ESTEGHLAL
Year:2004 | Volume:23 | Issue:1|Papers Count:14

Many Important problems In Artificial Intelligence can be defined as Constralnt Satisfaction Problems (CSP). These types of problems are defined by a limited set of variables, each having a limited domain and a number of Constralnts on the values of those variables (these problems are also called Consistent Labelling Problems (CLP), in which "Labeling" nuans assigning a value to a variable.) Solution to these problems is a set of unique values for variables such that all the problem constralnts are satisfied. Several search algorithms have been proposed for solving these problems, som of which reduce the need for bacJctracklng by doing some sort of looking to future, and produce more efficient solutions. These are the so-called Forward Checking (FC), Partialiy Lookahead (PL), and Fully Lookahead (FL) algorithms. They are different In terms of the amount of looking to the future, number of backtracks thaJ are performed, and the quality of the solution that they find. In this paper, wepropose a new search algorithm we call Modified Fully Lookahead (MFL) which is Shown to be more efficient than the original Fully Lookahead algorithm

Pulse Width Modulation (PWM) techniques are commonly used to control the output voltage and current of DC to AC converters. Space Vedor Modulation (SVM), of all PWM methods, has attraded attention because of its simplicity and desired properties in digital control of Three-Phase inverters. The main drawback of this PWM technique is its complex and time-consuming computations in real-dine implementation. The dme-consuming calculation as well as software and kardware complaides of the network grow dranwtically as the number of inverter levels increases. Therefore, it is necnsary to develop an exact,fast, and general computation SVM algorithm for muld-Ievel converters. This paper introduces such an algorithm. Spedflcally, the SVM computation algorithm based on a vector classiflcation technique, Introduced for 2-levellnverters in 1996, is developed and generalized to be applicable in determining the switching sequences and calculating the switching instants In m-Ievel inverters. The proposed technique reduces hardware and software complaides, decreases the computation dme, and increases the accuraq of the positioning of the swltchifig instants when compared with the convmtJonalimplementatlon of the SVM in multi-level converters.

In this paper, a modified segmentation algorithm for printed Forsi words is presented. This algorithm is basid on a previous work by Azmi that uses the conditional labeling of the upper contour toflnd the segmentation points. The main objective is to improve the segmentation resultsfor low quality prints. To achieve this, various modifications on local baseline detection, contour labeling and segmentation rules have been applied. In an experiment, the correct segmentation rate was 97%. Based on the results obtained, a detailed error analysis is presented which shouid be useful for funhur research un this topic.

In this paper, a new method for developing a lower bound on exact compietion time distribution function of stochastic PERT lIetworks is provided that is based on simplifying the structure of this type of network. The designed mechanism simplifies network structure by arc duplication so that network distribution function can be calculated only with convolution and multiplication. The selection of dupllcable arcs in this method differs from that of Dodin"s so that it must be considered a different method. In this method, best duplicable arcs are adopted using a new mechanism. It is proved that duplicating numbers is minimized by this method. The distribution function of this method is a lower bound on exact network distribution function and an upper bound on distribution function of Dodin"s and Kleindorfer"s methods. After the algorithm for the method is presented, its efficiency is discussed and illustration examples will be used to Compare numerical results from this method with those from exact network distribution and Dodin"s method.

In this paper, a new algorithm for solving the single loop routing problem is presented. The purpose of the single loop routing problem(SLRP) Is to find the shortest loop for an automated guided vehicle covering at least one edge of each department of a block layout. First it shown that this problem can be represented as a graph model Then a meta-heuristic algorithm based on and colony system is developed for ALRP by using the properties of the graph model. Computational results show the efficiency of the proposed algorithm in comparison with other techniques for solving SLRP.

Similar to random sea waves, forces on the offshore structures due to waves are random. These forces can be mainly divided into two componen.ts, namely, inline forces and transverse or lift forces. The random nature of lift forces is more complicated than that of inline forces and both should be combined for design purposes. In the present paper, two different approaches have been used to determine time series of lift forces. Along these lines, the determination of lift coefficients is discussed which have then been used to obtain transverse forces and compared with experimental data. The experimental data used in this study were coliected at Delft Hydraulics Laboratory on a full-scale rough vertical cylinder.

In the Present study, attempt will be made to propose a new method for prediction of long-term essential creep of concrete utilizing some short-term creep tests under high temperature. To do so, regarding the similarities between essential creep of concrete and creep In viscoelastlc materials, the tlme-temperature equivalence relation in viscoelastlc materials is evaluated for concrete. This relation states that experimental curves of creep at different temperatures fit into a single curve when shifted along the axis of logaritmic time To develop the model, an equation wasjlm developed taking into account the effect of temperatureand the maturity of concrete. Then an appropriate method was proposed for transmissionof the creepcurve of concreteunder aspeclflc temperaturetofit in the creep curve of the same concrete under a temperature. The proposed model was verified using existing experimental data which very good agreement was observed.

