new methods to create an adaptive Wind Driven Optimization (WDO) algorithm, both of which are shown to outperform the classical WDO method while eliminating the need for fine-tuning the coefficients of the update equations. While the classical WDO offers a simple and efficient meta-heuristic optimization algorithm, the coefficients that are inherent to the workings of the algorithm create an undesired level of complexity especially for the novice users. To alleviate this complexity and automate the coefficient selection, two adaptive Wind Driven Optimization (AWDO) methods are proposed in this paper. First method is to replace the fixed values of the coefficients with randomly generated numbers from a uniform distribution at each iteration and the second method is to optimize the selection of the coefficients with the Covariance Matrix Adaptation Evaluation Strategy (CMAES). To evaluate the performance of the proposed methods for AWDO, four well-known numerical benchmark functions from the literature are utilized and results are compared against the classical WDO. Both of new methods outperform the classical WDO while the AWDO using CMAES performs the best among of all.

Two Neural Network Methods in Estimation of Air Pollution Time Series

Air pollution is globally significant because pollutants not only affect their source region, but also through transport they can affect other regions. The purpose of this study is to show the stand point of Turkey for EU directives. Sampling days that exceed PM10 limit values of EU PM10 restrictions are chosen as episode days for this research (PM10 > 150 μg/m³). Two test cases representative of summer and winter are investigated in terms of stability conditions and synoptic fields. Considering these fields, the dominant phenomenon in generating pollution is defined as either the local sources contribution or the transport processes.

Estimation of Air Pollution Parameters Using Artificial Neural Networks

In this work, we introduce a new nature-inspired multiobjective numerical optimization algorithm where Pareto dominance is incorporated into Adaptive Wind Driven Optimization for handling multiobjective optimization problems and named as Multiobjective Adaptive Wind Driven Optimization (MO-AWDO) method. This new approach utilizes an external repository of air parcels to record the non-dominated Paretofronts found at each iteration via the fast non-dominated sorting algorithm, which are then utilized in the velocity update equation of the AWDO for the next iteration. The performance of the MO-AWDO is tested on five different numerical test functions with two objectives and results indicate that the MO-AWDO offers a very competitive approach compared to well-known methods in the published literature even performing better than NSGA-II for ZDT4 test function.

Intercomparison of the cloud water phase among global climate models

The measurement of air pollution parameters is a costly process. Due to several reasons, the devices may not take measurements for certain days. In such cases robust estimation methods are quite necessary in order to fill the gaps in the time series. Artificial neural networks have been employed successfully for this purpose for hydrometeorological time series, as reported in literature. In this study, modelling of the time series of air pollution parameters was investigated using two ANN methods; a radial basis function algorithm (RBF) and feed forward back propagation method (FFBP). The ANN methods were employed to estimate the PM10 values using the NO and CO values. The data were from a measurement station in Istanbul, Turkey. The results of an initial statistical analysis were considered in the determination of the input layer node number. In the estimation study, values corresponding to other air pollution parameters were included in the input layer. The results were compared to those obtained with a conventional multi-linear regression (MLR) method.

Inter-comparison of the phase partitioning of cloud water among global climate models

The modeling of air pollution parameters is an issue investigated using different techniques. The pollution time series, however, are not continuous and contain gaps. Therefore, methods to infill the gaps providing satisfactory estimations are quite significant. In the presented study two ANN methods, feed forward back propagation, FFBP, and radial basis functions, RBF, were presented to estimate the SO2 values using the NO and CO values. It was seen that both ANN methods provided superior performances to conventional multi linear regression, MLR, method. The ANN performances were found satisfactory considering the selected performance criteria and the testing stage plots.