Year:2004 | Volume:38 | Issue:1 (83)|Papers Count:15

One of the important parameters to enhance the efficiency of Fresnel Zone Plate Antenna (FZPA) is the phase efficiency. This efficiency deepens on the number of layers in the multi-layer FZPA and dielectric materials in the multi-dielectric FZPA. Because the number of Fresnel Zone is limited by the mechanical and practical restrictions, this efficiency is also related to these parameters. Therefore, to obtain the maximum gain or perform optimum design of FZPA one must choose suitable values for the number of Fresnel zones and their radii, focal length, characteristics of feed and also must consider practical and mechanical restrictions. In this paper, methods for analysis of these structures are first introduced, then, a proper method for analyzing the structure is chosen. Second, for a given FZPA diameter and a given feed with symmetrical radiation pattern, a procedure is introduced to obtain optimum values for the focal length, radii of Fresnel Zones and simulated feed parameter, so that they yield the maximum gain and desired side lobe level of radiation pattern. Finally the simulated and measurement results are compared.

In this paper, a simple but accurate method for increasing the accuracy of relative localization of wheeled mobile robots with indoor applications is introduced. In contrast to existing methods, the proposed localization algorithm not only increases the accuracy of robot orientation, but also reduces the robot position error -that is essential for accomplishing object manipulation task with cooperative mobile robots. The proposed method benefits from combining the information of gyroscope, odometery and dead-reckoning of a passive wheel which is located between the active wheels. The information of passive wheel is compared with the dead-reckoning and gyroscope data to determine the roughness of ground or the wheel slippage. The experimental and simulation results show that, the proposed method is highly effective in reducing dead-reckoning errors even when the roughness of terrain is not negligible.

Practicability of driving an induction motor that has turn faults, and permissible operating points, at which no fault enlargement occurs, is shown in literature. In this paper, the permissible domain of operating points (PDOP) for a faulty motor are classified to steady and transient PDOP, which respectively means working points for arbitrary and short periods of times. As expected, for a certain size of fault, higher speed and torque as transient are tolerable, and this idea can be used for a more suitable control of a faulty motor. Dependency of PDOP on ambient temperature and capability of neural networks in shortening the time consuming process of determining steady PDOPs are shown. A thermal model for faulty motor is proposed which assists in following the temperature of fault location, and can be used for obtaining transient PDOPs and steady PDOPs at different ambient temperatures. The parameters of the model are obtained with the aid of a GA, neural network and the results of some experiments. The validity and accuracy of results are shown with the aid of many experimental tests.

In this paper, a method of dimensionality reduction for MR image sequence data of brain, to a three-dimensional space is presented. Reduction of dimensions is motivated by the need to visualize the distribution of data and interactively segment the MR image sequence. New feature space is extracted using a non-linear neural network. A genetic algorithm is used to search for the neural network parameters such that the transformed data in the network output optimize a specific objective function. Three objective functions are proposed based on Sammons cost function where in two of them a constraint is added to the cost function. The data in the three-dimensional (3-D) output feature space is visualized using the perspective image of the 3-D histogram. MR images are segmented interactively by determining the cluster centers in the perspective image. The results of the proposed methods are compared with those of Linear Transformation and Back-Propagation Neural Network (BPNN) methods. For simulated MR images, two of the proposed criterions produce better results in terms of cluster separation. Based on 10 real MR image sequence data, the same two criterions result in lower segmentation error rates.

In this paper an objective orientation is presented for controlling multi-objective systems. The principles of this method is based on an emotional learning and temporal difference learning, and has a neuro-fuzzy structure. The proposal method can control the system in a way that objectives such as the present conditions, the system action in the part and the controlling aims are attained in the best way and least amount of time.The holistic structure of this paper is as follows: first, in the third unit object oriented control is studied. The fourth unit deals with emotional learning as a method for intelligent control. In the fifth unit the emotional control of temporal difference is defined and discussed. The sixth unit presents and discusses a new critic structure with temporal difference learning and then Some Multivariable systems, which are one of the most important controlling problems because of coupling, are defined. Then in the last unit the functional control structures used in this paper are studied and the experimental results are compared.

The Surface Equivalence principle and Multiport Network method (SEMN) have been previously developed for analysis of arbitrary shaped micro strip antennas. However, the computation time for analysis of finite and infinite micro strip arrays by the SEMN method in spite of its generality makes it impractical. Even, analysis of single element micro strip antennas without considering geometrical and methodological symmetries takes some days by a 2GHz computer. In this paper, new algorithms that speeds up the calculation spectacularly are introduced that enables us to analyze large structures and arrays with no loss in accuracy and generality of SEMN method. The modified SEMN method is then applied to a four-element micro strip array antenna and the calculated results are compared with measurement. Additionally, an infinite periodic micro strip array antenna is analyzed by the modified SEMN method where the convergence rate of the involved double infinite summations is increased by the Shanks method. The infinite periodic micro strip patch array antenna analysis is in excellent agreement with the MoM (Method of Moment) solutions.

A distribution feeder reconfiguration method has been developed. Feeder reconfiguration is an easy and in-expensive method, for distribution system loss reduction, which is done by maneuver in the network. In this work, Graph Theory approach has been adopted for analysis and optimization of distribution system. Distribution system partitioning and optimization is modeled as a linear programming problem, using this theory. The distribution system is modeled as a hyper-graph, and specific weights are assigned to every branch of the hyper-graph. Then the distribution system is partitioned to sub-graphs. Feeder reconfiguration for loss reduction can be performed in all of the sub-graphs, simultaneously. The most important feature of the presented method, is improving the performance of reconfiguration for large scale networks. Noticeable reduction in calculation time is the other advantage of network partitioning that makes it suitable for fast and real-time applications. Performance of the method has been presented on a standard distribution system. Solution algorithm and discussion is included.