In this paper, development of a semi-two- dimensional mathematical model called SDARis introduced. The model is composed of two principal modules of hydraulics and sediment transport. The newly developed SDAR model has a number of capabilities including determination of sediment transport rate, aggradation - degradation calculation, longitinal profile and lateral cross-sectional variation simulation, and tracing armor iayer development phenomena. SDAR is a semJ- two dimensional (S-lD) model, in the sense that the lateral variation of velocity, shear stress, and rate of sediment transport are achieved by subdividing the channel course into several stream tubes carrying equal discharges. Sediment routing and bed variation calculations are accomplished along each stream tube designated by virtual interfaces.

In this paper, an elastic constitutive model based on the Eulerian corotalional rale of the logarithmic strain tensor is proposed. Using this model, the large deformation of a closed cycle containing tension, shear, compression and Inverse shear is analyzed. Since the deformation path Includes a closed cycle and the material is considered as an isotropic elastic material,the normal and shear components of the stress at the end of the cycle must vanish. Using conventional constitutive models, the non-zero solution for the stress components is obtained. Using the proposed constitutive model, the normal and shear components of stress at the end of the cycle are obtained to be exadly equal to zero.

Several experiments Involving two-dimensional and axisymmetric bodies have been carried out In a Trisonic wind tunnel at supersmuc speeds to Investigate and analyse the measured values of base pressure and to compare them with those from the theoretical methods. The objective of the experiments was to obtaln an appropriate method for processing the results of wind tunnel tests on rockets or alrcraft having base area, hence, base drag. Among numerous semi-empirical methods avallablefor two-dimensional and axisymmetric geometry, the methods presented by Chapman, Korst and Tanner have been utilized to calculate values of base pressure for comparison with the experimental flndings for the same modeL The results indicate that the pressure increasefor the 1-D model with a steplike base is greater that of the axisymmetric model with the same conditions. The experimental results for all cases considered in this investigation compare weli with thosefram existing the theoretical method developed by Tanner

In this research work, the possibility of semi industrial production of AI-TiB2 and AI-ZrB2 composites, using reactive slag in a flanu furnace have been investigated. For this purpose, commercial pure aluminum and powder mixturt of TiO2 (ZrO-2 , KBF4 and Na3AlF6 were used. The results showed that using a proper ratio of slag forming materials as well as proper amounts of the above-mentioned compounds make it possible to produce good quality Al- TiB2 and Ai-ZrB2 compounds employing the conventional melting equipment such as aflame furnace.

Among the titanium alioys, Ti-6AI-4V is the most widely used. In the present work, the uniaxial hot compressive behavior of Ti-6Al-4 V has been investigated under constant strain rates. A series of dilatometery experiments were carried out to determine the transformation temperatures at different cooling rates. Specimens were homogenized at 1050 °C for 10 minutesfallowed byfast cooling to different straining temperatures from 1050 to 850°C. The cooling rate was chosen fast enough to prevent high temperature transformation during cooling. A series of isothermal compression tests were conducted at different temperatures of850, 900, 950, 1000, 1050°C at constant true strain rates of 0.1, 0.01 and 0.001 s-1, respectively. Sampies were uniaxialy compressed to a true strain of 0.55 followed by water quenching to room temperature. The apparent activation energy for compression in two phase regions was caiculated at 840 KJmol-1. The partiai globular/zation of a phase was observed in the specimens deformed at low strain rates and at temperatures near the transformation zone followed by annealing.

Physkal properties of cotton yarns are affected by the characteristics of cotton fibers such as fineness, length, maturity and strength. This relationship has been worked out by means of multivariable regression and stepwise method for an open-end spun (Nee 20) cotton yarn. Moreover, with the help of linear programming, it was made possible to determine the percentage of different cottons in the blend with the aim of reducing the yarn price to a minimum while keeping the yam quality to a certain level.

Process optimization is one of the most important adivities in today"s competitive industries. the rather high cost of research and development has necessitated the development of experimental methods by which the factors affecting processes could be determined with minimum number of experiments. Over the last two decades, various types of experimental designs have been used. Among the different methods of experimental design such as complete and partial factorial and Latin squares design, the Taguchi method has found wide applications in some industrial divisions because of its comprehensive nature. In this paper, the laboratory scale flotation of the Sarcheshmeh copper ore which mainly consists of chalcocite was studied using the Taguchi method. The effects of seven factors namely collectors, Z11 (Xanthate) and R407 (Methyl Isoboutyl carbonyl),frothers, pine oil and A65 (Polypropylene glycof), particle size, pH and flotation time were investigated. In addition to a significant reduction in cost and time of experimentation, the results indicated that a 5% increase in copper recovery couid be obtalned if all the factors are tested at their high experimental levels, as suggested by the Taguchi method. The optimal flotation time was also found to be 11 minutes.