Information Fusion compared with Data Fusion concept stands in a higher level of hierarchy on decision making process. As a result, this makes it possible to achieve better and more complete results.Today intelligent systems are able to act as personal assistant of human, and provide the needed information among the available information sources. The research work on this field is essential and there are still many deficiencies on this respect to be sorted out. In this article, we have focused on the Information Fusion concept with a novel approach to design and implement an architecture of a customized intelligent agent, which will be able to retrieve the information needed by user with respect to his/her interests from various information servers (such as internet and intranet search engines and databases). The designed agent uses the different information fusion techniques to achieve this goal. The OWA operator which is the subset of the great family of Fuzzy Integral Operators is used to design the operational model of above mentioned agent. This agent fuses the Data, Features, and Decisions in three distinct levels.The agent is able to cluster categories, and extracts the behavioral model of information servers with respect to each cluster and uses these models in final decision making process.

The existing Gravity model, in trip distribution forecasting, involves iterative adjustments and time consuming, because it is not able to perform adjustment simultaneously. This drawback has been considered in this article.In the real world problems, the transportation parameters such as traveling time are not crisp. Therefore, it has been addressed by possibility distribution here.In this paper a new mathematical model is proposed to consider the above shortcomings. Finally the superiority of the new method has been shown by a numerical example.

In this paper, cellular manufacturing (CM) problems in dynamic and stochastic states have been solved by using a genetic algorithm. In fact, cellular manufacturing system is an application of group technology in a production system. The aim of such system is to cluster parts and machines into the related families in such a way that operational similarity in different production and planning aspects is considered. In most previous researches, cellular manufacturing problems in static production or deterministic demand have been considered continuously. However in a real-world situation, dynamic production and demand for products are non-deterministic. So in order to adapt a CM model with real situation, a great number of variables and restrictions are required. Thus, solution to the above model requires more computational time, memory, and high processing power of computer with traditional optimization methods. Consequently nowadays, modern heuristic methods such as genetic algorithms has been applied in a group of stochastic search techniques in order to solve such a NP-complete problem. In this paper, first a non-linear integer model is developed and then it is solved by genetic algorithms. Finally, obtained results are compared with optimal solutions to show the validity of the proposed algorithm.

In this paper we study a two-echelon inventory system with maximum profit objective for perishable items in which inflation effects are considered. We consider random product life and deterministic demand. Shortage is not allowed. Based on the stage in which perishable items are stocked, two models are developed. For each model, optimum order quantity is obtained and a practical example is solved. The models are evaluated through sensitivity analysis.

One of the known problems in the field of operation research, especially production planning, is cutting problem. In industrial field, two-dimension cutting problem is more compatible with real problem. Principal objective is to reduce losses resulted from cutting. In this paper, according to this objective, we develop solution space through an "and-or" graph. Then, we provide an improved branch-and-bound method in order to reducing solution space according to a stochastic optimization algorithm. Obtained results show the algorithm suitable capability in comparison with current algorithms.

This paper focuses on the formulation and solution of the problem of just in time supply of required materials or parts from suppliers for a manufacturing system. The unique characteristic of this problem is that the quantity of required material to be delivered to a central depot, is a function of the route taken by a vehicle assigned to serve some selected suppliers. Minimizing number of required vehicles and minimizing average inventories at the central depot are the objectives of this problem. The problem is formulated as a mathematical programming model and then an ant algorithm with recursive improvement is suggested. The performance of the ant algorithm is evaluated in comparison with the mathematical model. The ant algorithm with recursive improvement is shown to provide good results.

This paper aims at developing a manufacturing planning methodology for an Automated Guided Vehicle (AGV) based flexible manufacturing system by simultaneously dealing with material processing and transportation functions based on the Petri Nets (PN). As a graphical tool, PN provides visualization of a dynamic system and as a mathematical tool, it checks the properties related to the behavior of the underlying system, by using a set of linear algebraic equations. The integrated model is developed based on three modules: AGV module, Product module and machine module. The AGV module models the movement of AGVs from a current location to a destination. For each part type, the product module provides information to perform the operations of the products in accordance with the precedence relations. The machine module models the assignment of machines to tasks by using Sequential Petri Net (SPN). Each of these modules can model partial properties of the system and we can get a total picture of system by integrating them. In order to obtain an effective schedule while avoiding the NP-complete computing complexity, scheduling is performed as a heuristic search in the reach ability tree, which is guided by heuristic dispatching rules of AGVs and machines. The algorithm combines on-line dispatching rules of AGVs and machines based on the information provided by the PN simulation with the Limited Expansion A Algorithm. Experimental results for a practical example from IRAN KHODRO HEMMING SHOP indicate the effectiveness of PN- based heuristic search for FMS scheduling.

The aim of this paper is to present a methodology for using a data envelope analysis method at research organizations with intangible quantitative and qualitative inputs and outputs. This paper covers the way of applying changes and limitations used in the data envelope analysis model. It also shows the manner of transferring limitations and factors into the mathematical equations and terms. With respect to the attributes of research organizations and the existing different conditions in data analysis method, the use of the above method will not be possible easily to measure the efficiency of research centers. One of the most important existing problems in this respect is to quantify and qualify inputs and outputs at research organizations. Consequently, it is so difficult task to convert quality cases to quantity ones and to create uniform inputs and outputs. This activity in particular has been used at one of the research organizations in ministry of energy practically. Then, the efficiency of different engineering departments has been measured in three consecutive periods with respect to the similarity of activities. It is concluded that the trend of efficiency changes at each department is defined by linear programming models and is quantified by the use of LINGO software. The obtained result has been reported to the top manager as management information for making a decision.